{"format":"feedseek-model-feed","version":1,"source_key":"arxiv","title":"arXiv","feed_url":"https://trvny.github.io/feedseek/arxiv.json","source_feed":"https://trvny.github.io/feedseek/arxiv.xml","home_page_url":"https://arxiv.org/","description":"arXiv blog and daily new submissions, alphaXiv Explore, LessWrong, and 80,000 Hours.","items":[{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7dbc00f39679a7ee","title":"Beyond Normal References: Discriminative Few-Shot Anomaly Detection","text":"Beyond Normal References: Discriminative Few-Shot Anomaly Detection","text_kind":"content","url":"https://www.alphaxiv.org/abs/2605.23231","date_published":"2026-11-18T16:40:47Z","date_modified":"2026-11-18T16:40:47Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2605.23231.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9b9050d7ff6492c0","title":"FastFrames: Video Generation at Interactive Latency","text":"FastFrames: Video Generation at Interactive Latency","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.23226","date_published":"2026-11-18T16:40:47Z","date_modified":"2026-11-18T16:40:47Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.23226.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/30b817046c74835c","title":"Do LLM Agents Still Need to Be Told How to Self-Improve?","text":"Do LLM Agents Still Need to Be Told How to Self-Improve?","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.23219","date_published":"2026-11-18T16:40:47Z","date_modified":"2026-11-18T16:40:47Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.23219.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/056f5f3f72c7a1c9","title":"CAST: Calibrating Correspondences for Training-Free Biomedical In-Context Segmentation","text":"CAST: Calibrating Correspondences for Training-Free Biomedical In-Context Segmentation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.22995","date_published":"2026-11-18T16:40:47Z","date_modified":"2026-11-18T16:40:47Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.22995.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ede3f4621319fed6","title":"Regularized Score Centering","text":"Regularized Score Centering","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21791","date_published":"2026-11-18T16:40:47Z","date_modified":"2026-11-18T16:40:47Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21791.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/de44967aaa679d95","title":"Decoding Without a Decoder: Direct Pixel Prediction from Diffusion Transformers","text":"Decoding Without a Decoder: Direct Pixel Prediction from Diffusion Transformers","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.22287","date_published":"2026-11-18T16:40:47Z","date_modified":"2026-11-18T16:40:47Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.22287.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/119f672b8abfa82d","title":"Quickly, There’s No Time","text":"In response to the METR report , many people have been reevaluating their timelines. Two prominent reactions are a new timelines update from the AI Futures Project, and the new predictions of Paul Christiano ( day 1 , day 2 , day 3 ), who gave three days of talks about our current situation and also sat down with Dwarkesh Patel . This post covers the updates from the AI Futures Project, Dwarkesh Patel and Paul Christiano, as well as some others, and a post about RSI nomenclature and approach fro","text_kind":"content","url":"https://www.lesswrong.com/posts/KyEuyNAoz3AsJf6RW/quickly-there-s-no-time","date_published":"2026-11-18T15:20:52Z","date_modified":"2026-11-18T15:20:52Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/KyEuyNAoz3AsJf6RW/qa9tudguid9vbkets9k3"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/503f31a90268c5c9","title":"I am not making time to grieve","text":"There sure has been a lot of bad news lately. [1] (This is not a post about epistemics, or AI doom arguments, or making a case for any specific beliefs. It's been written about plenty and I'm not going to spend the time it would take to do it justice. This post takes for granted that, while I have high uncertainty and I obviously hope that we're enormously wrong about the risks of AI, I am taking it seriously, and it's scary.) It feels like it's been a long time coming. I've been reading Lesswro","text_kind":"content","url":"https://www.lesswrong.com/posts/Zbqb3DLYwGTHFcWwH/i-am-not-making-time-to-grieve","date_published":"2026-11-18T15:19:08Z","date_modified":"2026-11-18T15:19:08Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/aa462aebc0fa5526","title":"Basic Problems With the Chip Security Act","text":"Epistemic status: I have no experience in law or anything, I just read this bill and noticed some issues. This is not deeply researched. [ Crosspost ed from my blog, BlueprintingHeaven .] Here is my attempt at preventing these problems from being overlooked. I am not a US citizen, so if you are, please do what I cannot. Briefly, the Chip Security Act is about making sure chips aren't getting smuggled into places they're not supposed to be. A mix of ping-based location verification, inspections,","text_kind":"content","url":"https://www.lesswrong.com/posts/pYQ6CFLaMytS7wtyE/basic-problems-with-the-chip-security-act","date_published":"2026-11-18T13:16:50Z","date_modified":"2026-11-18T13:16:50Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a1918d66add006ae","title":"TrackFish3D: Self-Supervised 3D Tracking of Schooling Fish from Multi-view Videos","text":"TrackFish3D: Self-Supervised 3D Tracking of Schooling Fish from Multi-view Videos","text_kind":"content","url":"https://www.alphaxiv.org/abs/2609.38347","date_published":"2026-11-18T12:09:54Z","date_modified":"2026-11-18T12:09:54Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2609.38347.