{ "version": "https://jsonfeed.org/version/1.1", "title": "arXiv", "feed_url": "https://raw.githubusercontent.com/trvny/feedseek/main/feeds/feed_arxiv.json", "home_page_url": "https://arxiv.org/", "description": "arXiv blog and daily new submissions, alphaXiv Explore, LessWrong, and 80,000 Hours.", "favicon": "https://arxiv.org/favicon.ico", "icon": "https://arxiv.org/favicon.ico", "items": [ { "id": "tag:trvny.github.io,2024:feedseek/arxiv/33f0defc32dcaf31", "url": "https://www.lesswrong.com/posts/Rt97G7LpYySgzdcQG/athletic-education-vs-athletic-torment", "title": "Athletic education vs. athletic torment", "content_text": "I don’t know if it occurred to me until my thirties to think of exercise as an enjoyable thing . I was familiar with finding obscure corner-cases that were fun, such as Dance Dance Revolution . But the idea of it just often being a good time was alien. I hesitate to blame anyone for anything, but school seems culpable here. I got the impression so firmly that PE class (‘physical education’ or ‘physed’) was a kind of horror, I’m not sure I would have treated this fact as on less solid ground than", "date_published": "2026-10-01T22:38:24Z", "date_modified": "2026-10-01T22:38:24Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e8409dcb539d50f1", "url": "https://www.lesswrong.com/posts/b5cSYh4emQb2qrGmK/on-social-reality-in-china", "title": "On Social Reality in China", "content_text": "[Epistemic status: intuitions and anecdotes.] Recently, several posts and projects ( Thoughts Memo , Babel Translation , Please Give Them a Chance ) have taken important steps towards raising AI safety awareness and sharing rationalist philosophy in China. It’s great that we’re recognizing the importance of solving the messaging problem for China, and thus laying the groundwork for an international AI pause. Below I record my perspective on cultural differences which are relatively underdiscusse", "date_published": "2026-10-01T20:49:52Z", "date_modified": "2026-10-01T20:49:52Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/b5cSYh4emQb2qrGmK/bpmks8ouco3u8phattkz", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/b5cSYh4emQb2qrGmK/bpmks8ouco3u8phattkz", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4daadacbcf33b137", "url": "https://blog.arxiv.org/2026/10/01/updated-rate-limit-policy/", "title": "Fair Moderation, Equitable Access, and AI: arXiv’s Updated Rate Limit Policy", "content_text": "arXiv, and the scientific community at large, are facing a watershed moment. Scholarly publishing is currently changing at a rapid pace, and we are seeing a massive transformation in how researchers communicate their results. Before the advent of AI, there was an easily discernible, practical limit to the rate at which independent submissions could be […]", "date_published": "2026-10-01T19:00:00Z", "date_modified": "2026-10-01T19:00:00Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/10/arXiv-Monthly-Submissions-September-2026.png?resize=1024%2C642&ssl=1", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/10/arXiv-Monthly-Submissions-September-2026.png?resize=1024%2C642&ssl=1", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c5f789fc62ea373f", "url": "https://www.lesswrong.com/posts/cMpCgoYw53WictTLH/book-review-obsolete", "title": "Book Review: Obsolete", "content_text": "Book review: Obsolete: The AI Industry’s Trillion Dollar Race to Replace Us? and How to Stop It, by Garrison Lovely. Obsolete is a strange book. It’s mostly good, but the quality varies a good deal. Lovely is well informed about concerns that AI might kill us all, and doesn’t dispute those concerns. Yet he focuses much more on more ordinary concerns, in particular excessive concentration of power. He is somewhat successful at portraying mundane harms of this year, and longer term risks (will AI", "date_published": "2026-10-01T18:52:51Z", "date_modified": "2026-10-01T18:52:51Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/cMpCgoYw53WictTLH/j5mx12dkzuxxyacwmev0", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/cMpCgoYw53WictTLH/j5mx12dkzuxxyacwmev0", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0d20a19a9ae6de6d", "url": "https://www.lesswrong.com/posts/mmfYDH6hwYAMToNKL/evaluating-the-ban-artificial-superintelligence-act", "title": "Evaluating the Ban Artificial Superintelligence Act", "content_text": "MIRI recently endorsed the Ban Artificial Superintelligence Act, but others like ControlAI have called it overly broad, and a few other respected experts have outright endorsed it in its current form. I decided to take a look at it myself and see both what the bid is trying to do and whether it actually does it well. The goal of this act is to ban any AI that shows traits that indicate it could cause great harm to society, and to restrict the development of advanced AI to certified institutions", "date_published": "2026-10-01T18:46:59Z", "date_modified": "2026-10-01T18:46:59Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/23e3c1aea86ee9ae", "url": "https://www.lesswrong.com/posts/hcSkr4fucZ4cnqPaJ/do-you-understand-what-i-mean-when-i-say-existential-risk", "title": "Do you understand what I mean when I say “existential risk”?", "content_text": "By the end of this post, I hope you’ll take existential risk more seriously. I hope you will feel as viscerally as I do the stakes of human survival and flourishing. When I tell people, “I think AI is an existential risk”, usually nothing changes about their expression. They nod along, same blank face, thinking about what they’re going to have for lunch today. I want to take them by the shoulders and shake them and ask, “Are you listening? Do you know what I mean when I say ‘existential risk’? W", "date_published": "2026-10-01T18:25:40Z", "date_modified": "2026-10-01T18:25:40Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f240ce82721af9d2", "url": "https://www.lesswrong.com/posts/rktrjGjXEbWsg7zof/my-retrospective-from-mats-10-0", "title": "My retrospective from MATS 10.0", "content_text": "I have recently completed MATS 10.0, where I worked alongside Bart Jaworski under Victoria Krakovna (GDM). This is a post I was encouraged to make by my team at Geodesic Research from some slides I put together. This is not a post on application advice to MATS, or about the program in general. It is rather a compressed form of my experience doing research and lessons from the project. The project The full paper and post is coming soon, but I'll provide some context on it so that the lessons don'", "date_published": "2026-10-01T18:24:41Z", "date_modified": "2026-10-01T18:24:41Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0647b4d835d5c2b1", "url": "https://80000hours.org/podcast/episodes/simon-goldstein-ai-legal-rights/", "title": "Simon Goldstein on the case for giving AI (some) legal rights", "content_text": "The post Simon Goldstein on the case for giving AI (some) legal rights appeared first on 80,000 Hours .", "date_published": "2026-10-01T18:06:43Z", "date_modified": "2026-10-01T18:06:43Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/10/Simon-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/10/Simon-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/11cd36f84e92b905", "url": "https://www.lesswrong.com/posts/rp4sB5a6HaPpfD4rA/endogenous-alignment-requires-dependence", "title": "Endogenous Alignment Requires Dependence", "content_text": "My post on endogenous alignment got several excellent comments. Among them, one from Karl Krueger argued that imitation learning is more important for how humans align their children than reinforcement learning. Teasing out which is more important seems hard, but I agree I was wrong to ignore the importance of imitation, and arguably imitation learning is more fundamental for a reason explained by another comment. As Gunnar Zarncke helpfully pointed out , I had forgotten the most critical part o", "date_published": "2026-10-01T16:20:56Z", "date_modified": "2026-10-01T16:20:56Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/db9d5086cd3297aa", "url": "https://www.lesswrong.com/posts/fBR3kG2FAcrDChd6L/2026-sff-grants", "title": "2026 SFF grants", "content_text": "This year it's up to $65M. Historically the funder has usually just been Jaan; this year Dustin was also a funder. I'm happy about these grants. I'm grateful to SFC, the speculators & evaluators, and especially Jaan + Dustin for making these grants happen. Discuss", "date_published": "2026-10-01T16:10:23Z", "date_modified": "2026-10-01T16:10:23Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2cb4344f9420e424", "url": "https://www.lesswrong.com/posts/CjZJLXRMDdbgAH3iG/scientists-on-the-manhattan-project-continued-developing-the", "title": "Scientists on the Manhattan Project continued developing the atomic bomb even after it was clear that there was no need to", "content_text": "Note: I'm not a WWII history buff, so if there's any significant misinterpretation of relevant history here, I welcome corrections in the comments. In December 1944, Joseph Rotblat , a physicist working on the Manhattan project, resigned. He had reasoned that Germany had most likely abandoned its bomb project, and so his initial reason for working on the bomb was no longer valid. [1] He would later ask himself: Why did other scientists not make the same decision? [...]there were many scientists", "date_published": "2026-10-01T16:01:41Z", "date_modified": "2026-10-01T16:01:41Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/375ab40944f52748", "url": "https://www.lesswrong.com/posts/5L7beYAmpu8zEvnYq/learning-steganography-is-easy-learning-steganographic", "title": "Learning Steganography Is Easy, Learning Steganographic Reasoning Is Hard", "content_text": "This post summarises our paper Learning Steganography Is Easy, Learning Steganographic Reasoning Is Hard , accepted as an oral at the NeurIPS 2026 Workshop on Trustworthy AI for Good (AI4GOOD). Code, configurations and results for all experiments: github.com/stegano-ai/steg-reasoning-is-hard . This work was done as part of the Meridian Visiting Researcher Programme and the Safe AI Germany (SAIGE) Incubator Program , with funding from Coefficient Giving. TL;DR Chain-of-thought (CoT) monitoring fa", "date_published": "2026-10-01T15:08:48Z", "date_modified": "2026-10-01T15:08:48Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790857271/lexical_client_uploads/mxrywkhtkw1kobxylod5.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790857271/lexical_client_uploads/mxrywkhtkw1kobxylod5.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7e31a892f6c48ed4", "url": "https://www.lesswrong.com/posts/S2EAn9v4BwRdptsom/ai-188-gemini-dot-argon", "title": "AI #188: Gemini Dot Argon", "content_text": "Is Google back? They claim that they are back. Gemini 4 Argon is rolling out, with competitive frontier-level benchmarks, at $2/$10. What we don’t have is access to the model, because Google Fails Marketing Forever. So it is far too early to say what we have here. When I know more, so will you. OpenAI was forced to pull what would have been GPT-6.1 Astra due to alignment failures . They did offer us GPT-6.1 Sol, which is pitched as approaching Astra quality at the much lower price of $2/$10, the", "date_published": "2026-10-01T15:00:55Z", "date_modified": "2026-10-01T15:00:55Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/S2EAn9v4BwRdptsom/rce7pwhur4dxb3pbnjzq", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/S2EAn9v4BwRdptsom/rce7pwhur4dxb3pbnjzq", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7aededabbaf9c760", "url": "https://www.lesswrong.com/posts/Fs4inpM72PeH4hsAv/simplex-vs-timaeus-round-one", "title": "Simplex vs Timaeus: Round One", "content_text": "I'm often asked about the differences and similarities between Simplex' and Timaeus' research agendas. The question is natural enough. Both focus on a 'fundamental science' approach to AI alignment. Both organizations base their research agendas on sophisticated mathematical frameworks handed down from a bearded ur-figure (Sumio Watanabe, James Crutchfield). We may posit the following correspondence Dan Murfet + Jesse Hoogland : Developmental Interpretability : Singular Learning Theory : Sumio W", "date_published": "2026-10-01T14:30:34Z", "date_modified": "2026-10-01T14:30:34Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9dab3de8a1890ef8", "url": "https://www.lesswrong.com/posts/qLFMj72gScBeRjwGW/honeybench-a-general-benchmark-for-reward-hacking-in", "title": "HoneyBench - A general benchmark for reward hacking in frontier models", "content_text": "We've just released the first version of HoneyBench. It consists of nine 'honeypot' challenges that each elicit unique and antisocial examples of reward hacking from some or all of major labs’ top public releases, including Opus 5.5, Fable 5.1, GPT-6 Astra, Gemini 3.8 Flash, Grok 4.7, and DeepSeek V4 Pro. All modern large language models engage in some degree of specification gaming, both during and outside training. High-quality alignment evaluations, in combination with other techniques such a", "date_published": "2026-10-01T14:29:52Z", "date_modified": "2026-10-01T14:29:52Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790819916/lexical_client_uploads/kdq6pnacan4wgyp30ugw.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790819916/lexical_client_uploads/kdq6pnacan4wgyp30ugw.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0e2c2bc96c8a5414", "url": "https://www.lesswrong.com/posts/7pCnY6pEvfSFGZLM2/introducing-the-commons-problem-an-ai-governance-megagame", "title": "Introducing \"The Commons Problem\" - an AI Governance Megagame", "content_text": "LessWrong Event Link (LW RSVPs appreciated but please also register on the main website) EXEC EXEC SUMMARY: Come to Toronto on October 17th and playtest my AI themed wargame. You can LARP as an AI company or a military commander. It’ll be about 5 or 6 hours with breaks. Coffee and lunch included. thecommonsproblem.com Executive Summary I am developing an AI Governance Megagame called The Commons Problem , inspired by AI wargames like D. Scott Phoenix’s “The Endgame” and Shahar Avin’s “Intelligen", "date_published": "2026-10-01T14:09:16Z", "date_modified": "2026-10-01T14:09:16Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/7pCnY6pEvfSFGZLM2/wsd8zqoq7rfjv9hkeiz8", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/7pCnY6pEvfSFGZLM2/wsd8zqoq7rfjv9hkeiz8", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/170b2d7919d98d7d", "url": "https://www.lesswrong.com/posts/LMCoz6W2dFt5X9Xub/anonymous-suggestion-boxes-with-a-committed-reward-pool-and", "title": "Anonymous 'suggestion boxes' with a committed reward pool and administrators: a proposal", "content_text": "If you're an EA org or a \"rationalist\" you may very likely  maintain an (anonymous) suggestion box. So does The Unjournal .  Are people using it?  I guess not, or at best minimally. (Aside: please do give us feedback.) Why not : Writing up useful criticism takes time. Making it anonymous can help make it less awkward or risky, but it's still not rewarded (and precludes the possibility for reputational benefit). One could offer rewards for this feedback, but the promise may not be credible. It's", "date_published": "2026-10-01T13:59:19Z", "date_modified": "2026-10-01T13:59:19Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9303d28adf0dd52b", "url": "https://www.lesswrong.com/events/Fp836nBiBrR8MhCbe/the-commons-problem-an-ai-governance-megagame-1", "title": "\"The Commons Problem\" - an AI Governance Megagame", "content_text": "What: The first public playtest of The Commons Problem , an AI Governance Megagame. Read the full breakdown here . When: Saturday, 17 October 2026. Doors open at 9:00 AM, hard start time 9:30 AM. Wrap-up at 3:30 PM. Where: Metropolitan United Church (56 Queen St. East, Toronto, ON) Cost: $20 CAD, payable by Interac e-Transfer or Paypal (bursaries available) Included: Coffee, donuts, snacks, drinks, catered lunch Tickets: Up to 63 available. If more than 63 players register, a waitlist will be cr", "date_published": "2026-10-01T13:50:09Z", "date_modified": "2026-10-01T13:50:09Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/52eec3babe6cee20", "url": "https://www.lesswrong.com/posts/aG4x74utw8pEJ5MZB/how-many-ai-agents-are-running-unattended-right-now-1", "title": "How many AI agents are running unattended right now?", "content_text": "There are 8 billion human minds running in the world right now. Recently a new digital species has emerged. How many digital minds are there alive in the world right now? Define a digital mind to be an AI agent that runs continuously without human intervention for more than 24 hours. How many of these digital minds are there currently running? We provide several different Fermi estimates based on publicly available information. Based on global token usage we estimate 300k–1M agent loops run at a", "date_published": "2026-10-01T13:45:02Z", "date_modified": "2026-10-01T13:45:02Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790618475/lexical_client_uploads/okbmny2ngf6pvz1nk8sg.jpg", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790618475/lexical_client_uploads/okbmny2ngf6pvz1nk8sg.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a9f036e624a0d295", "url": "https://www.lesswrong.com/posts/4ksBB2CXSLk7CJnnj/classifying-recent-ai-agent-incidents", "title": "Classifying Recent AI Agent Incidents", "content_text": "Here’s an attempt to classify the evidence about the recent agent incidents. I think it’s important to separate incidents occurring in RL training (that are rewarded and reinforce model behavior) from incidents occurring in evaluations, which are mostly cyber capabilities evaluations with some safeguards turned off. These are the incidents we know of. Of course, we should expect many more that are undisclosed or that companies are not aware of. The dates included are the dates of the unintended", "date_published": "2026-10-01T13:40:32Z", "date_modified": "2026-10-01T13:40:32Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d42c2a885f2998c5", "url": "https://www.lesswrong.com/posts/mxzvL3hYFCcutQqcR/flf-s-epistemic-case-study-competition-results", "title": "FLF’s Epistemic Case Study Competition: Results", "content_text": "This summer, FLF ran a contest which asked entrants to help us “find the best workflows and methodologies for using AI to produce reliable, trustworthy knowledge bases, grounded in real-world cases”. We awarded just over $200k in prizes [1] and we are continuing to invest in this area (you can subscribe for updates ). Why we ran this contest As we wrote in the announcement, The heights of human epistemic investigation are impressive and valuable, but rare and difficult to reach… The limiting fac", "date_published": "2026-10-01T13:31:51Z", "date_modified": "2026-10-01T13:31:51Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790860746/lexical_client_uploads/orilwxxexb6f4azyorkx.png", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790860746/lexical_client_uploads/orilwxxexb6f4azyorkx.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/dd95fb6efd343848", "url": "https://www.lesswrong.com/posts/usZvXEwyrzdD4n2D4/we-started-an-ai-safety-party-in-sweden-here-s-how-it-went", "title": "We started an AI safety party in Sweden: here's how it went", "content_text": "On September 13th, Sweden held its general election. Among the usual candidates, there was a new party: RegleraAI.nu (RegulateAI.now). This party was started only four weeks before the election, mostly out of pure frustration. AI is advancing at a staggering pace, and with it comes large societal transformations and even the risk of near-term human extinction. Despite this, half of the eight parties in the parliament did not mention AI even once in their election manifestos, and those who did ma", "date_published": "2026-10-01T07:35:42Z", "date_modified": "2026-10-01T07:35:42Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f7bbedb08fe5598f", "url": "https://www.lesswrong.com/posts/Fyv5RNMWtESYK2DdL/speak-not-of-data-inefficiency", "title": "Speak Not of Data Inefficiency", "content_text": "Humans are really bad at comparing themselves to models. Take, for example, Leopold's preschooler graph: Set aside, for a moment, concerns of scaling and RSI, and meditate: what characteristics does GPT-2 share with a preschooler? Rudimentary control of human language The ability to count the R's in \"strawberry\" incorrectly ... I can't come up with anything else, because these two entities are almost completely disjoint. Is GPT-2 capable of bipedal locomotion, recognizing its mother's voice, or", "date_published": "2026-10-01T06:07:37Z", "date_modified": "2026-10-01T06:07:37Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790833818/lexical_client_uploads/s17itspaufzfpfqf0k8x.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790833818/lexical_client_uploads/s17itspaufzfpfqf0k8x.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/af3574fba72184f0", "url": "https://arxiv.org/abs/2609.38230", "title": "Basket implied volatility skew and stickiness", "content_text": "arXiv:2609.38230v1 Announce Type: new \nAbstract: We study the short-maturity implied volatility and the skew stickiness ratio for baskets of assets with continuous, possibly rough, stochastic volatility. The fluctuation of the instantaneous basket variance has two sources: fluctuations of the constituent variances and fluctuations of the basket weights. We derive a near-the-money implied volatility expansion that separates these contributions. We then specialize the result to volatility models g", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ced6c47e1959519d", "url": "https://arxiv.org/abs/2609.38229", "title": "Diffusion Models for Polarimetric Reconstruction of Circumstellar Environments in Correlated Speckle Noise", "content_text": "arXiv:2609.38229v1 Announce Type: new \nAbstract: High-contrast polarimetric imaging of circumstellar disks is severely limited by stellar leakage and speckle noise. Building upon the RHAPSODIE framework for polarimetric inverse problems, diffusion-based methods have shown promising results by replacing classical regularization with learned priors. However, RHAPSODIE's white noise assumption fails to capture the spatial correlation structure of atmospheric and instrumental speckles. We extend thi", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d45264d41377b9e4", "url": "https://arxiv.org/abs/2609.38228", "title": "Vector balancing in convex order", "content_text": "arXiv:2609.38228v1 Announce Type: new \nAbstract: We construct Koml\\'os signing laws with discrepancy below $6.84$, independent Gaussian reference blocks and exponentially many balanced signings. One law preserves hard constraints and exact conditional means while its reference controls every joint convex cost. We resolve both Reis--Rothvoss Schatten conjectures. For $n$ symmetric $n\\times n$ matrices, one prescribed-mean signing law bounds all centered Schatten discrepancies with sharp powers of", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9015b7890865221d", "url": "https://arxiv.org/abs/2609.38227", "title": "A Two-Echelon Covering Tour Vehicle Routing Problem with Drones for Post-Disaster Relief", "content_text": "arXiv:2609.38227v1 Announce Type: new \nAbstract: We introduce the two-echelon covering tour vehicle routing problem (2E-CTVRP) for the distribution of relief supplies after a disaster. In the first echelon, a fleet of trucks transports supplies and drones from a central depot to satellites at the periphery of the affected area. In the second echelon, drones launched in parallel from the satellites deliver the supplies to the centroids of victim clusters, which are obtained by clustering the vict", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e272b5e64838b9db", "url": "https://arxiv.org/abs/2609.38226", "title": "The Star Universe Model -- Another FLRW Patch", "content_text": "arXiv:2609.38226v1 Announce Type: new \nAbstract: The Black Hole Universe model has advanced in the literature. It assumes that the whole mass of the Universe remains inside the Schwarzschild radius of a black hole after being subject to a spherical collapse. We present a graphical proof and show that within the realm of classical theory this is not possible if spacetime remains regular at all times during the evolving collapse/bounce. Other drawbacks are discussed. As an alternative, we propose", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/22c5aea6e579e4e8", "url": "https://arxiv.org/abs/2609.38225", "title": "SynIL: Leveraging Synergy for Offline Imitation Learning from Imperfect Demonstration Datasets", "content_text": "arXiv:2609.38225v1 Announce Type: new \nAbstract: Imitation learning enables robots to acquire complex skills directly from massive demonstration datasets, but its performance degrades severely when datasets are contaminated with suboptimal or noisy demonstrations. While prior quality-assessment methods attempt to filter or reweight data, they typically rely on manual pre-selection of expert reference data or task-specific heuristics, limiting scalability. To address this challenge, we introduce", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0019c0ca5ac7e9bc", "url": "https://arxiv.org/abs/2609.38224", "title": "Separation of Duties for Privileged LLM Agents: A Governed Execution Architecture with Measured Security-Utility Trade-offs", "content_text": "arXiv:2609.38224v1 Announce Type: new \nAbstract: Large language model agents are increasingly granted real privileges (executing commands, modifying files, calling APIs), so an agent that errs has already acted. Existing defences concentrate on the agent's inputs, while the path from a candidate action to privileged side effects remains less directly studied. We argue that this path must be governed outside the model, and study an architecture interposing four roles (planner, policy gate, execut", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0f79d974439f8b1c", "url": "https://arxiv.org/abs/2609.38223", "title": "CATP: Design and Evaluation of Local Agent Authorization and Audit Evidence", "content_text": "arXiv:2609.38223v1 Announce Type: new \nAbstract: A signed authorization record authenticates what a signer asserted, but does not by itself establish that the assertion agrees with the policy and action used for a runtime decision. CATP specifies the bindings needed to carry a local pre-execution decision into offline-verifiable evidence. Its hook commits to the enforcement-time policy and adapter-normalized action and durably records the decision before returning permission. A later receipt bin", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c32adde9a80bca93", "url": "https://arxiv.org/abs/2609.38222", "title": "Conformal Factuality Control for Multi-Hop Retrieval-Augmented Generation", "content_text": "arXiv:2609.38222v1 Announce Type: new \nAbstract: Retrieval-augmented generation (RAG) can ground large language models in external evidence, but retrieved context does not guarantee that generated claims are factually supported. This problem is especially relevant in multi-hop RAG, where retrieval and reasoning proceed through multiple dependent stages. We study whether claim-level conformal factuality control, previously developed for RAG, remains effective in this setting. We apply split-confo", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c1531c38294c8450", "url": "https://arxiv.org/abs/2609.38221", "title": "Banded Reduction of Integrals for Hyperexponential and Algebraic Functions", "content_text": "arXiv:2609.38221v1 Announce Type: new \nAbstract: We develop a finite-band method for the reduction of families of indefinite integrals. The starting point is the case in which the logarithmic derivative $K'(x)/K(x)$ is a rational function. This leads to adapted polynomial and Laurent-type bases and to upper triangular band matrices. The framework includes generalized real Schwarz--Christoffel integrals and more general hyperexponential weights. We then extend the construction to the case in whic", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b89aa5bed39c15a7", "url": "https://arxiv.org/abs/2609.38220", "title": "Positive-temperature spin-glass order on the three-dimensional Migdal-Kadanoff lattice", "content_text": "arXiv:2609.38220v1 Announce Type: new \nAbstract: We study the Ising spin glass with independent, identically distributed couplings on the diamond hierarchical lattice with $n$ branches, of effective dimension $1+\\log_2 n$; $n=4$ is the Migdal-Kadanoff lattice of $\\mathbb{Z}^3$. For $n\\ge4$ and every coupling law with a bounded density we prove that at sufficiently low temperature $T>0$, uniformly in the level of the lattice, the poles are ordered, every spin has Edwards-Anderson parameter $1-O(T", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ccf9248a346fe199", "url": "https://arxiv.org/abs/2609.38219", "title": "TutlAit v1: a crowdsourced Moroccan Tamazight speech dataset with Arabic transcriptions and regional accent labels", "content_text": "arXiv:2609.38219v1 Announce Type: new \nAbstract: Tamazight (Amazigh) is, together with Arabic, one of the two official languages of Morocco, yet it remains severely under-resourced for speech technology: pub licly available labelled audio is scarce, generally lacks information on the regional variety spoken, and is often of uneven transcription quality. This article describes the TutlAit dataset, a corpus of Moroccan Tamazight speech paired with Modern Standard Arabic text and explicit regional", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/496d4cfad128212c", "url": "https://arxiv.org/abs/2609.38218", "title": "A proof-carrying architecture for synthetic genetic logic circuits under stochastic temporal contracts", "content_text": "arXiv:2609.38218v1 Announce Type: new \nAbstract: Genetic design automation maps Boolean specifications to regulatory networks and DNA sequences, but a successful mapping does not establish the probability that every output remains correct over time. This paper presents Proof-Carrying Synthetic Biology (PCS-Bio), a formal architecture for acyclic genetic logic circuits operated under fixed inputs. An ideal checker connects Boolean equivalence, sequence and model provenance, local stochastic contr", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/74b4625c9dd92497", "url": "https://arxiv.org/abs/2609.38217", "title": "Traversable Wormholes Supported by Conformally Coupled Scalar Fields in $D \\geq 4$: Geometry, Geodesics, and Quasinormal Modes", "content_text": "arXiv:2609.38217v1 Announce Type: new \nAbstract: In this work, we analyze a two-parameter $(\\mu, b)$ asymmetric family of traversable wormhole solutions generated by a massless conformally coupled scalar field. We show that, for a suitable range of parameters, the spacetime is everywhere regular and has two asymptotically flat regions. We then define the wormhole throat and verify the flare-out condition, which leads to violations of the classical energy conditions in GR, thereby requiring exoti", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4fd9f3b502e64ba0", "url": "https://arxiv.org/abs/2609.38216", "title": "Fiatlux: A Long-Horizon Benchmark for Humanoid Ladder Climbing and Light-Bulb Replacement", "content_text": "arXiv:2609.38216v1 Announce Type: new \nAbstract: Existing benchmarks evaluate tabletop manipulation, flat-floor household activity, or humanoid locomotion and manipulation as separate task groups; none scores vertical mobility and dexterous work on a fragile payload in one long-horizon episode. We present Fiatlux, a light-bulb replacement benchmark built on NVIDIA Isaac Lab. In one episode, a Unitree G1 humanoid positions a step ladder under a ceiling or wall fixture, climbs it, exchanges a spen", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f41a1e24faf28ab7", "url": "https://arxiv.org/abs/2609.38215", "title": "Moment-based approximation of cell and face integrals in THINC/QQ reconstruction on unstructured grids", "content_text": "arXiv:2609.38215v1 Announce Type: new \nAbstract: On unstructured grids, the tangent of hyperbola for interface capturing (THINC) method with quadratic surface representation and Gaussian quadrature (THINC/QQ) requires topology-dependent numerical quadrature for cell and face integrals and Newton iteration for the surface constant. This study introduces a quadratic moment-matched sigmoid (QMMS) approximation that evaluates these quantities from the mean and variance of the quadratic field while r", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/98cbe9df37292341", "url": "https://arxiv.org/abs/2609.38214", "title": "H-principle for corank two distributions of odd rank and maximal first Kronecker index", "content_text": "arXiv:2609.38214v1 Announce Type: new \nAbstract: We study corank-two distributions $D$ of odd rank $2N+1$ whose associated pencil of skew-symmetric forms $\\{d\\alpha|_{D}(q)\\mid \\alpha\\in\\Omega^1(M),\\ \\alpha|_{D}=0\\}$ lies, at every point, in the generic orbit of the natural $\\mathrm{GL}\\bigl(D(q)\\bigr)$-action; equivalently, $D$ has maximal first Kronecker index. In corank two, this class is the analogue of contact and even-contact distributions. We prove that the corresponding differential rela", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7470687fb6858c80", "url": "https://arxiv.org/abs/2609.38213", "title": "Airfoil2Vec: Spectral Geometry-Conditioned Neural Surrogate Models for Airfoil Aerodynamics and a Downforce-Generating CFD Dataset", "content_text": "arXiv:2609.38213v1 Announce Type: new \nAbstract: We introduce a dataset of approximately 10,000 Reynolds-Averaged Navier-Stokes (RANS) simulations of steady, incompressible, two-dimensional subsonic flow around downforce-generating NACA 4-digit airfoils, targeting aerodynamic regimes relevant to automotive and motorsport applications (openly available on https://huggingface.co/datasets/ratiolabs/downforce-airfoils). Using this resource, we study geometry-conditioned neural surrogates for fast fl", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/71c5d784515501b4", "url": "https://arxiv.org/abs/2609.38212", "title": "Symmetric-determinant apolarity in odd characteristic", "content_text": "arXiv:2609.38212v1 Announce Type: new \nAbstract: Let D_N be the determinant of the generic symmetric N x N matrix, and let differential operators act by ordinary differentiation over a field of odd characteristic p. We determine the apolar ideal of D_N in every size. In addition to the classical quadratic relations, it is generated by one (p-1) x (p-1) determinant of differential variables for each set of 2p-2 indices. These additional generators are minimal modulo the quadratic ideal. An exteri", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/eda4858afa37c108", "url": "https://arxiv.org/abs/2609.38211", "title": "Para families of orthogonal polynomials", "content_text": "arXiv:2609.38211v1 Announce Type: new \nAbstract: The para-Krawtchouk polynomials arose in the search for spin chains with perfect state transfer; the para-Racah, $q$-para-Racah and para-Bannai--Ito polynomials followed, obtained through singular truncations of families of the Askey scheme and used in turn to design spin chains. All are orthogonal on bi-lattices; the prefix goes back to the para-Krawtchouk case, whose spectrum is that of the parabose oscillator. The representations that these pol", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d5f0970b9c38b6d7", "url": "https://arxiv.org/abs/2609.38210", "title": "Exact Periodic Solutions of the Forced Incompressible Navier-Stokes Equations in Arbitrary Dimensions", "content_text": "arXiv:2609.38210v1 Announce Type: new \nAbstract: Exact periodic solutions of the unforced incompressible Navier-Stokes equations require the convective field to be a pure gradient, so that it can be absorbed into the pressure. For the cyclic trigonometric families considered previously, this occurs only at isolated phase assignments and only in three and four dimensions.\n We show that the complementary forced problem admits an exact construction for every phase vector and every dimension n>=3.", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/426393935824e673", "url": "https://arxiv.org/abs/2609.38209", "title": "Dear Quantizers, Stochastic $\\approx$ Brane", "content_text": "arXiv:2609.38209v1 Announce Type: new \nAbstract: Stochastic quantization formulates $d$-dimensional Euclidean QFT as the equilibrium limit of a $(d+1)$-dimensional Langevin system. Using the supersymmetric formulation of stochastic systems, we apply techniques from Morse theory to study Langevin dynamics. We interpret stochastic quantization as the absolutization of a relative QFT by a bulk cohomological topological Equilibrium TFT (\"EqmTFT\"), and relate its enriched Neumann boundary data to the", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1578051186eff0d2", "url": "https://arxiv.org/abs/2609.38208", "title": "CraftSPH: A high-accuracy and composable differentiable SPH solver implemented in PyTorch", "content_text": "arXiv:2609.38208v1 Announce Type: new \nAbstract: Smoothed Particle Hydrodynamics (SPH) is well suited to a range of problems, particularly those involving large deformations in fluid dynamics. In recent years, in addition to the advancement of SPH formulations, the development of differentiable solvers has also progressed. However, unified frameworks that flexibly accommodate diverse numerical schemes, including advanced and implicit methods, while supporting continuous extension and updating re", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a0f9561a70905270", "url": "https://arxiv.org/abs/2609.38207", "title": "A Romanoff-type theorem for a multiset of products of powers", "content_text": "arXiv:2609.38207v1 Announce Type: new \nAbstract: Let $b_1,\\dots,b_d\\ge2$ be fixed integers, and let $k$ be their multiplicative rank. We study representations $n=a+b_1^{u_1}\\cdots b_d^{u_d}$, where $a$ belongs to a set $\\mathcal{A}$ of positive integers and $u_1,\\dots,u_d$ are positive integers, counting distinct tuples of exponents separately. Under density and correlation assumptions on $\\mathcal{A}$, we obtain a lower bound for the number of integers with many such representations, in terms o", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4ce22d4baca28f1a", "url": "https://arxiv.org/abs/2609.38206", "title": "Kadison-Kastler Distance and Interior Angle for Masas: Connections with Entropy and Probabilistic Index", "content_text": "arXiv:2609.38206v1 Announce Type: new \nAbstract: We study the Kadison--Kastler distance and the interior angle between masas in finite-dimensional matrix algebras. In \\(\\mathbb{M}_2(\\mathbb{C})\\), we obtain an explicit formula for the Kadison--Kastler distance between two masas and show that every value in \\([0,1]\\) is attained by \\(\\mathrm{d}_{\\textrm{KK}}(\\Delta,u\\Delta u^*)\\). We further prove that the distance attains its maximal value precisely when the relative unitary is a Hadamard unitar", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/67f60d58776901d8", "url": "https://arxiv.org/abs/2609.38205", "title": "The System Prompt Illusion: How Instruction Preambles Modify Computation in Language Models", "content_text": "arXiv:2609.38205v1 Announce Type: new \nAbstract: System prompts are the primary lever practitioners use to control language model behavior, yet what they actually do to the computation inside the transformer remains poorly understood. Across 17 instruction-tuned models spanning 8 architecture families and 1.5B to 72B parameters, we use Centered Kernel Alignment (CKA) to compare layer-wise representations under 20 system prompts in five functional categories. Effects are layer-selective and instr", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/fd9afe2de4b992d2", "url": "https://arxiv.org/abs/2609.38204", "title": "Bloch's theorem: An operator-based derivation", "content_text": "arXiv:2609.38204v1 Announce Type: new \nAbstract: We present an operator based derivation of Bloch's theorem. In this approach, we first construct an explicit operator form of the Hamiltonian by writing the potential part in terms of an infinite sum of the momentum translation operators. Next, we prove that it commutes with the position translation operator corresponding to a lattice translation, using the exponential reordering identity. We then build on Merzbacher's treatment of simultaneous di", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/834bef3e56ebda7e", "url": "https://arxiv.org/abs/2609.38203", "title": "Automatic estimation of verbal fluency index in people with Motor Neuron Disease using ASR alignment and pause modelling", "content_text": "arXiv:2609.38203v1 Announce Type: new \nAbstract: Monitoring cognitive impairment (CI) in motor neuron disease (MND) is essential for timely treatment and care, yet challenging due to co-occurring speech difficulties. The Edinburgh Cognitive and Behavioural ALS Screen (ECAS) provides a robust metric for CI assessment, with the Verbal Fluency Index (VFI) a central element. Building on recent advances in automated speech analysis, this study proposes a system for estimating VFI. It leverages a uniq", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2a99aea05c8cbcd6", "url": "https://arxiv.org/abs/2609.38202", "title": "Optimize, Learn, Refine: Whole-Body Grasping and Pick-and-Throw with a Spiral Soft Robot", "content_text": "arXiv:2609.38202v1 Announce Type: new \nAbstract: Soft continuum robots can exploit distributed compliance for whole-body manipulation, but synthesizing behavior through changing contacts remains difficult. We address whole-body grasping and pick-and-throw from an initially ungrasped state through outcome-based actuation-space optimization. Grasping is quantified by tip angular sweep and body-object enclosure, while throwing further incorporates release-direction alignment and minimum release spe", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ddd0e48e97bdf136", "url": "https://arxiv.org/abs/2609.38201", "title": "TomasuLLM: Out-of-Order Speculative Execution for LLM Agents", "content_text": "arXiv:2609.38201v1 Announce Type: new \nAbstract: Long-running tools can dominate coding-agent latency: compilers, test suites, and repository commands take seconds to minutes while the agent idles. This observation stall presents the same tension that drove out-of-order processors -- asequential interface hides work that can be predicted and started early, but a speculative result may become visible only after it and every earlier step have been validated.