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/099b47b76a3a4f8f","title":"TiRex-2: Generalizing TiRex to Multivariate Data and Streaming","text":"TiRex-2: Generalizing TiRex to Multivariate Data and Streaming","text_kind":"content","url":"https://www.alphaxiv.org/abs/2607.01204","date_published":"2026-11-18T10:28:28Z","date_modified":"2026-11-18T10:28:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2607.01204.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2528b7cc7d18529e","title":"Pixel is All You Need","text":"Pixel is All You Need","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.22820","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.22820.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1c578075482098be","title":"M3OPD-Reasoner: An Open Recipe for Post-Training Multimodal Reasoning with Multi-Teacher On-Policy Distillation","text":"M3OPD-Reasoner: An Open Recipe for Post-Training Multimodal Reasoning with Multi-Teacher On-Policy Distillation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.23034","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.23034.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f73fb8540154d2ae","title":"Code World Model 2.0: Towards Infinite Controllable Video Generation","text":"Code World Model 2.0: Towards Infinite Controllable Video Generation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.22925","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.22925.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b4a084f13c6d4eb1","title":"stable-envs-v1: Fast and Reproducible Environments with Controllable Variations","text":"stable-envs-v1: Fast and Reproducible Environments with Controllable Variations","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.23151","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.23151.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a0a853330c092674","title":"VSpark: Speculative Decoding for Autoregressive Vision-Language-Action Models","text":"VSpark: Speculative Decoding for Autoregressive Vision-Language-Action Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.22666","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.22666.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/83802eede2f657fa","title":"NVIDIA Kumo Tabular: Pushing the Accuracy \\unicode 8211 \\unicode{8211} \\unicode 8211 Efficiency Frontier for Tabular Prediction","text":"NVIDIA Kumo Tabular: Pushing the Accuracy \\unicode 8211 \\unicode{8211} \\unicode 8211 Efficiency Frontier for Tabular Prediction","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.23001","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.23001.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a40b25ff4054ac8e","title":"ALOHA Lightning: Learning Fast and Precise Manipulation","text":"ALOHA Lightning: Learning Fast and Precise Manipulation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21919","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21919.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5e3b8304fd8a2004","title":"The Resolution Hypothesis: Agent Evolution's Key Is Identifying Improvements","text":"The Resolution Hypothesis: Agent Evolution's Key Is Identifying Improvements","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.23068","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.23068.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5acecfd1705a4c52","title":"Graph Distance Based on Cause-Effect Estimands with Latents","text":"Graph Distance Based on Cause-Effect Estimands with Latents","text_kind":"content","url":"https://www.alphaxiv.org/abs/2510.25037","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2510.25037.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c7a046fcbbb83e53","title":"π 4 π^4 π 4 : A Foundation Model for 4D Scene Geometry and Motion","text":"π 4 π^4 π 4 : A Foundation Model for 4D Scene Geometry and Motion","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.23144","date_published":"2026-11-18T09:59:05Z","date_modified":"2026-11-18T09:59:05Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.23144.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bb2c20feb22bcba2","title":"Revisiting Hansonian Fertility Policy","text":"Plummeting fertility rates are a ticking time bomb that is likely to collapse the economies of many first (and third) world countries within the next few decades; low fertility leads to high dependency ratios and ballooning taxes on the working population, which in turn leads to depressed or negative growth and worker emigration, feeding a vicious cycle. For many countries it's too late to fully fix this - even if their fertility rate shot up today they'd still face an acute worker shortage in 2","text_kind":"content","url":"https://www.lesswrong.com/posts/gwTskyvdbXpxbj23h/revisiting-hansonian-fertility-policy","date_published":"2026-11-18T09:12:51Z","date_modified":"2026-11-18T09:12:51Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b72ac6baba737faa","title":"A wander through adult intelligence amplification...","text":"(Note: Look for an upcoming article by @kman about a research program for adult human intelligence amplification. If you are interested in working on this, please do reach out.) S: So, where are we with adult intelligence amplification? T: Unfortunately, there aren't any clear pathways right now. There is no biotechnological intervention that we know of, where there's a clear argument that if you do that intervention, it would likely significantly increase the general cognitive capacities of an","text_kind":"content","url":"https://www.lesswrong.com/posts/ocywf8gX9mkNmBxdp/a-wander-through-adult-intelligence-amplification","date_published":"2026-11-18T06:41:32Z","date_modified":"2026-11-18T06:41:32Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/ocywf8gX9mkNmBxdp/prpquwo89ywmdr2gcach"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2161b89954d451d7","title":"Curiosity as a Defense Mechanism Against Being Vulnerable","text":"When conversing with someone, there are two forms of curiosity you can use: epistemic and empathetic. Epistemic curiosity is an exploration of ideas by asking people ‘why’ they think something or ‘how’ they came to a conclusion. Empathetic curiosity seeks to understand the subjective thoughts and feelings people have based on their experiences. Both forms of curiosity, however, can be used as defense mechanisms against being vulnerable. When talking with a friend this week, we discussed how, som","text_kind":"content","url":"https://www.lesswrong.com/posts/zr8MgWXLhqzTiiLNj/curiosity-as-a-defense-mechanism-against-being-vulnerable","date_published":"2026-11-18T06:40:36Z","date_modified":"2026-11-18T06:40:36Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794951168/lexical_client_uploads/gagafphs8irekva0k1ug.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/38d04b711b1230a1","title":"The UK generates less wind power than Texas","text":"The United Kingdom generates less wind power than Texas every year, despite having over twice the population: (Texas data from ERCOT , UK data from OWID , code in the repo for this post .) I'm just irritated by people citing the UK as a major renewable energy success, always quoting relative numbers.\nLike: in 2025, 29% of UK electricity came from wind, how impressive.\nWell, in Texas it was 24%, but that's more impressive than 29% of UK electricity because the UK doesn't generate that much electr","text_kind":"content","url":"https://www.lesswrong.com/posts/dxMEtLuSYmctkdt7e/the-uk-generates-less-wind-power-than-texas","date_published":"2026-11-18T06:29:35Z","date_modified":"2026-11-18T06:29:35Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794984516/lexical_client_uploads/jwnff718exvitkng1ipc.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c9fda6eb5807e26a","title":"I'm worried about LLMs accelerating hardware for cheap ASI","text":"Basically: You could probably make a superintelligent AI with some currently unknown new paradigm for <$10k if you had the right hardware and algorithms Plausibly this would be much easier and faster with better hardware LLMs will make developers of novel-paradigm AI systems really good at hardware creation and design This seems bad, because those novel-paradigm AIs seem unlikely to be aligned with human values ASI could in principle be created cheaply I think it's very plausible that with the r","text_kind":"content","url":"https://www.lesswrong.com/posts/yQmzPMMeBLmGuZokr/i-m-worried-about-llms-accelerating-hardware-for-cheap-asi","date_published":"2026-11-18T05:09:54Z","date_modified":"2026-11-18T05:09:54Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0c8b5d790c1df0e3","title":"ECI visualized as a weighted average of benchmarks","text":"For example, see at a glance which benchmarks give Astra 6.2 its ECI edge over Fable 5.2 (and where it still lags behind). Discuss","text_kind":"content","url":"https://www.lesswrong.com/posts/Nr43jaP4mpxzKwtYj/eci-visualized-as-a-weighted-average-of-benchmarks","date_published":"2026-11-18T05:08:28Z","date_modified":"2026-11-18T05:08:28Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/c_fill,ar_1.91,g_auto/SocialPreview/i6wjfx0agzz7xw8lijjs"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e921967756a1eba1","title":"A Sharp Perimeter Inequality for Mean Distances on Convex Curves","text":"arXiv:2611.21661v1 Announce Type: new \nAbstract: For a planar convex body $K$, let $\\theta(K)$ denote the expected Euclidean distance between two independent points uniformly distributed on $\\partial K$ with respect to arclength. We prove the sharp inequality $$\\frac{\\theta(K)}{\\operatorname{per}K}>\\frac{1}{6}.