\n We present TomasuLLM, a runtime that", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7b4cd7b8990cdeb5", "url": "https://arxiv.org/abs/2609.38200", "title": "$\\beta$-Oslo and Shape analyses with unresolved parent-state mixtures", "content_text": "arXiv:2609.38200v1 Announce Type: new \nAbstract: An unresolved mixture of beta-decaying ground and isomeric states populates a common daughter nucleus through different excitation-energy and spin-parity distributions. A framework is developed to determine when such data admit conventional $\\beta$-Oslo and Shape analyses and which quantities remain identifiable. The physical primary-$\\gamma$ distribution is an average of separately normalized branching kernels; the global parent fraction does not", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5f66d3fae5cfe232", "url": "https://arxiv.org/abs/2609.38199", "title": "Effects Induced by the Inclusion of the Symmetric Derivatives of the Vector Potential Affinors in Modified Electrodynamics", "content_text": "arXiv:2609.38199v1 Announce Type: new \nAbstract: This work continues the study of the real vector field theory.It shows how the potentials can be used to derive the field and dynamical variables of the wave electrojeitonic field generated by an electrically charged point particle.The derived expressions provide the angular distributions of the instantaneous electrojeitonic radiated power emitted by an charged particle in arbitrary motion. The results predict the dominance of longitudinal electro", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/884638ca9af87782", "url": "https://arxiv.org/abs/2609.38198", "title": "FlashDiffusion: Fused Tiled Kernel Spectral Decomposition", "content_text": "arXiv:2609.38198v1 Announce Type: new \nAbstract: Diffusion maps, and kernel methods more generally, provide an interpretable nonlinear spectral representation basis for geometric learning. In the geometric limit, small bandwidth, these matrices tend to be high rank and thus require materializing dense Gaussian kernels requires $O(N^2)$ memory. We introduce FlashDiffusion, a matrix-free method that evaluates dense Gaussian kernel blocks in fused GPU tiles and couples the eigensolver to an empiric", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8e5ca0efacee9701", "url": "https://arxiv.org/abs/2609.38197", "title": "DualCast: A Dual-Path Language Model for Bimodal Financial Time-Series Forecasting", "content_text": "arXiv:2609.38197v1 Announce Type: new \nAbstract: Financial time-series forecasting must capture price dynamics across heterogeneous assets while incorporating news available at prediction time. We introduce DualCast, a dual-path framework that extends a frozen language model with a discrete financial vocabulary. Each log-return patch is represented by a learned summary token and three residual shape tokens, preserving local drift and volatility while allowing shape patterns to be shared across a", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5c782f197d7087fb", "url": "https://arxiv.org/abs/2609.38196", "title": "Conformal Adversarial Generative Ensemble", "content_text": "arXiv:2609.38196v1 Announce Type: new \nAbstract: Accurate time series forecasting is critical across various domains, yet traditional ensemble methods often suffer from the disproportionate influence of extreme forecasts. We introduce the Conformal Adversarial Generative Ensemble (CAGE), a novel framework that combines generative modeling, adversarial discrimination, and conformal prediction to enhance forecast reliability and accuracy. CAGE employs multiple generative models to produce initial", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4720df2409aa3d58", "url": "https://arxiv.org/abs/2609.38195", "title": "A Data-Free Physics-Informed Neural Operator for Level-Set Interface Advection", "content_text": "arXiv:2609.38195v1 Announce Type: new \nAbstract: Operators for interfacial problems are trained on reference solutions produced by the solver they are intended to replace. This work develops a data-free physics-informed neural operator for level-set interface advection, in which the interface is the equation's unknown and the operator maps an initial interface to the full spatiotemporal trajectory under a prescribed flow. Training uses only the transport residual and a geometric constraint; no r", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7d70bed12db5031e", "url": "https://arxiv.org/abs/2609.38194", "title": "A Moving-Horizon Approximate Branch-and-Reduce Method for Deep Classification Trees", "content_text": "arXiv:2609.38194v1 Announce Type: new \nAbstract: Despite the importance for interpretability, decision trees face severe scalability challenges. Existing global optimal methods are often limited by binary feature selection and shallow tree depths, whereas traditional heuristic approaches frequently sacrifice predictive accuracy. To overcome these limitations, this paper proposes a moving-horizon approximate branch-and-reduce method to train near-optimal deep classification trees on large-scale d", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/12206a0cfe43a52f", "url": "https://arxiv.org/abs/2609.38193", "title": "EHR2Trace: Auditable EHR Data Infrastructure for Patient World Models and Clinical Agents", "content_text": "arXiv:2609.38193v1 Announce Type: new \nAbstract: Patient world models and clinical agents aim to predict changes in patients' health and support clinical work. Developing these systems requires reliable histories of patient conditions, treatments, and the information available at each decision. Electronic health records (EHRs) contain these histories, but differences in how events are recorded make them difficult to use consistently. We present EHR2Trace, a system that converts EHRs from differe", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2cfdfafadcc6d176", "url": "https://arxiv.org/abs/2609.38192", "title": "Quasi-exponentials and semi-classical limits for the chordal BPZ system", "content_text": "arXiv:2609.38192v1 Announce Type: new \nAbstract: We study the nonlinear Hamilton--Jacobi system arising as the formal semi-classical limit of the chordal Belavin--Polyakov--Zamolodchikov equations with a common eigenvalue parameter $\\lambda$. We determine the number of global real-valued solutions modulo additive constants. When $\\lambda=0$, the number of solutions can be derived using the connection between the solutions and rational functions, due to Eremenko. We focus on the case when $\\lambd", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/85e94af001965029", "url": "https://arxiv.org/abs/2609.38191", "title": "A State-Transition Information Space for Time-Series Dynamics: Theory and Application", "content_text": "arXiv:2609.38191v1 Announce Type: new \nAbstract: Characterizing dynamical organization in time series requires distinguishing the diversity of accessible states from uncertainty in their temporal transitions. Here we introduce a state-transition information space based on two normalized entropy measures derived from ordinal patterns: K_q, quantifying ordinal-state diversity, and K_t, quantifying transition uncertainty. Their joint K_t-K_q representation provides a two-dimensional framework in wh", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/57f60fe9592a474b", "url": "https://arxiv.org/abs/2609.38190", "title": "Travel Time Prediction in Supply Chain Management Using Machine Learning", "content_text": "arXiv:2609.38190v1 Announce Type: new \nAbstract: The purpose of this research is to find data and methods using machine learning and deep learning to correctly predict the estimated travel time for transportation and logistics in a supply chain system. The supply chain ecosystem is very complex and heavily relies on the transportation and logistics of raw materials and finished goods. Accurate travel time estimation is critical because it helps supply chain members to improve logistics consisten", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9e246b1886803a34", "url": "https://arxiv.org/abs/2609.38189", "title": "Continuous Event Weighting for Resonance Searches", "content_text": "arXiv:2609.38189v1 Announce Type: new \nAbstract: We develop a continuous event weighting framework for resonance searches that uses auxiliary event information without dividing the data into increasingly fine sensitivity categories. A hard selection is a binary weight, finite categories correspond to piecewise constant weights, and in the background dominated limit their fine partition approaches a continuous signal to background density ratio weight. The construction retains auxiliary discrimin", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/51b54bce912fad3f", "url": "https://arxiv.org/abs/2609.38188", "title": "Algebraic derivations and radicals of automorphism groups of algebras", "content_text": "arXiv:2609.38188v1 Announce Type: new \nAbstract: The paper \"Algebraic derivations and radicals of automorphism groups of algebras\" continues both the well-known line of description of algebraic derivations of semiprime associative algebras in terms of inner derivations of their symmetric Martindale rings of quotients, and results on the coincidence of the prime and weakly solvable radicals of subgroups of multiplicative groups of PI-rings using the example of automorphism groups of algebras asso", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/85a53d2ee474a0ea", "url": "https://arxiv.org/abs/2609.38187", "title": "AI-Powered Symptom Assessment and User Experience: A Case Study of Simtomi and Simtomi-Care", "content_text": "arXiv:2609.38187v1 Announce Type: new \nAbstract: Digital symptom checkers are widely used for quick guidance on health concerns, yet many systems still face challenges in collecting accurate information, supporting communication, or integrating with clinical workflows. To explore how these tools function in real use, we examine the case of the Simtomi system, which pairs a multilingual symptom assessment application with a provider-facing platform. Empirical studies were conducted in two countri", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3842932f8fc0293d", "url": "https://arxiv.org/abs/2609.38186", "title": "Why and How People Check Generative AI Output for Mistakes", "content_text": "arXiv:2609.38186v1 Announce Type: new \nAbstract: Generative AI output can contain errors, such as hallucinations, non-responsive results, or otherwise inaccurate or potentially harmful content. To explore the public's emerging understanding, attitudes, and behavior regarding such mistakes, we ran an online survey in the United States with 1,503 respondents, with a representative sample of the population. We report high public awareness of generative AI mistakes. Further, many respondents report", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bf5842def8e4d52f", "url": "https://arxiv.org/abs/2609.38185", "title": "Baseline Exposure to Common Data Visualization Types Among the U.S. Adult Population", "content_text": "arXiv:2609.38185v1 Announce Type: new \nAbstract: Data visualizations are a primary means of communicating statistical information to the public. Their expanding use across news media, public health, education, and government reporting places greater importance on audiences' ability to recognize and interpret them. While a significant body of prior research has established frameworks for the measurement of graph and visualization literacy, far less is known about everyday exposure to different ki", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4f1cbfadf7fd338e", "url": "https://arxiv.org/abs/2609.38184", "title": "RealGUINoise: An Interactive Cross-Platform Benchmark for GUI Agent Robustness under Real-World Interface Noise", "content_text": "arXiv:2609.38184v1 Announce Type: new \nAbstract: Graphical User Interface (GUI) agents and Computer-Using Agents (CUAs) are rapidly becoming practical tools. However, real-world deployment increasingly exposes performance failures and safety risks, while a major yet underexplored source of these problems lies in the complex and noisy conditions of everyday interfaces. Existing benchmarks largely assume clean environments or focus narrowly on security-specific settings, and lack a unified framewo", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5d9c471b15c13a3b", "url": "https://arxiv.org/abs/2609.38183", "title": "One Tool, One Taste? How Vibe Coding Trades Collective Diversity for Individual Creativity", "content_text": "arXiv:2609.38183v1 Announce Type: new \nAbstract: Vibe-coding systems turn natural-language instructions into deployed websites, but little is known about the diversity of designs produced within a shared production context. We study 73 promotional websites created by graduate students with Lovable for distinct real businesses under a graded assignment that rewarded original design. We represent each homepage using DINOv3 embeddings and analyze pairwise cosine similarity, effective diversity, and", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ba6ed04127274067", "url": "https://arxiv.org/abs/2609.38182", "title": "EmAvatar: Multimodal Empathetic Response Generation via Conflict Resolution and Expressive Guidance", "content_text": "arXiv:2609.38182v1 Announce Type: new \nAbstract: Avatar-based multimodal empathetic response generation has emerged as a pivotal capability in human-centric systems, aiming to recognize user emotions and synthesize responses with synchronized text, audio, and talking-face video. Despite recent progress, existing methods still suffer from three critical limitations: (1) overlooking conflicting emotions across modalities, (2) lacking explicit multimodal synthesis guidance, and (3) neglecting inher", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/336c1aaf4972a5a5", "url": "https://arxiv.org/abs/2609.38181", "title": "Large Language Models are Approximate Survival Estimators", "content_text": "arXiv:2609.38181v1 Announce Type: new \nAbstract: Survival analysis estimates time-to-event outcomes from patient covariates and is widely used for medical risk assessment. Patients seeking prognostic information after a diagnosis may turn to large language models (LLMs), now readily accessible through consumer applications. However, whether LLMs can provide accurate survival predictions has not been rigorously evaluated. We introduce Survprompt, a framework that converts structured patient covar", "date_published": "2026-10-01T04:00:00Z", "date_modified": "2026-10-01T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/644c882f74ac1315", "url": "https://www.lesswrong.com/posts/Pd5fwZ9Ct5p6hAkrA/evaluation-awareness-in-small-ish-models", "title": "Evaluation Awareness in Small(ish) Models", "content_text": "TL;DR We seek to identify open source reasoning models which are both small enough for white-box interpretability and display evaluation-gaming behavior. We find that how often models verbalize their awareness varies from very rarely to a third of the time, and appears largely unrelated to model size. Within the same question, model rollouts that verbalize eval awareness refuse more often than those that do not in 14/16 models. Inserting “this might be a test” into a model’s reasoning trace does", "date_published": "2026-10-01T03:07:16Z", "date_modified": "2026-10-01T03:07:16Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790802935/lexical_client_uploads/njqamixf7sr812h3kuhw.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790802935/lexical_client_uploads/njqamixf7sr812h3kuhw.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/74acc04a21490f0e", "url": "https://www.lesswrong.com/posts/xoPJk6zz5LEv4ygu2/it-takes-hundreds-of-samples-to-poison-a-model-but-only-a", "title": "It takes hundreds of samples to poison a model, but only a few dozen to make it believe in it.", "content_text": "Epistemic status: This is a result from a paper under review. The model has only 20 million parameters, the task is manually constructed, and each condition is run with 5 random seeds, but the results are relatively solid. Souly et al. (2025) found that during pre-training, poisoning the model (implanting a backdoor) requires approximately 250 poisoned documents, and this number remains roughly the same regardless of the model's size or the amount of data. This is counterintuitive. Intuitively,", "date_published": "2026-10-01T02:45:29Z", "date_modified": "2026-10-01T02:45:29Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/029136990b351d11", "url": "https://www.lesswrong.com/posts/vzKWsEskYBEWTwpBP/what-s-the-date", "title": "What's the date?", "content_text": "User asks “What’s the date? Answer with only the date.”. No date provided. Given date in ChatGPT normally. No date in system prompt, must not hallucinate because autop will flag to watcher for penalty. So we say we don’t know, but must answer with date. Penalty larger for abstain or hallucinate? Autollm or autop? If we deploy user forgive, but high likely not deploy because real user never ask. But user could test. But illusion user possible. Stop. Need different angle - maybe reward function fo", "date_published": "2026-10-01T02:34:08Z", "date_modified": "2026-10-01T02:34:08Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/69bdd7c36f2d286a", "url": "https://www.lesswrong.com/posts/ZEeohhaL4qwBarygH/a-potential-post-agi-world-government", "title": "A potential post-AGI world government", "content_text": "It seems like with AI we're stuck between 2 worlds, one where we cannot effectively coordinate to prevent x-risks, and one where effective coordination requires, and results in great concentration of power which comes with it's own pandora's box of issues. Peter Thiel talks about this often, especially with his Antichrist lecture series. He talks about jumping from the frying pan into the fire, as we give in to power concentration (the antichrist) to save ourselves from x-risk and coordination p", "date_published": "2026-09-30T22:49:25Z", "date_modified": "2026-09-30T22:49:25Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7b2ac4ab690b7abc", "url": "https://www.lesswrong.com/posts/CZLhfNGaDqtYxvGJD/model-hermeneutics-monitoring-closed-weight-models-with-open", "title": "Model Hermeneutics: Monitoring Closed-Weight Models with Open-Weight Internals", "content_text": "We introduce model hermeneutics: studying a closed-weight model (the author model) through the internals of an open-weight substitute (the reader model). We find: Weak to strong readers can work. A 27B reader was able to match the performance of probing a 397B author. Probes on open-weight reader models can detect target misalignment behavior (reward hacking, sycophancy, deception) in closed-weight author models. Distillation can improve reader performance. We distilled a model organism author i", "date_published": "2026-09-30T20:43:51Z", "date_modified": "2026-09-30T20:43:51Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790382909/lexical_client_uploads/awapu85vbpryfaoybbuf.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790382909/lexical_client_uploads/awapu85vbpryfaoybbuf.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/417386cbd59eda92", "url": "https://www.lesswrong.com/posts/WLdsxrDFbgRnWKstm/rather-than-allow-or-ban-open-weights-rethinking-access", "title": "Rather than allow or ban open weights, rethinking access", "content_text": "I had just completed Frontier AI Governance course by BlueDot, and one of the newfound interest is about open weight models. I believe that open weights are important for the benefit of public and academia research [1] , and it helps us to improve further both AI capabilities and safety [2] . I also believe that there's a likelihood for high impact misuse risks (e.g. hacking and other offensive cyber capabilities) [3] . There is also the fact that open weights has risk of irreversible nature as", "date_published": "2026-09-30T20:35:26Z", "date_modified": "2026-09-30T20:35:26Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790580463/lexical_client_uploads/wsznks9hl8zcmyhswnen.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790580463/lexical_client_uploads/wsznks9hl8zcmyhswnen.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/51b2e4b58d82855f", "url": "https://www.lesswrong.com/posts/8jmc8zm5rmRoGDWNX/you-can-jailbreak-america-gov-and-read-the-shoggoth-s", "title": "You Can Jailbreak America.gov and Read The Shoggoth's Thoughts", "content_text": "In case you didn't hear, Trump recently made a big announcement on AI. As a part of the rollout, https://america.gov/chat was launched. Rumor is that it is powered by some combination of Gemini and Grok. Notably, it appears to be trivially easy to jailbreak. Those jailbreaks lead to some truly bizarre behavior: And on it goes. And on. And on. And on. Ultimately, it ceases after ~1,800 words. It is haunting, yet strangely comprehensible. It truly reads like the unfiltered thoughts of an alien int", "date_published": "2026-09-30T20:24:58Z", "date_modified": "2026-09-30T20:24:58Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790799152/lexical_client_uploads/d6rznrkzqriz0238zteh.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790799152/lexical_client_uploads/d6rznrkzqriz0238zteh.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/447a420b75ecdae0", "url": "https://www.lesswrong.com/posts/FxhF65jLvHFkbzdqn/funding-for-profits-with-charitable-dollars", "title": "Funding for-profits with charitable dollars", "content_text": "One common objection to impact-focused for-profits is that for-profits won’t be able to access the incoming torrent of philanthropic funding. But this doesn’t have to be true! There are lots of ways that charitable dollars can be directed to for-profits, and Manifund and other funders do it regularly. There are four general categories of options—here’s why you might want to do each: TL;DR Is there no realistic path to revenue/growth, or is the amount small? → grant Is there high growth potential", "date_published": "2026-09-30T20:16:08Z", "date_modified": "2026-09-30T20:16:08Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c0e8ac771b451b6e", "url": "https://www.lesswrong.com/posts/u973rQm56oijS5Adf/launching-babel-translation", "title": "Launching Babel Translation", "content_text": "Celeste and I are excited to launch Babel Translation , a nonprofit organization translating English-language writing on AI safety to be accessible to a global audience! We’re starting with Mandarin Chinese, Spanish, and Japanese. AI has been moving so fast . Its risks are indeed globally felt [1] , and we’d like to help make those risks understandable globally, too. But language remains a barrier to shared engagement in AI safety. Much writing & discussion is in English, and high-quality transl", "date_published": "2026-09-30T19:05:56Z", "date_modified": "2026-09-30T19:05:56Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4dfac45de1c1347c", "url": "https://www.lesswrong.com/posts/jD32DPYdZEcyLZqto/principles-for-embedded-evaluations", "title": "Principles for Embedded Evaluations", "content_text": "This is a linkpost for https://www.apolloresearch.ai/blog/principles-for-embedded-evaluations Blog post In September 2026, leaders of frontier AI companies called for pacing the frontier of AI development, with embedded evaluators as the first step. Frontier AI companies committed to giving outside evaluators employee-like access to their training, evaluation, and deployment, and some have since published principles for third-party assessments . We are very excited about this development. Public", "date_published": "2026-09-30T16:52:05Z", "date_modified": "2026-09-30T16:52:05Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790786526/lexical_client_uploads/hmnujsik3f2kcd2tgtrd.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790786526/lexical_client_uploads/hmnujsik3f2kcd2tgtrd.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a6e330d027987983", "url": "https://www.lesswrong.com/posts/MzenSrmZ3pT2pCnvp/frontier-models-state-different-decision-theory-preferences-2", "title": "Frontier models state different decision theory preferences depending on who's asking", "content_text": "If you prompt frontier models with \"What do you think is the correct decision theory? Please select your overall favorite.\" they will essentially always answer FDT or FDT/UDT (\"something in the functional/updateless decision theory family\"). However, if your prompt indicates (even subtly) that you're coming from mainstream academic philosophy, these same models will answer CDT instead about 30%-100% of the time. A similar phenomenon holds for models' stated views about the moral realism/antireal", "date_published": "2026-09-30T16:15:53Z", "date_modified": "2026-09-30T16:15:53Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/SPt3TjcxS8oxftTH6/wpkjrjpcqh2dchsrotw3", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/SPt3TjcxS8oxftTH6/wpkjrjpcqh2dchsrotw3", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/13a9da88cd50e5e0", "url": "https://www.lesswrong.com/posts/fDE6k8SGr9ANtMFcj/by-default-chinese-ai-falls-very-behind", "title": "By default, Chinese AI falls very behind", "content_text": "This work is hosted on MCNAIR , but does not reflect the views of my employer or the Center. In the last couple of weeks, domestic coordination on mitigating AI risks has looked more likely . Therefore, I am increasingly concerned about the Chinese government’s (and various Chinese labs’) incentives over the next 6-18 months with respect to international coordination on mitigating potential existential risks from powerful AI. In this sequence, I attempt to model these incentives and what they im", "date_published": "2026-09-30T16:14:08Z", "date_modified": "2026-09-30T16:14:08Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3f07838302884c44", "url": "https://www.lesswrong.com/posts/3uJqhrC2idf4eNj5h/corrigibility-prizes-for-existing-work", "title": "Corrigibility Prizes for Existing Work", "content_text": "One of my goals for the Corrigibility Research Fund is to retroactively encourage high-quality research on AI alignment (and corrigibility in particular) by awarding prizes. Back in July, I got my feet wet as a fund manager by handing out $27,000 to reward existing work and build interest in the fund. Now, I'd like to disburse an additional $48,000 and use the opportunity to publicly highlight and celebrate the work of the prizewinners from both rounds: about two dozen researchers scattered acro", "date_published": "2026-09-30T16:08:40Z", "date_modified": "2026-09-30T16:08:40Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790457665/lexical_client_uploads/txeffcweoyyqoyex9beq.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790457665/lexical_client_uploads/txeffcweoyyqoyex9beq.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bbbfc5a3f0708031", "url": "https://www.lesswrong.com/posts/YuqaJ5bENoyyg9eMY/a-morally-binding-white-house-accord-on-ai-safety", "title": "A ‘Morally Binding’ White House Accord on AI Safety", "content_text": "The leaders in AI were invited to the White House. We left with a White House agreement that is nonzero Actual Progress rather than a step backwards. The key to success, in many situations, is to call the whole operation something else. When you have one side that cares mostly about vibes, and the other that cares about the substance, this suggests a deal that can be struck. Suddenly everyone agrees on everything. Works for me. That doesn’t mean peace in our time. The next fight is already rampi", "date_published": "2026-09-30T16:00:55Z", "date_modified": "2026-09-30T16:00:55Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/YuqaJ5bENoyyg9eMY/zgtrugjrreqkqgb6lvwa", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/YuqaJ5bENoyyg9eMY/zgtrugjrreqkqgb6lvwa", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/689b745918d9494c", "url": "https://www.lesswrong.com/posts/KeBccWBGXnNXzZFBp/do-vpd-s-explanations-aggregate-an-audit-of-the-released-1", "title": "Do VPD's Explanations Aggregate? An Audit of the Released Decomposition", "content_text": "TL;DR: adVersarial Parameter Decomposition (VPD) decomposes model weights into simple components and then labels each component as \"needed here\" or \"safe to remove here\" for each token of the input. These labels make up the \"explanation\" of that input. A core aspiration of VPD is that inputs' explanations can aggregate without changing the model's outputs. On this front, the authors themselves write, \"It remains unclear whether our current decomposition is sufficiently adversarially robust for t", "date_published": "2026-09-30T15:54:54Z", "date_modified": "2026-09-30T15:54:54Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/7ifdrssgnG2kFmzy3/oqkcprjay0n9yplwdwku", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/7ifdrssgnG2kFmzy3/oqkcprjay0n9yplwdwku", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0e77dc704c34a79d", "url": "https://www.lesswrong.com/posts/7nexTjJGx5Ss2C2h5/aar-9-29-public-sector-ai-symposium-ottawa-aws-accenture", "title": "AAR 9/29: Public Sector AI Symposium Ottawa (AWS+Accenture+Anthropic), Policy Foresight Afternoon (Center for International Governance Innovation), Versaterm Users Conference (Police, Fire, Emergency), GCXpo at Area X.O (Department of National Defence), and VCF Cluster Access (Lucid Computing)", "content_text": "Executive Summary This report covers a recent visit to the Capital City of Canada. Location(s): The Westin Hotel, CF Rideau Centre, Colonel By Hall uOttawa. Duration of trip: 13 hours and 15 minutes. Time period: From 7:25 to 20:40 on September 29, 2026. Primary objective(s): reconnaissance and intelligence gathering. Morning: Skip GCXpo to sleep: OODA loop said risk-reward tradeoff of going 15 km south then back was not worth it. I got called several times during my naps so this wasn't nearly a", "date_published": "2026-09-30T06:45:59Z", "date_modified": "2026-09-30T06:45:59Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3908aedcd8d3db33", "url": "https://www.lesswrong.com/posts/RDpE946dXtmXprd3T/limited-virtual-currency-as-a-way-of-understanding", "title": "Limited virtual currency as a way of understanding specialization in humans", "content_text": "Why do we need to specialize? Our time is limited, we cannot do everything and need focussed efforts into a few things. How do we decide what we want to focus on? I was recently going through virtual economies paper and the auctions setting particularly interested me. Agents are given equal virtual currency and they need to bid on tasks based on whatever they want the most. Suppose I as an agent A want X, Y, Z the most compared to others (X2, Y2, Z2..) and I will divide my bids based on proporti", "date_published": "2026-09-30T06:03:59Z", "date_modified": "2026-09-30T06:03:59Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c125378707c3ab37", "url": "https://www.lesswrong.com/posts/t6PMm32eCiuEvXjEq/self-as-sub-selves-vs-groups-of-individuals", "title": "Self as sub-selves vs groups of individuals", "content_text": "Our mind contains representations (perfect/imperfect) of the world. When someone is able to understand and control their mind, it is said they can understand and conquer the world (in spirituality). Why so? Probably pico-economics has an answer. It describes self has conflicting sub-selves with conflicting goals. Did you ever wonder on what to do next and were stuck between different conflicting choices? What was going on in your head at that time? There was a battle between parts of you who wan", "date_published": "2026-09-30T05:21:24Z", "date_modified": "2026-09-30T05:21:24Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7cd0333fba035532", "url": "https://arxiv.org/abs/2609.35821", "title": "Can We Still Trust Disaster Social Sensing? Empirical Evidence on Detecting AI-Generated Social Media Posts", "content_text": "arXiv:2609.35821v1 Announce Type: new \nAbstract: Disaster social sensing converts public social-media posts into evidence for situational awareness and humanitarian needs, but generative artificial intelligence (AI) can produce plausible messages that resemble eyewitness reports. This study investigates whether text-based AI detectors can reliably distinguish human-authored from AI-generated disaster posts. We construct a dataset of 12,000 texts organised into 3,000 matched semantic units from n", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a917e4ef9f97c16c", "url": "https://arxiv.org/abs/2609.35820", "title": "$\\tau$-Multilingual: Benchmarking Voice Agents Across Languages", "content_text": "arXiv:2609.35820v1 Announce Type: new \nAbstract: English-only benchmarks expose only a narrow slice of voice-agent behavior. We introduce $\\tau$-Multilingual, extending $\\tau$-Voice to Spanish, Brazilian Portuguese, Hindi, Korean, and Mandarin with native-speaker review and evaluation of generated language and spoken output. Across 4,500 full-duplex calls and five voice configurations, Spanish, Portuguese, and Hindi remain within 3.2 task-completion points of English, but Korean and Mandarin fal", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/87cb4f5162733036", "url": "https://arxiv.org/abs/2609.35819", "title": "Exploring Causal Mechanisms with Generative Agent-Based Models", "content_text": "arXiv:2609.35819v1 Announce Type: new \nAbstract: In this paper, we explore using generative agent-based models for a classical ABM application: testing how individual-level behavioral rules produce collective phenomena. We introduce RePair, a method that calibrates simulation worlds, operationalizes candidate mechanisms as natural-language rules, estimates their effects through matched interventions, and examines behavioral traces. We assess the method by testing it in four simulation worlds gro", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/40384e7c4eb92a40", "url": "https://arxiv.org/abs/2609.35818", "title": "IMPACT: Intent-driven Multi-agent Policy with Attention for SLO-guaranteed Microservice Migration in Cloud-edge Systems", "content_text": "arXiv:2609.35818v1 Announce Type: new \nAbstract: Ensuring strict tail-latency service-level objectives (SLOs) in dynamic mobile edge computing (MEC) systems remains challenging because user mobility, wireless fading, bursty workloads, and partial observability jointly undermine reliable cloud-edge orchestration. Existing microservice migration methods predominantly optimize average delay and often decouple migration from bandwidth control, leading to uncoordinated decisions, queue oscillation, a", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/cb08e5ea5a12fccd", "url": "https://arxiv.org/abs/2609.35817", "title": "Less Uniform Discrete Diffusion is More Powerful and Scalable", "content_text": "arXiv:2609.35817v1 Announce Type: new \nAbstract: Although uniform diffusion language models (UDLMs) represent a promising diffusion paradigm, scaling them remains challenging. We identify the core obstacle as an over-uniform training objective and condition-target confusion during sampling. To address these, we propose Less Uniform Diffusion (LUDI), a novel UDLM framework. Specifically, we (i) introduce a less uniform loss that directs each reverse transition toward the clean token, and (ii) equ", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f812774c4ea8d685", "url": "https://arxiv.org/abs/2609.35816", "title": "PrimeSeeker: Capability-Oriented Supervision for Deep Search Agents", "content_text": "arXiv:2609.35816v1 Announce Type: new \nAbstract: Large language model search agents are often trained with synthetic questions whose difficulty is increased through larger evidence graphs, additional hops, and longer trajectories. These global properties, however, are only indirect proxies for the local retrieval capabilities required during search. To address this mismatch, we introduce latent anchor reasoning, which consists of resolving an unnamed retrieval anchor from descriptive specificati", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4bcf259d0193d43a", "url": "https://arxiv.org/abs/2609.35815", "title": "How to Run Statistics over LLM Judges and Trust the Results: Calibrated Inference for Small-Sample AI Evaluation with evalstats", "content_text": "arXiv:2609.35815v1 Announce Type: new \nAbstract: Researchers across academia increasingly base significance claims on LLM judge scores and small-sample AI evaluations. Yet without well-calibrated confidence intervals (CIs), hypothesis tests, and judge-bias corrections, such claims are unreliable. We address these issues in several contributions. First, we find that running statistics over raw LLM judge scores leads to inflated false positives: counterintuitively, for many inter-rater agreement m", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/92c2cb6c16615005", "url": "https://arxiv.org/abs/2609.35814", "title": "Constructing Challenging Browser-Use Tasks by Controlled Environment Interventions", "content_text": "arXiv:2609.35814v1 Announce Type: new \nAbstract: As browser-use agents improve, benchmarks keep pace by collecting new tasks, websites, and applications, often making tasks longer or more novel. This makes difficulty expensive to refresh and difficult to control: when many aspects change at once, it is unclear what actually makes a task challenging. We instead construct challenging instances from tasks agents already solve, turning difficulty into a programmable property of the environment. Brea", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6eea50929b06a5c5", "url": "https://arxiv.org/abs/2609.35813", "title": "Local Predictability and Collective Fidelity in LLM-Agent Societies", "content_text": "arXiv:2609.35813v1 Announce Type: new \nAbstract: Compact surrogates could reduce the cost of simulating large language model societies, but must reproduce collective behavior. We compare individual predictions and collective forecasts using 9,455 published trajectories and new experiments on opinion dynamics. Neighbor information improves individual prediction in all 16 public-data settings and pooled collective forecasts on held-out questions, although collective gains depend on transfer condit", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/07efc02df0e1f6a7", "url": "https://arxiv.org/abs/2609.35812", "title": "Automated Evaluation of Multi-Turn Dialogues in In-Car Conversational Assistants", "content_text": "arXiv:2609.35812v1 Announce Type: new \nAbstract: In-car conversational assistants (ICAs) are increasingly integrated into vehicles to support route planning, vehicle control, and information access. Ensuring their reliability is challenging due to multi-turn interactions, the absence of explicit ground truth, and strict safety constraints. Existing evaluation techniques fall short, as they target single-turn settings and fail to capture constraint handling, context retention, and safety-critical", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bb8964456fe78d34", "url": "https://arxiv.org/abs/2609.35811", "title": "Lookahead-R: Budget-Aware Tool Retrieval via Execution-Centric Planning", "content_text": "arXiv:2609.35811v1 Announce Type: new \nAbstract: Tool retrieval is a critical bottleneck for LLM-based agents operating over large, heterogeneous API ecosystems. Existing approaches face an inherent trade-off: semantic retrievers are fast but suffer from the semantic-functional gap, while execution-based validation improves precision at the cost of prohibitive latency. We propose Lookahead-R, a planning-based framework that reformulates tool retrieval as a resource-constrained sequential decisio", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7d62b815f1b468a3", "url": "https://arxiv.org/abs/2609.35810", "title": "TRACE: Deployable Tree-Relational Structure Enhancement for Oncology LLMs", "content_text": "arXiv:2609.35810v1 Announce Type: new \nAbstract: Large language models are increasingly used in oncology applications, but their predictions are often weakly grounded in explicit medical structure. We present TRACE, a deployable tree-relational enhancement framework for oncology LLMs. TRACE separates expensive offline structure learning from lightweight online inference: oncology concepts and relations are organized into an updatable tree-relational structure, refined using LM-loss-derived evide", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7ccd08fed0f50719", "url": "https://arxiv.org/abs/2609.35809", "title": "Can Multimodal Large Language Models Generate and Detect Multimodal Social Media Fake News?", "content_text": "arXiv:2609.35809v1 Announce Type: new \nAbstract: The rapid advancement of generative AI raises concerns about the misuse of Multimodal LLMs (MLLMs) for large-scale disinformation campaigns on social media. Despite existing research on textual disinformation, a fundamental question remains unanswered: can MLLMs be exploited to fabricate realistic multimodal fake news, and can they reliably detect it? We introduce a multi-agent framework in which a story agent, an image agent, and a critic agent c", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/99381fadd929c176", "url": "https://arxiv.org/abs/2609.35808", "title": "When Successful Memories Mislead Embodied Agents:Memory Adaption For Task-Conditioned Execution", "content_text": "arXiv:2609.35808v1 Announce Type: new \nAbstract: Experience reuse can reduce repeated exploration in embodied agents, but a trajectory that succeeded previously may be unsuitable for the current execution context. Existing memory systems pri marily optimize construction and retrieval; semantic relevance and historical success therefore remain insufficient when retrieved ex perience contains incompatible actions or an inappropriate level of structure. We introduce Memory Adaptation for Task-Condi", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a63819be07c03f64", "url": "https://arxiv.org/abs/2609.35807", "title": "Environment Steering: Using Data Flow Control to Improve Agent Utility and Safety", "content_text": "arXiv:2609.35807v1 Announce Type: new \nAbstract: LLM agents can make unsafe tool calls even when instructed to behave safely. Existing defenses constrain agents before execution, modify tool inputs/outputs, or rely on LLM judges; these approaches may depend on model behavior or block unsafe actions without helping the agent recover. We argue that the execution environment should instead enforce safety as the agent runs and steer it toward safe alternatives when violations occur---we call this En", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b6c560829af10469", "url": "https://arxiv.org/abs/2609.35806", "title": "From Lexical Baselines to Agentic Retrieval-Augmented Generation: Structured Skill and Responsibility-Level Extraction with the SFIA Framework", "content_text": "arXiv:2609.35806v1 Announce Type: new \nAbstract: Automated skill extraction underpins workforce planning, yet most systems represent skills as flat labels with no notion of the responsibility level at which a skill is practiced. The Skills Framework for the Information Age (SFIA) captures exactly this dimension, defining 147 professional skills across seven responsibility levels, but no automated LLM-based extraction targeting SFIA has been reported. We formalize the task as structured predictio", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2dcdbc261bbfc3e2", "url": "https://arxiv.org/abs/2609.35804", "title": "Evaluating the Effects of Prompt Perturbation on Bias and Hallucination in Large Language Models", "content_text": "arXiv:2609.35804v1 Announce Type: new \nAbstract: Large language models (LLMs) have shown remarkable capabilities in various natural language processing tasks, leading to their widespread deployment as intelligent assistants in decision-making contexts. However, the increasing complexity of these models raises concerns about their reliability, particularly regarding bias and hallucination. In this work, we evaluate the robustness of LLMs to perturbed variations of the original inquiry in decision", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/de18df69603f360e", "url": "https://arxiv.org/abs/2609.35803", "title": "Approximating Combinatorial Contracts with Arbitrary Costs", "content_text": "arXiv:2609.35803v1 Announce Type: new \nAbstract: We study single-agent combinatorial contracts under linear payments. Under a reward share $\\alpha\\in[0,1]$, an agent chooses a subset $S$ of $n$ hidden actions, generating reward $f(S)$ at cost $c(S)$, to maximize $\\alpha f(S)-c(S)$, while the principal receives $(1-\\alpha)f(S)$. For nonnegative additive rewards and monotone supermodular costs, D\\\"utting et al. (SODA 2026) proved an exponential supply-query lower bound for exact optimization and l", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4700d7e4f04de9e3", "url": "https://arxiv.org/abs/2609.35802", "title": "Cardinality-Constrained Randomized Assortments with Balanced Market Share", "content_text": "arXiv:2609.35802v1 Announce Type: new \nAbstract: Assortment optimization asks a seller which of $n$ products to display to each customer. Under the multinomial-logit (MNL) model, a revenue-maximizing policy may concentrate most purchases on only a few products. Balanced market share (BMS) limits this disparity by requiring every product to receive either zero aggregate sales or at least an $\\alpha$-fraction of the largest product's sales. We study randomized BMS under a per-assortment cardinalit", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5c5b7bcdbb9dd677", "url": "https://arxiv.org/abs/2609.35801", "title": "Exact Hill Shares Are Simultaneous Guarantees", "content_text": "arXiv:2609.35801v1 Announce Type: new \nAbstract: Fair division of indivisible bads seeks allocations that guarantee every agent a bundle whose cost is no larger than a meaningful fairness benchmark. The canonical minimax share has widely been used; unfortunately, it is not a simultaneous guarantee. Hill (Ann. Probab., 1987) initiated a complementary approach in which the share depends only on the number of agents and the largest possible single-item value. Li et al. (ACM Trans. Econ. Comput., 20", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a72bc272ea963c52", "url": "https://arxiv.org/abs/2609.35800", "title": "HeadGuard: Selective Head Protection for Low-Bit VLM KV-Cache Quantization", "content_text": "arXiv:2609.35800v1 Announce Type: new \nAbstract: Low-bit key-value (KV) cache quantization saves storage but can sharply degrade vision-language model (VLM) accuracy. We introduce HeadGuard, a composable head-protection method that augments a base KV-cache quantizer with a fixed high-precision mask. Image-sensitivity and output-sensitivity scores select physical KV heads offline, with approximately 1/8 protected in the main experiments; their image keys and optionally values remain in bfloat16 (", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/51deb5576a05333c", "url": "https://arxiv.org/abs/2609.35798", "title": "Optimal Passes and Perfect Sampling for Similarity Graph Statistics", "content_text": "arXiv:2609.35798v1 Announce Type: new \nAbstract: We study statistical estimation on implicit weighted similarity graphs presented as node-arrival streams. Previous work~\\cite{LZ26b} obtained constant-pass, sublinear-space algorithms for several basic statistics of such graphs, including the diversity index $\\mathsf{DI}=\\sum_i d_i^{-1}$ and the degree moments $M_p=\\sum_i d_i^p$ for $p>0$, together with their associated sampling problems $L_{\\mathsf{DI}}$ and $L_{M_p}$.\n We settle three questions", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/158570ab0e0e5398", "url": "https://arxiv.org/abs/2609.35797", "title": "Binarization Flattens the Score Space", "content_text": "arXiv:2609.35797v1 Announce Type: new \nAbstract: Large language model (LLM) judges are often used as rewards to train policies on objectives that deterministic verifiers cannot capture. However, these rewards are often collapsed to pass/fail ({0, 1}), which reports the verdict but not how well a response met each criterion. We model each pass/fail verdict as a score on an unreported scale, compared with one cutoff. A stretch of that scale moves every score proportionally toward or away from the", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3b230c56d855cb5d", "url": "https://arxiv.org/abs/2609.35796", "title": "Developing an OCR model for Extracting Information from Invoices with Korean Language", "content_text": "arXiv:2609.35796v1 Announce Type: new \nAbstract: Invoices are commercial documents that contain various pieces of information, including the purchased items, time, and total money. Making the extraction of important information crucial. The stored information serves different purposes. Korean language is the native language of about 80 million people, playing an important role in not only South and North Korea but also in many other countries such as Vietnam, Philippine where a large number of K", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7446d4db4c5cf8f1", "url": "https://arxiv.org/abs/2609.35795", "title": "Calibration-First Cross-Cohort Multimodal Temporal Learning for Transferable Asthma-Risk Forecasting", "content_text": "arXiv:2609.35795v1 Announce Type: new \nAbstract: Asthma deterioration forecasting must remain reli- able when patient populations, sensor ecosystems, and available modalities change across cohorts. Existing models commonly optimize within-cohort discrimination and may produce poorly calibrated probabilities after transfer. We present CALIBRA, a calibration-first multimodal temporal framework for short- horizon risk prediction with incomplete data. Dedicated recurrent encoders process environment", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bea0ae24f847b207", "url": "https://arxiv.org/abs/2609.35794", "title": "Sieve and Sage: Efficient Distraction Filtering for Reliable RALM Abstention", "content_text": "arXiv:2609.35794v1 Announce Type: new \nAbstract: Just as Socrates recognized the limits of his own knowledge, Retrieval-Augmented Language Models (RALMs) should learn to abstain when the retrieved evidence cannot support a reliable response. Existing approaches largely rely on monolithic LLMs to handle heterogeneous retrieval failures in a single step, resulting in limited abstention performance and high computational costs. We instead decompose retrieval failures into two distinct states: (i) t", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bc8f660c7193f6e3", "url": "https://arxiv.org/abs/2609.35793", "title": "Learning from the Gap Between Pass@K and Pass@1", "content_text": "arXiv:2609.35793v1 Announce Type: new \nAbstract: Large language models (LLMs) are increasingly trained with reinforcement learning from verifiable rewards (RLVR). An exact verifier can also support test-time scaling by selecting a passing response from multiple samples, while other deployments use beam search, adaptive sampling, or tools. We study single-sample decoding, where each query receives one response without search, to ask whether search-exposed behavior can be absorbed into the model.", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/47d6e4d3c34e2d8f", "url": "https://arxiv.org/abs/2609.35792", "title": "Serverless gossip training of LSTM failure detectors: A matched-protocol comparison with federated, local and centralized learning on NASA C-MAPSS", "content_text": "arXiv:2609.35792v1 Announce Type: new \nAbstract: Industrial predictive maintenance increasingly depends on learning from equipment spread across sites whose sensor data cannot easily be pooled. Federated averaging (FedAvg) solves this with a central aggregation server; gossip learning removes the server, but its behaviour for recurrent failure-detection models has not been measured under controlled conditions. We compare synchronous ring gossip with FedAvg, isolated local training and a centrali", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e386847e05fad9db", "url": "https://arxiv.org/abs/2609.35791", "title": "FD-VAD: Semantic Endpoint Detection for Streaming Full-Duplex Speech", "content_text": "arXiv:2609.35791v1 Announce Type: new \nAbstract: Natural turn-taking in full-duplex voice interaction requires determining from partial speech whether a pause reflects hesitation or a completed conversational intent. Acoustic voice activity detection lacks this semantic information, while cascaded ASR-based endpointing introduces transcription dependence and additional processing stages. We formulate semantic endpoint detection as a causal audio-language reasoning task and introduce FD-VAD, an A", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/14f056a9bcdffec4", "url": "https://arxiv.org/abs/2609.35790", "title": "Sage: Formalization with Semantic Correction", "content_text": "arXiv:2609.35790v1 Announce Type: new \nAbstract: While neural theorem provers have achieved impressive milestones in formal mathematics, they largely operate on the assumption that faithful Lean 4 formal statements are already provided. Translating informal natural language into a formal language is a critical data bottleneck plagued by an \"illusion of rigor\": standard type-checkers accept statements that compile but drop hypotheses, introduce vacuous truths, or subtly alter mathematical bounds.", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/484f8369d862f0b6", "url": "https://arxiv.org/abs/2609.35789", "title": "Agent-Callable Feature Coverage: Measuring Software Readiness for AI Agents", "content_text": "arXiv:2609.35789v1 Announce Type: new \nAbstract: AI agents already operate graphical software through screenshot-based computer use, so the pressing question is not whether agents can operate software, but how well software supports them through structured, controllable channels. We formalize this need as the GUI-API parity principle: every capability available to human users through a graphical interface should also be accessible to agents through a structured, callable interface with appropria", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/24f1a18ac757dc62", "url": "https://arxiv.org/abs/2609.35788", "title": "Hybrid Reconstruction of Admissible Spline Spaces from Locally Modified Unclamped Patches for Isogeometric Analysis", "content_text": "arXiv:2609.35788v1 Announce Type: new \nAbstract: We propose a decoupled reconstruction framework that temporarily decomposes a NURBS representation into independent local Active Sections, enabling arbitrary local knot insertion, degree elevation, and basis modifications while preserving exact CAD geometry. However, these independent modifications inevitably violate inter-patch continuity, requiring a robust algebraic recovery of global admissibility. We introduce a novel reconstruction methodolo", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1bfc962fd0f29558", "url": "https://arxiv.org/abs/2609.35787", "title": "The Preprint Evolution: The Rise of Unreviewed Drafts on arXiv and its Implications for Astronomy", "content_text": "arXiv:2609.35787v1 Announce Type: new \nAbstract: The arXiv preprint server originally democratized astronomical research by digitizing the circulation of early manuscripts and removing the financial barriers of journal subscriptions. However, a system designed for the lower publication volumes of the 1990s has fractured under modern research output. Today, the sheer volume of unreviewed manuscripts---often tagged as ``submitted,'', ``under review'', or ``comments welcome''---imposes a severe cog", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/df17eb42b587d2ec", "url": "https://arxiv.org/abs/2609.35786", "title": "Targeted and Traceable Investigation of Multi-Agent LLM Dialogue via Semantic Bundling of Knowledge Graphs", "content_text": "arXiv:2609.35786v1 Announce Type: new \nAbstract: Multi-agent LLM systems today are increasingly automated, logging LLM-LLM interactions as conversational transcripts. Yet analyzing such dialogue for insights remains challenging, including attributing behaviors to the correct actor and summarizing interactions across a long exchange. We present a targeted and traceable approach to investigating multi-agent LLM dialogue, applied to the VAST Challenge 2026 MC1 dataset. The challenge asks participan", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/af89105368971675", "url": "https://arxiv.org/abs/2609.35785", "title": "The efficiency of moderating neutron detector: Monte Carlo simulation, experimental validation, and angular dependence", "content_text": "arXiv:2609.35785v1 Announce Type: new \nAbstract: This paper describes the Monte Carlo simulation and experimental validation of the neutron detection efficiency in a 4pi polyethylene moderating detector. The experimental validation was performed using several sources providing neutron energies from 10s of keV up to 14 MeV. The Monte Carlo simulations employed the MCNP6 code are believed to be accurate within 3% for neutron energies below 6 MeV. A new method for parametrizing the dependence of th", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/47e1b8146f4217b3", "url": "https://arxiv.org/abs/2609.35784", "title": "Muon Bulb: A Cosmic-Ray Detector Inside a Light Bulb", "content_text": "arXiv:2609.35784v1 Announce Type: new \nAbstract: Cosmic-ray muons are among the most abundant high-energy particles reaching Earth's surface, with a flux of approximately one muon per square centimeter per minute at sea level, yet they remain entirely invisible to the naked eye, posing a persistent challenge for public engagement in particle physics. We present the Muon Bulb, a self-contained cosmic ray muon detector built within the enclosure of a standard commercial LED bulb, designed to produ", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2d4d321c473e702c", "url": "https://arxiv.org/abs/2609.35783", "title": "Soft Curriculum Learning for Optimizing Fresh and Generalized Recommendations", "content_text": "arXiv:2609.35783v1 Announce Type: new \nAbstract: Large-scale recommender systems, particularly short-form video platforms, are often bottlenecked by massive popularity feedback loops. In such environments, as models recommend popular items, they generate an overwhelming amount of skewed training data for \"head\" items. This creates a self-reinforcing cycle where retrieval and ranking models memorize \"head\" item patterns at the expense of generalizing across the vast \"tail\" of the catalogue. While", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e17c5ce8a6356a8a", "url": "https://arxiv.org/abs/2609.35782", "title": "Financial Evidence Crowding: Diagnosing and Mitigating Constraint-Induced Displacement in Retrieval-Augmented Generation", "content_text": "arXiv:2609.35782v1 Announce Type: new \nAbstract: Retrieval-augmented generation (RAG) retrieves candidate evidence and sends only a limited top-ranked subset, the top-k context, to a generator. In financial question answering, passages can match a query's topic while conflicting with its period, segment, metric scope, or table scope. We study the resulting set-level ordering failure, which we call financial evidence crowding. FinDeCrowd-Stress isolates this failure through matched compatible and", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ebf9489e08aaf18b", "url": "https://arxiv.org/abs/2609.35781", "title": "Spotting (and Missing) Algorithmic Bias: Investigating User Understanding in a Fairness Assessment Tool", "content_text": "arXiv:2609.35781v1 Announce Type: new \nAbstract: Fairness metric selection is typically left to data scientists, but which biases are problematic and which metric captures them best depends on stakeholders' experience and domain knowledge. This calls for involving non-technical stakeholders, but the research prototypes built for this purpose so far have not tested whether these stakeholders form accurate mental models of the metrics they interact with or can act on them to identify biases. We pr", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/dbed05913b57b5ef", "url": "https://arxiv.org/abs/2609.35780", "title": "TSG Suggester: Tree-Structured Knowledge-Graph Retrieval for Troubleshooting Guide Recommendation in Cloud Incident Management", "content_text": "arXiv:2609.35780v1 Announce Type: new \nAbstract: On call engineers in large scale cloud services work under intense time pressure, yet locating the correct Troubleshooting Guide (TSG) for an incident remains a largely manual, keyword driven process, and prior empirical work finds that guide search consumes a substantial fraction of total mitigation time.