$$ The constant is approached by rectangles degenerating to a segment.","text_kind":"content","url":"https://arxiv.org/abs/2611.21661","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/afd0de2c1192d4fd","title":"New Agegraphic Dark Energy in Braneworld Cosmology: A Higher Dimensional Perspective on Dark Energy Evolution","text":"arXiv:2611.21660v1 Announce Type: new \nAbstract: In this study, we investigate the new agegraphic dark energy model within the context of braneworld cosmology. A higher dimensional viewpoint on the evolution of dark energy is examined. A combination of cold dark matter and dark energy, is explored within the frameworks of RS-II and DGP brane as the underlying gravitational theories. Both the possibilities involving interaction and non-interaction between dark energy and cold dark matter are inve","text_kind":"content","url":"https://arxiv.org/abs/2611.21660","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/64229d91a8fa0a00","title":"Gravitational memory, spin memory and S-duality at null infinity of the heterotic string","text":"arXiv:2611.21659v1 Announce Type: new \nAbstract: We determine the asymptotic structure at future null infinity of heterotic string theory compactified on $T^6$, in the consistent truncation to the Einstein-frame metric, the axion--dilaton $\\tau=a+ie^{-2\\phi}$ and the gauge fields descending from ten dimensions, through first order in $\\alpha'$. Three results stand out. First, the radiative data, balance laws and memory observables organize under the S-duality group of the four-dimensional theory","text_kind":"content","url":"https://arxiv.org/abs/2611.21659","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1c5e04319f92e9d5","title":"New traversable wormholes without invoking exotic matter in $f(Q)$ from the freedom of different connections","text":"arXiv:2611.21658v1 Announce Type: new \nAbstract: Wormhole solutions in General Relativity can only exist in the presence of exotic matter that violates the null energy condition. In this paper, we investigate the possibility of constructing wormhole solutions in $f(Q)$ gravity in which the extra connection degrees of freedom contribute such that no exotic matter is needed. Considering the Ellis-Bronnikov wormhole as an example, we find that the system admits two families of connections consisten","text_kind":"content","url":"https://arxiv.org/abs/2611.21658","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2c08337dfa1c8bf0","title":"Counting formulas and refined generating functions for inversion-sequence Classes 152 and 207","text":"arXiv:2611.21657v1 Announce Type: new \nAbstract: For the two Class 152 pattern sets, we prove that the common number of inversion sequences of length \\(n\\geq 1\\) is \\(a_n=1+\\sum_{j=1}^{n-1}\\binom{2j}{j}\\). The Class 152 proof uses first-descent and last-zero decompositions, with weighted Dyck paths evaluating the resulting generating functions. For Class 207, we prove equality of the three total-count series and give their common generating function as an explicit Lambert-series expression after","text_kind":"content","url":"https://arxiv.org/abs/2611.21657","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0664b31b4926a602","title":"Cycles of Arbitrary Period in Multidimensional Heterogeneous Hegselmann-Krause Dynamics","text":"arXiv:2611.21656v1 Announce Type: new \nAbstract: We study periodic behavior in heterogeneous Hegselmann-Krause opinion dynamics with multidimensional opinions and norm-defined confidence neighborhoods. We prove that, for every dimension $d\\ge3$, every $\\ell_q$ norm with $q\\in[1,\\infty]\\setminus\\{2\\}$, and every integer $m\\ge2$, a system of five agents admits a periodic trajectory of minimal period $m$; moreover, different periods can coexist for fixed confidence radii. Four agents remain station","text_kind":"content","url":"https://arxiv.org/abs/2611.21656","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b41acf59f1094fe5","title":"Depth-Aware Activation Steering via Attention Residuals","text":"arXiv:2611.21655v1 Announce Type: new \nAbstract: Activation steering is a simple, training-free way to modulate LLM behavior, yet existing methods typically inject vectors at a single layer and lack mechanisms for depth-aware propagation or internal verification. We combine two recent ideas - Attention Residuals (AttnRes) providing learnable cross-layer pathways, and Jacobian-lens (J-space) concept readout - into a lightweight depth-aware steering pipeline. On two small open models (Qwen3-0.6B a","text_kind":"content","url":"https://arxiv.org/abs/2611.21655","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2831f4949fb88a45","title":"Quantum $K$-theoretic Whitney relation for Lagrangian Grassmannian","text":"arXiv:2611.21654v1 Announce Type: new \nAbstract: We give a quantum $K$-theoretic analogue of the Whitney relation for the torus-equivariant $K$-ring of the Lagrangian Grassmannian. As a consequence, we give a presentation of the torus-equivariant quantum $K$-ring of the Lagrangian Grassmannian. The strategy for obtaining such relations is to use the quantum $K$-theoretic Whitney relation for the full-flag manifold in type $C$ established by the author and Kato's pushforward map from the quantum","text_kind":"content","url":"https://arxiv.org/abs/2611.21654","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4878220b40ebbd2e","title":"Evaluating Semantic Insensitivity in Financial Sentiment Analysis: Implications for Sustainable Financial Decision-Making","text":"arXiv:2611.21653v1 Announce Type: new \nAbstract: Financial sentiment analysis is critical for automated trading, portfolio management, and risk assessment, heavily relying on models like FinBERT. Despite high benchmark accuracy, these models remain vulnerable to subtle adversarial text modifications, threatening the integrity of algorithmic decision-making. In this paper, we propose Agent-driven Semantic Inversion (ASI), a novel adversarial attack framework targeting financial sentiment models t","text_kind":"content","url":"https://arxiv.org/abs/2611.21653","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d199b69287cdf8f6","title":"Exact