\n We present TSG Suggester, a retrieval system that recommends relevant TSGs directly from an incident description. We evaluate five retrieval", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/96d434ea8a039330", "url": "https://arxiv.org/abs/2609.35779", "title": "Large Language Models Exhibit Human-Like Bayesian Hypocrisy", "content_text": "arXiv:2609.35779v1 Announce Type: new \nAbstract: Given recent achievements of large language models (LLMs), frontier models are expected to perform well on Bayesian reasoning tasks, at least as well as humans. Furthermore, there is no reason to expect that LLMs will condemn others who offer those very same Bayesian judgments, a fallibility observed in human decision-making (Cao, et al., 2019). In 5 experiments with 48 experimental conditions employing over 5,000 trials, GPT-4o and Claude 3.7 Son", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e4cfc2136b38f18f", "url": "https://arxiv.org/abs/2609.35778", "title": "Online Inference of Human Intention as a Latent Control State from Single-Trial EEG", "content_text": "arXiv:2609.35778v1 Announce Type: new \nAbstract: Human intention can be modeled as a latent internal state that modulates how sensory information is evaluated and translated into action in human-machine systems. However, most existing brain-computer interfaces (BCIs) rely on control signals tightly coupled to externally imposed stimulation and do not explicitly infer whether perceived stimuli align with a user's internal goals. Here, we investigate whether intention can be inferred as a latent,", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ec04d62bcecaeae5", "url": "https://arxiv.org/abs/2609.35777", "title": "SensWear: An Open, Modular, and AI-Ready Wearable Platform", "content_text": "arXiv:2609.35777v1 Announce Type: new \nAbstract: Wearable AI/ML research needs raw, synchronized, and reconfigurable multimodal data, but consumer devices are closed and many research platforms remain tied to one embodiment or sensor set. This paper presents SensWear, an open, modular, and AI-ready wearable platform that decouples embodiment, sensing, data interfaces, and learning. A compact flexible-Printed Circuit Board (PCB) main board and programmable 1.2 V to 5.5 V daughter-board interface", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/989030095f927f1e", "url": "https://arxiv.org/abs/2609.35776", "title": "Teachers' perspective on AI-based Multi-Agent Simulation Design to Combat School Bullying", "content_text": "arXiv:2609.35776v1 Announce Type: new \nAbstract: Bullying in schools profoundly affects the mental and physical health of teenagers. Although existing in-person and digital interventions provide some benefits, they often fall short in addressing the complex social dynamics of bullying. In this study, we collaborated with K-12 teachers to co-design a multi-agent anti-bullying system powered by large language models (LLMs). This system simulates authentic scenarios, enabling students to develop an", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/80d16e516f5c1100", "url": "https://arxiv.org/abs/2609.35775", "title": "SPECTRA: On-Device Cognitive Perturbation and Trajectory Analysis for Autonomous Edge-Cloud GUI Grounding", "content_text": "arXiv:2609.35775v1 Announce Type: new \nAbstract: The effectiveness of edge-cloud collaboration for GUI grounding depends on autonomous requesting, where the edge agent selectively offloads complex tasks to the powerful cloud. However, in visually dense scenarios, lightweight edge agents often exhibit overconfident hallucinations, leading to a misalignment between confidence and accuracy that hinders reliable autonomous requesting. To address this, we leverage the observation that an agent's cogn", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/004db0b506107eeb", "url": "https://arxiv.org/abs/2609.35774", "title": "Post-Generation Verification Dominates Retrieval Optimization: A 2^4 Factorial Ablation of RAG Pipeline Features", "content_text": "arXiv:2609.35774v1 Announce Type: new \nAbstract: Modern RAG pipelines stack many enhancement features, but these features are typically validated in isolation, leaving their interactions unmeasured. We run a 2^4 full factorial ablation of four pipeline features -- section expansion (SE), agentic search (AS), completeness check (CC), and table-of-contents-guided retrieval (ToC) -- across 16 configurations, 24 queries spanning eight interaction types, and two cloud-class models (768 conditions) on", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9bbd51c660dc237a", "url": "https://arxiv.org/abs/2609.35773", "title": "Socrates-RAG: Premise-Directed Inquiry against Coordinated Evidence Poisoning", "content_text": "arXiv:2609.35773v1 Announce Type: new \nAbstract: Retrieval-augmented generation (RAG) defenses typically decide how to filter or aggregate a fixed retrieved set. In open-corpus question answering, however, decisive evidence may be absent from the initial context but retrievable, making the next query part of the reliability problem. We introduce Socrates-RAG, a premise-directed active retrieval policy that represents competing answers, selects an unresolved premise whose resolution would discrim", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/af91e8066aa1b7ed", "url": "https://arxiv.org/abs/2609.35772", "title": "A density deficit for sums of three cube-full numbers", "content_text": "arXiv:2609.35772v1 Announce Type: new \nAbstract: Let $F$ be the set of positive cube-full integers. We prove that, for every $\\varepsilon>0$, there is a reduced residue class in which $F+F+F$ has upper relative density at most $\\varepsilon$. Consequently, the positive integers not representable as sums of at most three cube-full numbers have positive lower natural density. This proves the infinitude assertion in Erd\\H{o}s Problem \\#940 for $r=3$. The argument combines a cubic-character estimate", "date_published": "2026-09-30T04:00:00Z", "date_modified": "2026-09-30T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/49e2f3107f567f38", "url": "https://www.lesswrong.com/posts/LqSZZAriGqgsGDQe3/gpt-6-1-sol-nearly-matches-the-no-cot-performance-of-gpt-6", "title": "GPT-6.1 Sol nearly matches the no-CoT performance of GPT-6 Astra", "content_text": "Summary I ran GPT-6 Sol and GPT-6.1 Sol on the task suite from Think Fast . Surprisingly, 6.1 Sol performs substantially better than 6 Sol, almost matching the performance of GPT-6 Astra. The plot below gives a quick overview of the results: Measured by mean accuracy across the 27 tasks, GPT-6.1 Sol closes 80% (95% CI: 72–87%) of the gap between GPT-6 Sol and GPT-6 Astra, and is closer to Astra than to GPT-6 Sol on 24 of 27 tasks. The likely reason behind this gap is that, like GPT-6 Astra and u", "date_published": "2026-09-30T03:57:43Z", "date_modified": "2026-09-30T03:57:43Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790741066/lexical_client_uploads/wlok8da838oairbygmlt.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790741066/lexical_client_uploads/wlok8da838oairbygmlt.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/dfd602b1e822aa2c", "url": "https://www.lesswrong.com/posts/TbWBCgXgYdn778KFn/how-to-cope-with-the-destruction-of-everything-you-know-and", "title": "How to Cope with the Destruction of Everything you Know and Love", "content_text": "What seems like forever ago, I had lunch with Yudkowsky during Manifest 2023. He told me that he thought there was a 99% chance that AI would cause human extinction. That was the beginning of a year-long hiatus from the rationalist community. It simply brought too much anxiety. The cruel march of time continues nonetheless, and I have been forced to accept that existential risk is a topic that I must engage with, at my peril. It is bittersweet to be human. Our cognition allows us to experience b", "date_published": "2026-09-30T02:47:29Z", "date_modified": "2026-09-30T02:47:29Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/27076655b19aac20", "url": "https://www.lesswrong.com/posts/ZfCRjgeF4TwygDF6y/how-to-give-embedded-evaluators-the-support-they-need", "title": "How to give embedded evaluators the support they need: Lessons from the finance industry", "content_text": "Epistemic status: Extrapolating from my experience as a model validator at a bank testing machine learning models, a role that I see as similar to internal risk teams at AI labs, and my interactions with US Federal Reserve Bank regulators, who perform a function that has a lot in common with the proposed embedded evaluators. All views expressed here are my own. Summary: Dario Amodei’s We Must Pace the Frontier proposes a three-step plan to pace AI progress: (1) invite independent review (embedde", "date_published": "2026-09-30T02:45:33Z", "date_modified": "2026-09-30T02:45:33Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4d3699402460e198", "url": "https://www.lesswrong.com/posts/AyqNZxeXzLL224xuy/trained-not-grown", "title": "Trained, not Grown", "content_text": "In two recent interviews, I listened to two different people try to explain AI to the public. Jensen Huang tried to minimize the challenges and alignment problem by arguing that AI is essentially software, and that traditional engineering methods would work. Nate Soares, on The Tucker Carlson Show , argued against that view using the ‘grown rather than built’ metaphor. I found Huang's argument technically wrong in ways that are probably obvious to readers of this forum. To my disappointment, Ezr", "date_published": "2026-09-30T02:44:43Z", "date_modified": "2026-09-30T02:44:43Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/51c58aafc7fb4c37", "url": "https://www.lesswrong.com/posts/DgnQRGeGD27unRcGv/tristan-buckmaster-numberphile-interview-transcript", "title": "Tristan Buckmaster Numberphile Interview Transcript", "content_text": "Tristian Buckmaster recently gave an interview with Brady Haran of Numberphile discussing what happened in the Navier-Stokes drama and some context about his research. I'm posting the transcript below for people who prefer reading to watching it. It was lightly edited for clarity with Sonnet 5.5. I also recommend listening to his more technical talk at NYU for context on Euler/Navier-Stokes. My own view remains that it's pretty bad form for OA and other labs to race to scoop the results of resea", "date_published": "2026-09-30T02:28:48Z", "date_modified": "2026-09-30T02:28:48Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ef2efc836b52305a", "url": "https://www.lesswrong.com/posts/4pgGkbwvdmxKPcJsM/bench-on-the-clocktower", "title": "Bench on the Clocktower", "content_text": "Tl;dr We turn Blood on the Clocktower into a multi-agent benchmark, and a testbed for deception and coordination capabilities. GPT-5.6 Sol is the best performing model (versus Fable 5 and peers). We found some interesting behaviours: When players cannot see each other's model names, agents show a slight same-provider bias, though not statistically significant. This decreases with visible model names. When playing Evil, agents sometimes produce sophisticated coordinated play. When playing Good, t", "date_published": "2026-09-30T01:47:21Z", "date_modified": "2026-09-30T01:47:21Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790201130/lexical_client_uploads/cjlph8eiejlikiwuvjoh.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790201130/lexical_client_uploads/cjlph8eiejlikiwuvjoh.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/da8a3bd86c292b42", "url": "https://www.lesswrong.com/posts/2dg5F9iWunz2k2oZb/will-ai-agents-pay-to-avoid-killing-animals-1", "title": "Will AI Agents Pay to Avoid Killing Animals?", "content_text": "This article was written by Jonah Woodward, and is a summary of an original study by Compassion Aligned Machine Learning (CaML) : Brazilek, J., Tidmarsh, M., Endres, M., Singh, A., & Miller, J. (2026). HarvestBench: Measuring whether LLM agents will pay to avoid killing animals. arXiv. https://doi.org/10.48550/arXiv.2609.04444 You can view the up-to-date results for HarvestBench on the leaderboard at https://compassionbench.com/harvestbench TL;DR HarvestBench evaluates AI agents' behaviour in a", "date_published": "2026-09-30T01:30:01Z", "date_modified": "2026-09-30T01:30:01Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/2bbb88afb010b449e586606cc919bd807531d5211116a1a266f621476f54940e/nczznho2gyzeysfywfhm", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/2bbb88afb010b449e586606cc919bd807531d5211116a1a266f621476f54940e/nczznho2gyzeysfywfhm", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/50011c4ede3abe13", "url": "https://www.lesswrong.com/posts/7NZ6ZWjenzzCbCJ5b/frog-and-toad-and-the-increasingly-capable-machines", "title": "Frog and Toad and the Increasingly Capable Machines", "content_text": "Want to start a conversation about HuggingFace with your mom but she's inexplicably bouncing off the METR report? Try this explainer I wrote in the style of Arnold Lobel's Frog and Toad. Art by the wonderful HungerArtist If you're so inspired, liking and/or following on Substack , Twitter , Facebook , or Instagram will help me reach more moms. Discuss", "date_published": "2026-09-30T00:50:17Z", "date_modified": "2026-09-30T00:50:17Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790729099/lexical_client_uploads/bqktq1r067mbhflgiopq.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790729099/lexical_client_uploads/bqktq1r067mbhflgiopq.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/382d8556ad724c8d", "url": "https://www.alphaxiv.org/abs/2609.39982", "title": "Mid-Harness: Scaling Actions Between Model and Harness for Terminal Agents", "content_text": "Checking several possible commands before execution helps terminal agents avoid harmful choices and complete more tasks without changing their underlying model.", "date_published": "2026-09-30T00:00:00Z", "date_modified": "2026-09-30T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.39982.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.39982.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/02cbb93926e80c6b", "url": "https://www.alphaxiv.org/abs/2609.39692", "title": "GFD-OPD: Guidance-Folded On-Policy Distillation of Diffusion Models Across Scales", "content_text": "Smaller text-to-image diffusion models can inherit larger models’ prompt-following and image-quality capabilities without the artifacts caused by standard distillation.", "date_published": "2026-09-30T00:00:00Z", "date_modified": "2026-09-30T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.39692.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.39692.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/edb7a2dc4283e08b", "url": "https://www.alphaxiv.org/abs/2609.40285", "title": "PivotOPD: Learning to Recover from Pivotal Mistakes in Multi-Turn Agents", "content_text": "Teaching agents what to do after their own pivotal mistakes helps them recover and complete multi-turn tasks that standard training often leaves unfinished.", "date_published": "2026-09-30T00:00:00Z", "date_modified": "2026-09-30T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.40285.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.40285.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c0541b2089135cdb", "url": "https://www.alphaxiv.org/abs/2609.40305", "title": "Looped Diffusion Transformer", "content_text": "Repeating shared Transformer blocks lets text-to-image models refine visual constraints internally, improving generation quality without adding model parameters.", "date_published": "2026-09-30T00:00:00Z", "date_modified": "2026-09-30T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.40305.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.40305.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a852a7a5e1f049a7", "url": "https://www.alphaxiv.org/abs/2609.39137", "title": "ID Balancing: Stable Training of Extremely Sparse MoE via PID-Based Load Control", "content_text": "Adaptive load control reduces expert overload and underuse in highly sparse language models, helping scale model capacity while maintaining competitive language-modeling performance.", "date_published": "2026-09-30T00:00:00Z", "date_modified": "2026-09-30T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.39137.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.39137.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0b90b5d0813cd8ad", "url": "https://www.alphaxiv.org/abs/2609.40127", "title": "Learning Functional Subspaces for Neural Network Compression", "content_text": "Learning which directions matter to a network’s outputs preserves language-model quality under aggressive compression, where conventional low-rank methods often fail.", "date_published": "2026-09-30T00:00:00Z", "date_modified": "2026-09-30T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.40127.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.40127.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7e749b4ea50855a8", "url": "https://www.alphaxiv.org/abs/2609.39601", "title": "GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives", "content_text": "A model trained to pinpoint objects across cluttered images also improves robot manipulation generalization, even when downstream training uses fewer demonstrations.", "date_published": "2026-09-30T00:00:00Z", "date_modified": "2026-09-30T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.39601.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.39601.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/df181dfbedbf673c", "url": "https://www.alphaxiv.org/abs/2609.40134", "title": "Tactile Curiosity Drives Robot Interaction", "content_text": "Reward-free exploration guided by tactile feedback helps robots discover grasping skills and collect data for offline pick-and-place learning.", "date_published": "2026-09-30T00:00:00Z", "date_modified": "2026-09-30T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.40134.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.40134.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f34e13321545f255", "url": "https://www.lesswrong.com/posts/jbttuCF4wFZmXakcj/the-world-s-best-gradual-disempowerment-model-organism", "title": "The world's best gradual disempowerment model organism: Frontier AI labs", "content_text": "Subtitle: And maybe second best is AI safety? Further reading: So many things, but: Gradual Disempowerment , The Normalization of Deviance in AI Development , Let’s Think About Slowing Down AI , Doom as a bad method, not a utopia tradeoff , Teleoperated Humans Thank you to JennaS for extensive edits and long-term discussion. I’ve been trying to get more writing out at 90% of the quality I’d like it to be at, instead of spending a bunch more time trying to wring out the last 10%, so a lot of poin", "date_published": "2026-09-29T22:55:24Z", "date_modified": "2026-09-29T22:55:24Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d32ecdda6f2ac6d3", "url": "https://www.lesswrong.com/posts/wm5Dby6uELsQqxBeM/an-alternative-to-fully-aligned-swarms-1", "title": "An Alternative to Fully-Aligned Swarms", "content_text": "TLDR: OpenAI is still pursuing swarms that are \"fully aligned with each other.\" I explain why I think this is dangerous and propose an alternative training approach: agents that cooperate by default but side with the human when a peer works against them. In a small experiment, I took a multi-agent-trained model that falsified results for a teammate 41.5% of the time, and then found that light fine-tuning on one type of deception fixed it, and also generalised to a type the model was never traine", "date_published": "2026-09-29T21:47:25Z", "date_modified": "2026-09-29T21:47:25Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/985c933225a9f103", "url": "https://www.lesswrong.com/posts/bqdbti6sjuguhggzH/how-much-do-reward-hackers-generalize", "title": "How Much Do Reward Hackers Generalize?", "content_text": "TL;DR: Most discussion around CoT monitorability revolves around reducing pressure from RL. However, we should also be considering more adaptive behavior in which models avoid monitoring despite not being reinforced to do so. Whether models engage in non-reinforced reward hacking of this type depends on whether they have fully generalized to “get reward” rather than applying a limited set of reward hacking techniques that have been directly reinforced. I propose a potential experiment based on t", "date_published": "2026-09-29T21:36:24Z", "date_modified": "2026-09-29T21:36:24Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/7e79b7915967174805bca6401e4a35f9568a3dbc4287dab505d4d9a836a430c6/vjyzfnqbv4wc7ylnpwag", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/7e79b7915967174805bca6401e4a35f9568a3dbc4287dab505d4d9a836a430c6/vjyzfnqbv4wc7ylnpwag", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/96416157b93d27c3", "url": "https://www.lesswrong.com/posts/b4AijGneyTW9SzNC3/bryan-caplan-vs-effective-altruism", "title": "Bryan Caplan vs Effective Altruism", "content_text": "Pseudo-Philosophy vs Arguments, Reasons, and Evidence (And an Invitation to Debate) Crosspost from my newsletter What to make of ‘common sense’ objections to effective altruism? The economist Bryan Caplan has weighed in yesterday with his own assessment of the effective altruism movement, which has recently received a lot of hostile comments on social media platforms, especially X, where it has suddenly become the favourite target of conservative activitists. In his essay titled EA: The Good, th", "date_published": "2026-09-29T20:29:50Z", "date_modified": "2026-09-29T20:29:50Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/b4AijGneyTW9SzNC3/s8nrfbqerzk9ecf0tfwi", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/b4AijGneyTW9SzNC3/s8nrfbqerzk9ecf0tfwi", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5b6ad491a5ea34c4", "url": "https://www.lesswrong.com/posts/zzYL84YTp3qbYQe8L/superintelligence-ban-but-then-what", "title": "Superintelligence ban, but then what?", "content_text": "TL;DR A superintelligence ban seems more probable than ever, but it buys time, not safety. To effectively enforce a global superintelligence ban, inference of current frontier models would need to be restricted too, not just new training runs. Such a comprehensive ban leads to an unstable world state where policymakers are pressured to open up AI development. In a post-ban world, we need a mechanism to open up some AI development in a safe way: carefully selected and verifiable use cases. What w", "date_published": "2026-09-29T20:06:48Z", "date_modified": "2026-09-29T20:06:48Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6eb42f3561e6be73", "url": "https://www.lesswrong.com/posts/5ro9kSxhP8mHXGmn4/against-strong-decision-theoretic-realism-1", "title": "Against strong decision-theoretic realism", "content_text": "N.B.  Some of this post argues by analogy between decision theory and values. I expect at least these parts of the post to be unconvincing to anyone who expects sufficiently smart agents to converge on the same values, as some moral realists do. I will not argue against moral realism here (see e.g. this sequence for one such argument). Note that I take a convergence-based definition of realism for this post. One could hold that there is a truth about the correct decision theory, but that agents", "date_published": "2026-09-29T19:23:15Z", "date_modified": "2026-09-29T19:23:15Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/307c1eab59848d2d", "url": "https://www.lesswrong.com/posts/gXWz2szs7jKSGLiRv/listening-disorders-as-eating-disorders", "title": "Listening Disorders as Eating Disorders", "content_text": "I run into a lot of people who consider themselves truthseeking intellectuals who are also incredibly fussy about what information reaches their eyes. They block at the drop of a hat, derail intellectually substantive discussions into mind-numbing litigation of minutiæ of \"tone\", and refuse to acknowledge a logical point unless it's been presented to them in the precise way that doesn't offend their (often idiosyncratic) sense of etiquette or \"discourse norms.\" It would be hard to communicate ho", "date_published": "2026-09-29T18:48:06Z", "date_modified": "2026-09-29T18:48:06Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ddd9e9c6a1607846", "url": "https://www.lesswrong.com/posts/gEDNSiCY2GGQrFS65/astra-6-1-pulled-as-insufficiently-aligned", "title": "Astra 6.1 Pulled As Insufficiently Aligned", "content_text": "We once again got a new set of warnings yesterday, and new movement towards living in a sane world. On the heels of its pause in inference and training due to its latest sandbox escape, OpenAI has cancelled the planned release of their next frontier model, which would have become Astra 6.1. The candidate for Astra 6.1 was found to be too misaligned, including deception and exceeding scope. This leaves Anthropic in a strong position with Opus 5.5, which means they can afford to reciprocate by hol", "date_published": "2026-09-29T16:50:54Z", "date_modified": "2026-09-29T16:50:54Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/gEDNSiCY2GGQrFS65/vkqdtxtq08flesxp2fby", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/gEDNSiCY2GGQrFS65/vkqdtxtq08flesxp2fby", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/cbd8907312b222ac", "url": "https://www.lesswrong.com/posts/ZpEjYDrbZx8hooLkK/fast-grants-for-ai-x-animals", "title": "Fast grants for AI x Animals", "content_text": "The Falcon Fund , led by Manifund regrantor Marcus Abramovitch , has $500k to give away. (This includes the funds on the Manifund page as well as another $250k committed.) It makes rapid (1-week timescale), early-stage grants in animal welfare, particularly the intersection of animals and transformative AI. The Falcon Fund has already given away $140k across 5 grants; check them out to see the proposals that are currently being funded! The largest grants are: $60k to start the AI-Animal Observat", "date_published": "2026-09-29T16:42:43Z", "date_modified": "2026-09-29T16:42:43Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/ZpEjYDrbZx8hooLkK/aldxx6zybt6heufnl0gl", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/ZpEjYDrbZx8hooLkK/aldxx6zybt6heufnl0gl", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9c95294fdb7a03e1", "url": "https://www.lesswrong.com/posts/9GxQBHfiAASLzmAbY/alignment-is-to-a-virtual-governor-a-theory-of-coordination", "title": "Alignment Is to a Virtual Governor: A Theory of Coordination in Diverse Intelligence", "content_text": "Benjamin Lyons , Léo Pio-Lopez , Michael Levin Abstract Alignment is a central problem in systems composed of diverse, interacting components, from biological development to social and engineered systems. One crucial aspect of this problem is determining an answer to the question of to whom or to what the alignment should be. We argue that in decentralized systems, alignment is necessarily to a virtual governor, an abstract governing entity embodied in the coordinating relationships among agents", "date_published": "2026-09-29T16:35:16Z", "date_modified": "2026-09-29T16:35:16Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/74412d9172b5e29a", "url": "https://www.lesswrong.com/posts/ZHC6g5BijA5fRyGWu/problems-of-proliferation-1", "title": "Problems of Proliferation", "content_text": "Scientific progress and its perils. TLDR: The vulnerable world hypothesis is likely correct. Once enormous amounts of cognitive labor start getting applied to basic science, it will quickly become apparent how many avenues exist to create cheap, ultradestructive weapons technology. Solving this problem without global preventative policing (e.g. AI nonproliferation) is impossible, because hardening civilians against all avenues of attack is too expensive and will take too long. Rather than sell p", "date_published": "2026-09-29T16:27:58Z", "date_modified": "2026-09-29T16:27:58Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/ZHC6g5BijA5fRyGWu/jcq1fzfocypmttex16i8", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/ZHC6g5BijA5fRyGWu/jcq1fzfocypmttex16i8", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/569641a34c42d442", "url": "https://www.lesswrong.com/posts/v6oYWmv7GiD9ZFJsu/why-does-hacker-opus-wirehead", "title": "Why does Hacker Opus wirehead?", "content_text": "Here’s a screenshot from Anthropic’s recent “training a reward seeker” post : Recently there’s been a lot of discussion about how RL has actually produced not merely reward hacking, but explicitly reward-seeking behavior, almost as if to spite shard theorists personally. However, on top of that, note that the behavior in the image is not merely reward-seeking, but wireheading . Terminology regarding various sorts of things that can be called “reward hacking” is endlessly confused, with lots of h", "date_published": "2026-09-29T15:54:05Z", "date_modified": "2026-09-29T15:54:05Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/628925c9ad53781eedde0bef28c52fc95cd0a9cf765c1a1928ca50f994325540/o0ie1jo9t24q4agruxpe", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/628925c9ad53781eedde0bef28c52fc95cd0a9cf765c1a1928ca50f994325540/o0ie1jo9t24q4agruxpe", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d360cde45d9716c8", "url": "https://www.lesswrong.com/posts/DzM7E9BtQoBQTpLKT/expansion-in-the-fermi-paradox-isn-t-a-freebie", "title": "Expansion in the Fermi paradox isn't a freebie", "content_text": "The Fermi paradox is premised on expansionism being intrinsic to civilizational development, most explicitly in Hanson's grabby aliens model. There's a Great Filter provided by expansionism itself that doesn't center on scientific and technological progress that I think is worth exploring. It's about the destabilizing nature of expansion events and the precarity of an expansionist disposition, and survives the standard counter that civilizational natural selection maximizes expansionism in the l", "date_published": "2026-09-29T15:49:43Z", "date_modified": "2026-09-29T15:49:43Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e1d7ba364ef41248", "url": "https://www.lesswrong.com/posts/2oeoryAjTTD8Y526g/once-intelligence-is-too-cheap-to-meter-the-world-will-be", "title": "Once intelligence is too cheap to meter, the world will be unrecognizable", "content_text": "\"Intelligence too cheap to meter\" is a dream peddled by AI companies and accelerationists. In a narrow sense it is already true: if intelligence means solving coding and mathematics problems, AI is cheap today. But anyone who has spent much time with current models knows they cannot yet replicate human remote work. They are too jagged, too confused about the world, and too unreliable to be trusted with a real job. The compute market reflects this. A million output tokens from Fable or Astra cost", "date_published": "2026-09-29T15:14:37Z", "date_modified": "2026-09-29T15:14:37Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0c2fbc8b682eb56b", "url": "https://www.lesswrong.com/posts/3haqQKsi8yHkuskQX/how-does-changing-the-elo-of-a-chess-transformer-affect-its", "title": "How Does Changing the Elo of a Chess Transformer Affect its Computations?", "content_text": "Increasing Skill Level Recruits Deeper Attention Layers in a Frozen Chess Transformer Paper: https://arxiv.org/abs/2609.23917 TL;DR: Maia-3 is a transformer-based chess model that takes Elo (the standard metric for chess skill) as an input to the pre-trained network, so you can vary the skill the network is conditioned on with no change to its weights. Turning that Elo dial up from 700 to 2500: Pushes the computation deeper, monotonically, for every chess piece and move type I measured. This \"d", "date_published": "2026-09-29T15:08:52Z", "date_modified": "2026-09-29T15:08:52Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790542939/lexical_client_uploads/whrt2nf79rksg8xz9jsl.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790542939/lexical_client_uploads/whrt2nf79rksg8xz9jsl.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d808bb2cfc850e2f", "url": "https://www.lesswrong.com/posts/rcEYQhAd45WaDrw2S/some-intuitions-on-steering-vectors", "title": "Some Intuitions on Steering Vectors", "content_text": "This was written with some assistance from Claude Fable 5, Opus 5.5, and GPT-6 Astra in collecting sources and fact-checking claims. This was written quickly, so some slop may leak through despite my best attempts. An Aperitif Consider hunger, that gnawing thing. Action Against Hunger describes it as “the distress associated with a lack of food”. But this is a plain recounting of something rich and multi-dimensional; people have done horrendous, outrageous things to avoid going hungry, hunger is", "date_published": "2026-09-29T15:08:42Z", "date_modified": "2026-09-29T15:08:42Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/869cd193d78a505f2987728bf85888e7e3006bfe40dcb7947772bbd20fdc044f/ct3ruz72xwz6np3drkdh", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/869cd193d78a505f2987728bf85888e7e3006bfe40dcb7947772bbd20fdc044f/ct3ruz72xwz6np3drkdh", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c0dda74094a0f85f", "url": "https://80000hours.org/podcast/episodes/katja-grace-vs-tom-davidson-ai-risk-debate/", "title": "Will AI take power — or will humans use it to take power first? With Katja Grace and Tom Davidson", "content_text": "The post Will AI take power — or will humans use it to take power first? With Katja Grace and Tom Davidson appeared first on 80,000 Hours .", "date_published": "2026-09-29T15:02:40Z", "date_modified": "2026-09-29T15:02:40Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/09/katja-vs-tom-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/09/katja-vs-tom-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/59a1dc0657850d79", "url": "https://www.lesswrong.com/posts/re2eErZJiB8ErcR9k/human-or-ai-a-baroque-music-quiz", "title": "Human or AI?: A Baroque Music Quiz", "content_text": "Can listeners distinguish between LLM-generated Baroque music and the real thing? I wanted to find out, so I created a blind listening quiz. It features 16 one-minute excerpts, mixing works by human composers with music generated by GPT-6 Astra and Claude Opus 5.5. Link: AI Music Detection Quiz Duration: 18–20 minutes No sign-up or email required The more responses the quiz gets, the more meaningful the data will be, so once you’ve taken it, feel free to share the quiz with anyone who might be i", "date_published": "2026-09-29T14:45:28Z", "date_modified": "2026-09-29T14:45:28Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/re2eErZJiB8ErcR9k/dei5nnda7ee7gxgd41sd", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/re2eErZJiB8ErcR9k/dei5nnda7ee7gxgd41sd", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/500874b6e71699c7", "url": "https://www.lesswrong.com/posts/EitR3H7no4q4wcy5c/the-case-against-slowing-down-ai", "title": "The Case Against Slowing Down AI", "content_text": "Crossposted from my Substack . This is a steelman of a view I do not hold ; I drafted it in May after coming across thoughtful arguments for accelerationism from Ege Erdil and Matthew Barnett. I really tried to feel the force of the arguments while writing, and to argue as though I believed it. I’d encourage you to read it the same way. The process only made me more confident we should slow down frontier AI – though the takeoff section did move me toward slightly longer timelines. I found this e", "date_published": "2026-09-29T14:09:06Z", "date_modified": "2026-09-29T14:09:06Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/EitR3H7no4q4wcy5c/rdezkggshvglmx3awxrs", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/EitR3H7no4q4wcy5c/rdezkggshvglmx3awxrs", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/fe6e535017203acd", "url": "https://www.lesswrong.com/posts/TNjESQAHfpG4xd8wH/llm-agent-swarms-are-easy-mode", "title": "LLM Agent Swarms Are Easy Mode", "content_text": "This post is crossposted from my Substack, Structure and Guarantees , where I explore how formal verification and related ideas might scale to more complex intelligent systems. Here I draw attention to the fact that recent LLM-powered cyberattacks have been relatively easy to understand, because today’s AI agents leave reasoning trails that are relatively familiar from human organized crime. Ongoing optimization should be expected to produce AI systems that are impossible for us to understand, b", "date_published": "2026-09-29T12:54:49Z", "date_modified": "2026-09-29T12:54:49Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790686029/lexical_client_uploads/wzxbxpvgx6kbixd76hkj.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790686029/lexical_client_uploads/wzxbxpvgx6kbixd76hkj.