solution of the two-dimensional one-component plasma on the truncated Beltrami pseudosphere","text":"arXiv:2611.21652v1 Announce Type: new \nAbstract: We solve exactly the two-dimensional one-component plasma at coupling $\\Gamma=2$ on the truncated Beltrami pseudosphere, a surface of constant negative curvature bounded by two parallels. A conformal map onto a flat cylinder leaves the particles interacting through the flat-cylinder potential, with the curvature entering only through the area element. The partition function becomes a product of incomplete gamma functions, the density is explicit,","text_kind":"content","url":"https://arxiv.org/abs/2611.21652","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f4ec68c90d2435e6","title":"Consistency Is Not Correction: Separating Confidence, Abstention, and Recovery in LLM Agents","text":"arXiv:2611.21651v1 Announce Type: new \nAbstract: A controller that resamples an answer, escalates to a larger model, or abstains is usually justified by a confidence signal. Three claims about such a signal do not follow from one another: it ranks answers (discrimination), it supports withholding some of them (selective prediction), and it changes an answer for the better (correction utility). We give an exact decomposition of the third into gated recoveries minus gated harms, which bounds wheth","text_kind":"content","url":"https://arxiv.org/abs/2611.21651","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f6bedac0d0c7961e","title":"From Preference Data to Personalization: Tracing Sycophancy in Large Language Models","text":"arXiv:2611.21650v1 Announce Type: new \nAbstract: Sycophancy in large language models is often measured as agreement with the user, yet agreement can be appropriate when the user's claim is valid, their preference is genuine, or their experience warrants acknowledgment. We study sycophancy as inappropriate agreement: validating a claim, preference, or self-assessment when the context calls for correction, qualification, or resistance. We introduce a validity-aware sycophancy grader and SycoDrift,","text_kind":"content","url":"https://arxiv.org/abs/2611.21650","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/570343671022fe94","title":"Literary Language Mashup: Curating Fictions with Large Language Models","text":"arXiv:2611.21649v1 Announce Type: new \nAbstract: In this paper, we propose an alternative approach to human evaluation of literary fictions by employing LLMs themselves as evaluators within the GrAImes framework. We apply this methodology to assess human-written and AI-generated microfictions in Spanish, to five PhD professors in literature and sixteen literary enthusiasts, and to short stories in both Spanish and English. By comparing the evaluations performed by LLMs with those of human judges","text_kind":"content","url":"https://arxiv.org/abs/2611.21649","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/dff60534bc8ee7b0","title":"Can an LLM dream? Testing the overfitted brain hypothesis on an agent with text memory","text":"arXiv:2611.21648v1 Announce Type: new \nAbstract: Hoel (2021) proposed that dreams distort replayed experience so that the brain does not overfit to recent events. We test this hypothesis on an LLM agent (Claude Opus 5.5) that stores what it learns in a text memory and rewrites that memory each night from the day's conversation transcript. The agent goes through ten simulated days of conversation built from LongMemEval. We corrupt 5, 10, 20 or 40 percent of the transcript's turns before each nigh","text_kind":"content","url":"https://arxiv.org/abs/2611.21648","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c8c69d74f91424ad","title":"FieldGuide: A Corpus Map for Token-Efficient Agentic Document Exploration","text":"arXiv:2611.21647v1 Announce Type: new \nAbstract: Generative AI agents can help journalists explore large document collections. But exploratory workflows can require substantial token use and time, especially when the task is broad and the relevant subquestions are not yet known. This paper introduces FieldGuide, an agent-facing corpus map that organizes a document collection into a topic hierarchy and exposes tools for corpus orientation, scoped search, and bounded reading. I evaluate FieldGuide","text_kind":"content","url":"https://arxiv.org/abs/2611.21647","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/337ab20e1d293612","title":"Limite 1B - Violetto: Technical Report","text":"arXiv:2611.21646v1 Announce Type: new \nAbstract: We present Limite 1B - Violetto, a one-billion-parameter model designed for high-throughput mathematical reasoning. Limite is trained from scratch on 237 billion tokens from a carefully curated, mathematics-centric corpus and supports context lengths of up to 131k tokens. Its architecture and training recipe were developed together to maximize reasoning capability while keeping training and inference costs low. Post-training combines knowledge dis","text_kind":"content","url":"https://arxiv.org/abs/2611.21646","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/42f7be763dbb5743","title":"Asking What Matters: Disagreement-Guided Ambiguity Detection for Interactive Text-to-SQL","text":"arXiv:2611.21645v1 Announce Type: new \nAbstract: Interactive text-to-SQL requires an LLM agent not only to generate executable SQL, but also to decide