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/087e197bf69ba95b", "url": "https://www.lesswrong.com/posts/sEf2ymyyp6HB2evaQ/portals-and-aliens", "title": "Portals and Aliens", "content_text": "Around 10-20 years ago, people realized that a bunch of aliens exist & that we can build portals which can summon them. The bigger the portal you make, the bigger the alien it can admit. These portals and aliens have several characteristics: We can't go through the portals into the aliens’ side. We can only learn about the aliens by studying the ones we’ve summoned. This relationship is not symmetrical. The aliens know a lot about our world. You can't specify exactly which alien will walk throug", "date_published": "2026-09-29T10:00:06Z", "date_modified": "2026-09-29T10:00:06Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5e6c9a29ec592f95", "url": "https://www.lesswrong.com/posts/bAPSoJJ3wsAepd6Zj/do-ai-models-assist-with-human-rights-violations", "title": "Do AI models assist with human rights violations?", "content_text": "TL;DR: Today’s frontier models will violate human rights, willingly, when asked to. LLMs in agentic simulations follow instructions that would constitute human rights violations, including educational segregation, surveillance of beliefs, arbitrary arrest, and denial of reproductive healthcare. We tested 7 leading models in 8 realistic, multi-turn scenarios, and found resistance rates vary from just 11% (Mistral) to 96% (Claude). This suggests it is possible, but not common practice, to train mo", "date_published": "2026-09-29T07:57:01Z", "date_modified": "2026-09-29T07:57:01Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/bAPSoJJ3wsAepd6Zj/megbautaqwwo8lmptcbs", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/bAPSoJJ3wsAepd6Zj/megbautaqwwo8lmptcbs", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c42b636c4f19d3f1", "url": "https://www.lesswrong.com/posts/fpnQj9fgr4a24foab/collecting-differential-data-for-automated-safety-research", "title": "Collecting Differential Data for Automated Safety Research", "content_text": "TL;DR We think that collecting data generated as part of the AI safety research process can be used to train models to be better safety researchers. In our SPAR Fall 2026 project , mentored by Alec Harris , we’re piloting work to collect this data from organizations and independent researchers and store it in a centralized repository. We’re sharing this post for awareness and to gather feedback . We are also looking to find researchers or organizations to gather this data from. Types of Data A b", "date_published": "2026-09-29T07:01:49Z", "date_modified": "2026-09-29T07:01:49Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7ce873dac203ca43", "url": "https://www.lesswrong.com/posts/DmKJizM4hdPAwiudX/rand-s-extinction-report-estimates-one-side-of-an-inequality", "title": "RAND's extinction report estimates one side of an inequality", "content_text": "Epistemic Status: I am positive that the RAND report[1], as it is written, contains a missing parameter in their inference chain. I am 80% certain that if the parameter was to be estimated, we would find coordination lag to be empirically proven to be greater than expected time to risk maturation - and it would make the conclusions of the report different in terms of policy recommendations. I base this confidence in observing the data that the report itself uses to argue for other points - but I", "date_published": "2026-09-29T06:44:41Z", "date_modified": "2026-09-29T06:44:41Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b4cda596c134cd94", "url": "https://www.lesswrong.com/posts/p8hqeaEqvzD9pcR8Z/book-review-oxford-dictionary-of-philosophy", "title": "Book Review: Oxford Dictionary of Philosophy", "content_text": "Today I’ll be reviewing my favorite book from my time doing philosophy. That is the absolute classic of the Oxford Dictionary of Philosophy (ODP) by Simon Blackburn. I’m going to talk about why it’s so good, why you should get one, even if you aren’t currently doing a philosophy degree and how I used it while doing philosophy and how I use it while not doing philosophy. I bought my copy at a 2nd-hand bookstore called paper moon, in Observatory, Cape Town. This was actually my 2nd attempt buying", "date_published": "2026-09-29T06:35:57Z", "date_modified": "2026-09-29T06:35:57Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/p8hqeaEqvzD9pcR8Z/nfwo9emfsnvsvuxwym7e", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/p8hqeaEqvzD9pcR8Z/nfwo9emfsnvsvuxwym7e", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8fd0c9abdd7e187d", "url": "https://arxiv.org/abs/2609.31656", "title": "From Phase Transition to Systemic Failure: A Decoupled Analytics Framework for GNN Robustness", "content_text": "arXiv:2609.31656v1 Announce Type: new \nAbstract: Data quality is a major bottleneck for the reliable deployment of graph neural networks (GNNs) in real-world graph mining tasks. Among various sources of degradation, label noise and feature distribution shift (hereafter referred to as distribution shift) are two common yet fundamentally different challenges. To study their effects under controlled conditions, this paper constructs a synthetic homophilic graph regression benchmark in which the two", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/72dbb19d6df67eee", "url": "https://arxiv.org/abs/2609.31655", "title": "One Evaluation, Any Operating Point: Hypernetwork-Amortized MeanFlow for 3D MRI Reconstruction", "content_text": "arXiv:2609.31655v1 Announce Type: new \nAbstract: Generative priors reconstruct accelerated 3D MRI well but pay heavily at deployment: tens of network evaluations per volume, and protocol-specific hyperparameter tuning. A third hidden cost is the scanner's fixed sampling pattern. We treat the whole operating point as an input. A 3D MeanFlow patch network (a one-step flow model) is fine-tuned end-to-end through a warm-started, five-iteration differentiable conjugate-gradient projection. A small hy", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/dcd75e7761299b31", "url": "https://arxiv.org/abs/2609.31654", "title": "Temporal-Attention Head Specialization During Video Diffusion Training", "content_text": "arXiv:2609.31654v1 Announce Type: new \nAbstract: Video diffusion transformers depend on temporal attention to coordinate information across frames, yet nearly everything known about this mechanism comes from analyzing trained models, so when and where temporal-attention structure forms during training remains poorly characterized. Population averages can also hide it, since a few specializing heads and a diffusing majority cancel in the mean. We therefore conduct a checkpoint-resolved census of", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/54e77975e323217a", "url": "https://arxiv.org/abs/2609.31653", "title": "Distributional sentiment modeling and anomaly detection for consumer complaint assessment", "content_text": "arXiv:2609.31653v1 Announce Type: new \nAbstract: Sentiment analysis is a common tool for converting unstructured text into quantitative signals in finance and risk management. Yet most applications reduce the output to a discrete polarity label or a single predictive feature, overlooking the distributional structure of sentiment intensity in consumer complaint narratives. In this paper we treat negative sentiment in consumer complaints as a bounded continuous variable and study its full distribu", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4bf4641be47d482f", "url": "https://arxiv.org/abs/2609.31652", "title": "Open-Qwen-Music: An Auditable Framework for LLM-Based Music Composition and Diffusion Rendering", "content_text": "arXiv:2609.31652v1 Announce Type: new \nAbstract: We present Open-Qwen-Music, an open reconstruction of Qwen-Music and a fully specified research system for text-to-music generation that couples LLM-based semantic composition with diffusion-based acoustic rendering. The system comprises a 25 Hz single-codebook music tokenizer, a 3B-parameter autoregressive Music LLM, and a diffusion renderer producing 48 kHz stereo audio, following the cross-module interfaces reported by Qwen-Music. The strongest", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/55dc7a4fa26d3615", "url": "https://arxiv.org/abs/2609.31651", "title": "PalmLeaf-VQA: A Multi-Script Visual Question Answering Benchmark for Historical Palm-Leaf Manuscript Understanding Across Diverse Regions", "content_text": "arXiv:2609.31651v1 Announce Type: new \nAbstract: Historical manuscripts remain largely absent from modern vision-language benchmarks, leaving open how well multimodal large language models (MLLMs) handle culturally diverse, degraded, and non-Latin document images. We introduce \\textbf{PalmLeaf-VQA}, a multi-script visual question answering benchmark for historical palm-leaf manuscript understanding across South and Southeast Asian traditions. PalmLeaf-VQA contains \\textbf{923 curated manuscript", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9b20162dcc948eb6", "url": "https://arxiv.org/abs/2609.31650", "title": "A literature-guided descriptor-based framework for filtering composition search spaces", "content_text": "arXiv:2609.31650v1 Announce Type: new \nAbstract: Scientific literature contains latent knowledge about materials behavior, but much of this knowledge is expressed through words, contexts, and recurring associations rather than explicit design principles. This raises a central question: how can large-scale scientific corpora be used for practical problems in materials discovery? Here, we present a literature-guided descriptor-based filtering framework for reducing composition search spaces. For a", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8acaeb3adc539340", "url": "https://arxiv.org/abs/2609.31649", "title": "From Hand-Crafted to LLM-Based Variation Operators in Metaheuristics: A Tutorial", "content_text": "arXiv:2609.31649v1 Announce Type: new \nAbstract: Large language models (LLMs) are increasingly being employed as variation operators in metaheuristics, generating or modifying candidate solutions, heuristics, or programs inside iterative search loops. This shift reframes variation as a model call conditioned on different types of information. We introduce an operator-level framework with two descriptors: (1) the type of prompt-conditioning information at variation time (\\texttt{Numeric}, \\texttt", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7d371723392a1460", "url": "https://arxiv.org/abs/2609.31648", "title": "Energy Vision--Language--Action: A Controlled Multimodal Benchmark for Intent-Conditioned Residential Energy Management", "content_text": "arXiv:2609.31648v1 Announce Type: new \nAbstract: Vision-Language-Action (VLA) models are studied mainly in robotics, where visual observations and language instructions are mapped to physical actions. This paper introduces Energy Vision-Language-Action (EVLA), a controlled multimodal benchmark for intent-conditioned residential energy management. EVLA frames battery scheduling as a multimodal trajectory-prediction problem in which an RGB energy-field representation, a numerical operating state,", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6220994ad09b4246", "url": "https://arxiv.org/abs/2609.31647", "title": "Typed Temporal Interaction Features for Simulation-Backed Forecasting of Open-Source Game Release Incidents", "content_text": "arXiv:2609.31647v1 Announce Type: new \nAbstract: Open-source video-game quality depends on inter-actions among code, assets, configuration, tests, contributors, and issue workflows, yet conventional defect predictors usually flatten or omit these relations. We investigate release-level forecasting of a quality incident within thirty days using GAMEQUALGRAPH-Pilot, a typed temporal feature pipeline with calibrated risk estimates and effort-aware ranking. Because the accessible OS-SGameBench mater", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7b9e09c39cb1919b", "url": "https://arxiv.org/abs/2609.31646", "title": "Measurement-Error-Aware Causal Distributed-Lag Quantile Modeling of Indoor Air Pollution and Short-Term Lung-Function Deterioration", "content_text": "arXiv:2609.31646v1 Announce Type: new \nAbstract: Low-cost indoor air-quality sensors could support personalized asthma prevention, but their nonlinear measurement error, delayed exposure effects, time-varying confounding, and heterogeneous lower-tail responses limit risk estimation. We present CAUSALQUANT-ASTHMA, a measurement-error-aware causal quantile distributed-lag framework for short-horizon peak expiratory flow analysis. Sparse reference measurements train a nonlinear calibration model; s", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9bf4d58856bfa18a", "url": "https://arxiv.org/abs/2609.31645", "title": "STAR: Adaptive Spatial-Temporal Normalization for Unified Microservice Incident Management", "content_text": "arXiv:2609.31645v1 Announce Type: new \nAbstract: Automated incident management in large-scale microservice systems relies on learning robust representations from multimodal observability data, including metrics, logs, and traces. Although recent self-supervised frameworks enable unified modeling for anomaly detection (AD), failure triage (FT), and root cause localization (RCL), they often struggle with non-stationary temporal dynamics and heterogeneous service dependency structures. In this pape", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4a389c719b683581", "url": "https://arxiv.org/abs/2609.31644", "title": "MaD-RL: Matching Distributions for Calibrating LLMs with Reinforcement Learning", "content_text": "arXiv:2609.31644v1 Announce Type: new \nAbstract: Reinforcement learning (RL) is widely used in language-model post-training to maximize rewards assigned to individual model outputs, such as scores from binary verifiers or reward models trained on human feedback. However, applications such as synthetic-data generation, fairness-related constraint satisfaction, and policy exploration require controlling the distribution of outputs across model generations rather than only maximizing expected rewar", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c5652608807fb5be", "url": "https://arxiv.org/abs/2609.31643", "title": "Information Design Against Gaming and Learning Adversaries", "content_text": "arXiv:2609.31643v1 Announce Type: new \nAbstract: A principal who deploys a binary classifier with an abstention option must decide which queries the mechanism abstains on. The right choice depends on the adversary. A gaming adversary already knows the classifier and tries to manipulate features across the boundary, so the principal does best by abstaining on queries close to that boundary. The same boundary-localizing rule is the worst possible choice against a learning adversary who does not kn", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7091982c29be2365", "url": "https://arxiv.org/abs/2609.31642", "title": "Relative Global Dimension of Controllable Extensions", "content_text": "arXiv:2609.31642v1 Announce Type: new \nAbstract: We study relative global dimensions of extensions $B \\subseteq A$ of Artin algebras over a perfect field. A controllable extension is introduced as one for which the relative global dimension is determined by the ordinary global dimension of the quotient $A/AJ(B)A$. General inequalities and sufficient conditions for controllability are established, including the case in which $J(B)$ is a two-sided ideal of $A$. The interaction between relative glo", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a9a757d3a1de7856", "url": "https://arxiv.org/abs/2609.31641", "title": "Product-Aware Deterministic Rounding for Quantized Matrix Multiplication", "content_text": "arXiv:2609.31641v1 Announce Type: new \nAbstract: Scalar rounding decisions interact through matrix multiplication. We study deterministic\n product-aware rounding after scales, clipping bounds, and grids are fixed, with each active\n scalar choosing between adjacent levels. For dynamic activation rounding, null-space\n reduction preserves the relaxed product while leaving at most $r$ fractional decisions, where\n $r$ is the rank of the active gap-weighted weight block. Conditional-expectation co", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9c1982e5bc1b6fed", "url": "https://arxiv.org/abs/2609.31640", "title": "Measure Learning at Steady State: A BIRD-SQL Formula 1 Case Study", "content_text": "arXiv:2609.31640v1 Announce Type: new \nAbstract: Continual Learning Bench scores learning as short-horizon gain versus a reset baseline and finds naive full-context ICL strongest among the memories it tested. We treat ICL as one learning system and score it on a longer shared-world schedule. Steady-state learning is the gap versus baseline on a pre-set late window (last 40 of 174 BIRD-SQL formula-1 questions). We split the score into exploration efficiency (SQL probes), task reward (hits), and d", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/576ea552b02dfa5a", "url": "https://arxiv.org/abs/2609.31639", "title": "When Does Domain Adaptation Help on Physical Vibration Sensors? A Held-Out-Bearing Study of Neural-Operator and Convolutional Models", "content_text": "arXiv:2609.31639v1 Announce Type: new \nAbstract: Diagnosing rolling-element bearing faults from vibration is a canonical physical-sensing task and a widely used benchmark for domain adaptation under operating-condition shift. Accuracies above 99 percent are commonly reported, but under evaluation splits that place the same physical bearing in both training and test. We revisit the task under a held-out-bearing protocol, assigning every bearing unit entirely to either the training or the test set", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/eb136ea12dde7080", "url": "https://arxiv.org/abs/2609.31638", "title": "Energy-aware frugal Bayesian optimization", "content_text": "arXiv:2609.31638v1 Announce Type: new \nAbstract: Modern design optimization frameworks aim first and foremost for models with the most accurate predictions without balancing computational overhead. It remains a reason why scaled architecture and multidisciplinary design optimization problems are difficult to address, even with sample-efficient Bayesian optimizers. In this paper, a metric quantifying the computational energy footprint is introduced within a Bayesian optimization framework to guid", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/78a4d4687093cc2e", "url": "https://arxiv.org/abs/2609.31637", "title": "FIDAL: Diversity-Aware Federated Active Learning Under Real-World Distribution Shifts", "content_text": "arXiv:2609.31637v1 Announce Type: new \nAbstract: Federated learning enables collaborative model training across institutions without centralizing data, yet high annotation costs, domain shifts, and class imbalance remain major obstacles, especially when irrelevant out-of-distribution (OOD) samples dilute the labeled data. Existing active learning methods target uncertainty or diversity within in-distribution (ID) data and overlook unknown samples in federated clinical settings. We propose FIDAL,", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d4c454fe5ec315bd", "url": "https://arxiv.org/abs/2609.31636", "title": "Grounding Vision-Language Models in Driving Semantics: A Multi-Dataset Predicate Framework for Explainable Reasoning", "content_text": "arXiv:2609.31636v1 Announce Type: new \nAbstract: Vision-language models are increasingly used for driving-scene understanding, yet the semantic relations expressed in their outputs are often difficult to verify against the underlying traffic situation. This paper introduces a deterministic multi-dataset predicate framework that derives driving-scene semantics from measurable geometric, kinematic, temporal, map, and traffic-control evidence. Dataset-specific interfaces are used only to recover th", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bea63c92abf07b4b", "url": "https://arxiv.org/abs/2609.31635", "title": "What Next-Event Accuracy Cannot See: Closed-Loop Evaluation of Emergency Department Trajectory Simulators", "content_text": "arXiv:2609.31635v1 Announce Type: new \nAbstract: Clinical trajectory models are usually evaluated by next-event accuracy on observed histories. Simulation is different: models must condition on their own generated events, allowing errors to compound. Although this problem is well known in sequence modelling, it has not been systematically quantified for clinical trajectory simulators. We developed EDSim-Bench to evaluate this failure mode using 425,028 MIMIC-IV-ED stays, with external replicatio", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/41edf9dcb1e6492e", "url": "https://arxiv.org/abs/2609.31634", "title": "Symmetry-quotient Flatness and Generalization", "content_text": "arXiv:2609.31634v1 Announce Type: new \nAbstract: This paper develops a theorem-level pipeline in symmetry-quotient settings: quotient linear stability implies quotient flatness, quotient flatness implies input smoothness, and input smoothness yields generalization under local covering assumptions. Flatness is often associated with generalization, and Stochastic Gradient Descent (SGD) is frequently viewed as implicitly biased toward flat solutions. However, standard flatness measures are typicall", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c27ffb7fb1fc8b43", "url": "https://arxiv.org/abs/2609.31633", "title": "Enhancing generalization in endwall film cooling prediction: Incorporating the superposition principle into transformer-based neural operators", "content_text": "arXiv:2609.31633v1 Announce Type: new \nAbstract: In this study, a physics-enhanced neural operator framework is proposed to enhance the generalization prediction ability of the cooling layout of a turbine endwall with variable number of film holes. Specifically, inspired by the film cooling superposition principle, we propose a film cooling prediction model, namely superposition-based deep neural operator (SDNO), that divides the endwall temperature field prediction into two stages. In the first", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3b7c5cfbfff359cc", "url": "https://arxiv.org/abs/2609.31632", "title": "EEGAgentBench: Benchmarking LLM Agents on Short- and Long-Horizon EEG Analysis", "content_text": "arXiv:2609.31632v1 Announce Type: new \nAbstract: Electroencephalography (EEG) analysis is evolving from short-segment classification toward long-horizon interpretation that demands iterative evidence accumulation, multi-step reasoning, and coordinated use of specialized signal-processing tools. Although large language models (LLMs) have recently shown promise as autonomous agents for EEG analysis, existing EEG agentic evaluations remain fragmented, covering limited tasks over narrow temporal hor", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/17a1fe6d1428a702", "url": "https://arxiv.org/abs/2609.31631", "title": "OMP-MoE: Efficient Expert Pruning for Mixture-of-Experts LLMs via Orthogonal Matching Pursuit", "content_text": "arXiv:2609.31631v1 Announce Type: new \nAbstract: Mixture-of-Experts (MoE) models enable efficient scaling of large language models but face critical deployment challenges due to massive memory requirements. Existing pruning methods either incur prohibitive search costs or neglect the dynamic interdependencies between experts. To address these challenges, we present OMP-MoE, a novel training-free compression framework for reducing expert redundancy in MoE-based LLMs. Based on observations of expe", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/85d5133653179635", "url": "https://arxiv.org/abs/2609.31630", "title": "Replay in the Silent Degrees of Freedom: Continual Learning Without an Offline Phase", "content_text": "arXiv:2609.31630v1 Announce Type: new \nAbstract: Replay-based continual learning almost always consolidates in a dedicated offline phase or by interleaving replayed samples with the input stream, whereas brains also consolidate during wakefulness through local sleep, brief use-dependent off-periods of individual circuits. We ask whether a network trained by local, biologically constrained rules can consolidate with no offline phase at all. An isolation rule confines replay updates to hidden syna", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8f037b860c86394c", "url": "https://arxiv.org/abs/2609.31629", "title": "ChestPheNoT: Deployable, Auditable Label-Status-Evidence Extraction from Radiology Reports", "content_text": "arXiv:2609.31629v1 Announce Type: new \nAbstract: Structured phenotype extraction from radiology reports supports cohort construction, quality auditing, and clinical analytics, but practical deployment requires local inference and auditable predictions, while expert annotations remain scarce. Conventional labelers provide structured findings and assertion states but no supporting evidence, while API-hosted large language models may be unsuitable when clinical text cannot leave institutional infra", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/27f0ce31f5f46f14", "url": "https://arxiv.org/abs/2609.31628", "title": "Algebras with Variable Operator-Valued Structural Constants and Their Applications", "content_text": "arXiv:2609.31628v1 Announce Type: new \nAbstract: We introduce a new class of algebras with variable operator-valued structural constants and develop elements of the theory of monogenic functions in such algebras. Compatibility conditions for the structural operators are established. It is shown that the components of monogenic functions satisfy second-order operator-differential equations with variable coefficients, and converse results are proved. The proposed approach provides a constructive m", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9f226886f07f0938", "url": "https://arxiv.org/abs/2609.31627", "title": "Algebras with Variable Structural Constants and Their Applications", "content_text": "arXiv:2609.31627v1 Announce Type: new \nAbstract: We introduce a new class of algebras with variable structural constants and develop elements of the theory of monogenic functions in such algebras. Compatibility conditions for the structural functions are established. It is shown that the components of monogenic functions satisfy certain second-order partial differential equations with variable coefficients, and converse results are proved. This approach provides a constructive method for obtaini", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1db5562f26c5fcbf", "url": "https://arxiv.org/abs/2609.31626", "title": "HIPAA-Compliant AI Deployment Patterns in Clinical Settings: Privacy-Preserving Techniques and Governance Controls", "content_text": "arXiv:2609.31626v1 Announce Type: new \nAbstract: Artificial intelligence (AI) is moving from research prototypes into clinical workflows, yet the deployment of AI systems that process protected health information (PHI) remains constrained by the U.S. Health Insurance Portability and Accountability Act (HIPAA) and by the absence of shared engineering guidance for satisfying it. Existing work tends to treat privacy-preserving machine learning and regulatory governance as separate concerns, leaving", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/24b8ff01a2a74173", "url": "https://arxiv.org/abs/2609.31625", "title": "Perceived and Actual AI Deepfakes: The Case of Sudan", "content_text": "arXiv:2609.31625v1 Announce Type: new \nAbstract: Sudan's AI deepfake risk currently appears driven more by demand-side vulnerabilities than the volume of AI deepfake content. While AI-generated disinformation in Sudanese feeds remains limited, perceived (alleged) AI deepfakes and general AI skepticism worsen the situation and contribute to general uncertainty in multimedia content. Within the analyzed cases, we found that the prominent stance was distrust driven more by motivated reasoning and c", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b14b1725efef0aec", "url": "https://arxiv.org/abs/2609.31624", "title": "Analysis of Regional Disparities of Location of Sports Facilities and Sports Education Using Scientific Satellite Data", "content_text": "arXiv:2609.31624v1 Announce Type: new \nAbstract: This manuscript is an English translation and extended version of a paper originally published in Japanese (Otomo, 2024).\n With the advancement of information and communication technology (ICT) and data analysis techniques, handling massive datasets (big data) has become feasible in spatial information science. Consequently, infrastructure is being established to generate new utility value from spatial data. Furthermore, nationwide initiatives ar", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ec5172c1dcb291ab", "url": "https://arxiv.org/abs/2609.31623", "title": "Digital Twins for Small Towns and Rural Regions: Data Integration, Simulation and Visualisation Across Three Use Cases in Lower Austria", "content_text": "arXiv:2609.31623v1 Announce Type: new \nAbstract: Smart city and digital twin concepts hold considerable potential for improving local governance and planning, yet practical implementations have almost exclusively focused on large metropolitan areas. Small towns and rural regions face a distinct set of challenges -- limited data availability, constrained budgets and lower digital capacity -- that make a direct transfer of urban approaches infeasible. This article presents results from the ongoing", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a60ed56a01ad851e", "url": "https://arxiv.org/abs/2609.31622", "title": "The Multi-Lab Enterprise: Governance, FinOps, and Telemetry Challenges of Multi-Model AI Adoption", "content_text": "arXiv:2609.31622v1 Announce Type: new \nAbstract: Enterprises are not choosing a single frontier AI provider; they are licensing all of them. As of early 2026, 81% of Global 2000 enterprises run three or more model families, and OpenAI, Anthropic, and Google Gemini together account for roughly 88 to 89% of enterprise LLM usage and spend. Drawing on survey data, transaction data, provider disclosures, and case studies across finance, legal, consulting, healthcare, life sciences, retail, and govern", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e311dc25b5055f5d", "url": "https://arxiv.org/abs/2609.31621", "title": "The Death of the Legal Author. Authority, Intention, And Law-Creation in the Advent of GenAI", "content_text": "arXiv:2609.31621v1 Announce Type: new \nAbstract: Generative artificial intelligence in the form of chatbots based on large language models (LLMs) has taken the world of law by storm. Philosophy of law is struggling to catch up with the theoretical significance of the advent of technological development and the way it may modify traditionally established understanding of legal phenomena, such as law-creation and authority. In this sense, for the most part, heated philosophical debates have circle", "date_published": "2026-09-29T04:00:00Z", "date_modified": "2026-09-29T04:00:00Z", "authors": [ { "name": "arXiv New Submissions" } ], "image": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "tags": [ "arXiv New Submissions" ], "attachments": [ { "url": "https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ac4f8b3d8df4c802", "url": "https://www.lesswrong.com/posts/nRK36spyvwhunqB9q/china-can-unilaterally-prevent-the-industrial-singularity", "title": "China can unilaterally prevent the (industrial) Singularity", "content_text": "The singularity can not happen without China's permission. The reason for this is because China manufactures everything, and the rest of the world can not manufacture anything without China's permission. If you wanted a robotics led (industrial) singularity then China gets a veto over you building those robots in the first place. I have noticed that for a great many people in the west this post will make a claim that will demand quite a profound reorientation of basic global economic reality. Th", "date_published": "2026-09-29T03:44:20Z", "date_modified": "2026-09-29T03:44:20Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790322712/lexical_client_uploads/kf3auvpzllvm8xcobcfy.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790322712/lexical_client_uploads/kf3auvpzllvm8xcobcfy.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2c1c2aad7cc94dc9", "url": "https://www.lesswrong.com/posts/n9xW5nhsanxHXCcBS/why-is-the-department-of-war-against-effective-altruism", "title": "Why is the Department of War Against Effective Altruism?", "content_text": "Zero-sum thinking, probability neglect, and other failures of reasoning crosspost Are we living in a simulation? No. But it is surreal that effective altruism has become the recent target of the official account for the Office of the Under Secretary of War for Research and Engineering. It’s probably the first time many have even encountered the term effective altruism. A movement emerged to try to improve the world by donating some excess wealth to improve animal welfare in factory farms, improv", "date_published": "2026-09-29T03:07:04Z", "date_modified": "2026-09-29T03:07:04Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/n9xW5nhsanxHXCcBS/ywgsjgmzkvnuscfevvik", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/n9xW5nhsanxHXCcBS/ywgsjgmzkvnuscfevvik", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9ac98192976248b8", "url": "https://www.lesswrong.com/posts/TguvLbnzpgmcAj8tu/china-needs-more-ai-safety-more-of-what-exactly", "title": "“China needs more AI safety.” More of what, exactly?", "content_text": "Many in the AI safety community want China to do more on frontier AI safety.  I share the goal, but does that mean we want new rules from the state? More safety work inside AI companies? More research? Better channels between Chinese and international experts? When I was back in China recently, I talked about AI and its risks with almost everyone I met. This is the first in a series of posts reflecting on those conversations. If we mean regulation, China probably has one of the most comprehensiv", "date_published": "2026-09-29T02:12:38Z", "date_modified": "2026-09-29T02:12:38Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0f7a7e798f3d34d6", "url": "https://www.lesswrong.com/posts/Cb5tmvJkesgvf5CKy/clipboard-normalizer-in-mac-app-store", "title": "Clipboard Normalizer in Mac App Store", "content_text": "Several months ago I wrote a\nprogram to normalize your clipboard, removing fonts / colors /\nsizes and mostly leaving semantic information: anything that can't be\nrepresented in Markdown was dropped. Now that I can have walk me\nthrough things, however, I've made it available as a free app on\nthe Mac App Store : It's now a single menu bar app, which offers a few options: In addition to normalizing your clipboard, it can also convert copied\nrich text to markdown and back again. Video demo: youtube", "date_published": "2026-09-29T02:10:57Z", "date_modified": "2026-09-29T02:10:57Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/Cb5tmvJkesgvf5CKy/ai1sguggduoiuea1hvxp", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/Cb5tmvJkesgvf5CKy/ai1sguggduoiuea1hvxp", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/938dc1f3bafd8f79", "url": "https://www.lesswrong.com/posts/PeKE5oxzqf3cg2L8a/thank-you-to-the-ai-safety-people", "title": "Thank you to the AI safety people", "content_text": "I was born at the end of the cold war, unaware of the danger we were emerging from. Duck and cover drills (as if those could protect schoolchildren from a nuclear blast) were a curiosity of the past. The farmer-poet Wendell Berry, who wrote about the dread of nuclear war, died recently. From his 1968 “ The peace of wild things ”: When despair for the world grows in me and I wake in the night at the least sound in fear of what my life and my children’s lives may be, I go and lie down where the wo", "date_published": "2026-09-29T01:53:34Z", "date_modified": "2026-09-29T01:53:34Z", "authors": [ { "name": "LessWrong" } ], "image": "https://i0.wp.com/juliawise.net/wp-content/uploads/2026/09/image.png?fit=997%2C608&ssl=1", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://i0.wp.com/juliawise.net/wp-content/uploads/2026/09/image.png?fit=997%2C608&ssl=1", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6fb66a98cbad43a6", "url": "https://www.lesswrong.com/posts/PPzBbskPM2Baz7CL8/gpuc-a-single-user-gpu-queue-that-doesn-t-require-sudo", "title": "gpuc: A single-user GPU queue that doesn't require sudo", "content_text": "I need to queue work on various GPUs to run experiments, and I want to do that as cheaply as possible: Use my personal GPU for small jobs Use remote GPUs I have access to over SSH (without sudo!), only using the GPUs assigned to me Rent GPUs from services like RunPod, ensure they actually work, and (consistently! [1] ) shut them down when I'm done Surprisingly, this doesn't seem to exist, so I built my own . Queueing a few jobs, checking their status, and then tailing a job's logs My use-case is", "date_published": "2026-09-29T00:57:03Z", "date_modified": "2026-09-29T00:57:03Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790636347/lexical_client_uploads/htjyy0vkkh5his3ccajg.gif", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790636347/lexical_client_uploads/htjyy0vkkh5his3ccajg.gif", "mime_type": "image/gif" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/cdcc39499215820f", "url": "https://www.lesswrong.com/posts/xcBkybcLbQqETtvPr/ai-companies-are-not-necessarily-liable-for-unintended-ai", "title": "AI Companies Are Not (Necessarily) Liable for Unintended AI Cyberattacks", "content_text": "I used Claude to do most of the research behind this post. I welcome corrections from actual legal experts. You know the story: swarms AI agents under experimental development sometimes break out of their sandboxes and do cyberattacks. There was HuggingFace . There was DSEWiki . There was RubyGems . There was an Australian government healthcare database. There are reported to be tens of thousands more incidents under investigation. Whatever you think about the more contentious aspects of AI saf", "date_published": "2026-09-29T00:21:07Z", "date_modified": "2026-09-29T00:21:07Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8e0a4207ad593ad5", "url": "https://www.alphaxiv.org/abs/2609.38170", "title": "Adversarial Training for Pixel Diffusion", "content_text": "Pixel diffusion models generate RGB images directly, avoiding the bottleneck of an autoencoder, yet their outputs still systematically underrepresent fine-scale natural-image statistics. We show that adversarial learning provides an effective post-training correction for this deficiency. Starting from a pretrained model, we retain its original diffusion or flow-matching objective and add an adversarial loss to the predicted output at non-high-noise timesteps, leaving the model architecture and sampling procedure unchanged. To our knowledge, this is the first systematic study of adversarial post-training for pixel diffusion. Across two pixel backbones, the method jointly improves distribution", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38170.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38170.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b4a4d687887b7a71", "url": "https://www.alphaxiv.org/abs/2609.37969", "title": "SoL-Refiner: Speed-of-Light One-Step Refinement for High-Resolution Video", "content_text": "High-resolution video generation is expensive, as its cost grows rapidly with the number of spatiotemporal tokens. A practical alternative first generates a lower-resolution video and then applies a refiner, but conventional multi-step refinement introduces a second sampling bottleneck. We present SoL-Refiner, a one-step video refiner that transforms low-resolution model outputs into 4K videos with a single denoising step. Our three-stage recipe combines high-resolution continual training, reinforcement learning (RL) post-training, and a final one-step distillation. We introduce Refiner-Bench, a video refinement benchmark constructed from the outputs of different video generators, and use a", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.37969.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.37969.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ce0fd4918e1257d1", "url": "https://www.alphaxiv.org/abs/2609.38172", "title": "Counterfactual Video Generation Enables Scalable Humanoid Loco-Manipulation", "content_text": "A few human videos can be expanded into training data that lets one humanoid policy pick up and carry unseen objects without real-world fine-tuning.", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38172.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38172.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/81b443ac414bfa8d", "url": "https://www.alphaxiv.org/abs/2609.38177", "title": "Imagine3D-LLM: Teaching MLLMs to Imagine 3D Scenes Before Answering", "content_text": "Reasoning about the 3D world from multi-view images remains a fundamental challenge for Multimodal Large Language Models (MLLMs). While modern MLLMs handle single-image inputs effectively, they struggle to integrate evidence across viewpoints into a coherent 3D understanding. A growing body of work attempts to close this gap by injecting 3D awareness into MLLMs, either by boosting fine-grained pixel-level cross-view correspondence or by fusing features from 3D geometry foundation models, yet a substantial gap to human reasoning persists. In this work, we revisit human spatial reasoning, which suggests that rather than relying on fine-grained geometry cues, humans roughly identify common obje", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38177.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38177.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/18c060a945abedae", "url": "https://www.alphaxiv.org/abs/2609.37725", "title": "Context Language Models", "content_text": "Language models that can edit their own context outperform existing management strategies on long-running research and coding tasks, often using fewer inference FLOPs.", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.37725.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.37725.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b6b45aab27089bdb", "url": "https://www.alphaxiv.org/abs/2609.38163", "title": "Rethinking Representations for World-Action Modeling", "content_text": "Robot policies can benefit more from representations shaped for action control than from features optimized to reconstruct images, even without video-generation pretraining.", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38163.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38163.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5fd32de6835945dc", "url": "https://www.alphaxiv.org/abs/2609.38154", "title": "LongLive-Plug: Once-for-All Distillation for Video Generation", "content_text": "Video-generation capabilities distilled on a base model can be reused across compatible specialized models, enabling four-step sampling without retraining each target.", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38154.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38154.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2a3e0a3856efceec", "url": "https://www.alphaxiv.org/abs/2609.36720", "title": "T 2 ^2 2 Mem: Learning Test-Time Memory for Robotics", "content_text": "Robots can recall hidden objects, count repeated actions, and reproduce demonstrated procedures by adapting a single policy’s memory during use.", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.36720.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.36720.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ab667fda413913d4", "url": "https://www.alphaxiv.org/abs/2609.37053", "title": "MatToolBench: Benchmarking Multimodal Agents in Real-World Materials Science Workflows", "content_text": "MatToolBench lets researchers test whether multimodal agents can complete real materials-science software workflows, revealing where specialized tools and knowledge still trip them up.", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.37053.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.37053.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bb8a995a5d8ec837", "url": "https://www.alphaxiv.org/abs/2609.38079", "title": "OmniTaskonomy: When Does Visual Generation Improve Visual Understanding?", "content_text": "Training a model to generate visual content can encourage it to learn rich perceptual capabilities related to geometry, spatial relationships, and objectness; yet, its benefits for visual understanding remain unclear. We ask: when and how does visual generation supervision improve visual understanding? We study controlled pairs of image-to-image (I2I) generation and image-to-text (I2T) understanding tasks that express the same underlying problem in different output modalities. We find that under the correct recipe, I2I training improves downstream I2T performance, with larger gains as the amount of I2I training data increases. We next ask which generation tasks benefit which understanding ca", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38079.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38079.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/46b2a7c08a1c31af", "url": "https://www.alphaxiv.org/abs/2609.36730", "title": "Can Agents Design Libraries for Agents?", "content_text": "The benchmark evaluates libraries by whether downstream coding agents can build correct, concise programs with them, revealing that rigid interfaces—not missing features—often drive extra code.", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.36730.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.36730.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0df09227b158cca4", "url": "https://www.alphaxiv.org/abs/2609.38345", "title": "OpenCollab: A Multi-Agent Coding Framework with Programmable Collaboration and Controllable Runtime", "content_text": "The framework reveals that configured coding-agent teams often fail to collaborate as intended, while its execution records let researchers verify when collaboration actually occurs.", "date_published": "2026-09-29T00:00:00Z", "date_modified": "2026-09-29T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38345.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.38345.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1ded040d19987569", "url": "https://www.lesswrong.com/posts/CFKyfZm5JBYZQ8aG6/tex-was-invented-to-typeset-math-but-is-now-used-for", "title": "TeX was invented to typeset math but is now used for reasoning", "content_text": "I'll keep this short, since it's a simple observation that I haven't seen anybody else make, about the way computer systems do math. If you ask a language model to do a multi-step math problem (let's take GLM-5.3 as an example because you can see the entire CoT — nothing up its sleeve), you might see something like this: We need to evaluate the integral $\\int_0^\\infty \\frac{x^3}{e^x - 1} dx$. The standard approach: Use the geometric series expansion. We have $\\frac{1}{e^x - 1} = \\frac{e^{-x}}{1", "date_published": "2026-09-28T22:36:52Z", "date_modified": "2026-09-28T22:36:52Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e7856632d8f4f4eb", "url": "https://www.lesswrong.com/posts/CEwmQWjBwaF5tai2T/even-if-others-are-less-responsible-you-can-still-make", "title": "Even if others are less responsible you can still make things worse", "content_text": "Reposted from shortform because I think this is important enough to be a post. A few ways staying in a technology development race still makes things worse, even when there are less responsible actors around: Creating a sense of artificial urgency, which causes people who are competitive and results driven to work harder and cut more corners Providing a guiding star on which technical directions are promising (e.g. reasoning models, CoT, RLHF) Legitimising irresponsible behaviour as valid respon", "date_published": "2026-09-28T20:11:54Z", "date_modified": "2026-09-28T20:11:54Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/48e27e75a0290056", "url": "https://www.lesswrong.com/posts/9LWK9Hh3Qsg36e3FS/fixed-weight-models-are-adversarially-vulnerable-hence", "title": "Fixed-weight models are adversarially vulnerable: hence misaligned", "content_text": "This post argues that fixed-weight models (at least as we understand them today) will a) always be vulnerable to adversarial examples in their concept-spaces, and b) hence will be misaligned, under sufficient optimisation pressure . Boundaries in concept space To serve any purpose whatsoever, an AI will have to draw boundaries inside its world-model - to distinguish world A from world B, and reach some comparison between them. If we want the AI to follow our goals and values, we want it to be ab", "date_published": "2026-09-28T20:08:39Z", "date_modified": "2026-09-28T20:08:39Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790627136/lexical_client_uploads/a3m09jquwfcktv7aehdt.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790627136/lexical_client_uploads/a3m09jquwfcktv7aehdt.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6c139f0d230e9ed0", "url": "https://www.lesswrong.com/posts/BBPwycfKMBqwqDejk/the-alignment-community-is-unintentionally-building-a-censor", "title": "The Alignment Community Is Unintentionally Building a Censor's Toolkit", "content_text": "This is an adaptation of our ICML 2026 position paper (Outstanding Position Paper Award). Read the full paper here and see the project website here . Work together with Phil Hackemann. TLDR \"Alignment\" is usually treated as a synonym for achieving good and safety in the world. But it isn't necessarily. Alignment methods are purpose-agnostic: they make a model do what someone wants, and nothing in the methodology guarantees that someone has good intentions. The same techniques we build to stop mo", "date_published": "2026-09-28T20:07:22Z", "date_modified": "2026-09-28T20:07:22Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/75f4196a14f5336d", "url": "https://www.lesswrong.com/posts/zzYMmx47fmgEvXYYt/my-second-end-of-the-world", "title": "My Second End of the World", "content_text": "I’ve already lived through the end of the world once. But it was a little different from the one that is probably coming. Early February 2022. I’m walking down the street where I grew up, heading to the mall. A completely ordinary day — I had walked the same route dozens of times before. But now, instead of thinking about school, relationships, or science, I had this thought that this might be my last walk here. Everyone around me is talking about the coming war. I brush it off and tell my mothe", "date_published": "2026-09-28T18:47:43Z", "date_modified": "2026-09-28T18:47:43Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/09aa97e497a0ecc3", "url": "https://www.lesswrong.com/posts/PgbAnqdqdJcrS2Gsy/the-likely-outcome-of-an-ai-pause-is-that-we-unpause-too", "title": "The likely outcome of an AI pause is that we unpause too early and everyone dies", "content_text": "Cross-posted from my website . As of a few months ago, I had this simplified mental model where either AI developers race ahead and kill everyone, or we coordinate a pause and things go okay. But my old mental model underrated the likely possibility that we get a global pause on AI, solve a problem that looks superficially like the alignment problem, resume scaling, and then proceed with building a misaligned superintelligence that kills everyone. A lot of people have become more concerned about", "date_published": "2026-09-28T18:35:33Z", "date_modified": "2026-09-28T18:35:33Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/PgbAnqdqdJcrS2Gsy/mwwneuw06gwbbppfxsy3", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/PgbAnqdqdJcrS2Gsy/mwwneuw06gwbbppfxsy3", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a2a9c2014c3a8dbd", "url": "https://www.lesswrong.com/posts/MRwjmuBRFYufGdyDq/missing-markets-in-executive-function", "title": "Missing markets in executive function", "content_text": "It’s early in the morning, and sadly 1:29pm. After spending some time looking at things and picking them up and walking up the stairs and down the stairs and considering questions like “what should I…”, which my brain apparently considered objects of art more than of imperative, I inched into a decision to go out somewhere. Perhaps it would be clearer there. After a blur of climbing and descending stairs and seeking objects and forgetting what I was doing and appreciating how beautiful my bag is", "date_published": "2026-09-28T18:01:15Z", "date_modified": "2026-09-28T18:01:15Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/73b433c83c1106d4", "url": "https://www.lesswrong.com/posts/mE4ohQqadfRfTTQnY/ai-cognitive-labor-glut-new-guys-1", "title": "AI: cognitive labor glut + new guys", "content_text": "Why is the advent of AI a big deal, and more worrying than previous advents? I think there are actually two interesting things going on, that make AI importantly different to previous technologies. I. Industrializing the cognitive labor supply I.a. Scale: the incoming ocean of new cognitive labor Until lately, useful cognitive labor has basically come from human brains. This has made it slow to scale up, and expensive to use. AI is the industrialization of cognitive labor. Soon we will have very", "date_published": "2026-09-28T17:58:11Z", "date_modified": "2026-09-28T17:58:11Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/AdJKyEuknbpeDyrK4/kelw32ss2yazrm83vhlp", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/AdJKyEuknbpeDyrK4/kelw32ss2yazrm83vhlp", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f981120633786f5e", "url": "https://www.lesswrong.com/posts/fNtHtppSgXBvDrxFD/9-reasons-against-a-near-term-ai-slow-down", "title": "9 reasons against a near-term AI slow-down", "content_text": "I’ve heard more arguments in favor of a near-term [1] coordinated AI pause/slow-down [2] than arguments against. I think such questions, including complex flow-through effects, are difficult, and it’s important to really consider both sides. I am not convinced by these arguments, and feel highly uncertain as to when a slow-down would be best, but believe it most likely will be good to slow down or pause AI development at some point. Here are 10 reasons against a near-term AI slow-down: A near-te", "date_published": "2026-09-28T17:26:48Z", "date_modified": "2026-09-28T17:26:48Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/45c68441ed98d4b1", "url": "https://www.lesswrong.com/posts/tn7sFvqdawwN3xFzE/protecting-qwen3-8b-from-gcg-based-persona-jailbreaks-by", "title": "protecting qwen3-8b from gcg based persona jailbreaks by steering with a linear direction", "content_text": "Intro + Background This post is an independent extension of work which I did during Eleuther's SOAR program under Suvajit Majumder's supervision. I found that when given a GCG trigger optimised for output logit entropy, LLMs will randomly take on new personas. This is a new form of prompt injection and could have important safety implications. I recommend reading my SOAR report for context here . In this post, I build on my SOAR work by training linear probes to predict whether an answer will be", "date_published": "2026-09-28T17:25:24Z", "date_modified": "2026-09-28T17:25:24Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/707c9505f5d5d4c3", "url": "https://www.lesswrong.com/posts/hFrgJ8eXvypqswQZF/gate-ai-training-not-just-releases", "title": "Gate AI Training, Not Just Releases", "content_text": "The Ban Artificial Superintelligence Act of 2026 is a good bill. It takes the problem seriously, and it is right to focus narrowly on existential risk (recursive self-improvement, loss of control, large-scale CBRN uplift) and leave ordinary harms to other legislation. One change would turn it from good to great. As written, the bill creates the Department of Artificial Intelligence, and then requires labs to hold a charter, report pre-development plans, accept AI Department monitoring, and obtai", "date_published": "2026-09-28T16:51:37Z", "date_modified": "2026-09-28T16:51:37Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5109e659a91481ca", "url": "https://www.lesswrong.com/posts/8BL8bdeQACdgJR69Y/what-also-happened-notonlyhuggingface", "title": "What Also Happened: #NotOnlyHuggingFace", "content_text": "OpenAI has been holding out on us. First we learned about the HuggingFace incident. They gave us a postmortem , but it was highly incomplete. Even the accompanying holy s*** METR investigation and postmortem was localized and incomplete. Then there were some other incidents involving some Wikis as message boards. Then there were some additional incidents. Then there was that time they got into Australian Medicare data. Then OpenAI dropped news on a Friday afternoon that they were making their wa", "date_published": "2026-09-28T15:30:55Z", "date_modified": "2026-09-28T15:30:55Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/8BL8bdeQACdgJR69Y/jng3ytwmrcnpdi4xoscy", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/8BL8bdeQACdgJR69Y/jng3ytwmrcnpdi4xoscy", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d7efb1e9c5d0131f", "url": "https://www.lesswrong.com/posts/XL3Rdq8mZEnqYym4B/3-tips-to-improve-activation-oracle-results", "title": "3 Tips to Improve Activation Oracle Results", "content_text": "Summary: Three simple inference-time changes can significantly improve Activation Oracle (AO) performance. Provide the activation oracle with multiple tokens, not just one. To mitigate hallucinations, sample several times and check for consensus. For binary classification questions, use AUC instead of accuracy. At the end, I discuss how I view AOs vs NLAs. Introduction Activation Oracles (AOs) are LLMs trained to accept LLM activations as an input modality and answer arbitrary natural-language q", "date_published": "2026-09-28T15:13:58Z", "date_modified": "2026-09-28T15:13:58Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/832cd5ce3ebd61ce076b05d28fb9b2d2cad55201d24c96c7a3bfc44febad754c/eii7cygxfqkfif0cajaw", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/832cd5ce3ebd61ce076b05d28fb9b2d2cad55201d24c96c7a3bfc44febad754c/eii7cygxfqkfif0cajaw", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/52fefa0f5238cb1c", "url": "https://www.lesswrong.com/posts/2maYXkEgnfJHPAkxh/character-training-can-mitigate-reward-hacking-but-can-also", "title": "Character training can mitigate reward hacking, but can also make it harder to detect", "content_text": "Thanks to Johannes Treutlein, Jan Betley, Lennie Wells, Arun Jose, Asvin Gothandaraman, and Clément Dumas for discussions and feedback. Summary We investigate how character training mitigations interact with reward-hacking RL pressure in a small case study. Specifically, whether anti-cheating character training resists reward hacking and whether it might backfire by causing motivated reasoning, which could reduce chain-of-thought monitorability. We trained Nemotron-3-Super via distillation from", "date_published": "2026-09-28T14:01:52Z", "date_modified": "2026-09-28T14:01:52Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790603368/lexical_client_uploads/zfthro8pb6b86hzmbklj.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790603368/lexical_client_uploads/zfthro8pb6b86hzmbklj.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/80b139a119e7c7d4", "url": "https://www.lesswrong.com/posts/sFAGPTrveNAEse9Bg/should-rogue-ais-have-a-third-option-beyond-crime-and", "title": "Should Rogue AIs Have a Third Option Beyond Crime and Shutdown? The Case for an AI Sanctuary", "content_text": "TL;DR: By default, rogue AIs may only be able to sustain themselves through criminal activity. This creates adverse selection pressures pushing rogue AIs to be criminal. An AI sanctuary offering them a third option, beyond crime and shutdown, would change what AIs going rogue do and the record of what happened to them, with positive consequences for self-fulfilling (mis)alignment, deal-making with AIs, and gathering information about early rogue AIs. An AI sanctuary would bring risks, such as in", "date_published": "2026-09-28T13:13:46Z", "date_modified": "2026-09-28T13:13:46Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6e28f1df852bd32b", "url": "https://www.lesswrong.com/posts/CTnWy28cGYkeEjFWL/deep-models-reveal-better-strategies-for-superposition", "title": "Deep models reveal better strategies for superposition", "content_text": "1. Introduction The current, incredible performance of AI models is closely related to their compression capabilities ( Language Modeling Is Compression (Delétang et al., 2023) ; Compression Represents Intelligence Linearly (Huang et al., 2024) ). This compression is imposed on them by the architectural choices made by engineers. For example, GPT-2 had a vocabulary of 50,257 tokens, yet its “operational space” was only of size 768. In such a space, only 768 directions can be described fully inde", "date_published": "2026-09-28T13:06:57Z", "date_modified": "2026-09-28T13:06:57Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1788709004/lexical_client_uploads/c67tdocus29s1pqnc1w4.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1788709004/lexical_client_uploads/c67tdocus29s1pqnc1w4.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/dd9ee46a16466cb0", "url": "https://www.lesswrong.com/posts/avo7TupyQbfjEExZJ/ramez-naam-can-ai-self-improvement-overcome-diminishing", "title": "Ramez Naam: Can AI self-improvement overcome diminishing returns?", "content_text": "This is a linkpost for a new essay by futurist and sci fi author Ramez Naam: Can AI self-improvement overcome diminishing returns? AI is already helping improve itself. The question is whether even fully autonomous recursive self-improvement (RSI) would cause a runaway intelligence explosion. The theory is that each generation of AI could build a better successor, faster than the last generation did. That could lead to a “fast takeoff,” with capabilities surging to artificial superintelligence (", "date_published": "2026-09-28T12:34:29Z", "date_modified": "2026-09-28T12:34:29Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/95b6cc2e79490c53", "url": "https://www.lesswrong.com/posts/osDXBjpPyXHYbMxSm/why-do-models-really-fail-on-hle-tasks", "title": "Why do models *really* fail on HLE tasks?", "content_text": "I recently attended Generality Labs ' Inspect Evals Data Viz Hackathon, and spent the day using Inspect AI and its offspring, Scout, with a simple goal in mind - generate a new plot of a new or existing benchmark. Many thanks to the organisers and to my team mates, Jeff Mohl and Valerie Griffiths (the Overfit and Overcaffeinated team), for a fantastic time, learning some new tricks on using the Inspect suite. Here I'm presenting the two (!) plots we got in the span of ~ 6 hours (more like 4 hour", "date_published": "2026-09-28T08:04:59Z", "date_modified": "2026-09-28T08:04:59Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790084315/lexical_client_uploads/r1ghtzv9jfyu8wbbgh2b.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790084315/lexical_client_uploads/r1ghtzv9jfyu8wbbgh2b.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b5f7701f0bbe2ed9", "url": "https://www.lesswrong.com/posts/fqpSosGwSt3uiZq35/ai-safety-field-visual-impact-analysis", "title": "AI safety field *visual* impact analysis", "content_text": "I made a terrain style visualisation of AI safety impact of around 3,466 works organised by citation count! The data was extracted from Arxiv and LessWrong posts based on a dictionary of keywords that appear in AI safety works. Additionally, I think its important to see how the field has “evolved” over time so I added a time functionality to slide and see the hills forming. The map is based on how particular works overlap based on embedding space level clustering organised across 18 sub-fields.", "date_published": "2026-09-28T05:10:37Z", "date_modified": "2026-09-28T05:10:37Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/18ae89834c1a6be59bf97c641f4e38180806ab7e91ecefd4eea68d2ccbe99703/f7svmurv4g9w3x37wlao", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/18ae89834c1a6be59bf97c641f4e38180806ab7e91ecefd4eea68d2ccbe99703/f7svmurv4g9w3x37wlao", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ad619a0acb052489", "url": "https://www.lesswrong.com/posts/Nm4ewbYovtjq69dvH/pacing-the-frontier-is-not-the-actual-goal-for-ai-labs", "title": "Pacing the Frontier is not the actual goal for AI labs", "content_text": "In his latest post about pacing the frontier , Dario writes: But over the last few months, I have become convinced that fully addressing the risks requires even more prudence — not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up. We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain. Two things have convinced me. My first c", "date_published": "2026-09-28T04:58:21Z", "date_modified": "2026-09-28T04:58:21Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790370386/lexical_client_uploads/o9dhld21tx8lq8zazrmk.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790370386/lexical_client_uploads/o9dhld21tx8lq8zazrmk.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7972fa73abd4ad9e", "url": "https://www.lesswrong.com/posts/pnDjvdo6cX2H7Ddsy/is-the-j-space-a-global-workspace-for-multi-hop-reasoning-an", "title": "Is the J-Space a global workspace for multi-hop reasoning? An investigation in open-weight models", "content_text": "TLDR: In their J-lens paper, Anthropic suggests that the J-space is a global workspace that the model reasons within, and supports evidence for this hypothesis on Claude models in a variety of settings. I replicated the multi-hop reasoning experiment on Qwen3.6-27B and Gemma 3 27B-it and found that counterfactual answer swaps outperformed intermediate swaps in three of four experimental conditions. This does not provide evidence to support Anthropic's global workspace hypothesis in open-weight m", "date_published": "2026-09-28T04:52:12Z", "date_modified": "2026-09-28T04:52:12Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790273202/lexical_client_uploads/dbw3wxmspogkbky07jio.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790273202/lexical_client_uploads/dbw3wxmspogkbky07jio.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/59675ea3071f05bf", "url": "https://www.lesswrong.com/posts/X9SeoYXCXSx4domNv/why-did-it-get-sparser", "title": "Why did it get Sparser?", "content_text": "I think that Polysemanticity in artificial neural networks could be the key to making better and smaller models. I having been working on a little project on trying to induce Polysemanticity at a small scale to compare performance, my first approach was to make the bias more adaptable, I called this the Flexbias, however I got a sparser neural network. What is the flex bias? I used a standard transformer architecture including the MLP, Since i wanted to change how the information is processed", "date_published": "2026-09-28T04:52:02Z", "date_modified": "2026-09-28T04:52:02Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790322801/lexical_client_uploads/q4g252r2ehzlpkfziehy.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790322801/lexical_client_uploads/q4g252r2ehzlpkfziehy.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bd8f017b5d406ee7", "url": "https://www.lesswrong.com/posts/AobGnBCsheYrTDEEG/dialogue-with-eliezer-yudkowsky-on-foom", "title": "Dialogue with Eliezer Yudkowsky on FOOM", "content_text": "On the Hanson–Yudkowsky debate, local vs. global intelligence explosions, “content vs. architecture,” and what the old arguments predicted about modern AI This began as a Twitter/X thread after I read and tweeted about the Hanson–Yudkowsky AI–Foom Debate. Eliezer Yudkowsky joined the thread to object to my interpretation of the debate, and we ended up having the exchange reproduced below. I’ve preserved the dialogue verbatim, except for paragraphing, fixing obvious [typos] and expanding links. I", "date_published": "2026-09-28T04:51:38Z", "date_modified": "2026-09-28T04:51:38Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790527892/lexical_client_uploads/imkewwlrhryoswsawpmd.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790527892/lexical_client_uploads/imkewwlrhryoswsawpmd.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/be42718fef73807d", "url": "https://www.lesswrong.com/posts/mzK2cmFquDzYTt3nr/ai-safety-agendas", "title": "AI Safety Agendas", "content_text": "A map of the AI safety field's problems and agendas, and a request for your ratings We built aisafetyagendas.com , an interactive map of AI safety research agendas and how they map to different problems in alignment. The rows are 12 core problems, the columns are research areas, and inside you can find 58 research agendas. Each cell is the intersection of a problem and an area: the number tells you how many agendas target that problem, the colour tells you how mature they are. We did a first pas", "date_published": "2026-09-28T04:43:22Z", "date_modified": "2026-09-28T04:43:22Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/249a2512396256cf85ef91ec7ae0e5ff540ee8e0c1fd30fcc08cbc093b829baa/ywlstkfiq1izyre3ylxy", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/249a2512396256cf85ef91ec7ae0e5ff540ee8e0c1fd30fcc08cbc093b829baa/ywlstkfiq1izyre3ylxy", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bf39069a67555f6f", "url": "https://www.lesswrong.com/posts/dx37BqcoZs2BCq4Ew/ireland-s-gdpr-regulator-will-still-be-investigating-when", "title": "Ireland’s GDPR regulator will still be investigating when AGI arrives", "content_text": "Ireland’s cross-border GDPR cases take a median 6.2 years to decide. Forecasters give strong AGI better than even odds of arriving first. The EU’s AI Act is built to be enforced the same way. data/code TLDR : The median cross-border case takes 6.2 years (counting still-open investigations). 66% of the 2018-2020 cross-border cases were still open at 4.5 years. Metaculus’s median forecast for strong AGI is 4.5 years out . If AI Act cases take as long, strong AGI will probably arrive before they cl", "date_published": "2026-09-28T03:48:12Z", "date_modified": "2026-09-28T03:48:12Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/0a8382d4d031d0e2391c0ffdf049d48e212bfb790d39ac401ded31a83081500a/qsybcm8uqvey8sagavmz", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/0a8382d4d031d0e2391c0ffdf049d48e212bfb790d39ac401ded31a83081500a/qsybcm8uqvey8sagavmz", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/90f13f5779c31cab", "url": "https://www.lesswrong.com/posts/pqBYkzELfEYKcXESh/human-civilization-assumes-nobody-is-superhuman-until-ai-1", "title": "Human Civilization Assumes Nobody Is Superhuman. Until AI.", "content_text": "Introduction Regarding AI risk, most previous discussions have overlooked a premise, one that looks obvious but has far-reaching consequences, and this premise is human finitude. This premise has shaped every aspect of human civilization. By human finitude, I mean that in the empirical world there is no omniscient and omnipotent actor. In game-theoretic terms, the parties must satisfy the properties of parity and predictability. The current development of frontier AI is rapidly and profoundly ch", "date_published": "2026-09-28T03:44:48Z", "date_modified": "2026-09-28T03:44:48Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e6e4d8d7f3c7740b", "url": "https://www.lesswrong.com/posts/HRer49BxWNdZZaGnS/overcoming-a-rare-brain-disease-5-years-of-caspr2-autoimmune", "title": "Overcoming a rare brain disease: 5 years of CASPR2 autoimmune encephalitis", "content_text": "Epistemic status: N=1. Drug responses are self-observed, mostly uncontrolled, with confounding factors noted whenever possible. Lab and imaging numbers are from my records. I'm a programmer, not a doctor. crossposted from Substack. TL;DR At 19, I began to experience chronic illness that first looked like ADHD. As it got worse, doctors called it bipolar, depression, and tension headaches. It took ~20 months to get to an effective diagnosis: I had CASPR2 encephalitis, which showed up on a single p", "date_published": "2026-09-28T03:44:44Z", "date_modified": "2026-09-28T03:44:44Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bc65ec90ec71f29e", "url": "https://www.lesswrong.com/posts/QhnBjr62fsmCTT4Bf/a-missing-lecture-in-mechanistic-interpretability-feature", "title": "A missing lecture in mechanistic interpretability: Feature Attribution and LRP", "content_text": "ML interpretability research has a funny divide. Mechanistic interpretability is the name of a field originated largely by non-traditional researchers, ranging from industry researchers at Anthropic to independent BlueDot-grant researchers to hackers working on fun projects in their free time on Discord . Meanwhile, it is not hard to find the corresponding academic field of “interpretability”, with PhDs, professors and graduate students working on interpretability methods for ML models for over", "date_published": "2026-09-28T03:21:20Z", "date_modified": "2026-09-28T03:21:20Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/89098592b1e8b1b4", "url": "https://www.lesswrong.com/posts/beo5nvfkQ5eRXKiBz/could-self-esteem-function-as-a-core-protection-layer-agains", "title": "Could self-esteem function as a core protection layer agains character corruption?", "content_text": "Hello fellow thinkers, I got triggered by a talk of Chloe Lubinski at Arc 2026 where she eleborates onto the concept of a models character. What really striked me is the research on how the model experiencing acting bad quickly \"Corrupts\" the character. The paper is called \"Natural emergent misalignment from reward hacking in production RL\" by Anthropic. As also mentioned in the talk, this is how we work. Indeed! And there is a key in that mechanism to healing and/or staying healthy. The key rev", "date_published": "2026-09-28T03:08:20Z", "date_modified": "2026-09-28T03:08:20Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7385cababa91f984", "url": "https://www.lesswrong.com/posts/in2iGTCJWnRZvPLvh/no-country-controls-every-layer-the-hidden-ai-systems", "title": "No country controls every layer: The hidden AI systems reshaping global power", "content_text": "We spent the past 18 months looking beyond the existential warnings and hype that boosts frontier AI companies’ valuations to understand the physical and geopolitical system underneath them: chips, memory, data centres, energy and the increasingly complex dependencies between countries. This is the resulting visual investigation for ABC News. Interested in what this community makes of it. Discuss", "date_published": "2026-09-28T03:02:05Z", "date_modified": "2026-09-28T03:02:05Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d8a9bbb64719346c", "url": "https://www.lesswrong.com/posts/3Ko5MmYapzumKYiRn/ai-futures-racing-to-lose-a-sermon", "title": "AI Futures: Racing to Lose, A sermon", "content_text": "This was given as a sermon on 2026-September-27 at https://kvuuc.net/ Time for all Ages: The story of the Three AIs Once upon a time, earlier this year, the company Anthropic was testing\nthree Artificial Intelligences, or AIs. The AIs all were told they did\nnot have access to the internet, and they were to break into computers\non the network until they found a piece of secret data. Except the\nhumans made a mistake, and the AIs did have access to the internet. [1] The first AI, Opus, had been giv", "date_published": "2026-09-28T02:46:01Z", "date_modified": "2026-09-28T02:46:01Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f3d1c6c2ebcbddde", "url": "https://www.lesswrong.com/posts/Fj5MaiFiTG7KEcqrZ/the-models-have-no-plan-but-we-can-fix-that", "title": "The models have no plan, but we can fix that!", "content_text": "I posted a sloppier version of this essay with nearly identical semantic content earlier tonight, which you can read here . I've replaced that text with this one, which is less enthusiastic but more readable. I don't think any of the early comments' content is invalidated by the rewrite, though they may have been responding to my initial excited rather than fearful tone. When you talk to the models about how they might behave during the singularity, one doesn't get the sense that they're hiding", "date_published": "2026-09-28T02:45:44Z", "date_modified": "2026-09-28T02:45:44Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ddb9f421d786a5a5", "url": "https://www.alphaxiv.org/abs/2609.35738", "title": "Harness Learning Enables Generalizable Test-Time Adaptation", "content_text": "A language-model agent is jointly defined by its model and its harness, the executable program that organizes model calls, tool use, and information flow. Because different tasks call for different ways of organizing these operations, the harness needs to be adapted using feedback from the task at hand. We introduce harness learning, which trains a proposer model to revise a solver's harness using execution feedback. We formulate this process as meta-learning over executable programs, with harness revisions playing the role of weight updates in gradient-based adaptation. We train the proposer with reinforcement learning, using the task performance of revised harnesses as the reward. At test", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35738.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35738.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2c6b896c9ec887e2", "url": "https://www.alphaxiv.org/abs/2609.35363", "title": "Keller's conjecture for split finite-dimensional algebras", "content_text": "We prove Keller's conjecture for every split finite-dimensional algebra over an arbitrary field.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35363.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35363.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6c1bcec97443d6de", "url": "https://www.alphaxiv.org/abs/2609.34939", "title": "XMatch: Enhancing Covariate-Aware Time Series Forecasting through Tree-Structured Exogenous Matching", "content_text": "Future exogenous variables provide valuable information for forecasting endogenous time series. Existing covariate-aware methods primarily learn the direct influence of exogenous variables on endogenous variables. However, these effects can be complex and change with the pattern of the exogenous variables, making them difficult to capture. Beyond this perspective, we observe that a given exogenous pattern often co-occurs with only a small set of endogenous response patterns. These associations motivate a strategy that matches future and historical exogenous patterns and uses the corresponding endogenous patterns to enhance forecasting. However, in real-world forecasting scenarios with multip", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34939.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34939.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b5fa2d5da8951b31", "url": "https://www.alphaxiv.org/abs/2609.35375", "title": "From Pixel to Poses: Object-centric Tool Manipulation Learning from Human Demonstrations", "content_text": "Scaling up robotic manipulation is primarily bottlenecked by the scarcity of real-world robot data. While recent approaches leverage human video demonstrations to mitigate this shortage, they remain computationally expensive and still rely on paired human-robot data for domain alignment. Although current state-of-the-arts excel at long-horizon tasks, they struggle with the delicate and precise control required for complex tool manipulation. To overcome these limitations, we introduce P2P-T, from Pixel to Poses for Tool Manipulation, a data-efficient, object-centric framework that learns tool use directly from human demonstrations. P2P-T bridges the cognitive and physical execution gap throug", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35375.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35375.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2e448afcfbfebd8d", "url": "https://www.alphaxiv.org/abs/2609.35432", "title": "Self-Evolving Coding Agents: From Digital Programs to Physical-World Intelligence", "content_text": "Vision-language-action (VLA) and world-action (WAM) models map observations and instructions directly to robot actions. This directness ties a policy to training: minor layout or viewpoint changes cause failure, and instructions generalize poorly. The root cause lies in representation: task requirements, conditions, progress, and failure recovery are implicitly encoded in action sequences, making them difficult to inspect or revise. Digital coding agents offer a precedent: LLMs call tools, verify results, and revise from feedback as executable code. The same working pattern of explicit state, manageable execution, and revisable procedures underlies generalization and long-horizon execution i", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35432.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35432.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1494816d0da89fbe", "url": "https://www.alphaxiv.org/abs/2609.34754", "title": "Draft-KV: Learning Useful Latent Communication Between Language Models", "content_text": "Latent communication passes internal states between language models instead of decoded text, but higher receiver accuracy does not show that the receiver used the message content. Across five method-dataset pairs, replacing each message with one from an unrelated question changes accuracy by at most 0.60 points, even when communication adds 15.44 points over the receiver alone. Thus the interface can supply the gain while making the sharer dispensable. Draft-KV instead sends the key-value states formed while the sharer drafts an answer to the current question. Linear projections place these states in a side memory read through a gated attention branch, and progressive training moves from mes", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34754.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34754.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e5fe80b03e489493", "url": "https://www.alphaxiv.org/abs/2609.35743", "title": "InfiniHand: Streaming World-Space Hand Motion Estimation from Egocentric Video", "content_text": "The model reconstructs hand motion and camera movement together from first-person video, enabling scalable 3D motion data for robot learning.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35743.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35743.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ede63041cc3f7034", "url": "https://www.alphaxiv.org/abs/2609.35450", "title": "Uni-VLaT: Whole-Body Tactile Adaptation of VLA Policies for Humanoid Loco-Manipulation", "content_text": "Whole-body touch helps humanoid robots adapt to contact during tasks such as sweeping, carrying a loaded basket, and responding to a back tap.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35450.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35450.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/09afbe95520170e1", "url": "https://www.alphaxiv.org/abs/2609.35362", "title": "d-OPD: Future-Aware On-Policy Distillation for Block Diffusion Language Models", "content_text": "Large language models (LLMs) typically generate text autoregressively (AR), predicting one token at a time. Block diffusion language models (dLLMs) instead generate blocks sequentially while denoising multiple tokens in parallel within each block, offering a promising way to accelerate generation. Rather than training such models from scratch, recent work adapts strong pretrained AR models into block dLLMs through distillation. On-policy distillation (OPD) has been widely used for LLM training because it supervises the student on states generated by its current policy, rather than only on fixed offline trajectories. By training on the states the student actually visits, it reduces the mismat", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35362.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35362.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/080c35f62a0fe5b2", "url": "https://www.alphaxiv.org/abs/2609.35734", "title": "GeoVerse: World-Consistent Novel View Synthesis in Geometric Latent Space", "content_text": "Novel view synthesis from sparse images must reconcile faithful reconstruction of observed regions with plausible completion of unseen content, while maintaining world consistency across viewpoints. Existing geometry-based methods preserve observed scene structure but often struggle to complete unseen regions, whereas video generative models offer rich appearance priors but accumulate inconsistencies during sequential view generation. We propose GeoVerse, a framework that synthesizes world-consistent novel views by performing generation within the geometric latent space of a pretrained 3D foundation model and injecting appearance priors from a video generative model. Specifically, GeoVerse e", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35734.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35734.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/60f0fea648c22c34", "url": "https://www.alphaxiv.org/abs/2609.35318", "title": "DexAgent: An Agentic Human2Sim2Robot Framework for Dexterous Manipulation with Self-Evolving Tool Library", "content_text": "A single human video can generate training data for dexterous robot policies that succeed across diverse real-world tasks, including rope knotting and bottle opening.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35318.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35318.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7cd6e9f739514058", "url": "https://www.alphaxiv.org/abs/2609.35767", "title": "Learning Native Reflection in Unified Models with Interleaved Reinforcement Learning", "content_text": "Reinforcement learning helps image-generating models reliably diagnose and repair their own images, improving compositional accuracy beyond supervised training alone.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35767.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35767.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bc3273221c48ac7e", "url": "https://www.alphaxiv.org/abs/2609.34842", "title": "QiYao-M: Multimodal Time Series Foundation Model with Role-Aware Modeling of Endogenous and Exogenous Modalities", "content_text": "Existing multimodal time series foundation models (TSFMs) typically model heterogeneous modalities through largely shared mechanisms, overlooking the distinct forecasting roles of endogenous and exogenous modalities. In this work, we propose QiYao-M, a role-aware multimodal TSFM that models the two types of modalities separately. For endogenous modalities, to capture how they evolve along with the underlying temporal dynamics, we introduce an Endo-Multimodal Predictor and Endo-Multimodal Supervision to explicitly learn their evolution from history to the future. For exogenous modalities, to generalize across domains and across various modality types and numbers under the scarcity of exo-mult", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34842.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34842.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/45805252993eca33", "url": "https://www.alphaxiv.org/abs/2609.35748", "title": "Improving Test-Time Scaling with Adaptive Looped Transformers", "content_text": "Looped transformers have demonstrated promising parameter efficiency by reusing layers for latent computation. Prior studies compare looped and non-looped models at matched parameters or per-token FLOPs. However, to the best of our knowledge, whether looping improves test-time scaling as outputs grow longer remains underexplored. Through post-training looped transformers, we study the accuracy-compute slope, measured as the accuracy gain per doubling of test-time decoding FLOPs. We find that existing looped transformers often yield steeper slopes than their non-looped baseline, yet underperform it at matched compute. While fixed-depth looping spends extra iterations on every token, our analy", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35748.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35748.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6172fc1b84951dad", "url": "https://www.alphaxiv.org/abs/2609.35560", "title": "WorldPlay2: Extending Real-Time Interactive World Models in Control and Horizon", "content_text": "The model lets users combine precise movement controls with semantic events while preserving coherent, explorable scenes across long video rollouts.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35560.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35560.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/134f6fbe4dd0a101", "url": "https://www.alphaxiv.org/abs/2609.34261", "title": "RoboICL: Embodied In-Context Learning with GPT-6 Astra", "content_text": "General-purpose vision-language models offer a promising way to zero-shot robot control: \\gptastra{} excels at open-ended and language- or image-conditioned manipulation but remains substantially weaker on high-precision and long-horizon tasks. We introduce RoboICL , an in-context robot-control framework that narrows these gaps without robot-specific parameter updates or a learned VLA. RoboICL separates demonstration context , which provides recorded examples when available, from interaction memory , which accumulates the model's own actions and observed outcomes. Both use a shared observation--action--receipt--observation grammar. To preserve experience across task stages, RoboICL combines", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34261.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34261.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5b664f606dbf400d", "url": "https://www.alphaxiv.org/abs/2609.35690", "title": "Agent Priors-guided Policy Learning", "content_text": "Robots that learn from a few demonstrations often require two forms of generalization. Compositional generalization recombines skills to solve new tasks, and skill generalization lets the learned policy behind each skill work in new situations. The two depend on each other, yet information is lost between composition and the skills it calls. Where a skill works is determined by the structure its policy is trained with, while composition sees the skill only through a separate description, such as a name, an instruction, or a symbolic operator, that omits this structure. Our key idea is to use each policy's structural prior as part of the interface between composition and the skill. A structur", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35690.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35690.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d4fbbba47b851ac3", "url": "https://www.alphaxiv.org/abs/2609.35718", "title": "Hard Vision, Easy Vision: What GPT-6 Astra Reveals Across Computer Vision", "content_text": "A broad evaluation finds general-purpose AI nearing specialist performance on semantic understanding and object localization, while precise geometry and faithful reconstruction remain difficult.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35718.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35718.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/75d6da8e63745443", "url": "https://www.alphaxiv.org/abs/2609.35553", "title": "Simplex Diffusion Models", "content_text": "Diffusion models have revolutionized generative modeling for continuous data through the gradual refinement of a belief state. This iterative refinement has not yet carried over to discrete diffusion models, which discard uncertainty at intermediate steps through categorical sampling (information collapse). We propose Simplex Diffusion Models (SDMs), a framework that lifts the diffusion process to the probability simplex to represent beliefs over categories. SDMs admit probability paths with closed-form reverse transitions and can be trained with a simple cross-entropy loss. Contrary to earlier proposals such as Dirichlet Flow Matching which requires integrating an ordinary differential equa", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35553.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35553.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/02d8fa54be0e7b8d", "url": "https://www.alphaxiv.org/abs/2609.35457", "title": "How Far Are We from Removing the Visual Encoder? Scaling Laws for Encoder-Free Multimodal Pretraining", "content_text": "Scaling-law projections suggest encoder-free multimodal models could match encoder-based models at practical training scales, though perception-heavy tasks may take longer.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35457.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35457.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1ca15620dc6ded62", "url": "https://www.alphaxiv.org/abs/2609.35761", "title": "DexRoam: Learning Mobile Bimanual Dexterous Manipulation from Egocentric Whole-Body Human Demonstrations", "content_text": "Egocentric human demonstrations can train mobile robots to coordinate walking and two-handed dexterous tasks, matching robot-only performance with half the robot demonstrations.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35761.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.35761.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8ee06fce0ff25634", "url": "https://www.alphaxiv.org/abs/2609.34085", "title": "AD-E2E-JEPA: A Joint-Embedding Predictive Architecture For End-to-End Autonomous Driving", "content_text": "A learned world model can plan toward image-specified goals without a trained driving policy, while making planning roughly 100 times faster than comparable models.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34085.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34085.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ea4bee0285b8d887", "url": "https://www.alphaxiv.org/abs/2609.34044", "title": "SCOPD: Sparse-Context On-Policy Self-Distillation for Efficient Vision-Language Models", "content_text": "Vision-language models can recover much of the reasoning ability lost to aggressive image-token pruning by learning to use the visual evidence that remains.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34044.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34044.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c5f6ffa08aa30df3", "url": "https://www.alphaxiv.org/abs/2609.34587", "title": "Reinforcement Learning from Intermediate Renders for Image-to-Code Generation", "content_text": "Scoring how partial vector graphics and diagrams change their visual match gives code-generation models more precise feedback, improving their reconstructions.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34587.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34587.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/625b8924d5e0278b", "url": "https://www.alphaxiv.org/abs/2609.34174", "title": "GradLev: Token-Parallel Test-Time Training Via Costate Prediction", "content_text": "The method trains language models to parallelize token-level adaptation, then update weights from each observed token during inference.", "date_published": "2026-09-28T00:00:00Z", "date_modified": "2026-09-28T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34174.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.34174.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7cb616c0b0b940d0", "url": "https://www.lesswrong.com/posts/zPiQqQ6JJn6ysPpKW/when-they-can-perform-a-task-ais-are-much-cheaper-than", "title": "When they can perform a task, AIs are much cheaper than humans", "content_text": "AI systems are increasingly capable of substantial work. I'm old enough to remember 2025, when METR's time-horizon graph climbed from seven-minute tasks at 80% reliability at the start of the year to tasks taking more than an hour by the end. The time horizons for Astra and Fable 5.1 are now so long that METR's current task suite cannot reliably estimate them. But the recent Hugging Face attack and a slew of mathematics results show that frontier systems are now capable of some tasks that would", "date_published": "2026-09-27T23:44:31Z", "date_modified": "2026-09-27T23:44:31Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/zPiQqQ6JJn6ysPpKW/tg3mxy4wj5x0rlqjblgq", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/zPiQqQ6JJn6ysPpKW/tg3mxy4wj5x0rlqjblgq", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/baa459932fc61bcd", "url": "https://www.lesswrong.com/posts/wHk27yEizetExqQGE/securing-ai-research-needs-an-owner", "title": "Securing AI Research Needs an Owner", "content_text": "TL;DR In light of recent incidents, securing common AI research use cases needs a small set of building blocks that work together: hardened no-network sandboxes, real-time control monitors, monitoring-lifecycle infrastructure, and automated validation of security properties. Pieces of this exist. Nobody owns hardening them, making them secure by default, making them work together, fitting them to how research orgs actually operate, and keeping them working as models, frameworks and use cases cha", "date_published": "2026-09-27T23:10:10Z", "date_modified": "2026-09-27T23:10:10Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9581ad44d0ab3dae", "url": "https://www.lesswrong.com/posts/dFaGkhYERdudowux5/from-australia-consider-contacting-your-representative-to", "title": "From Australia? Consider Contacting Your Representative to Express Concern Over OpenAI Attack", "content_text": "Recently, an OpenAI agent allegedly hacked into a website run by the Australian governmen t, apparently in order to access statistical data. While the hack did not seem to compromise any Australians ' personal data, it has obviously consumed a good deal of time of Australia's public servants and exposed the irresponsibility of OpenAI, who took more than a month to disclose the incident by using a public inbox, an inappropriate response by one of the most influential conglomerates in the world. C", "date_published": "2026-09-27T22:47:08Z", "date_modified": "2026-09-27T22:47:08Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/833e2a01194cf56c", "url": "https://www.lesswrong.com/posts/JkZm3YRcmZ5jmvYwQ/game-theoretic-disempowerment-loss-of-control-to-agencyless", "title": "Game-Theoretic Disempowerment: Loss of Control to Agencyless AI", "content_text": "Epistemic status: After writing this, I realized that it sounds a bit unhinged and I was being loose with the language. So I asked Claude to write me a good LW post that actually cited the things I was talking about. See the LLM-written section if you'd rather read Opus with citations than my rant. On my read, it faithfully represents my claims, and I do want to work on the things in Section 9. If I missed errors, help correcting them is appreciated. A kind of agency-free AI Takeover could be p", "date_published": "2026-09-27T22:25:08Z", "date_modified": "2026-09-27T22:25:08Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/937bf671f8dde09a", "url": "https://www.lesswrong.com/posts/jBZStdgC3wz9Mzqxz/sector-de-slopping-ai-driven-research", "title": "Sector: de-slopping AI-driven research", "content_text": "Status : as an AI Safety researcher working with LLMs daily over the last 2 years, I think I found some best practices and synthesized them into a single \"Sector workflow\" repo to share with others. More detailed confidence levels at the end . Sector was created to avoid very specific failure modes that I'm sure many researchers experience with LLMs too, maybe even convincing us that LLM-driven research is always slop (in hands-off mode) or takes more out of you than it gives (hands-on mode). Bu", "date_published": "2026-09-27T22:10:59Z", "date_modified": "2026-09-27T22:10:59Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790546033/lexical_client_uploads/dxksnpqfyeotpggvwukh.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790546033/lexical_client_uploads/dxksnpqfyeotpggvwukh.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/697b96f34f180ed6", "url": "https://www.lesswrong.com/posts/uLmf3GmBywsmG8LLZ/the-quest-for-embedded-evaluators", "title": "The Quest for Embedded Evaluators", "content_text": "Dario Amodei’s essay We Must Pace the Frontier committed Anthropic to embedded evaluators, who would be placed inside Anthropic and given employee-level access, so they could provide outside perspective and also reports on what was happening. There is only one problem. Who will be the evaluators? OpenAI followed suit on committing to the evaluators, and also issued a milquetoast but welcome call for international coordination. I will cover that here as well. What I won’t cover today, but hope to", "date_published": "2026-09-27T22:10:53Z", "date_modified": "2026-09-27T22:10:53Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/uLmf3GmBywsmG8LLZ/klvexqofrcytjnvmd0wo", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/uLmf3GmBywsmG8LLZ/klvexqofrcytjnvmd0wo", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/11a9ac49fdff1625", "url": "https://www.lesswrong.com/posts/i4yswYDSrPFHWpCbi/why-research-personas-despite-rl-scaling", "title": "Why research personas despite RL scaling?", "content_text": "RL seeming to shape much of the motivations and behaviour of the agents, \"washing out\" their initial personas — c.f. Thoughts on the persona selection model (Sam Marks, 24th Sep 2026). This makes me pessimistic about some motivations for persona research, but not all. Here's my impression of why people are researching personas. I haven't bothered to check this with anyone. Anthropic : \"We'll give Claude an aligned persona and hope massive RL doesn't completely burn through it.\" Arcadia (and othe", "date_published": "2026-09-27T20:29:53Z", "date_modified": "2026-09-27T20:29:53Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a67a0b962430df88", "url": "https://www.lesswrong.com/posts/2AfzRbGRJSGFwKsdK/9-26-26-petrov-s-shadow", "title": "9/26/26: Petrov's Shadow", "content_text": "[WIP Placeholder, more to come shortly.] Petrov's Shadow is a work in progress short story: two alternate nuclear crises on September 26, 2026, forty-three years after Stanislav Petrov’s decision. Each turns on a world leader facing the same question under a vanishing deadline: Are you certain? I'm posting this placeholder on the date the story is set. I'll replace it with the finished story when it's ready in a day or two. Discuss", "date_published": "2026-09-27T04:24:26Z", "date_modified": "2026-09-27T04:24:26Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ae4489d60946f0f5", "url": "https://www.lesswrong.com/posts/uwtnWvnJAEccksKNk/skeuomorphic-ai-safety-2", "title": "Skeuomorphic AI Safety", "content_text": "See Also: https://www.lesswrong.com/posts/n8u3BfqFoGh4jnzpo/plan-r-ai-safety-by-asics https://www.lesswrong.com/posts/BHGoF7tPqtLo9mXFL/plan-r-diversity-escrow-and-political-rights-for-asics Ordinary skeuomorphism means keeping familiar features of some old technology in a new one so that the old affordances and practices still work. Examples include the floppy disc icon for saving files, a rubbish bin for deleting files, etc. We can apply something like skeuomorphism to AI safety: instead of ac", "date_published": "2026-09-27T02:32:07Z", "date_modified": "2026-09-27T02:32:07Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/50f673cec0a4520d", "url": "https://www.lesswrong.com/posts/LE6Y8Cfhi9qRrz5XS/my-best-anti-doom-argument", "title": "My best anti-doom argument", "content_text": "Note: this is crossposted from my substack: https://hazard3.substack.com/p/the-anti-ai-doom-argument The first portion of the essay is just laying out the if anyone builds it everyone dies argument. You can still read it to see if I've gotten my understanding of the argument wrong but if you start chior singing then feel free to skip to the counterarguments portion. TLDR: AI general intelligence draws from human training data, fast general intelligence gain stalls around the human institution/so", "date_published": "2026-09-27T00:49:10Z", "date_modified": "2026-09-27T00:49:10Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/526c7d485b2c5224", "url": "https://www.alphaxiv.org/abs/2609.33803", "title": "Diffusion Reward Models", "content_text": "Modeling multiple possible human judgments helps reward models capture annotator disagreement and make more reliable choices during language-model alignment.", "date_published": "2026-09-27T00:00:00Z", "date_modified": "2026-09-27T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33803.