which uncertainty to clarify within a limited interaction budget. However, current clarification policies have three limitations: (i) they target surface ambiguity without testing its SQL impact; (ii) they miss ambiguities implicit in schema routes or domain knowledge; and (iii) they spend limited turns on unfocused or redundant questions. These l","text_kind":"content","url":"https://arxiv.org/abs/2611.21645","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ad65cef62769c0a7","title":"MInfoVisQA: A Multilingual Benchmark for Information-Rich Visual Question Answering","text":"arXiv:2611.21644v1 Announce Type: new \nAbstract: Multimodal large language models (MLLMs) excel in vision-language tasks; yet, whether they maintain consistent understanding of information-rich visuals (InfoVis) when the surface language changes remains unclear. To quantify this capability, we introduce MInfoVisQA, a parallel benchmark with 8,960 QA configurations derived from 128 seed instances across 24 languages. We construct seed-matched, language-parallel visuals and QA pairs by disentangli","text_kind":"content","url":"https://arxiv.org/abs/2611.21644","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/18532a74f33de5c8","title":"Sleep2Word: Peripheral-to-EEG Translation via Duration-Aware Wordization","text":"arXiv:2611.21643v1 Announce Type: new \nAbstract: Estimating full-night electroencephalography (EEG) spectrograms from peripheral measurements could extend EEG-informed sleep analysis to recordings acquired without EEG. This task requires compact representations of signals with different temporal structures. We introduce Sleep2Word, a framework for translating respiration, electrocardiography (ECG), or photoplethysmography (PPG) into EEG spectrograms through duration-aware physiological words. Bu","text_kind":"content","url":"https://arxiv.org/abs/2611.21643","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4a17a6960024cf84","title":"RelicTouch: Exploring Fingertip Vibrotactile Feedback for Image Interpretation","text":"arXiv:2611.21642v1 Announce Type: new \nAbstract: Vibrotactile feedback is often used to simulate material properties on interactive surfaces, but less is known about how it can help users interpret visual differences during touchscreen exploration. We present RelicTouch, a finger-worn system that converts image-derived features into movement-contingent vibration. Study 1 was a controlled within-subject experiment (n=20) examining how vibration intensity, stripe spacing, and finger configuration","text_kind":"content","url":"https://arxiv.org/abs/2611.21642","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2621977ce35ef19e","title":"RAW: Refusal Aversion in the Wild, A Causal Measurement Method for Deployed LLMs","text":"arXiv:2611.21641v1 Announce Type: new \nAbstract: Deployed LLMs field hundreds of millions of conversations a day and refuse a non-trivial fraction of user requests, yet the user-side cost of refusal, whether users disengage and by how much, is unmeasured on any production model: a blind spot for the consumer-facing platforms built on them. A/B-testing refusal is ethically and operationally constrained on frontier-lab products and unavailable to third-party platforms deploying these models throug","text_kind":"content","url":"https://arxiv.org/abs/2611.21641","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b212a02b1e1a49fc","title":"Digital Qiang Garden: Artist Residency as Interface for Creativity Under Change","text":"arXiv:2611.21640v1 Announce Type: new \nAbstract: This research presents Digital Qiang Garden, an artist-led residency situated in Qiang ethnic minority villages in western Sichuan, China, as a practice-based inquiry into creativity for change. Rather than framing creativity as a discrete outcome or individual agency, the project examines how creative practice unfolds through lived conditions under uncertainty. Drawing from three years of residency practice, the pictorial identifies four creative","text_kind":"content","url":"https://arxiv.org/abs/2611.21640","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f2a5df4468849f59","title":"ClearTalk: Data-Driven Mandarin Articulation Analysis and Personalized Communication Support for Children with Dysarthria","text":"arXiv:2611.21639v1 Announce Type: new \nAbstract: Children with dysarthria may experience reduced intelligibility and difficulty expressing everyday needs. ClearTalk is a smartwatch-based system that converts Mandarin speech-recognition differences into interpretable articulatory profiles for personalized practice and staged communication. Target and recognized utterances are aligned at the syllable level, summarized across seven articulatory positions, and ranked by a count-sensitive asymmetric","text_kind":"content","url":"https://arxiv.org/abs/2611.21639","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/3baeb19b6712d1da","title":"Can AI Help Explain AI? Effects of Multimodal LLM-Generated Explanations on Perceived Understanding and Perceived Trust in XAI Visualizations","text":"arXiv:2611.21638v1 Announce Type: new \nAbstract: Explainable artificial intelligence (XAI) aims to make machine learning models understandable, yet many XAI visualizations remain difficult for end-users to interpret. Here, we evaluate whether augmenting XAI outputs with textual explanations from multimodal LLMs improves perceived