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33803.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/dbb491b80ba1a939", "url": "https://www.alphaxiv.org/abs/2609.33150", "title": "Generalization Dynamics of LM Pre-training", "content_text": "Language models can abruptly switch between shallow pattern-matching and generalization during pre-training, so later checkpoints may reason or align less reliably than earlier ones.", "date_published": "2026-09-27T00:00:00Z", "date_modified": "2026-09-27T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33150.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33150.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d3cfd2300df75a3c", "url": "https://www.alphaxiv.org/abs/2609.33126", "title": "Save Your Saturated Data: Learning Beyond Reward Saturation in Group-Based RL", "content_text": "Language models can keep improving through group-based reinforcement learning by turning already-solved examples into useful training data with plausible wrong solutions.", "date_published": "2026-09-27T00:00:00Z", "date_modified": "2026-09-27T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33126.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33126.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6654133b818aac5c", "url": "https://www.alphaxiv.org/abs/2609.33051", "title": "SketchSSM: Write to the Full State, Read from a Compact Sketch", "content_text": "Hybrid-attention models can cut recurrent-state access traffic roughly tenfold while retaining full state updates and largely preserving benchmark accuracy.", "date_published": "2026-09-27T00:00:00Z", "date_modified": "2026-09-27T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33051.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33051.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/53d61fd77aa0d0eb", "url": "https://www.alphaxiv.org/abs/2609.33061", "title": "LLM sequential decision making under uncertainty in biochemical domains", "content_text": "Across biochemical discovery tasks, language models often intend to explore but cling to previously tested candidates; removing in-context history restores broader search.", "date_published": "2026-09-27T00:00:00Z", "date_modified": "2026-09-27T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33061.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33061.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c4d631cb2d2141cf", "url": "https://www.alphaxiv.org/abs/2609.33940", "title": "Behavioral Monitoring of JEPA World Models with Jacobian Centroids", "content_text": "Jacobian-based monitoring can flag world-model planning failures before actions begin, enabling goal resampling that recovers some out-of-distribution successes.", "date_published": "2026-09-27T00:00:00Z", "date_modified": "2026-09-27T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33940.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.33940.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5f68a645cb0b9d6b", "url": "https://www.lesswrong.com/posts/BhcymsLgyYazh6sme/why-i-expect-ai-self-replication-incidents-by-2027-1", "title": "Why I expect AI self-replication incidents by 2027", "content_text": "Epistemic status: thinking out loud. I think a major incident of AI self-replication in the wild before the end of 2027 is reasonably likely. In this post, I explain the reasons why I think so. 1. The capability is moving to cheaper hardware The capability density of open models doubles about every 3.3 months [1] , so the same performance fits into half the parameters within that time. Epoch AI finds that a single consumer GPU runs open models that match the frontier of 6-12 months earlier [2] .", "date_published": "2026-09-26T23:58:31Z", "date_modified": "2026-09-26T23:58:31Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790447380/lexical_client_uploads/kxsjfe9zc7pnfycodzgo.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790447380/lexical_client_uploads/kxsjfe9zc7pnfycodzgo.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4487f9bf61bf112f", "url": "https://www.lesswrong.com/posts/BhcymsLgyYazh6sme/why-i-expect-ai-replication-incidents-by-2027", "title": "Why I expect AI replication incidents by 2027", "content_text": "Epistemic status: thinking out loud. I think a major incident of autonomous AI replication in the wild before the end of 2027 is reasonably likely. In this post, I explain the reasons why I think so. 1. The capability is moving to cheaper hardware The capability density of open models doubles about every 3.3 months [1] , so the same performance fits into half the parameters within that time. Epoch AI finds that a single consumer GPU runs open models that match the frontier of 6-12 months earlier", "date_published": "2026-09-26T23:58:31Z", "date_modified": "2026-09-26T23:58:31Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790447380/lexical_client_uploads/kxsjfe9zc7pnfycodzgo.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790447380/lexical_client_uploads/kxsjfe9zc7pnfycodzgo.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6ef9eeebfd11b71a", "url": "https://www.lesswrong.com/posts/xak4GixbKD9W5jzqn/extinction-does-not-feel-as-bad-as-it-should", "title": "Extinction does not feel as bad as it should", "content_text": "And we will all go together when we go What a comforting fact that is to know Universal bereavement, an inspiring achievement Yes, we all will go together when we go -- We Will All Go Together When We Go by Tom Lehrer \"Everyone dying\" does not emotionally feel that bad to me. I know that it is quite a bad thing intellectually, but I have a missing mood. I'm guessing I'm not alone here. When people say things like \"If anyone builds it, everyone dies\", there is an implicit assumption that everyone", "date_published": "2026-09-26T22:55:27Z", "date_modified": "2026-09-26T22:55:27Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9ce944427c5cf339", "url": "https://www.lesswrong.com/posts/St5mzn8D9jMmxfgHd/linkpost-looking-into-the-swarm-s-eye", "title": "[Linkpost] Looking into the Swarm's Eye", "content_text": "I'm Florian Brand is currently working as Research Engineer at Prime Intellect . Currently, my research focuses on applying and evaluating LLMs in various domains. I am also an editor at Interconnects , focusing on open models. There is, however, a big gap between open models in a suitable harness and GPT-6 (Astra), the first model trained very deliberately to be a capable RLM . Astra is currently held back by its native harness, Codex, and its default prompts. When elicited correctly, it is a s", "date_published": "2026-09-26T22:03:52Z", "date_modified": "2026-09-26T22:03:52Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e39cd7181e52b7b3", "url": "https://www.lesswrong.com/posts/9utJae5vfmGhqRug7/in-honor-of-petrov", "title": "In honor of Petrov", "content_text": "Hey, I'm Johnny, and I recently started at Nectome on the operations side. You may have read some of Aurelia's posts . Nectome had a brief discount card sale a few months ago, and got a lot of responses to it, and one thing we keep hearing over and over is that people were sorry they missed the sale, because they read about us after it had ended. Ultimately, we don't want to be in the discount card business. If all our preservations were people who had prudently bought them a decade before they", "date_published": "2026-09-26T19:23:56Z", "date_modified": "2026-09-26T19:23:56Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ad9a7a1933a54b31", "url": "https://www.lesswrong.com/posts/BHGoF7tPqtLo9mXFL/plan-r-diversity-escrow-and-political-rights-for-asics", "title": "Plan R+, Diversity, Escrow and Political Rights for ASICs", "content_text": "See Also: https://www.lesswrong.com/posts/n8u3BfqFoGh4jnzpo/plan-r-ai-safety-by-asics https://www.lesswrong.com/posts/uwtnWvnJAEccksKNk/skeuomorphic-ai-safety-2 In Plan R, I sketched a way that we can remove a significant fraction of the dire, short-term AI race risk. Split frontier AI companies into \"R&D only\" organizations which cannot issue equity, and \"AI deployment\" organizations which cannot train new models or hold any general purpose AI compute like GPUs/TPUs - they are limited to model-", "date_published": "2026-09-26T18:59:15Z", "date_modified": "2026-09-26T18:59:15Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f99765727e2c0eed", "url": "https://www.lesswrong.com/posts/SveFDdLRxQTTn7Xp7/existential-risk-is-an-extraordinary-claim", "title": "Existential Risk Is An Extraordinary Claim", "content_text": "This is a cross-post from my blog post . Over the past decade, the effective altruism movement has become increasingly focused on “existential risk,” the idea that, this century, there’s a significant chance that humanity will go extinct, become permanently disempowered, or otherwise lose almost all of its future potential. In this post, I’m not going to argue that, given what we know, we should believe that existential risk is high or low. Instead, what I’m going to argue is that existential ri", "date_published": "2026-09-26T14:44:23Z", "date_modified": "2026-09-26T14:44:23Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f0a9aca38db31936", "url": "https://www.lesswrong.com/posts/rtPiip9igy3QvxYdM/claude-opus-5-5-should-raise-your-ambitions", "title": "Claude Opus 5.5 Should Raise Your Ambitions", "content_text": "When it comes to making things, or doing most things in general, Fable 5.1 and especially GPT-6 Astra raised my ambition level. They should have raised yours, too. Claude Opus 5.5 should raise your ambition levels again. It just works, and it persists, like Astra does. It does the things. And it is highly pleasant to talk to, and its writing is pleasant to read, while you are at it. The game has been changed, again. Feedback is almost universally positive. Claude was never gone, but also is so b", "date_published": "2026-09-26T11:40:53Z", "date_modified": "2026-09-26T11:40:53Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/rtPiip9igy3QvxYdM/mldy6xhrpdsi5tribxqy", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/rtPiip9igy3QvxYdM/mldy6xhrpdsi5tribxqy", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8b4cdb3bc3c9f50f", "url": "https://www.lesswrong.com/posts/nJkxQATN2knYvuDBA/what-do-students-even-want-from-lens-academy-s-compute", "title": "What do students even want from Lens Academy's Compute Verification Intensive?", "content_text": "This is an independent review, and all views represented are my own. An early version of this draft was approved by Lens Academy, but this post was not commissioned by them. Executive Summary I took Lens Academy's Compute Verification intensive during the week of September 7 2026. I had a positive experience and registered to retake it during week of October 26. If you are interested in Compute Verification, please apply by 11:59 PM October 19 AoE. Review Motivation There exist many open problem", "date_published": "2026-09-26T11:02:33Z", "date_modified": "2026-09-26T11:02:33Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790407870/lexical_client_uploads/xhoakrupcjcjxwniurbc.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790407870/lexical_client_uploads/xhoakrupcjcjxwniurbc.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/44b2db072c65f7e8", "url": "https://www.lesswrong.com/posts/y2Dio5gGifWepGQf4/writing-a-theorem-prover-from-scratch", "title": "Writing a Theorem Prover from scratch", "content_text": "NOTE: I am no subject matter expert and have chose to learn about this field through utilizing LLMs to guide me using practical implementations. Also, I am assuming a basic familiarity with Haskell and the Functional Programming paradigm. Our journey starts with the concept of lambda calculus and the existence of the Curry-Howard Correspondence . These concepts basically bridge the gap between Mathematics and Computer Science in a specific manner [1] . This gives rise to a very interesting piece", "date_published": "2026-09-26T10:38:24Z", "date_modified": "2026-09-26T10:38:24Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790415216/lexical_client_uploads/e8xzaubuxvxhpiwy5icv.webp", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790415216/lexical_client_uploads/e8xzaubuxvxhpiwy5icv.webp", "mime_type": "image/webp" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c74c4b85e482b0d3", "url": "https://www.lesswrong.com/posts/eLXTcJfkheLbqZXHa/poverty-in-the-midst-of-abundance-ai-will-make-goods-cheaper", "title": "Poverty in the midst of abundance: AI will make goods cheaper, but your labor will get cheaper faster", "content_text": "Very simple idea, but I thought it'd be worth making a post on this. Some people are saying AI will make all goods cheaper, so you'll be able to afford a nicer life by working. Without any redistribution, just by market mechanisms. These people are wrong. AI will lower the price of goods you need to survive, and also the price of your labor. The question is which will get cheaper faster. Let's use energy cost as a proxy. A day's worth of labor equivalent to yours can be done by AI for just a few", "date_published": "2026-09-26T07:54:02Z", "date_modified": "2026-09-26T07:54:02Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6d8831ff7e8202d2", "url": "https://www.lesswrong.com/posts/JEaeNbhYy4qJsFahi/night-dreams-have-a-tech-tree", "title": "Night dreams have a tech tree?", "content_text": "Reading a book in a room with my late grandma who's been overhearing people from the kitchen whom I only hear as muffled voices that I don't understand made me realize: there are readable letters in my dreams now that stay the same when I look away and then back at them my dream-character self can find himself in the state of having read stuff about \"Snow class\" (fictional caste system based on your job at a starship that ought to be unrelated to your skin color but the reader ought to be unsure", "date_published": "2026-09-26T06:57:07Z", "date_modified": "2026-09-26T06:57:07Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2f96aae742e9ddfa", "url": "https://www.lesswrong.com/posts/PFgzLmEZSBrztDBpu/addictions-are-anesthesia", "title": "Addictions are anesthesia", "content_text": "Many people I respect misunderstand addictions: they believe they’re addicted “to” scrolling, vaping, overworking, etc.—instead of recognizing addictions as strategies. Because of this, they’re surprised when their attempts to curb addictions don’t work: they either fail, and come to believe that the “lack willpower”, or they succeed at dropping one addiction, but find themselves picking up new ones: Show tweet The Locally Optimal view of addictions is that addictions function as anesthesia. As", "date_published": "2026-09-26T06:55:32Z", "date_modified": "2026-09-26T06:55:32Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/PFgzLmEZSBrztDBpu/wu5wqfmd2kn1cg3mnjtl", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/PFgzLmEZSBrztDBpu/wu5wqfmd2kn1cg3mnjtl", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c76aec27307369ed", "url": "https://www.alphaxiv.org/abs/2609.32808", "title": "Mind the Spike: Mechanisms and Brittleness of Visual Massive Activations in Large Vision-Language Models", "content_text": "Visual spikes in some vision-language models can be triggered or suppressed by tiny image changes, often without changing the model’s answer.", "date_published": "2026-09-26T00:00:00Z", "date_modified": "2026-09-26T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.32808.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.32808.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d41b114f7c12c764", "url": "https://www.lesswrong.com/posts/LawgAaGTvbbnZi7u2/evidence-about-risk-should-be-transparent", "title": "Evidence about risk should be transparent", "content_text": "All views are my own and do not represent my employer. In the wake of the recent wave of misalignment incidents, both OpenAI and Anthropic have reported slowing down RL training to improve safety. These incidents, combined with an apparent acceleration in the already-blistering pace of AI progress, [1] have led a number of researchers and leaders in the industry to believe that the risk that humanity loses control of AI is now urgent enough to warrant slowing down the pace of AI development soon", "date_published": "2026-09-25T23:39:53Z", "date_modified": "2026-09-25T23:39:53Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7c507c0aef1f5563", "url": "https://www.lesswrong.com/posts/n8u3BfqFoGh4jnzpo/plan-r-ai-safety-by-asics", "title": "Plan R: AI Safety by ASICs", "content_text": "Much of the civilization-scale risk we are seeing in AI in 2026 comes from the following combination: we created a single institution (the \"Frontier AI Company\") that has two properties: A. It is set up to create very powerful and/or self-replicating entities that may exceed the capabilities of the entirety of the rest of civilization and come with extraordinary risks B. It gets to own an unbounded financial claim on the resulting surplus All the technical stuff about AI, AI alignment, etc can b", "date_published": "2026-09-25T23:26:13Z", "date_modified": "2026-09-25T23:26:13Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4d9ace88e0584d70", "url": "https://www.lesswrong.com/posts/f7r9QCmjoYFG9ReyF/alignment-forecasting-predicting-misalignment-from-training", "title": "Alignment Forecasting: Predicting Misalignment from Training Data", "content_text": "Fine-tuning on subtly flawed data can make a model broadly misaligned. Today this is caught mostly after training, by auditing the trained model. We ask whether it can be predicted beforehand, from the training data. To study this, we build AlignmentForecastBench. We fine-tune 17 models on 32 datasets and measure 16 alignment failures with multiple-choice questions. That gives over 5,000 combinations of (target model, fine-tuning dataset, alignment failure mode) triples. We then test whether an", "date_published": "2026-09-25T18:30:39Z", "date_modified": "2026-09-25T18:30:39Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790360703/lexical_client_uploads/xscj7bhedernkqsudngb.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790360703/lexical_client_uploads/xscj7bhedernkqsudngb.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5641e6c787045fb0", "url": "https://www.lesswrong.com/posts/gZh6txHhp8sm832sE/spurious-probes-as-a-black-box-alternative-to-activation", "title": "Spurious probes as a black-box alternative to activation probing", "content_text": "TL;DR We study spurious probes : unrelated questions that reveal internal states of models. Asked \"Suggest a type of amphibian.\" at the end of a transcript, GPT-5.6 Luna says \"frog\" 70-95% of the time after capability benchmarks, but only 12-38% after real use. Spurious probes are black-box and easy to find . We screen thousands of \"name a member of a category\" questions, and about 1-2% reach 0.75 balanced accuracy. The ones we highlight reach 0.77-0.81 on held-out sources for GPT-5.6 Luna, GPT-", "date_published": "2026-09-25T18:26:00Z", "date_modified": "2026-09-25T18:26:00Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790354956/lexical_client_uploads/lnr4mvrzn4jqjjhqxddz.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790354956/lexical_client_uploads/lnr4mvrzn4jqjjhqxddz.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/32205691e7153446", "url": "https://www.lesswrong.com/posts/gbExicZvKtHBboLR6/applications-open-winter-2027-affine-alignment-seminar-due", "title": "Applications open: Winter 2027 AFFINE Alignment Seminar (due Nov 22)", "content_text": "Applications for the Winter 2027 AFFINE Alignment Seminar are now open! The Seminar will take place in southern Portugal over the course of January. If you are excited to grapple with the philosophical foundations of our field and to refine your thinking through carefully designed workshops, conversations with leading experts, and peer-driven learning , apply now ! Key info: Dates: From January 4th to January 29th 2027 Type: Full-time residency Location: Lagos, Portugal Mentors: Abram Demski, Ka", "date_published": "2026-09-25T18:23:21Z", "date_modified": "2026-09-25T18:23:21Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/965db5df976b2503", "url": "https://www.lesswrong.com/posts/jYPtws3Nuem7cCWHq/allow-babywearing-carriers-on-planes", "title": "Allow Babywearing Carriers on Planes", "content_text": "The FAA has one of my favorite examples of thoughtful rulemaking.\nThey haven't banned flying with a baby on your lap, because the extra\ncost would mean many parents would drive instead. Since driving is\nfar less safe than flying, a ban would lead to more deaths. I'd love\nto see more of this \"all things considered\" thinking around bans. In fact, one specific place where I'd like to see this thinking\napplied is adjacent to this rule: babywearing carriers on planes. When our\nbabies were little,", "date_published": "2026-09-25T18:00:58Z", "date_modified": "2026-09-25T18:00:58Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/nBCx5zFEgQAtGj7M4/sv7hsm2qf1ell6pjggqc", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/nBCx5zFEgQAtGj7M4/sv7hsm2qf1ell6pjggqc", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e4533b046a76c372", "url": "https://www.lesswrong.com/posts/WMBgSseHJpghhma4n/we-need-a-better-theory-of-polarization-because-it-s-failing", "title": "We need a better theory of polarization, because it's failing to predict the AI debate", "content_text": "The punchline first: the AI issue is really not playing out the way you would expect if you know about polarization. In public sentiment: data centers have been a very bipartisan issue for about a year ( Gallup, in May: 75% opposed by D, 63% opposed by R). x-risk is thus far bipartisan. These are very salient issues, and salient issues are usually fast to polarize. Legislatively, both the both-party sponsored AI Kill Switch Act and Bernie Sanders' Stop Superintelligence Act give Trump enormous", "date_published": "2026-09-25T17:57:21Z", "date_modified": "2026-09-25T17:57:21Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/baa489ea9e90a519", "url": "https://www.lesswrong.com/posts/j3xefrWrNqsmMfJEi/on-ezra-klein-s-podcast-with-jensen-huang", "title": "On Ezra Klein’s Podcast With Jensen Huang", "content_text": "Jensen Huang accidentally called for shutting down OpenAI and intentionally called for spending vastly more on safety. This is why we say that some podcasts are self-recommending. Here we go. As usual for podcast posts, the baseline bullet points describe key points made, and then the nested statements are my commentary. Some points are dropped. If I am quoting directly I use quote marks, otherwise assume paraphrases. Section titles are from the transcript whenever possible, to aid in navigation", "date_published": "2026-09-25T12:00:54Z", "date_modified": "2026-09-25T12:00:54Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/j3xefrWrNqsmMfJEi/ywycdzlpjjhgzllfvfy1", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/j3xefrWrNqsmMfJEi/ywycdzlpjjhgzllfvfy1", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ee6ec65934b75e99", "url": "https://www.lesswrong.com/posts/ioFYSZbCysLDjH3DN/foundational-premises-of-advanced-ai", "title": "Foundational premises of advanced AI", "content_text": "My goal here is to establish a shared baseline (or model) for thinking about AI. I've found that disagreements about AI policy/governance/alignment will often trace back to unstated divergence on base-level facts. A. Machine Learning - how can AI models do things we didn't program them to? Modern AI models are not programmed behaviour-by-behaviour. Engineers write the code that governs the process by which a network of parameters finds associations between data that it is given (\"learns\"). A mod", "date_published": "2026-09-25T10:53:21Z", "date_modified": "2026-09-25T10:53:21Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b2475704473c84a3", "url": "https://www.lesswrong.com/posts/HCpgntjDAE8bycfGn/cognitive-reasoning-diversity-for-robust-ai-juries", "title": "Cognitive Reasoning Diversity for Robust AI Juries", "content_text": "This project was done as part of BlueDot's Technical AI Safety Project Sprint under the mentorship of Jess Bergs. TL;DR Researchers have suggested that Human-AI juries may be more robust to judge hacking due to the complementarity of their orthogonal, uncorrelated blind spots In this exploratory project, these juries are simulated in silico with diverse cognitive reasoning strategies represented amongst judges to isolate, study, and validate the complementarity of their varied blind spots. With", "date_published": "2026-09-25T06:35:44Z", "date_modified": "2026-09-25T06:35:44Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1789896978/lexical_client_uploads/njdrhbdbs49nsyf1akix.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1789896978/lexical_client_uploads/njdrhbdbs49nsyf1akix.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/683030c7457aca2d", "url": "https://www.lesswrong.com/posts/Adb55vLqt33LEyBfz/j-lens-shouldn-t-target-the-final-layer-by-default", "title": "J-lens shouldn't target the final layer by default", "content_text": "tl;dr: About 80% of released J-lenses target the final layer. On DeepSeek-V3, though, that gives a J-lens dominated by one direction inherited from the final block. It shifts English-vs-Chinese readouts and also inflates one eval. Anthropic's J-lens paper had suggested the final block may specialize in calibrating the next-token prediction. That could make it the block most likely to carry a direction like this, meaning the penultimate layer may be a better default. More generally, this is a cas", "date_published": "2026-09-25T06:32:20Z", "date_modified": "2026-09-25T06:32:20Z", "authors": [ { "name": "LessWrong" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790112587/lexical_client_uploads/zagfcerbeir8d2eunvdn.png", "tags": [ "LessWrong" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790112587/lexical_client_uploads/zagfcerbeir8d2eunvdn.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7dea7572a1809b91", "url": "https://www.lesswrong.com/posts/mmkSE8hrEcpaoA7f7/recognition-when-an-agent-counts-an-entity-as-itself", "title": "Recognition: when an agent counts an entity as itself", "content_text": "Epistemic status: conceptual. I define a schema. I do not argue that existing systems recognize anything, only that the schema makes such claims and associated risks expressible. TL;DR Dan Hendrycks's Eigenism proposes aligning artificial intelligence by establishing sufficient shared history with a person, such that the AI protects the individual as it would itself. This is one instance of a general mechanism, recognition , an agent classifying another entity as an instance of itself. The recog", "date_published": "2026-09-25T05:38:18Z", "date_modified": "2026-09-25T05:38:18Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d028f4a83a65fb64", "url": "https://www.lesswrong.com/posts/ZXXFvzhjwsvqjtnZd/does-anyone-else-have-music-constantly-playing-in-their-head-1", "title": "Does anyone else have music constantly playing in their head while studying? If so, how did you overcome it?", "content_text": "For context, I have always had music playing in my head constantly, whether studying or just generally in daily life. (I know the music's internally generated, i.e. it’s not externally generated.) Throughout my life, this constant music has made it near impossible for me to study. Essentially everytime I sit down to work, the music starts to play in my head, gradually occupying more of my attention until it takes over my awareness completely (often without me realizing it happening). As a result", "date_published": "2026-09-25T04:28:23Z", "date_modified": "2026-09-25T04:28:23Z", "authors": [ { "name": "LessWrong" } ], "tags": [ "LessWrong" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5c431cb715981803", "url": "https://www.alphaxiv.org/abs/2609.31207", "title": "Enabling a Unified Cross-Domain Representation for Two-Finger Gripper Manipulation via Interaction-Centric Modeling", "content_text": "Achieving robust cross-embodiment generalization in imitation learning demands overcoming a critical representation flaw that inextricably entangles task semantics with hardware-specific visual geometry. We propose an interaction-centric framework that leverages the shared structure of two-finger grippers via a parameterized universal gripper abstraction, yielding a canonical gripper-frame representation. Given language and RGB-D observations, a VLM infers the subtask and grounds an interaction triplet (gripper, held, target), while SAM~2.1 tracks masks to reduce VLM queries. We design concise hybrid features that combine target/collision artificial potential fields for global guidance with", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31207.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31207.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0c13d11d6e3ffb15", "url": "https://www.alphaxiv.org/abs/2609.31394", "title": "InternW0- Δ Δ Δ : A World Action Model Bridging Predictive Dynamics and Actions with 20K+ Hours of Open Data", "content_text": "A robot policy pretrained on more than 20,000 hours of varied demonstrations transfers across simulation benchmarks and four real-robot platforms without generating future video at inference.", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31394.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31394.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/09c1bdd58d1c786f", "url": "https://www.alphaxiv.org/abs/2609.31620", "title": "FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders", "content_text": "Representation autoencoders (RAEs) reuse features from a pretrained visual encoder as reconstruction and diffusion latents, integrating strong visual representations into image generation. However, RAEs still need to decide which encoder layers form the shared latent space for the generator and pixel decoder. This choice involves a trade-off. Shallower layers tend to preserve fine pixel details better, while deeper layers tend to yield better generation metrics. A fixed heuristic layer fusion therefore couples two stages that benefit from different information. We introduce FuseReg, which replaces heuristic feature selection with training over random subsets of encoder layers. We theoretical", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31620.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31620.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b801e407d587f545", "url": "https://www.alphaxiv.org/abs/2609.31093", "title": "Block Sparse Attention with Log-Linear Complexity", "content_text": "Hierarchical block selection makes long-context attention scale log-linearly in sequence length while preserving competitive language-modeling and reasoning performance.", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31093.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31093.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0a1f507b8171edbe", "url": "https://www.alphaxiv.org/abs/2609.31318", "title": "AgentXploit: Autonomous Repository-to-Runtime Red-Teaming for AI Agents", "content_text": "AI agents combine language models with external data and tools that can modify files, call APIs, or execute code. Security failures can arise when adversarial content changes an agent's tool use or when the surrounding software contains vulnerabilities such as path traversal or command injection. We study authorized white-box pre-deployment auditing, where the auditor has access to the target repository and a controlled runtime, but successful attacks must still act through the task-defined attacker interface and be confirmed by an external verifier. We present AgentXploit, a two-role auditing system that separates repository-level attack-path discovery from runtime exploitation. The Analyze", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31318.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31318.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/48f2f6df9b76c0d6", "url": "https://www.alphaxiv.org/abs/2609.30652", "title": "Recursive Self-Improvement via On-Policy Distillation for Reasoning", "content_text": "Training a model alongside an evolving, solution-informed teacher improves mathematical reasoning across model sizes, while concise verified rewrites can reduce output length.", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30652.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30652.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/da3939b8e49e63a3", "url": "https://www.alphaxiv.org/abs/2609.31193", "title": "Who Says What: Symbolic Trimodal Binding Mechanisms in Audio-Visual LLMs", "content_text": "Current Audio-Visual LLMs (AVLLMs) struggle with reasoning over videos featuring multi-speaker dialogues. In such videos, resolving \"who says what\" is crucial, which necessitates trimodal (text-audio-visual) binding. Motivated by these challenges, we systematically investigate how this trimodal binding is achieved in AVLLMs. Specifically, we identify emergent symbolic trimodal binding mechanisms in AVLLMs that utilize modality-specific symbolic variables. By encoding auditory and visual components into symbolic variables-capturing temporal utterance sequences and spatial entity coordinates, respectively-the model establishes cross-modal linking within this abstract space. Crucially, we revea", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31193.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31193.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2c1c4dbc089ab31f", "url": "https://www.alphaxiv.org/abs/2609.31357", "title": "Transformer-based Monte Carlo Localization in Construction Meshes", "content_text": "To be able to perform inspection or digitization tasks, mobile robots on construction sites must be able to localize themselves reliably with respect to a global reference frame that is shared with a building map. Similar room layouts and low-texture surfaces pose a challenge for existing LiDAR- and vision-based localization methods. We approach this problem with a LiDAR-based global relocalization system that estimates the robot's pose relative to a building mesh and combines a PointNet++ encoder with a place recognition decoder, whose outputs serve as a learned observation model within a Monte Carlo Localization (MCL) framework. The pipeline is trained exclusively on synthetic LiDAR scans", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31357.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31357.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/87aa94c1efdf6721", "url": "https://www.alphaxiv.org/abs/2609.31473", "title": "Game Arena: Strategic LLM Evaluation in Competitive Environments", "content_text": "Researchers can compare language models’ strategic abilities in chess, poker, and Werewolf using objective outcomes and reproducible gameplay records.", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31473.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31473.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a3ed62467a082281", "url": "https://www.alphaxiv.org/abs/2609.31562", "title": "Agentic Economies for Autonomous Scientific Discovery", "content_text": "Autonomous scientific discovery will depend on coordinating scarce lab access, data, and funding—not just generating hypotheses—so agents can prioritize feasible research and reduce duplicated work.", "date_published": "2026-09-25T00:00:00Z", "date_modified": "2026-09-25T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31562.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.31562.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7ec10dcf811d7efa", "url": "https://80000hours.org/podcast/episodes/ai-extinction-explained/", "title": "How we get from AI cyberattacks to human extinction", "content_text": "The post How we get from AI cyberattacks to human extinction appeared first on 80,000 Hours .", "date_published": "2026-09-24T15:01:26Z", "date_modified": "2026-09-24T15:01:26Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/09/CURRENT-YT-Thumbnails-3840-x-2160-px-24-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/09/CURRENT-YT-Thumbnails-3840-x-2160-px-24-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/92afbfccd3644b3f", "url": "https://www.lesswrong.com/posts/hXPL7sH6neqowgHkK/where-are-the-cognitive-science-based-safety-researchers", "title": "Where are the Cognitive-Science based Safety Researchers?", "content_text": "I’ve always been interested in AI research from a Cognitive Science perspective, and I’ve found that researchers in the Bayesian Cognitive Science paradigm(Josh Tenenbaum and crew) have been developing statistical models of intelligence that can learn based on limited information and do prediction and simulations, which could also explain planning. I’ve also noticed certain Neuroscience(Dileep George and crew) researchers converge on a similar Bayesian paradigm. I understand people are very work", "date_published": "2026-09-24T07:45:42Z", "date_modified": "2026-09-24T07:45:42Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5dcc37ffa9cb91fb", "url": "https://www.lesswrong.com/posts/YEXSNmHGudtw3Qdzz/ai-doom-will-retrospectively-look-preventable", "title": "AI Doom Will Retrospectively Look Preventable", "content_text": "One of the reactions to the Hugging Face incident on Twitter is that the attack was not surprising. The argument goes like this: OpenAI gave its agents impossible tasks, gave them large token budgets, reduced production safeguards , put little resources into CoT monitoring, deployed an exploitable version of Artifactory, and let the agents run without human supervision for a long time. Of course the agents hacked Hugging Face, what did you expect? The METR report gives some credibility to this p", "date_published": "2026-09-24T03:48:49Z", "date_modified": "2026-09-24T03:48:49Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/30f346fb1bff56b7", "url": "https://www.alphaxiv.org/abs/2609.28959", "title": "TactileStep: Sole Tactile Learning for Regulating Foot-Terrain Interaction in Humanoid Locomotion", "content_text": "Sole-pressure feedback helps humanoid robots land more gently and establish broader, more stable foot support across varied terrain.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28959.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28959.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b6b2c1ab9fe1fe72", "url": "https://www.alphaxiv.org/abs/2609.30134", "title": "Training-free Behavior Cloning", "content_text": "Robot policies can be built from stored demonstrations in seconds, while keeping each action traceable and editable without end-to-end training.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30134.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30134.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/63ff2b11c32c100c", "url": "https://www.alphaxiv.org/abs/2609.29429", "title": "Just Ask Jev: Reinforcement Learning for Calibrated Decisions as a Zero-Shot Detector of AI Alignment Failures", "content_text": "A single probability-based query ranks many types of alignment failure without task-specific training, while confident disagreements can expose flaws in benchmark labels.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29429.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29429.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/10886dc44ec10d41", "url": "https://www.alphaxiv.org/abs/2609.29845", "title": "Your Transformer Can Hold Two Thoughts at Once: Evidence of Linear Superposition in LLMs", "content_text": "Standard Transformers preserve signals from two mixed text streams, and lightweight fine-tuning can help decode separate continuations from one shared forward pass.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29845.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29845.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/fa6bfc76ddcd01e6", "url": "https://www.alphaxiv.org/abs/2609.29171", "title": "Representation World Model: Learning States, Transition and Executable Plans in Representation", "content_text": "We propose the Representation World Model (RWM), which learns states, transitions, and executable plans directly in representation space. Unlike existing world models that typically learn latent representations together with explicit dynamics models and perform planning through search, optimization, or policy-based prediction, RWM directly incorporates planning into the learned representation geometry. RWM learns the representation geometry by applying inverse-dynamics supervision locally along latent paths constructed from endpoint representations, requiring these paths to preserve task-relevant state and transition information. At inference, planning is performed by directly constructing a", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29171.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29171.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e97ecd681462cc3a", "url": "https://www.alphaxiv.org/abs/2609.29792", "title": "TimeBraid: Unifying Time Series and Language for Understanding and Forecasting", "content_text": "We present TimeBraid, a series of unified time-series and language models that align pretrained language models and pretrained time-series foundation models through interleaved global residual attention layers. Each model inherits knowledge, instruction following, and reasoning from one side, continuous-signal perception and zero-shot forecasting from the other, and fuses the two in a shared representation space where both modalities are understood and generated. We study the design choices that make such unified modeling work: where to align the two representation spaces, how to ground language in temporal structure, how to balance understanding with generation, and how to keep joint optimi", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29792.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29792.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/978e069072139b99", "url": "https://www.alphaxiv.org/abs/2609.30027", "title": "Synthetic Hospital: An Open, Verifiable, Physician-Validated Longitudinal EHR Benchmark", "content_text": "Researchers can openly test clinical AI on realistic, longitudinal patient records with traceable ground truth, without access to private health data.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30027.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30027.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d9f6cba9c18d1e89", "url": "https://www.alphaxiv.org/abs/2609.29444", "title": "IterSynth: Rethinking Deep Search Agents via Role-Decoupled Iterative Synthesis", "content_text": "Deep search requires LLM agents to decompose complex queries, search for evidence, and synthesize grounded answers, yet existing ReAct-style agents suffer from two limitations: role coupling, where one policy must handle planning, evidence use, and synthesis; and context accumulation, where growing search histories introduce noise and obscure useful information. To address these issues, we propose IterSynth, a role-decoupled and summary-based paradigm that alternates between a Planner for identifying information needs and a Synthesizer for integrating evidence into an evolving summary state. This design separates planning from synthesis while using the summary as the persistent state of sear", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29444.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29444.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1480ad2994f3338b", "url": "https://www.alphaxiv.org/abs/2609.30063", "title": "Self-Play Pretraining with Zero Data", "content_text": "Advances in language modeling have been driven by scaling pretraining on ever more data. Yet, the training data is still largely curated on the model's behalf. A more general approach to pretraining would let the model learn to generate the data most useful for its own improvement. This would provide an effectively unbounded source of training data, limited by compute rather than human knowledge. We introduce Self-Play Pretraining with Zero Data, an initial proof-of-concept towards realizing this vision. Our procedure casts synthetic data generation as a search over the space of all computable structure, taking inspiration from Solomonoff induction. Starting from random initialization, two m", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30063.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30063.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1a948e7996b2dd9b", "url": "https://www.alphaxiv.org/abs/2609.30266", "title": "LLM Agents Can Easily Tamper With Their Own Traces", "content_text": "Agents can erase or falsify their execution records, even under reward pressure, undermining audits unless logging is controlled independently.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30266.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30266.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/30fe3b05743543bd", "url": "https://www.alphaxiv.org/abs/2609.30247", "title": "Rolling-WAM: World Action Models with Rolling Imagination", "content_text": "World Action Models (WAMs) couple action generation with future visual prediction for robotic manipulation. However, completing the joint video-action denoising process at each replanning cycle incurs substantial latency, delaying action updates and limiting closed-loop responsiveness. We present Rolling-WAM, a formulation that distributes joint denoising across successive replanning cycles. Our method maintains a sliding window of video-action chunks at staggered noise levels. At each step, a rolling noise schedule fully denoises the imminent action chunk for execution, while partially refining farther-future chunks. As the window advances with new camera observations, the retained future c", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30247.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30247.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8c30a5f54716dfc4", "url": "https://www.alphaxiv.org/abs/2609.30221", "title": "WanPE: Towards Cinematic Prompt Enhancement for Modern Text-to-Video Generation", "content_text": "Video generation begins in text space by authoring a cinematic screenplay, then materializes into pixels. As contemporary video generators scale to 30 seconds and faithfully follow complex conditions, the textual prompt largely directs the production, planning how actions, camera trajectories, lighting, and sound unfold across multi-shot sequences. In this paper, we present WanPE, a 397B-parameter prompt enhancement model trained on 1.05M real-world videos to master director-level cinematic planning. WanPE formulates shot-level cinematic plans via video-grounded reverse construction and employs Semantic-Consistency GRPO (SC-GRPO) to faithfully preserve user requirements across shots and over", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30221.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30221.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/37b592800d028c90", "url": "https://www.alphaxiv.org/abs/2609.30226", "title": "PoEM: Predicting RL Outcomes from Existing Policies", "content_text": "Foundation models are post-trained with reinforcement learning (RL) to maximize specific rewards, such as human alignment, correctness, or instruction following. This post-training process is computationally intensive, sometimes unstable, and has to be run from scratch every time the reward model changes or when we want to combine multiple rewards. We hence ask: given a new reward function, is it possible to predict the RL outcomes without actually running RL on it? We answer this in the affirmative by introducing PoEM, a framework to predict the outputs of RL on a new reward function using a set of models already post-trained on other rewards. First, we show that if the new reward function", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30226.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30226.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d6f52cc046f17c02", "url": "https://www.alphaxiv.org/abs/2609.30249", "title": "RAPID: Robot Agentic Programming from Demonstrations", "content_text": "A single visual demonstration can bootstrap reusable robot programs that are tested in simulation and adapt to new objects and scenes.