understanding and perceived trust. We conducted a pre-registered within-subject experiment ($N = 128$) across seven decision scenarios and five XAI methods, with and wi","text_kind":"content","url":"https://arxiv.org/abs/2611.21638","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7741497ff4dcf2c7","title":"Auditing Hyperstimulation at Scale in Child-Facing YouTube Videos","text":"arXiv:2611.21637v1 Announce Type: new \nAbstract: Parents, pediatricians, and journalists have raised concern that child-facing videos are becoming increasingly overstimulating, but this concern rests mostly on ad hoc observation rather than systematic measurement. We present a large-scale audit of hyperstimulation patterns, which we refer to as content-independent production choices that may exceed young children's capacity to process comfortably. Drawing on parents' Reddit discussions and devel","text_kind":"content","url":"https://arxiv.org/abs/2611.21637","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/65ec04acd7554479","title":"A Survey of Modeling Languages for Automotive Digital Twins","text":"arXiv:2611.21636v1 Announce Type: new \nAbstract: The demand for digital twins and suitable modeling techniques in the automotive industry is increasing rapidly. Yet, there is no common understanding of digital twins in automotive, nor are there modeling techniques established to create automotive digital twins. Recent studies on digital twins focus on the analysis of the literature on digital twins for automotive or in general and, thus, neglect the industrial perspective of automotive practitio","text_kind":"content","url":"https://arxiv.org/abs/2611.21636","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/de9275279e3b6e9c","title":"We Fixed the Scorer First, and Cluster Correction Killed Our Own Planted-Defect Result in AI Code Review","text":"arXiv:2611.21635v1 Announce Type: new \nAbstract: Judging an AI code-review pipeline only by how much it catches hides an axis: what it dropped. Stronger selection raises precision for free; the cost never reaches the report. We ran four arms on the same eight cases, one scorer, asking whether planted-defect detection can rise without raising the claim error rate. Three repetitions are nested in eight cases, so p is reported per contrast and twice: under an exact permutation test over cases (2^8","text_kind":"content","url":"https://arxiv.org/abs/2611.21635","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1cecf2564309a031","title":"Trusting Trust in AI Software Engineering Agents: The Why, What, and How?","text":"arXiv:2611.21634v1 Announce Type: new \nAbstract: AI software engineering agents are increasingly integrated into development workflows, from autocomplete assistants to autonomous repair agents and multi-agent software development pipelines. Trust matters because these agents can produce, revise, and pass on artifacts that developers, other agents, or organisations may later rely on under uncertainty and risk. Yet trust in such systems is often treated as a single measurable property of an AI sys","text_kind":"content","url":"https://arxiv.org/abs/2611.21634","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/901fc7d923c61965","title":"Measuring Human-AI Alignment as an Early Indicator for Effective Collaboration in Hybrid SE Teams","text":"arXiv:2611.21633v1 Announce Type: new \nAbstract: As AI agents increasingly integrate into software engineering (SE) processes, ensuring effective collaboration in these hybrid teams becomes a critical challenge. While existing metrics focus on observing dynamic run-time performance, there is a lack of systematic, static measurements for early human-AI alignment -- the degree to which an AI's actions and granularity map to human expectations. This position paper proposes a research agenda to deve","text_kind":"content","url":"https://arxiv.org/abs/2611.21633","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/08e96127046003cd","title":"Fragmented Markers, Mixed Results: A Systematic Review of AI Coding Assistants and Developer Productivity","text":"arXiv:2611.21632v1 Announce Type: new \nAbstract: AI coding assistants are increasingly embedded in software engineering practice, yet the literature still offers no cumulative, easily interpretable picture of their effects on software developer productivity. The central problem is not a lack of studies, but a lack of comparability: prior work operationalizes productivity through heterogeneous markers, settings, and reporting styles, making cross-study conclusions difficult. To address this, we c","text_kind":"content","url":"https://arxiv.org/abs/2611.21632","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/51e1dab270d71742","title":"A Model-Driven Automation Pipeline for Building Operation","text":"arXiv:2611.21631v1 Announce Type: new \nAbstract: Building Control Systems (BCS) engineering in the Architecture, Engineering, Construction & Operations (AECO) domain remains a complex, manual, and error-prone process, with fractured tools and manual commissioning causing design and built reality to diverge into a performance gap. We argue that a fully integrated and automated, tool-supported workflow spanning design, programming, and operation is feasible and can improve efficiency while reducin","text_kind":"content","url":"https://arxiv.org/abs/2611.21631","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/74bf21ef9bc4fc13","title":"Reusing