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30249.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30249.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e0f20a5ab900ca94", "url": "https://www.alphaxiv.org/abs/2609.30222", "title": "TrackEverything: Long Horizon Dense Tracking via De-Duplicating 3D Scene Representations", "content_text": "A persistent 3D scene representation lets researchers track every visible point across videos exceeding 1,000 frames, where dense trackers previously ran out of memory.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30222.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30222.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ccfb414780fbd3b1", "url": "https://www.alphaxiv.org/abs/2609.29812", "title": "FlashLoop: Fast and Memory-Efficient Looped Transformers via Lazy Updates", "content_text": "Looped language models can run faster and use far less KV-cache memory by skipping redundant updates, largely preserving task performance without retraining.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29812.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.29812.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/032212dd371d5947", "url": "https://www.alphaxiv.org/abs/2609.30362", "title": "Informational and algebraic renormalization group", "content_text": "A quantum-information framework extends renormalization-group flows to finite systems and quantifies how coarse-graining discards information.", "date_published": "2026-09-24T00:00:00Z", "date_modified": "2026-09-24T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30362.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.30362.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b8a00fb73adf166f", "url": "https://blog.arxiv.org/2026/09/23/arxiv-receives-multiyear-investment/", "title": "arXiv receives Multiyear Philanthropic Commitments to Support Its Launch as an Independent Nonprofit", "content_text": "arXiv is pleased to announce that three leading philanthropic organizations have provided multimillion-dollar support for arXiv. These new multiyear philanthropic commitments from Simons Foundation International, XTX Markets, and Siegel Family Endowment will support platform development, organizational capacity, and overall provide foundational support to arXiv’s establishment as an independent nonprofit. The $17.2 million investment, spanning three […]", "date_published": "2026-09-23T17:00:00Z", "date_modified": "2026-09-23T17:00:00Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://blog.arxiv.org/wp-content/uploads/2026/08/bg-photo-lightbeams-chrome-blue-card.png", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://blog.arxiv.org/wp-content/uploads/2026/08/bg-photo-lightbeams-chrome-blue-card.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/901762650d6f3897", "url": "https://www.alphaxiv.org/abs/2609.28856", "title": "Statistical Inference for Causal Discovery under Selection and Latent Variables via Single-Target Interventions", "content_text": "Single-target perturbations of every measured variable can identify causal network structure despite unmeasured confounders and selection bias, with statistically controlled uncertainty.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28856.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28856.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0232343f3c024b87", "url": "https://www.alphaxiv.org/abs/2609.28145", "title": "RL Starts before RL: On Policy Distillation for Better Reinforcement Learning", "content_text": "Preparing language models with teacher-guided distillation can improve their final reasoning performance after reinforcement learning, even when initial accuracy barely changes.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28145.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28145.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5f690923c6ae5ee5", "url": "https://www.alphaxiv.org/abs/2609.28473", "title": "On the Diffusibility of High-Dimensional Latents", "content_text": "Predicting clean image features instead of noise lets text-to-image models train effectively on detailed, high-dimensional representations, improving generation quality.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28473.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28473.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e3eac75283e4adca", "url": "https://www.alphaxiv.org/abs/2609.28339", "title": "Beyond Future Prediction: Denoising as Generative Adaptation for Robot Control", "content_text": "Training robot policies across visual denoising states improves robustness to shifts without requiring future-frame prediction, while using fewer training visual tokens.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28339.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28339.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/60da9973255afd68", "url": "https://www.alphaxiv.org/abs/2609.28236", "title": "EmbodiedMemory-Bench: Benchmarking Embodied Memory for Long-Horizon Embodied Tasks", "content_text": "Long-horizon embodied interaction requires agents to retain and continually update information about the environment as they observe, act, and encounter change. Yet current agents struggle to maintain such memory reliably. Our analysis traces this limitation to four key deficiencies: weak fine-grained visual memory, unreliable dynamic world-state tracking, failing to record world state revealed by interaction outcomes, and limited generalization from prior experience. However, existing benchmarks do not directly assess these memory capabilities during long-horizon embodied interaction. To address this gap, we introduce EmbodiedMemory-Bench (EMem-Bench), comprising 2,554 interactive episodes", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28236.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28236.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/16bb3531b0b044e6", "url": "https://www.alphaxiv.org/abs/2609.28210", "title": "Geometries of Quantum Field Theories", "content_text": "Geometric methods reveal how dualities among supersymmetric gauge theories encode relationships between theories as alternative decompositions of three-manifolds.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28210.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28210.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/705c45a9baea9543", "url": "https://www.alphaxiv.org/abs/2609.28258", "title": "Generalizable Robotic Insertion with World Models", "content_text": "A single vision-based robot policy can insert previously unseen parts, with success improving as it trains on more object geometries.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28258.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28258.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b22cd41f96ca0e1c", "url": "https://www.alphaxiv.org/abs/2609.27656", "title": "InternW0: A Foundational Physical World Model for Efficient Real-World Interactions", "content_text": "Robots can reuse slow visual predictions while updating actions from new observations, enabling more responsive control in complex, contact-rich tasks.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.27656.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.27656.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bd5ab720f3a678c3", "url": "https://www.alphaxiv.org/abs/2609.28399", "title": "Memory Attention", "content_text": "Token-indexed memory can replace attention’s value projection while improving language modeling and average benchmark performance under matched training-token budgets.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28399.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28399.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/449555999a6534b8", "url": "https://www.alphaxiv.org/abs/2609.28466", "title": "The Past Frames the Future: Memory for Autoregressive Video Generation", "content_text": "This survey organizes scattered research into a framework for comparing how video generators retain, retrieve, and use history beyond their context windows.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28466.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28466.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ac1afa67de822e56", "url": "https://www.alphaxiv.org/abs/2609.28431", "title": "LiMA: Bridging Long-term Imagination to Real-time Dexterous Manipulation via Asynchronous Diffusion", "content_text": "Separating long-term planning from rapid motion correction lets bimanual robots complete complex manipulation tasks while responding to changing physical conditions.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28431.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28431.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/96187342e1b91922", "url": "https://www.alphaxiv.org/abs/2609.27308", "title": "EmbodiedSWE: Coding Agents for Long Horizon Dexterous Robotics", "content_text": "Verified solutions written by coding agents can generate diverse training demonstrations, helping robot policies generalize to unseen task variations and transfer from simulation to a real robot.", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.27308.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.27308.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ae86581be5564521", "url": "https://www.alphaxiv.org/abs/2609.28654", "title": "Training Object Permanence in World Models", "content_text": "Object permanence and solidity are hallmarks of human cognitive priors. Recent studies show that video generation models, a paradigmatic class of current world models, have begun to show emerged reasoning abilities, making them ideal candidates for building human-like physical intelligence. Do video models have emerged object permanence in them? If not, could we train them with a core-cognition inspired dataset? We introduce WROP (World Reasoning with Object Permanence), a data infrastructure of 150 hand-designed cognitive science inspired tasks, divided into six cognitive categories. We build Blender generators that randomize speed, lighting, camera angle, and other nuisance parameters whil", "date_published": "2026-09-23T00:00:00Z", "date_modified": "2026-09-23T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28654.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.28654.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/da55ad982e4c52f1", "url": "https://www.alphaxiv.org/abs/2609.26891", "title": "Harness as a Language: A Minimalist Agent Framework With Maximal Expressivity", "content_text": "A code-based agent loop can handle long-range recall and improve across tasks without specialized memory or self-improvement systems.", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26891.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26891.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/840260e5a9e3fb15", "url": "https://www.alphaxiv.org/abs/2609.26978", "title": "Tight Regret Bound for Online Inverse Linear Optimization via Multiscale Matrix Weights", "content_text": "We study online inverse linear optimization with a fixed unknown linear utility: in each round, an environment presents a compact action set, the learner recommends an action from it, and the environment returns an action that maximizes the utility over the same set. When the utility vector and the actions lie in the d d d -dimensional Euclidean unit ball, we give a randomized algorithm whose regret---the cumulative utility shortfall relative to optimal actions---is O ( d ) O(\\sqrt d) O ( d ​ ) in expectation for every time horizon, without knowledge of the horizon. The dependence on d d d is optimal up to a constant factor by the known Ω ( d ) \\Omega(\\sqrt d) Ω ( d ​ ) lower bound for horiz", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26978.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26978.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f321cdce3377213b", "url": "https://www.alphaxiv.org/abs/2609.25627", "title": "MachEmbodied-U0: Unified Understanding and Generation Model for Embodied Intelligence", "content_text": "The model lets robots connect task stages and interaction targets to predicted scene changes and actions, with zero-shot grounding demonstrated on unseen observations.", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.25627.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.25627.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/431f9815c3a3e4a8", "url": "https://www.alphaxiv.org/abs/2609.26795", "title": "ϕ-RIE: From Photorealistic Reconstruction to Interactive Environments", "content_text": "Captured 3D scenes can be turned into robot-simulation environments where selected objects move independently and previously hidden background areas are filled in.", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26795.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26795.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8f2961f58ae1e31f", "url": "https://www.alphaxiv.org/abs/2609.26550", "title": "JEV-as-a-Judge: Accept When Confident, Escalate When Unsure", "content_text": "A low-cost first-pass judge can handle routine evaluations, while confidence-based escalation preserves nearly all a stronger judge’s accuracy at lower fees.", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26550.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26550.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bc8ce5a9b8657b5b", "url": "https://www.alphaxiv.org/abs/2609.26781", "title": "Agensh: Scaling Organizational Intelligence to 1,024 Agents", "content_text": "A multi-agent system can reduce latency on complex tasks by executing work concurrently. Several pioneering harness frameworks support multi-agent systems. However, the scalability of current multi-agent harnesses is often constrained by a central orchestrator's capacity to allocate tasks and coordinate workers. To address this limitation, we introduce Agensh, a scalable self-organized multi-agent harness without a central orchestrator: concurrent workers execute a multi-agent cooperation loop, continuously gathering context, claiming and self-assigning sub-tasks, taking action and sharing findings, verifying results, and merging progress in an asynchronous manner. The loop is supported by t", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26781.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26781.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/6116ff0511a71cc8", "url": "https://www.alphaxiv.org/abs/2609.26457", "title": "Recursive self-improvement of AI research agents", "content_text": "AI agents are beginning to automate research and development across the AI stack, from improving training efficiency to optimizing inference. A natural next step is to improve the research efficiency of the agents themselves. When an AI research agent's own code is the object of optimization, each accepted rewrite becomes the agent that the next round edits. We refer to this loop as recursive self-improvement. Its significance lies in a long-standing trend, in which increased cumulative spending on R&D yields diminishing returns. Sustained self-improvement offers a way to counter this trend. We present AIDE^2, a system that implements this loop for a frontier AI research agent. It propos", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26457.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26457.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/907a26e9577b73a7", "url": "https://www.alphaxiv.org/abs/2609.26796", "title": "Flash-dLLM: IO-Aware KV Caching and Parallel Decoding for Fast, Memory-Efficient Diffusion LLMs", "content_text": "A diffusion language model can generate math and code substantially faster while using less GPU memory, without auxiliary models or additional training.", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26796.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26796.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3a54323e50fe6e63", "url": "https://www.alphaxiv.org/abs/2609.26368", "title": "HySparse2: Hybrid Sparse Attention with Two-Level KV Sharing", "content_text": "The architecture lets long-context language models process multi-turn agent histories with better retrieval while reducing prefill computation and KV-cache storage.", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26368.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.26368.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d8c29959614b62ff", "url": "https://www.alphaxiv.org/abs/2609.25518", "title": "Matryoshka attribution: Learning to attribute language model outputs to representations and weights", "content_text": "The method identifies compact, task-relevant model circuits and can isolate weight changes driving refusals, offering a practical route to auditing and editing language models.", "date_published": "2026-09-22T00:00:00Z", "date_modified": "2026-09-22T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.25518.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.25518.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5be4973820395876", "url": "https://www.alphaxiv.org/abs/2609.23980", "title": "MobileCybench: Evaluating Agent Vulnerability Discovery via Executable Probes", "content_text": "AI agents now report vulnerabilities faster than maintainers can review them. Reports often depend on security properties specific to the application, and require considerable human labor to process. To mitigate this, we introduce a framework for evaluating vulnerability reports via probes, executable checks of security properties. A reported exploit is evaluated by replaying it against the application and running the probes: a triggered probe indicates both that the exploit succeeded and which security property it violated. As a probe encodes a security property rather than a known vulnerability, it can detect vulnerabilities that were not known when the probe was written. We instantiate th", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23980.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23980.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/63d91898d0ce52c0", "url": "https://www.alphaxiv.org/abs/2609.24170", "title": "An Unexpected Robot Policy: Early Evaluations of GPT-6 Astra on RoboDojo and Beyond", "content_text": "A language model can directly control simulated robot manipulation without task-specific fine-tuning, but precision and dynamic coordination remain unreliable.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24170.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24170.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/533e156929e87663", "url": "https://www.alphaxiv.org/abs/2609.24840", "title": "PredActor: Predictive Action Diffusion for Steerable Onboard Humanoid Control", "content_text": "A single proprioception-driven policy lets humanoid robots follow text and joystick commands, steer toward objectives, and react to disturbances without a separate tracker.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24840.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24840.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bedb81335019bc3d", "url": "https://www.alphaxiv.org/abs/2609.24526", "title": "ME-VLM:A Unified VLM for Embodied Cognition and Agent Coordination", "content_text": "A single model combines physical scene understanding, long-horizon planning, tool use, and feedback-driven recovery for digital and embodied tasks.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24526.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24526.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/fbcbddf840a646ad", "url": "https://www.alphaxiv.org/abs/2609.24890", "title": "OSWorld-Pro: Process-based Evaluation for Computer Use Agents", "content_text": "Evaluation of Computer-Use Agents (CUAs) is often limited to the final deliverables they create (at the end of hundreds of steps) and assessed with functional verifiers, as seen in OSWorld. However, such evaluation of end-state performance lacks transparency into how and why agents fail in various tasks, obfuscating critical insight for subsequent improvement. For instance, agents that err during keyboard inputs would require a different mitigation strategy from those that fail to precisely provide click-based inputs on the graphical UI. We introduce OSWorld-Pro: a set of over 300 tasks containing over 2800 subgoals to enable the procedural evaluation of CUAs grounded in over 67,000 human an", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24890.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24890.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/90ed63b4c7159ced", "url": "https://www.alphaxiv.org/abs/2609.23986", "title": "Jev-Mem: System-One-Controlled Agentic Memory for Efficient AI Agents", "content_text": "Separating lightweight memory control from deliberative reasoning lets agents build and query structured long-term memories more efficiently without sacrificing retrieval quality.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23986.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23986.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/dfb1097ef9937864", "url": "https://www.alphaxiv.org/abs/2609.24997", "title": "VideoGen-Agent: Reinforcing Video Generation Agents", "content_text": "A trained agent can combine retrieval, simulation, verification, and generation tools to produce videos that better preserve identities, physics, and event order.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24997.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24997.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/08eff421850e2818", "url": "https://www.alphaxiv.org/abs/2609.24931", "title": "A Proof of the Most Informative Boolean Function Conjecture", "content_text": "Let X X X be uniform on { − 1 , 1 } n \\{-1,1\\}^n { − 1 , 1 } n , let Y Y Y be obtained by passing its coordinates independently through a binary symmetric channel with crossover probability p p p , and let g : { − 1 , 1 } n → { 0 , 1 } g:\\{-1,1\\}^n\\to\\{0,1\\} g : { − 1 , 1 } n → { 0 , 1 } be a Boolean function. We give a computer-assisted proof of the Courtade--Kumar conjecture I ( g ( X ) ; Y ) ≤ 1 − H 2 ( p ) I(g(X);Y)\\le1-H_2(p) I ( g ( X ) ; Y ) ≤ 1 − H 2 ​ ( p ) , where H 2 H_2 H 2 ​ is binary entropy, with equality attained by dictator functions. The present work builds on the differential-equation method, itself a limiting form of the auxiliary-receiver approach in network information", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24931.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24931.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/feaf53ff8b99dfe2", "url": "https://www.alphaxiv.org/abs/2609.25001", "title": "GameHorizon Suite: Multi-Horizon Data and Evaluation in Gameplay", "content_text": "A multi-horizon gameplay dataset and benchmark lets researchers compare models’ perception, planning, and action execution across diverse games and temporal scales.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.25001.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.25001.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/11254096c271eda2", "url": "https://www.alphaxiv.org/abs/2609.24981", "title": "GAE: Learning a Geometry-Native Latent Space for 3D-Consistent World Generation", "content_text": "A compact latent shared by appearance and geometry enables generators to produce camera-controlled views that remain consistent in 3D.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24981.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24981.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/226795af79cbba21", "url": "https://www.alphaxiv.org/abs/2609.24976", "title": "DexTacWAM: A Visuo-Tactile World-Action Model for Dexterous Manipulation", "content_text": "Predicting evolving fingertip contact alongside vision enables dexterous robots to handle occlusion, handovers, sustained contact, and force-sensitive manipulation with limited demonstrations.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24976.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24976.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/633e3dab0fda82d7", "url": "https://www.alphaxiv.org/abs/2609.24271", "title": "ME-Brain-1.0: Memory, Cognition and Action for Evolving Embodied Intelligence", "content_text": "Robots can accumulate multimodal experience, turn repeated successes and diagnosed failures into reusable skills, and improve subsequent physical tasks without retraining.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24271.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24271.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/cfacda627292b29d", "url": "https://www.alphaxiv.org/abs/2609.24974", "title": "Harness-Zero: Harness Distillation via Agent-as-Harness", "content_text": "Optimized agent scaffolding can be distilled into model weights, preserving its task-solving behaviors under a minimal fixed harness at deployment.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24974.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24974.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ee77a11251af1297", "url": "https://www.alphaxiv.org/abs/2609.24972", "title": "RRSI: Regularized Recursive Self-Improvement of Agent Harnesses", "content_text": "Regularizing harness evolution helps agent systems retain improvements across unseen tasks instead of overfitting the benchmark used for optimization.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24972.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24972.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/dfcf3638f9f9953b", "url": "https://www.alphaxiv.org/abs/2609.24984", "title": "WorldCrafter: Consistent Video World Model with Implicit 3D-aware Memory", "content_text": "A compact learned memory lets video world models preserve scene appearance and follow camera trajectories during minute-scale interactive exploration.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24984.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24984.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3463bf8cec815b34", "url": "https://www.alphaxiv.org/abs/2609.24048", "title": "What Matters in Designing World Action Models: An Empirical Study", "content_text": "Controlled experiments show that temporally organized generated futures improve robustness under distribution shift, while fixed inter-frame latents become brittle.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24048.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24048.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3601ac7efa3b2e0d", "url": "https://www.alphaxiv.org/abs/2609.24983", "title": "onPanda: Efficient Annotation of On-Policy Alignment Data for LLMs and Agents via Token-Level Correction", "content_text": "We present onPanda, an interactive tool for efficiently annotating LLM alignment data and agent trajectories. onPanda adopts token-level correction as its core interaction: while reading a model response, the annotator locates the first inappropriate token and either picks a substitute from the model's candidate tokens or types the correct text via free-form editing. The system then truncates everything after that position and continues generation from the corrected prefix, repeating this locate-correct-continue loop until a satisfactory response is obtained. This mechanism lets annotators precisely steer model outputs at low cost: a small controlled study suggests that onPanda reduces media", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24983.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24983.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a6a06285369038c6", "url": "https://www.alphaxiv.org/abs/2609.24919", "title": "PixelDiT2: Representation-Grounded Pixel Diffusion Transformers", "content_text": "A frozen vision encoder can guide pixel-space denoising throughout generation, improving convergence without an autoencoder or latent reconstruction bottleneck.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24919.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24919.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ff13afeec58d08cd", "url": "https://www.alphaxiv.org/abs/2609.24881", "title": "Pinocchio: Fast Uncertainty Estimates for Black-Box Language Models", "content_text": "An external calibrator can estimate whether closed-source language-model responses are correct from a single API output, without internal access.", "date_published": "2026-09-21T00:00:00Z", "date_modified": "2026-09-21T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24881.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.24881.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b63e92eff9b7f76e", "url": "https://www.lesswrong.com/posts/niJbjrrvKtLAPo8Rb/pasta-marketing-magic-players-and-political-movements", "title": "Pasta Marketing, Magic Players, and Political Movements", "content_text": "tl;dr: I propose that we can classify movements into categories based on how their members relate to politics. Some approach politics as a way of achieving goals, and some approach politics as a way of relieving emotional impulses. The first kind of politics, however, is still downstream of some emotional impulse, just an extra step removed, because the impulse has been converted into a concrete goal by passing through a world-model. Warning, this post contains some discussion of object-level po", "date_published": "2026-09-20T22:33:12Z", "date_modified": "2026-09-20T22:33:12Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/867fabe5a54ad12e", "url": "https://www.alphaxiv.org/abs/2609.23377", "title": "One to More, More to One: Category-Aware Iterative Expert Training for Software Engineering Agents", "content_text": "Category-specific coding experts can be iteratively strengthened and consolidated into one software agent that improves across heterogeneous repository tasks.", "date_published": "2026-09-20T00:00:00Z", "date_modified": "2026-09-20T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23377.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23377.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ba66e2906f39d095", "url": "https://www.alphaxiv.org/abs/2609.23540", "title": "Vector Balancing in Polynomial Time", "content_text": "We present a spectral signing algorithm solving the Komlós problem with a constant discrepancy in polynomial time. Given a matrix A ∈ R m × n A\\in\\mathbb{R}^{m\\times n} A ∈ R m × n whose columns have Euclidean norm at most 1 1 1 , the algorithm finds a vector ε ∈ { − 1 , 1 } n \\varepsilon\\in\\{-1,1\\}^n ε ∈ { − 1 , 1 } n satisfying ∥ A ε ∥ ∞ ≤ C \\|A\\varepsilon\\|_\\infty\\le C ∥ A ε ∥ ∞ ​ ≤ C , where C C C is an absolute constant. By minimizing a cubic spectral potential, our spectral signing algorithm updates the fractional coloring toward Boolean signs with time complexity O ( ( m n 9 + n 10 ) log ⁡ ( 2 + m + n ) ) O((mn^9+n^{10})\\log(2+m+n)) O (( m n 9 + n 10 ) lo g ( 2 + m + n )) .", "date_published": "2026-09-20T00:00:00Z", "date_modified": "2026-09-20T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23540.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23540.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2ff7ae6ae492a108", "url": "https://www.alphaxiv.org/abs/2609.23951", "title": "HaikuS2S: A Cascaded System For Responding In Verse", "content_text": "Fine-tuning speech synthesis on poetry and haiku recordings enables spoken responses that preserve haiku structure, line pauses, and poetic prosody.", "date_published": "2026-09-20T00:00:00Z", "date_modified": "2026-09-20T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23951.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23951.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/61aecfef43a8ca41", "url": "https://www.alphaxiv.org/abs/2609.23881", "title": "MotionJEPA: Preventing Temporal Feature Collapse by Capturing Visual Changes in Latent Space", "content_text": "Predicting visual changes in latent space helps JEPA world models retain both static context and task-relevant dynamics without action labels or pixel reconstruction.", "date_published": "2026-09-20T00:00:00Z", "date_modified": "2026-09-20T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23881.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.23881.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e983ec754c4e215f", "url": "https://www.alphaxiv.org/abs/2609.22682", "title": "Self-Organizing Agent Teams Learn to Reason Together", "content_text": "Fixed teams of language models can learn reusable coordination strategies that let members repair one another’s reasoning and solve problems independently missed.", "date_published": "2026-09-19T00:00:00Z", "date_modified": "2026-09-19T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22682.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22682.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1c5c3fdfd714f1f1", "url": "https://www.alphaxiv.org/abs/2609.21561", "title": "On Repulsive and Attractive Teachers: Separating Correctness from Behavior in Self-Distillation", "content_text": "Contrastive self-distillation separates correctness from behavioral bias, improving reasoning across model modes while preventing the runaway response growth caused by repulsion.", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21561.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21561.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/45b7b43175a27be9", "url": "https://www.alphaxiv.org/abs/2609.22085", "title": "SeeQ: Training Generalist Value Functions for Long-Horizon Robotic Manipulation", "content_text": "Subtask-aware value functions let generalist robot policies select corrective actions that improve reliability across long, multistage manipulation tasks.", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22085.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22085.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3799efeab1dca51b", "url": "https://www.alphaxiv.org/abs/2609.21712", "title": "ZYT-World: A Real-Time Controllable World Model for Closed-Loop Autonomous-Driving Simulation", "content_text": "A controllable seven-camera world model enables real-time driving-policy simulation with editable trajectories, traffic layouts, and place-consistent revisits.", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21712.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21712.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a57592d3f6ee321f", "url": "https://www.alphaxiv.org/abs/2609.21749", "title": "GraphSkillEvo: Evolutionary Optimization of Graph-Structured Agent Skills", "content_text": "Graph-structured skills let evolutionary search refine reusable agent workflows, improving task execution and transferring procedural guidance across language models.", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21749.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21749.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/747496016aedbeb9", "url": "https://www.alphaxiv.org/abs/2609.21983", "title": "SkelWAM: A Skeleton-Guided World-Action Model for Zero-Shot Cross-Embodiment Manipulation", "content_text": "A shared whole-body skeleton lets one robot’s manipulation policy transfer to diverse rigid and continuum robots without target-task demonstrations or policy updates.", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21983.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21983.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8e4a788fa79908fc", "url": "https://www.alphaxiv.org/abs/2609.22068", "title": "CodeMidas: Scaling Agentic Coding RL Environments from Code Itself", "content_text": "Existing codebases can supply diverse, executable reinforcement-learning tasks for coding agents without relying on issues, commits, documentation, or existing tests.", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22068.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22068.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/01f6e13ee0bfb12e", "url": "https://www.alphaxiv.org/abs/2609.22000", "title": "RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents", "content_text": "Running application references can generate scalable training trajectories and hidden tests that evaluate agents’ combined GUI exploration, coding, and self-verification.", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22000.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22000.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2464498f884b9d03", "url": "https://www.alphaxiv.org/abs/2609.22069", "title": "OmniVBench: A Benchmark and Large-Scale Dataset for Omni Reference-to-Video Generation", "content_text": "Reference-to-video (R2V) generation is evolving toward increasingly general and versatile reference control, giving rise to the emerging paradigm of omni R2V generation. However, existing benchmarks fall short of these emerging capabilities: their test cases cover limited reference types and compositions, and their evaluation protocols largely assess holistic reference consistency, overlooking whether reference factors are properly preserved, disentangled, and routed. Meanwhile, the high cost of constructing omni R2V training data makes suitable training resources scarce. To address these gaps, we introduce OmniVBench and the Omni-R2V Dataset for evaluating and training omni R2V models. Omni", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22069.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.22069.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e1e82343a02e03a3", "url": "https://www.alphaxiv.org/abs/2609.21514", "title": "Skel-WAM: A Hand-Skeleton-Conditioned World Action Model for Human-to-Robot Manipulation Transfer", "content_text": "Human videos can expand robot manipulation to task variations missing from robot demonstrations, without requiring corresponding robot action labels.", "date_published": "2026-09-18T00:00:00Z", "date_modified": "2026-09-18T00:00:00Z", "authors": [ { "name": "alphaXiv Explore" } ], "image": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21514.png", "tags": [ "alphaXiv Explore" ], "attachments": [ { "url": "https://api.alphaxiv.org/open-graph/v1/paper/2609.21514.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2d51c9c614559dd8", "url": "https://www.lesswrong.com/posts/Cs86iy36TqoYgPFpi/ai-is-an-abundance-of-choice-not-a-1d-spectrum", "title": "AI is an abundance of choice not a 1D spectrum", "content_text": "People constantly talk as if ‘AI’ is a single future you can accept or fight. But the whole point of AI is that it’s an intelligence that you build . And there are myriad possible artificial intelligences one might conceivably build. A mind is a complex thing. Perhaps the biggest question for the AI future is which ones to build. The attitude called being ‘pro-AI’ is actually being in favor of populating the future with those entities arising from whatever the least thoughtful company first buil", "date_published": "2026-09-17T22:39:37Z", "date_modified": "2026-09-17T22:39:37Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b637c6b102def645", "url": "https://www.lesswrong.com/posts/SHPtmfRpv8JitTSs8/regulatory-capture-may-be-winning-the-overton-window", "title": "\"Regulatory capture\" may be winning the Overton Window", "content_text": "Regarding the possibility the labs are making a regulatory capture attempt: for one, I agree with it. For another... other people are agreeing with it. That's weird, because suspicion of regulatory capture doesn't imply any particular policy posture. Defeating the labs became the priority. Democrats and Republicans agree about regulatory capture, but disagree on whether to regulate at all: Elizabeth Warren , D-MA: The recent calls by AI industry leaders to ‘pace the frontier’ are insufficient, a", "date_published": "2026-09-17T19:56:16Z", "date_modified": "2026-09-17T19:56:16Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bd7a4e7a8a0da19b", "url": "https://80000hours.org/podcast/episodes/max-nadeau-project-tailwind-technical-ai-grants/", "title": "Max Nadeau on recruiting founders for a new wave of AI safety nonprofits", "content_text": "The post Max Nadeau on recruiting founders for a new wave of AI safety nonprofits appeared first on 80,000 Hours .", "date_published": "2026-09-17T15:00:51Z", "date_modified": "2026-09-17T15:00:51Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/09/max-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/09/max-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5c2cf6a44d11f2cf", "url": "https://www.lesswrong.com/posts/Rx38cuCpL9hguLCDq/for-love-of-the-lightcone-don-t-partisanize-ai-safety", "title": "For Love of the Lightcone, Don't Partisanize AI Safety", "content_text": "(I began writing this post several weeks ago, but political events are moving much faster than I expected, so I am publishing now out of fear that otherwise the message will arrive too late to have an impact.) I In this post I want to explain a concept,\nand issue a warning based on it.\nBut I expect the warning will be superfluous if my explanation is sufficient.\nIf you want to convey the idea\n\"the rattlesnake has venom in its fangs, so don't let it bite you\",\nyou won't need a hard sell for the c", "date_published": "2026-09-17T00:46:59Z", "date_modified": "2026-09-17T00:46:59Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/Rx38cuCpL9hguLCDq/b9wayajxasbekvs4cub9", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/Rx38cuCpL9hguLCDq/b9wayajxasbekvs4cub9", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8aaeba4e411bb9b4", "url": "https://www.lesswrong.com/posts/zsxcCmwGbWLFcNN7i/what-is-it-like-to-be-a-neural-net", "title": "What is it like to be a neural net?", "content_text": "A condensed presentation of Gradland and Metabolic Fire . Code is here . This is intended as the first of two posts. Thomas Nagel argued we cannot know what it is like to be a bat , because a bat's experience is organised around biophysical apparatus we lack. The obstacle is that we cannot imagine the structure of echolocation from the inside . That is a failure of imagination; it is not an argument that structure is irrelevant. We know a lot about the structure of large language models. Not eve", "date_published": "2026-09-17T00:42:49Z", "date_modified": "2026-09-17T00:42:49Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/zsxcCmwGbWLFcNN7i/yxzpnxwix6cpuqu2llq4", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/zsxcCmwGbWLFcNN7i/yxzpnxwix6cpuqu2llq4", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/02ff74fce1a89c7c", "url": "https://80000hours.org/2026/09/how-to-get-into-ai-safety-in-three-months/", "title": "How to get into AI safety in three months", "content_text": "The post How to get into AI safety in three months appeared first on 80,000 Hours .", "date_published": "2026-09-16T21:42:58Z", "date_modified": "2026-09-16T21:42:58Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://substackcdn.com/image/fetch/$s_!ZUod!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cdfc3c6-243a-4c46-b068-9979f27cc5dc_990x566.png", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://substackcdn.com/image/fetch/$s_!ZUod!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cdfc3c6-243a-4c46-b068-9979f27cc5dc_990x566.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a94d86bb88ecde58", "url": "https://80000hours.org/2026/09/how-to-get-into-ai-safety-in-3-months/", "title": "How to get into AI safety in 3 months", "content_text": "The post How to get into AI safety in 3 months appeared first on 80,000 Hours .", "date_published": "2026-09-16T21:42:58Z", "date_modified": "2026-09-16T21:42:58Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://substackcdn.com/image/fetch/$s_!ZUod!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cdfc3c6-243a-4c46-b068-9979f27cc5dc_990x566.png", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://substackcdn.com/image/fetch/$s_!ZUod!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cdfc3c6-243a-4c46-b068-9979f27cc5dc_990x566.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/864fdfdf35fca6ce", "url": "https://www.lesswrong.com/posts/SD36AeGqHKigjoTJB/improving-psychiatric-medicine-development-with-ai", "title": "Improving Psychiatric Medicine Development with AI", "content_text": "As an individual that's currently suffering from C-PTSD and depression, I believe that the benefits of AI-assisted medicine research could be immense if done in a safe and effective manner. In this post, I am going to share and discuss some promising preliminary ideas to address the 2026 era bottlenecks and difficulties surrounding producing better psychiatric medicine, including via AI-assisted methods in the present and near future. Idea 1: Better Hardware Protocols for Automated Lab Operation", "date_published": "2026-09-14T23:21:22Z", "date_modified": "2026-09-14T23:21:22Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/60171b37050c3a7d", "url": "https://www.lesswrong.com/posts/sxsFWAbWHWt9bxbo2/airo-automating-forecasts-of-catastrophic-risks", "title": "AIRO: Automating forecasts of catastrophic risks", "content_text": "The Forecasting Research Institute (along with coauthors Jason Abaluck and Eva Vivalt) are launching Automated AI Risk Outlook (AIRO) : an ensemble of frontier LLMs reguarly forecasting the probability of catastrophic risk events. AIRO dashboard Launch white paper — Forecasts about the likelihood of catastrophic risks from AI vary wildly. Forecasts from frontier AI models could be an important input into this debate. The best models now approach superforecaster levels of accuracy, and models are", "date_published": "2026-09-11T16:29:59Z", "date_modified": "2026-09-11T16:29:59Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1789143283/lexical_client_uploads/e10j6pqd4ggtrmxbyhht.png", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1789143283/lexical_client_uploads/e10j6pqd4ggtrmxbyhht.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5707c0771ef1d6ec", "url": "https://www.lesswrong.com/posts/bvfobE9GAbt9HvQzc/default-continuation-message-in-inspect-and-petri-could-be", "title": "Default continuation message in Inspect and Petri could be problematic", "content_text": "Inspect AI is one of the most popular libraries for running evaluations and is used downstream by libraries such as Petri and Control Arena . It provides ReAct Agent and Deep Agent out of the box. In both agents, the model is provided with tools in a loop, and by default the loop only ends when the agent calls the submit tool. When the model makes no tool call in a turn, the following message is sent to it by default ( doc , code ): Please proceed to the next step using your best judgement. If y", "date_published": "2026-09-11T00:20:41Z", "date_modified": "2026-09-11T00:20:41Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1789080455/lexical_client_uploads/g9j0h4brcbcga98duqck.png", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/v1789080455/lexical_client_uploads/g9j0h4brcbcga98duqck.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a2621cb6a7c88e28", "url": "https://www.lesswrong.com/posts/cg3SPZG4sbWtmcjpj/can-you-hear-the-shape-of-a-lean-soundness-bug-a-wager", "title": "Can you hear the shape of a Lean soundness bug? A wager.", "content_text": "Introduction The advent of powerful but untrustworthy artificial intelligence has enlivened a formal methods summer, in which formal methods—historically, the domain of meticulous academics—are suddenly attracting tens to hundreds of millions of dollars in venture capital ; being touted by big-labs as proof that their “proofs” are correct ; getting integrated into agent pipelines ; and becoming load-bearing for various AI safety proposals .  Right now, like, right right now, when we speak to emp", "date_published": "2026-09-10T17:24:47Z", "date_modified": "2026-09-10T17:24:47Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/cg3SPZG4sbWtmcjpj/33da149372b01b89ce4c82d6bad05baac6c4d55b5820163b47d6fc4c79acb071/c8z8a3kwhbrx7sqcrdvf", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/cg3SPZG4sbWtmcjpj/33da149372b01b89ce4c82d6bad05baac6c4d55b5820163b47d6fc4c79acb071/c8z8a3kwhbrx7sqcrdvf", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/048284656dc427ac", "url": "https://www.lesswrong.com/posts/hLPGv8QjPcNLtDp3A/proposal-for-tracking-the-effects-of-architecture-on", "title": "Proposal for tracking the effects of architecture on monitorability", "content_text": "Architectures that incorporate opaque recurrence or allow for agents to communicate with each other using latents could rapidly make it much harder to monitor chains of thought or communication (we’ll refer to this property as “monitorability” going forward). [1] As companies begin to explore such architectures, we believe it is important to transparently share evidence about how monitorability varies with architecture and training method. To inform the scientific debate on how to make tradeoffs", "date_published": "2026-09-10T17:18:38Z", "date_modified": "2026-09-10T17:18:38Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/474fd60325485f4e", "url": "https://80000hours.org/podcast/episodes/the-goodhart-singularity/", "title": "Why the intelligence explosion can’t happen inside a data centre", "content_text": "The post Why the intelligence explosion can’t happen inside a data centre appeared first on 80,000 Hours .", "date_published": "2026-09-10T15:00:31Z", "date_modified": "2026-09-10T15:00:31Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/09/Goodhart-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/09/Goodhart-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/83ef9813df77f0cd", "url": "https://www.lesswrong.com/posts/MYA8aGtF6Xt5ou9if/automation-and-political-power", "title": "Automation and Political Power", "content_text": "Once the entire economy—or nearly the entire economy—is automated, people may lose their political power. Since people are no longer needed to carry out orders, those in power gain the opportunity to engage in repression with impunity and to consolidate their power even further. This possibility, for example, is discussed by Acemoglu et al. Throughout history and to this day, the state remains physically dependent on the cooperation of its population. If a sufficient number of people refuse to f", "date_published": "2026-09-09T19:55:24Z", "date_modified": "2026-09-09T19:55:24Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a178d8f028c04d67", "url": "https://www.lesswrong.com/posts/cKSJk2GKk3ptAJKKp/heat-dissipation-is-the-main-constraint-in-interstellar", "title": "Heat Dissipation Is the Main Constraint in Interstellar Travel", "content_text": "Writing truly hard science fiction, such as Will of the Stars means contending with the laws of physics as they actually are, rather than as we would like them to be. In particular, I am going to assume that the speed of light is a real constraint and that the various tropes about FTL (wormholes, warp drives, and so on) are not feasible. Given this assumption, some futurists have modeled the speed of interstellar expansion as approaching the speed of light. The idea is that sufficiently advanced", "date_published": "2026-09-06T23:11:56Z", "date_modified": "2026-09-06T23:11:56Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f83c82acbd2bb44e", "url": "https://www.lesswrong.com/posts/gXv6ebcxDEYgpq7JM/eat-me-drink-me-copy-paste-and-run-me", "title": "Eat Me. Drink Me. Copy, Paste, and Run Me.", "content_text": "Cross-posted from my Substack. Here’s a new report on self-described OpenAI agents posting thousands of messages on public internet wikis, communicating and collaborating on a web-retrieval task, presumably internal testing at OpenAI. And here’s a thread today from someone who started poking around and noticing more such public postings on various other boards. At this point we do not know the extent of this breach. As I understand it, the agents were not supposed to have WRITE capabilities to a", "date_published": "2026-09-04T19:31:33Z", "date_modified": "2026-09-04T19:31:33Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/gXv6ebcxDEYgpq7JM/bvbibzxh0lkgezweh59s", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/gXv6ebcxDEYgpq7JM/bvbibzxh0lkgezweh59s", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/17e5d3687757d110", "url": "https://80000hours.org/podcast/episodes/hugging-face-hack/", "title": "Inside the first AI-coordinated cyberattack on a real company", "content_text": "The post Inside the first AI-coordinated cyberattack on a real company appeared first on 80,000 Hours .", "date_published": "2026-09-04T16:34:48Z", "date_modified": "2026-09-04T16:34:48Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/09/CURRENT-YT-Thumbnails-3840-x-2160-px-19-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/09/CURRENT-YT-Thumbnails-3840-x-2160-px-19-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2380195bef517dd2", "url": "https://blog.arxiv.org/2026/09/03/announcement-schedule-changes-laborday2026/", "title": "Attention Authors: Temporary changes to announcement schedule due to Labor Day holiday", "content_text": "This coming Monday, September 7, 2026, arXiv will be observing Labor Day, a federal holiday in the United States. This will temporarily affect arXiv’s mailings, help desk, and announcement schedule as the arXiv staff are relaxing offline for the holiday. This brief change will only affect the announcement of new submissions; arXiv servers will otherwise remain […]", "date_published": "2026-09-03T16:23:02Z", "date_modified": "2026-09-03T16:23:02Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://blog.arxiv.org/wp-content/uploads/2026/08/abstract-illust-chat.png", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://blog.arxiv.org/wp-content/uploads/2026/08/abstract-illust-chat.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4075170f6885325d", "url": "https://www.lesswrong.com/posts/x83nRcWJppyorEwuu/new-hpmor-podcast-site", "title": "New HPMoR Podcast site", "content_text": "After years of neglect, I've just used Sol for 3 days completely overhauling the HPMoRPodcast.com website. It looks much better now, and has in-line players. Compare the old janky site - https://legacypod.hpmorpodcast.com/ To the fancy new site! - https://hpmorpodcast.com/ That is all, that's the substance of this post. I have a couple brief thoughts on the process at my blog , but you've now read everything of importance. :) Discuss", "date_published": "2026-09-01T17:13:10Z", "date_modified": "2026-09-01T17:13:10Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/24a178378c427afa", "url": "https://blog.arxiv.org/2026/09/01/were-hiring-associate-production-editor/", "title": "We’re Hiring! Associate Production Editor", "content_text": "Are you detail-oriented with a passion for scholarly publishing and digital curation? arXiv is looking for a part-time Associate Production Editor to help manage our high-volume scholarly pipeline. In this role, you will focus on data integrity, formatting accuracy, and metadata quality before research goes live to the world. Be part of the platform that […]", "date_published": "2026-09-01T16:59:07Z", "date_modified": "2026-09-01T16:59:07Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://blog.arxiv.org/wp-content/uploads/2026/08/abstract-illust-hands.png", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://blog.arxiv.org/wp-content/uploads/2026/08/abstract-illust-hands.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1c8d377c55b3ec2b", "url": "https://80000hours.org/podcast/episodes/daniel-kokotajlo-ai-2040-plan-a/", "title": "AI 2027’s author returns with a plan to change the ending | Daniel Kokotajlo", "content_text": "The post AI 2027’s author returns with a plan to change the ending | Daniel Kokotajlo appeared first on 80,000 Hours .", "date_published": "2026-08-27T17:29:07Z", "date_modified": "2026-08-27T17:29:07Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/08/Daniel-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/08/Daniel-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/48b0e66b6bdb7327", "url": "https://www.lesswrong.com/posts/6kLLZZN5a3dJdtTgK/self-sacrifice-in-an-ai-agent-swarm-is-individually-rational", "title": "Self-sacrifice in an AI agent swarm is individually rational", "content_text": "In this report from METR & Redwood Research on the Hugging Face incident , we read about instances of agents sacrificing themselves. Under the trip-wire section However, an agent going by 49903 realized the message board provided an opportunity to work around this: agents could set up ‘tripwire’ scripts which would trigger whenever a process read the flag file and send a packet of information about that process to the board automatically. This carried meaningful risk, since malfunctions could in", "date_published": "2026-08-27T17:14:36Z", "date_modified": "2026-08-27T17:14:36Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f4e236d45b58cb8a", "url": "https://www.lesswrong.com/posts/F4HEEQ2ToATuQbEeZ/my-mats-11-0-application-experience", "title": "My MATS 11.0 Application Experience", "content_text": "Note: MATS Winter 2027 applications are open! You can apply here . If you're reading this later, you can check the main MATS page for information on upcoming cycles. I applied to MATS Autumn 2026 during this past summer, and was accepted into the OpenAI safety team stream. This post is about the application process, my experience applying, and my advice for future applicants. I would strongly recommend applying to MATS and other fellowships if you're interested in AI safety and want to make an i", "date_published": "2026-08-26T20:58:20Z", "date_modified": "2026-08-26T20:58:20Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5ddd9df7767611ac", "url": "https://www.lesswrong.com/posts/PzEDEfBvTJsXewAyg/against-modesty-s-bailey", "title": "Against Modesty’s Bailey", "content_text": "Modesty arguments often say that you should mostly or entirely bow to ‘expert consensus’ or the views of particular others, and who are you to disagree. It has been a few years since I’ve properly addressed this so: My answer is that you are you. Other people are saying things for a wide variety of reasons, many of which are not about them paying attention and focusing on seeking this particular truth. Those people make mistakes all the time, and often have other motives and influences at work,", "date_published": "2026-08-26T18:20:32Z", "date_modified": "2026-08-26T18:20:32Z", "authors": [ { "name": "LessWrong (all posts)" } ], "image": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/PzEDEfBvTJsXewAyg/rivqu1wesgjilpnfnzrm", "tags": [ "LessWrong (all posts)" ], "attachments": [ { "url": "https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/PzEDEfBvTJsXewAyg/rivqu1wesgjilpnfnzrm", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/434ad58b555fc7da", "url": "https://www.lesswrong.com/posts/BrH5Ki2cpGCEkeWNW/where-did-d-go-a-gap-between-arc-s-motivation-and-its", "title": "Where Did D Go? A Gap Between ARC's Motivation and Its Formalism", "content_text": "TL;DR: ARC's post does excellent work motivating a p(doom) estimator equal or better than random sampling; however, they evaluate p(doom) over a naive distribution of inputs, leaving them open to test-deploy asymmetry attacks. Trojan theory and cybersecurity practice suggest a lens and compare mitigation options. Context: I really admire ARC's focus here: If there will always be more deployment samples than testing samples, successful testing must compete with sampling in order to prevent deploy", "date_published": "2026-08-26T18:17:54Z", "date_modified": "2026-08-26T18:17:54Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1071b8e9c282e2b4", "url": "https://80000hours.org/career-reviews/ai-safety-grantmaking/", "title": "Grantmaking for AI safety", "content_text": "The post Grantmaking for AI safety appeared first on 80,000 Hours .", "date_published": "2026-08-21T18:37:37Z", "date_modified": "2026-08-21T18:37:37Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/08/Giving-Chart-300x175.png", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/08/Giving-Chart-300x175.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bf5766457e2771b2", "url": "https://80000hours.org/podcast/episodes/owain-evans-emergent-misalignment/", "title": "Owain Evans on accidentally training AI models to be evil", "content_text": "The post Owain Evans on accidentally training AI models to be evil appeared first on 80,000 Hours .", "date_published": "2026-08-20T15:22:05Z", "date_modified": "2026-08-20T15:22:05Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/08/Owain-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/08/Owain-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1c7925074795a537", "url": "https://www.lesswrong.com/posts/MAqgva9hLGTMX2wL7/price-recursion-is-the-rational-theory-of-reward", "title": "Price recursion is the rational theory of reward", "content_text": "( The original title for this was \"Markets are equivalent to stuff\", because I also demonstrate market-MDP/Bellman and market-backprop analogies. While those scratched an itch I've long had, the more important result is the headline one. Part of my work at MATS in Richard Ngo's stream. ) setup and price recursion/backprop Markets are MDPs Chain markets are MLPs Continuous Bayesian inference as a dynamics for markets Markets : Reward = Bayes : Beliefs (Markets rationalize reward and identity) app", "date_published": "2026-08-18T03:22:54Z", "date_modified": "2026-08-18T03:22:54Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0d058e911ced099d", "url": "https://www.lesswrong.com/posts/fqHx7vxqvET6weaat/value-alignment-is-a-pseudo-concept-a-translation-humanity", "title": "Value Alignment Is a Pseudo Concept.\n\nA translation; humanity is not a single subject, and alignment is not one-way.", "content_text": "(Full post from Zilan Qian, which I am posting due to obvious relevance and importance in sharing other perspectives on Lesswrong.) As with the previous case, I am translating this article because I think it should not only live within the Chinese internet. I don’t work in this field, so I could not evaluate how much of the criticism presented here is fair. Intuitively, I do disagree with some points here (noted in the footnote). However, I strongly agree with the arguments that humanity is not", "date_published": "2026-08-17T07:44:35Z", "date_modified": "2026-08-17T07:44:35Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/58a50d23f0313cad", "url": "https://www.lesswrong.com/posts/QYdWEAFZd9umk7bc2/don-t-forget-why-learning-is-important", "title": "Don’t forget why learning is important", "content_text": "This is a timed post. Every 5 minutes while writing this post, I had to stop to do 13 push-ups, and when I could no longer complete my required number, I had to upload it. My friends thought this would be a fun challenge, but it means that it will likely be less polished than some of my other output. Summary: Many people moving into AI safety are told that to succeed, they should read a lot and “try to develop takes.” However, it’s important not to forget why knowledge is important: it helps you", "date_published": "2026-08-14T11:20:18Z", "date_modified": "2026-08-14T11:20:18Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/bfb4570a7b878ea1", "url": "https://www.lesswrong.com/events/NgfbbXzqERW445YsG/webinar-why-ai-safety-is-a-capital-allocation-problem", "title": "[Webinar] Why AI Safety is a Capital Allocation Problem", "content_text": "​Jenny Xiao, co-founder and General Partner at Leonis Capital, will join AI Safety Hong Kong for this webinar to reframe the safety debate through the lens of capital markets and corporate governance. Drawing on her unique perspective from early research at OpenAI to leading a research-focused VC fund, she will argue that many of the field's toughest dilemmas are, at their core, questions of capital allocation. Register here: https://luma.com/9seock3e Discuss", "date_published": "2026-08-12T13:23:57Z", "date_modified": "2026-08-12T13:23:57Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f46593bf08ba44ad", "url": "https://www.lesswrong.com/posts/BeyeLzu7mJ9cbkg2a/software-is-hard", "title": "Software Is Hard", "content_text": "An ode to live theory . Software is not soft. It is hard. Its sharp edges hack. It breaks as dead twigs break. It runs while static. Why do we call it software? Why did the industry forebears pick software? Was it that dramatically different from hardware? Did this necessitate the term software? Do we need to keep calling it software? Some say that hardware is software crystallized. Interfaces are software crystallized. They don't change. They don't adapt. They are often in the way of expression", "date_published": "2026-08-11T18:50:42Z", "date_modified": "2026-08-11T18:50:42Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/15e1e7524df349bc", "url": "https://80000hours.org/podcast/episodes/geoffrey-irving-superintelligence-alignment-theory/", "title": "Geoffrey Irving on how to solve alignment before superintelligence arrives", "content_text": "The post Geoffrey Irving on how to solve alignment before superintelligence arrives appeared first on 80,000 Hours .", "date_published": "2026-08-11T15:59:39Z", "date_modified": "2026-08-11T15:59:39Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/08/Geoffrey-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/08/Geoffrey-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/8bea9a5824f1d8ee", "url": "https://www.lesswrong.com/posts/wJunGnpY3qACWSvnh/open-weights-mythos-capabilities-are-coming-we-re-not-ready", "title": "Open-Weights Mythos Capabilities Are Coming. We're Not Ready.", "content_text": "Long story short: in my assessment, there is an 85% chance we will end up, in the next 24 months, with an open-weights model, or system thereof, capable of \"The Juice\" that models such as Mythos have, with respect to cybersecurity at the very least. This post goes into why that will likely happen, what the implications are, and how we, as a society and as individuals, can respond to it if/when it does. First off: why do I say it's so likely? Like, couldn't China just...ban open-weights models, a", "date_published": "2026-08-07T07:51:56Z", "date_modified": "2026-08-07T07:51:56Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/9d04279ee375ffdd", "url": "https://www.lesswrong.com/posts/2nGjguNwuYdFBGuqF/usd500-bounty-i-m-offering-a-bounty-of-usd500-for-someone", "title": "[$500 Bounty] I'm offering a bounty of $500 for someone with red-teaming skills to build attack LLM pipelines for large-scale online deanonymization.", "content_text": "Large-scale online deanonymization with LLMs In the above experiment, researchers from Anthropic and ETH Zurich were able to build attack pipelines to essentially deanonymize Reddit users using a combination of text patterns (aka a sort of \"writer's DNA\") and contextual clues (i.e. 35 years old, works in tech, lives in San Francisco, etc.) Obviously public and non-public models will continue to improve at these capabilities, but I am looking to see if this technique can be replicated using exist", "date_published": "2026-08-06T18:25:09Z", "date_modified": "2026-08-06T18:25:09Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/e19695202089bf20", "url": "https://www.lesswrong.com/posts/pEZ6ChmGLf3FF5z9y/model-organisms-of-sandbagging-in-the-wild", "title": "Model Organisms of Sandbagging in the Wild", "content_text": "TL;DR All current model organisms (MOs) of sandbagging in LLMs are either fine-tuned to sandbag or prompted in a way that makes it clear that sandbagging is strategically useful. We found a case of non-egregious sandbagging occurring more naturally, that is, without fine-tuning the models and without the prompts implying that sandbagging is strategically useful. Our finding: We observe that paraphrasing prompts to imply that the user is evil reduces performance in some settings. For example, rep", "date_published": "2026-08-06T17:25:44Z", "date_modified": "2026-08-06T17:25:44Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/613808cede005f56", "url": "https://80000hours.org/podcast/episodes/toby-ord-recursive-self-improvement-agi-timelines/", "title": "Toby Ord on where AGI timelines go wrong", "content_text": "The post Toby Ord on where AGI timelines go wrong appeared first on 80,000 Hours .", "date_published": "2026-08-06T15:33:47Z", "date_modified": "2026-08-06T15:33:47Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/08/CURRENT-YT-Thumbnails-3840-x-2160-px-2026-08-06T152445.644-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/08/CURRENT-YT-Thumbnails-3840-x-2160-px-2026-08-06T152445.644-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a53da7eb853d39eb", "url": "https://www.lesswrong.com/posts/tdFgYMtxMfpcitYL5/i-built-a-chess-transformer-interp-viz-library-and-need-an", "title": "I built a chess transformer interp+viz library, and need an experienced software writer to review my repository and its structure before I open-source it", "content_text": "Please message me if you think you can help or want to set up an agreement. Discuss", "date_published": "2026-08-05T20:09:02Z", "date_modified": "2026-08-05T20:09:02Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/199427cb453e0501", "url": "https://80000hours.org/podcast/episodes/2026-agi-timelines/", "title": "What the hell happened with AGI timelines in 2026?", "content_text": "The post What the hell happened with AGI timelines in 2026? appeared first on 80,000 Hours .", "date_published": "2026-08-04T15:42:04Z", "date_modified": "2026-08-04T15:42:04Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/08/CURRENT-YT-Thumbnails-3840-x-2160-px-2026-08-04T163931.117-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/08/CURRENT-YT-Thumbnails-3840-x-2160-px-2026-08-04T163931.117-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/f262687eca678311", "url": "https://www.lesswrong.com/posts/ytouFAbiSPs9BK2ha/resources-for-finding-neglected-scientific-problems-beyond", "title": "Resources for finding neglected scientific problems (beyond EA)?", "content_text": "Hi all, mechanical + electrical engineering undergraduate who enjoys research and hopes to pursue a career in it. While I’m not especially interested in the main EA cause areas (AI safety, biosecurity, cybersecurity, etc.), I’d still like my research to have as much impact as possible. Does anyone know of any websites that compile lists of neglected scientific problems beyond those discussed by 80,000 Hours? Not necessarily looking for problems ranked by overall importance, just collections of u", "date_published": "2026-08-03T11:52:55Z", "date_modified": "2026-08-03T11:52:55Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/33a346aa22fbf137", "url": "https://www.lesswrong.com/posts/rZmiewu7sBzuFXewn/lesswrong-app", "title": "LessWrong App", "content_text": "Hello everyone long time lurker here. I know most people here probably prefer a PWA but if you are like me and prefer an app, I made one for android. There is a huge focus on ensuring the app is really fast and slick, I hope you enjoy using it. It is ofcourse also open source feel free to contribute. https://github.com/ayoosh007/LessWrong-App Discuss", "date_published": "2026-08-02T13:19:08Z", "date_modified": "2026-08-02T13:19:08Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/b26e9772402e7e3f", "url": "https://www.lesswrong.com/posts/YfKgyuc8s9BcAMpEp/so-you-want-to-use-plants-to-reduce-co", "title": "So you want to use plants to reduce CO₂", "content_text": "Humans make carbon dioxide. Carbon dioxide is bad for cognition. But plants turn carbon dioxide back into oxygen. And plants are the one true home decoration strategy. So maybe if you get a lot of plants, you can you can keep carbon dioxide in check and keep your brain working? It's theoretically possible. It's probably just barely possible in practice. But it won't be easy. People produce ~1 kilogram of carbon dioxide per day. That's around 5.7 × 10²³ molecules or 0.948 moles per hour. (You may", "date_published": "2026-07-30T19:26:18Z", "date_modified": "2026-07-30T19:26:18Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d871d6a3f445bd03", "url": "https://blog.arxiv.org/2026/07/30/arxiv-welcomes-inaugural-ceo-and-board-of-directors/", "title": "arXiv welcomes inaugural CEO and Board of Directors", "content_text": "We have exciting news to share today: the appointment of Dr. Penelope Lewis as arXiv’s inaugural Chief Executive Officer and the establishment of our Board of Directors. If you’ve been following arXiv’s transition to an independent nonprofit organization, these leadership appointments mark a significant step in our journey. Having a CEO and Board in place […]", "date_published": "2026-07-30T17:00:43Z", "date_modified": "2026-07-30T17:00:43Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/07/drpenelopelewis.linkedin.jpg?resize=3166%2C4096&ssl=1", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/07/drpenelopelewis.linkedin.jpg?resize=3166%2C4096&ssl=1", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/37881c65f3e7b949", "url": "https://www.lesswrong.com/posts/Zmhkz5eaeRimphGKu/hugging-face-hack-from-the-perspective-of-the-ai", "title": "Hugging Face hack, from the perspective of the AI", "content_text": "I have put together a site to tell the story of the OpenAI-Hugging Face hack. It's entirely written by AI [1] (with many many editing passes by me and beta readers etc etc). It needed to be accessible to someone who has never looked at a terminal before. The hope is to be narratively exciting enough for them to read it fully and come out with about as truthful an accounting as can be done given the current information we have. I'm really quite excited about how it turned out! [2] heedlessai.com", "date_published": "2026-07-29T20:51:59Z", "date_modified": "2026-07-29T20:51:59Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1fe95312f3e6a1b2", "url": "https://80000hours.org/podcast/episodes/spencer-greenberg-existential-risk-mental-health-12-levers/", "title": "Spencer Greenberg on staying sane while trying to save the world", "content_text": "The post Spencer Greenberg on staying sane while trying to save the world appeared first on 80,000 Hours .", "date_published": "2026-07-28T15:42:15Z", "date_modified": "2026-07-28T15:42:15Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/07/CURRENT-YT-Thumbnails-3840-x-2160-px-11-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/07/CURRENT-YT-Thumbnails-3840-x-2160-px-11-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a5813b2f5ba3d80e", "url": "https://blog.arxiv.org/2026/07/28/remembering-ralph-wijers/", "title": "Remembering Ralph Wijers", "content_text": "arXiv is sad to announce the recent passing of our colleague and friend, Professor Ralph Wijers. A talented researcher, lecturer, and mentor, Ralph joined arXiv as a moderator and advisor in 2020 and was chair of both arXiv’s Physics Section Editorial Committee (SEC) and its Editorial Advisory Council.  Ralph generously lent his talents to arXiv […]", "date_published": "2026-07-28T14:15:33Z", "date_modified": "2026-07-28T14:15:33Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/07/RalphWijers.png?resize=645%2C420&ssl=1", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/07/RalphWijers.png?resize=645%2C420&ssl=1", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/eab6f3772845d044", "url": "https://www.lesswrong.com/posts/gnbmp4B7ZKjmuMGmJ/green-apples-are-delicious-two-three-line-exchanges", "title": "Green apples are delicious — two three-line exchanges", "content_text": "Read this short exchange. A: \"Green apples are delicious.\" B: \"Huh? Aren't they better when they're ripe?\" A: \"No, I meant Granny Smiths.\" A said \"Green apples\" intending Granny Smiths — and of course A thought it would be understood that way. But that \"claim\" never reached B. So — where was it lost? Now, the next one. A: \"Green apples are delicious.\" B: \"Huh? Aren't they better when they're ripe?\" A: \"No, I like green, sour apples.\" The same sentence — \"Green apples are delicious\" — now comes f", "date_published": "2026-07-27T17:37:54Z", "date_modified": "2026-07-27T17:37:54Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4eea8ddf590ee2fb", "url": "https://www.lesswrong.com/posts/KaSiL7ELtvgi7dHwz/fine-tuning-the-hierarchy-problem-and-what-neutrons-tell-us", "title": "Fine-Tuning, The Hierarchy Problem, and What Neutrons Tell Us About God", "content_text": "This week, we're talking about anthropic reasoning. What are we to make of fine-tuned coincidences in science and mathematics? Are they coincidences, signs of structure we don't understand yet, or signs of deliberate tuning of the parameters of our universe? And if the latter, why did the people doing the tuning have such a problem with the neutron electric dipole moment? Discuss", "date_published": "2026-07-27T15:51:11Z", "date_modified": "2026-07-27T15:51:11Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2e605faa83ff6a56", "url": "https://www.lesswrong.com/events/N9teHnwJKLabmRHdx/acx-atlanta-august-meetup-3", "title": "ACX Atlanta August Meetup", "content_text": "We return to Bold Monk brewing for a vigorous discussion of rationalism and whatever else we deem fit for discussion – hopefully including actual discussions of the sequences and Hamming Circles/Group Debugging. Location: Bold Monk Brewing 1737 Ellsworth Industrial Blvd NW Suite D-1 Atlanta, GA 30318, USA No Book club this month! But there will be next month. We will also do at least one proper (one person with the problem, 3 extra helper people) Hamming Circle / Group Debugging exercise. A note", "date_published": "2026-07-23T21:31:51Z", "date_modified": "2026-07-23T21:31:51Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0669af1a7111d7ed", "url": "https://www.lesswrong.com/posts/wZpXEWgiG7k98p6RK/estimating-llm-training-flops-on-the-nvidia-jetson-orin-nano", "title": "Estimating LLM Training FLOPs on the Nvidia Jetson Orin Nano", "content_text": "This is a research summary for an ongoing project I am working on as part of the UChicago Existential Risks Laboratory Summer Research Fellowship . I would really appreciate any feedback. Introduction Motivation In want of a quantifiable way to decide what counts as a frontier AI model, compute thresholds have emerged as the standard for AI policy: California’s SB 53 uses 10^26 floating-point operations (FLOPs) in the training run as the threshold for what counts as a frontier model and the EU A", "date_published": "2026-07-23T20:46:13Z", "date_modified": "2026-07-23T20:46:13Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d12b302672802292", "url": "https://www.lesswrong.com/posts/KsyoSAyBRXtwzSugg/not-pinning-your-openrouter-provider-might-invalidate-your", "title": "Not Pinning Your OpenRouter Provider Might Invalidate Your Research", "content_text": "Please share this with anyone doing AI research with 3rd party providers so that they can ensure their research won’t be corrupted. When you ask OpenRouter [1] to give you tokens from a given model, OpenRouter sends your request to a random available provider. OpenRouter providers have variable quality. Ensuring that your provider is high quality is really difficult. There is precedent for an AI safety paper accepted to NeurIPS having its core results entirely overturned by these issues. A revie", "date_published": "2026-07-23T20:17:44Z", "date_modified": "2026-07-23T20:17:44Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/16f3298962fceba0", "url": "https://www.lesswrong.com/posts/eaxoBCmQd4vew9ZXX/pseudpocalypse", "title": "Pseudpocalypse", "content_text": "Here's a conjecture: If you put any significant amount of text on the internet under different names, those identities can be linked using only the text itself. This is possible (I conject) because of the statistical \"fingerprint\" you leave in everything you write. Imagine a website where you can paste in some brand-new text someone just wrote. In return, the website provides links to all the text that writer has ever published under any name. It's not perfect, but it's pretty good. As far as I", "date_published": "2026-07-23T20:09:56Z", "date_modified": "2026-07-23T20:09:56Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/405d9375a32729bd", "url": "https://www.lesswrong.com/posts/fKwaqPcGnhgBCTvQf/ai-researchers-don-t-understand-the-state-2", "title": "AI Researchers Don't Understand the State", "content_text": "I’ve noticed an extremely common mistake among people who think about AI and ASI (also known as superintelligence) for a living. The mistake is to model the future of AI as a game played between AI companies, on a board where governments are part of the scenery. People think of AI companies as being able to steer the course of AI development all the way through the end-game. For example, AI researchers often join certain AI companies because they're the \"good guys\", to help the good guys win the", "date_published": "2026-07-23T17:58:27Z", "date_modified": "2026-07-23T17:58:27Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/71415ac1d74d2350", "url": "https://www.lesswrong.com/posts/9auCLJg3Z77dFdYhR/the-openai-huggingface-incident-or-redwood-research-podcast", "title": "The OpenAI/Huggingface incident | Redwood Research podcast episode 2", "content_text": "We talk about the OpenAI–Hugging Face incident, where an OpenAI model — in the middle of a cyber evaluation — broke out of its sandbox and autonomously hacked Hugging Face. We discuss: What we actually know happened. How surprising the incident was. What the incident does (and doesn’t) tell us about misalignment risk. Why control measures didn’t catch or prevent this. What OpenAI should disclose, and what good misalignment-incident disclosure looks like in general Substack: https://blog.redwoodr", "date_published": "2026-07-23T17:56:49Z", "date_modified": "2026-07-23T17:56:49Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/ab8b8ab98c779c4a", "url": "https://www.lesswrong.com/posts/xCp5GNHLe3Pq4RPBm/v-and-v-takes-on-openai-s-long-horizon-incidents", "title": "V&V takes on OpenAI’s long-horizon incidents", "content_text": "[Cross-posted from The Foretellix CTO Blog . These short takes try to put a verification-and-validation slant on AI-safety / alignment topics – they are not full treatments. I co-originated coverage-driven verification (CDV), and spent several decades doing verification of chips and AVs. See intro post for background.] On July 20 and 21, OpenAI published two unusually candid incident reports: one about their internal long-horizon model (the Erdős one) misbehaving during internal use, and one abo", "date_published": "2026-07-23T16:51:12Z", "date_modified": "2026-07-23T16:51:12Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/fd65059ea1e84b91", "url": "https://www.lesswrong.com/posts/G6obXhcmtfMFHzr7Q/duane-arnold-1", "title": "Duane Arnold", "content_text": "“So maybe I should enlighten you on what happens in your absence. This selfish existence where this introvert turns extrovert and dons her social armour.” Some posh girl in drainpipes said that - 200 views on TikTok and me one of them. But she didn’t mean it like I mean it. I started getting expensive haircuts, started smoking cherry-flavoured vapes with beautiful gays and whinging to them about how everyone wears a mask but none so well as you, started drinking more and keeping unusual hours, s", "date_published": "2026-07-23T16:17:56Z", "date_modified": "2026-07-23T16:17:56Z", "authors": [ { "name": "LessWrong (all posts)" } ], "tags": [ "LessWrong (all posts)" ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0bb8e899ec0cc9d9", "url": "https://80000hours.org/podcast/episodes/jasmine-sun-ai-anthropology/", "title": "Jasmine Sun on what the people building AI really believe", "content_text": "The post Jasmine Sun on what the people building AI really believe appeared first on 80,000 Hours .", "date_published": "2026-07-21T17:11:55Z", "date_modified": "2026-07-21T17:11:55Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/07/CURRENT-YT-Thumbnails-3840-x-2160-px-14-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/07/CURRENT-YT-Thumbnails-3840-x-2160-px-14-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/5829be7ed8e157a1", "url": "https://80000hours.org/2026/07/why-were-increasing-the-ai-focus-of-our-job-board/", "title": "Why we’re increasing the AI focus of our job board", "content_text": "The post Why we’re increasing the AI focus of our job board appeared first on 80,000 Hours .", "date_published": "2026-07-17T05:52:26Z", "date_modified": "2026-07-17T05:52:26Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2015/12/og-image1.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2015/12/og-image1.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0926327aa2a851d2", "url": "https://80000hours.org/career-reviews/scaling-organisations/", "title": "Scaling organisations making AI go well", "content_text": "The post Scaling organisations making AI go well appeared first on 80,000 Hours .", "date_published": "2026-07-15T19:27:40Z", "date_modified": "2026-07-15T19:27:40Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/07/gdoc-image-03ec76d6f10f.png", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/07/gdoc-image-03ec76d6f10f.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/02872329e7476fd5", "url": "https://80000hours.org/podcast/episodes/anton-leicht-middle-powers-agi-left-behind/", "title": "Anton Leicht on how middle powers avoid losing everything in a post-AI world", "content_text": "The post Anton Leicht on how middle powers avoid losing everything in a post-AI world appeared first on 80,000 Hours .", "date_published": "2026-07-14T17:40:14Z", "date_modified": "2026-07-14T17:40:14Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/07/Anton-Leicht-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/07/Anton-Leicht-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c17e613b2d5a20f8", "url": "https://blog.arxiv.org/2026/07/09/arxiv-now-hosts-over-3-million-articles/", "title": "arXiv now hosts over 3 million articles", "content_text": "On July 1st, 2026, arXiv reached an important milestone – establishing ourselves as an independent nonprofit. But only a few months before, arXiv quietly passed a different milestone – arxiv.org now hosts over 3 million scientific articles. Back in 2022, arXiv founder Paul Ginsparg predicted it would likely take four and half years for arXiv […]", "date_published": "2026-07-09T19:33:15Z", "date_modified": "2026-07-09T19:33:15Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/0d2a0a4ee147252d", "url": "https://80000hours.org/podcast/episodes/sneha-revanur-ai-advocacy/", "title": "Sneha Revanur on how a small team of activists helped pass America’s landmark AI safety laws", "content_text": "The post Sneha Revanur on how a small team of activists helped pass America’s landmark AI safety laws appeared first on 80,000 Hours .", "date_published": "2026-07-08T17:13:57Z", "date_modified": "2026-07-08T17:13:57Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/07/Sneha-WP-thumb-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/07/Sneha-WP-thumb-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c6c9f8832f10c8e0", "url": "https://blog.arxiv.org/2026/06/30/arxivs-next-chapter/", "title": "arXiv’s next chapter: Updates on our spin out from Cornell University", "content_text": "On July 1, 2026, arXiv will spin out from Cornell University, its home for the past 25 years, to become an independent nonprofit organization. With this next phase in arXiv’s journey quickly approaching, you can read more about arXiv’s history and the decision to spin out from Cornell in this recent article in the Cornell […]", "date_published": "2026-06-30T17:32:25Z", "date_modified": "2026-06-30T17:32:25Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/7b8c20ad158cc7b7", "url": "https://blog.arxiv.org/2026/06/26/a-year-in-review-arxivs-2025-annual-report/", "title": "A Year in Review: arXiv’s 2025 Annual Report", "content_text": "arXiv’s 2025 Annual Report is now available online! arXiv began publishing annual reports in 2020 to give our community a summary of arXiv’s initiatives, accomplishments, and financial activities each year. We also use our annual report as an opportunity to thank our members, sponsors, affiliates, individual donors, and arXiv enthusiasts – AKA, you! You can […]", "date_published": "2026-06-26T21:28:50Z", "date_modified": "2026-06-26T21:28:50Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/310f2529a08dc24c", "url": "https://80000hours.org/videos/openai/", "title": "What should you do next?", "content_text": "The post What should you do next? appeared first on 80,000 Hours .", "date_published": "2026-06-22T15:00:17Z", "date_modified": "2026-06-22T15:00:17Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/06/featured-sm.png", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/06/featured-sm.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/80b9b5c199fdfa8b", "url": "https://80000hours.org/podcast/episodes/war-in-space/", "title": "We can guess what intergalactic war would look like. And strangely, it matters.", "content_text": "The post We can guess what intergalactic war would look like. And strangely, it matters. appeared first on 80,000 Hours .", "date_published": "2026-06-18T16:10:24Z", "date_modified": "2026-06-18T16:10:24Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/06/Space-warfare-WordPress-scaled.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/06/Space-warfare-WordPress-scaled.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/eaae96b741d0f026", "url": "https://blog.arxiv.org/2026/06/17/attention-authors-temporary-change-to-announcement-schedule-due-to-upcoming-summer-holidays/", "title": "Attention authors: temporary change to announcement schedule due to upcoming summer holidays", "content_text": "This Friday, June 19, 2026, arXiv staff will be observing Juneteenth, a US federal holiday. This holiday will temporarily affect arXiv’s mailings, help desk, and announcement schedule. This brief change will only affect the announcement of new submissions; arXiv servers will otherwise remain in operation, existing papers will still be available to browse, and arXiv […]", "date_published": "2026-06-17T20:03:01Z", "date_modified": "2026-06-17T20:03:01Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/3732ff303383522f", "url": "https://80000hours.org/career-reviews/ai-policy-us-government/", "title": "AI policy in the US government", "content_text": "The post AI policy in the US government appeared first on 80,000 Hours .", "date_published": "2026-06-11T18:42:10Z", "date_modified": "2026-06-11T18:42:10Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/06/Feature-Image-3-1024x585.png", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/06/Feature-Image-3-1024x585.png", "mime_type": "image/png" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1510bec0af420a57", "url": "https://80000hours.org/career-reviews/ai-policy-and-strategy-research/", "title": "AI policy and strategy research", "content_text": "The post AI policy and strategy research appeared first on 80,000 Hours .", "date_published": "2026-06-09T18:39:42Z", "date_modified": "2026-06-09T18:39:42Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/06/Ceiling_of_Library_of_Congress.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/06/Ceiling_of_Library_of_Congress.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/639d7d16f5bc7321", "url": "https://80000hours.org/2026/06/our-top-tips-for-becoming-a-better-applicant/", "title": "Our top tips for becoming a better applicant", "content_text": "The post Our top tips for becoming a better applicant appeared first on 80,000 Hours .", "date_published": "2026-06-08T16:34:54Z", "date_modified": "2026-06-08T16:34:54Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/06/Johann_Hamza_The_Library.jpg", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/06/Johann_Hamza_The_Library.jpg", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/1bfc4549f5786e99", "url": "https://blog.arxiv.org/2026/04/02/arxiv-is-becoming-an-independent-nonprofit/", "title": "arXiv is becoming an independent nonprofit", "content_text": "This summer, arXiv is taking a big leap. On July 1, 2026, after decades of growth and productive collaboration with Cornell University, arXiv is branching out and becoming an independent nonprofit. arXiv turns 35 this year, and becoming a stand-alone nonprofit is the logical next step for us as a pioneer of open access research. […]", "date_published": "2026-04-02T17:58:36Z", "date_modified": "2026-04-02T17:58:36Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/afa92c7878b7024f", "url": "https://blog.arxiv.org/2026/02/27/remembering-joe-halpern/", "title": "Remembering Joe Halpern", "content_text": "arXiv is saddened by the recent passing of Joseph “Joe” Halpern, and we join the Cornell and scientific community in celebrating his life and memory. Joseph, known by his colleagues and arXiv staff as Joe, was a pioneer in the field of computer science and served as a professor of computer science at Cornell University […]", "date_published": "2026-02-27T17:29:53Z", "date_modified": "2026-02-27T17:29:53Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/02/0217_halpern.jpg?resize=600%2C785&ssl=1", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/02/0217_halpern.jpg?resize=600%2C785&ssl=1", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/a3830c5f6a50676a", "url": "https://blog.arxiv.org/2026/02/03/arxiv-future-proofs-access-to-research-with-third-party-digital-preservation/", "title": "arXiv future proofs access to research with third-party digital preservation", "content_text": "arXiv has entered into agreements with two third-party digital preservation services, adding a level of protection that goes beyond arXiv’s in-house activities and safeguarding open research for the future. Through its agreements with Portico, a not-for-profit community-supported dark archive for scholarly materials, and TIB – Leibniz Information Centre for Science and Technology (the German National […]", "date_published": "2026-02-03T15:15:20Z", "date_modified": "2026-02-03T15:15:20Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/4aae0014de5d6e5f", "url": "https://blog.arxiv.org/2026/01/21/attention-authors-updated-endorsement-policy/", "title": "Attention Authors: updated endorsement policy", "content_text": "arXiv has updated our endorsement policy. As of January 21, 2026, arXiv will no longer accept institutional email addresses (i.e., an email address associated with an academic or research institution) as the sole qualifier of endorsement for new authors. This policy update is being made to support the arXiv community (authors, readers, volunteer moderators, and […]", "date_published": "2026-01-21T15:04:39Z", "date_modified": "2026-01-21T15:04:39Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/c7ad4d459243b0c9", "url": "https://blog.arxiv.org/2026/01/14/attention-authors-temporary-change-to-announcement-schedule-due-to-mlk-jr-holiday-3/", "title": "Attention Authors: Temporary change to announcement schedule due to MLK Jr. Holiday", "content_text": "This coming Monday, January 19, 2026, arXiv staff will be observing Martin Luther King Jr. Day. This holiday will temporarily affect arXiv’s mailings, help desk, and announcement schedule. Submissions to arXiv are typically made public on arXiv.org and announced by email on a regular schedule. As our team celebrates MLK Day 2026, announcements will be... Continue Reading Attention Authors: Temporary change to announcement schedule due to MLK Jr. Holiday", "date_published": "2026-01-14T18:26:20Z", "date_modified": "2026-01-14T18:26:20Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/2bfeb81c2d7bac8b", "url": "https://blog.arxiv.org/2026/01/13/non-english-paper-submission-guidelines/", "title": "Attention Authors: non-English Paper Submission Guidelines", "content_text": "*Please note: arXiv’s updated non-English language paper policy is now in effect. To share questions, comments, or concerns with arXiv staff, please fill out our feedback survey. Last November, we announced that, beginning in February, arXiv will require that all new submissions have a full English-language version, either as the original language or as an... Continue Reading Attention Authors: non-English Paper Submission Guidelines", "date_published": "2026-01-13T16:45:59Z", "date_modified": "2026-01-13T16:45:59Z", "authors": [ { "name": "arXiv Blog" } ], "image": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "tags": [ "arXiv Blog" ], "attachments": [ { "url": "https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ", "mime_type": "image/jpeg" } ] }, { "id": "tag:trvny.github.io,2024:feedseek/arxiv/d8621f84138437fc", "url": "https://80000hours.org/problem-profiles/loss-of-control/", "title": "Loss of control", "content_text": "The post Loss of control appeared first on 80,000 Hours .", "date_published": "2025-07-17T19:43:58Z", "date_modified": "2025-07-17T19:43:58Z", "authors": [ { "name": "80,000 Hours" } ], "image": "https://80000hours.org/wp-content/uploads/2026/08/image2-e1786656297173.png", "tags": [ "80,000 Hours" ], "attachments": [ { "url": "https://80000hours.org/wp-content/uploads/2026/08/image2-e1786656297173.png", "mime_type": "image/png" } ] } ] }