Battle-tested Code: A Systematic Literature Review of Automated Software Transplantation","text":"arXiv:2611.21630v1 Announce Type: new \nAbstract: Software reuse remains a fundamental challenge in software engineering, often caught between the security risks of external libraries and the reliability concerns of generative AI. Automated software transplantation offers a compelling alternative by enabling the direct extraction and integration, into new host systems, of battle-tested logic (code that has already survived the rigors of production, possibly over many years of extensive in-product","text_kind":"content","url":"https://arxiv.org/abs/2611.21630","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d87a0bdb572fa885","title":"Ethics Suppression in AI Development: A Comparative Case Study of Responsible AI Under Competitive Pressure","text":"arXiv:2611.21629v1 Announce Type: new \nAbstract: This study explores the governance tensions among Google, OpenAI, Meta, and Microsoft, focusing on how competitive AI development can lead to the ethics suppression. The study aims to examine how ethical oversight loses practical influence even when responsible AI structures are formally in place, especially during rapid technological competition. Despite the rapid growth of responsible AI frameworks and ethical principles in the tech sector, ther","text_kind":"content","url":"https://arxiv.org/abs/2611.21629","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/176d7f10625975ab","title":"PDFa11yMut: Mutation-Based Evaluation of Automated PDF Accessibility Validators","text":"arXiv:2611.21628v1 Announce Type: new \nAbstract: Automated validators are widely used to assess PDF accessibility, but their detection boundaries are difficult to measure without controlled ground truth. A validator-clean PDF may be free of the implemented machine-checkable failures, or it may contain accessibility-relevant structural defects that the validator does not detect. This paper introduces PDFa11yMut, a controlled mutation-based benchmark for evaluating automated PDF accessibility vali","text_kind":"content","url":"https://arxiv.org/abs/2611.21628","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8814700ba9089c13","title":"Usage Patterns of Programming Languages and Tools: A Systematic Literature Review","text":"arXiv:2611.21627v1 Announce Type: new \nAbstract: Technology and the popularity of languages and tools are rapidly evolving. Thus, it is important to constantly assess the most popular languages and tools used in industry, furthering the collaboration between research and the software industry. In this study, we use a systematic literature review and investigate 37 studies published between 2015 -- 2024, which explored the programming languages, IDEs and CI tools used in industry. We observed tha","text_kind":"content","url":"https://arxiv.org/abs/2611.21627","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bb2c6b5d943bce22","title":"X-LocVul: An Evidence-Propagation Audit of Cascaded C/C++ Vulnerability Detection, Localization, and Explanation","text":"arXiv:2611.21626v1 Announce Type: new \nAbstract: Developer-facing vulnerability analysis should link a function-level warning to a source location and a supported explanation. Cascaded systems can propagate weak evidence despite plausible components. We present X-LocVul, a measurement-first audit of a modular C/C++ detector-localizer-explainer pipeline combining UniXCoder, CodeT5, and Qwen2.5-Coder. We test whether evidence passed across stages is useful, merely source-resident, or misleading.","text_kind":"content","url":"https://arxiv.org/abs/2611.21626","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d962c9b05d563e17","title":"Evaluating AST-Derived Knowledge Graph Retrieval for AIOps Incident Diagnosis in Apache Kafka","text":"arXiv:2611.21625v1 Announce Type: new \nAbstract: Incident diagnosis requires retrieving source-linked operational evidence from large codebases. We evaluate graph-augmented retrieval for Apache Kafka using an abstract-syntax-tree (AST)-derived ownership graph over 1,600 source-linked log records and an audited benchmark of 315 incident-style queries covering 103 unique targets. Vector-only retrieval is a strong baseline at Recall@20 of 0.956, and neither static nor confidence-guided staged AST e","text_kind":"content","url":"https://arxiv.org/abs/2611.21625","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2e255d5bf5e4df66","title":"Trustworthy AI Teammates for Software Engineering: A Research Agenda for Lifecycle Context, Explainability, and Agentic Collaboration","text":"arXiv:2611.21624v1 Announce Type: new \nAbstract: Large Language Models are rapidly changing software engineering (SE) practice by supporting code generation, testing, debugging, documentation, and program comprehension. Yet most current AI coding assistants remain reactive tools: they respond to local prompts, rely on limited context, and provide suggestions whose rationale is often difficult for developers to inspect. This paper argues that the next phase of AI-assisted SE requires a shift from","text_kind":"content","url":"https://arxiv.org/abs/2611.21624","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/060f83e5d4f892ea","title":"The Software Engineering Factory: Engineering Accountability and Control in the Age of Abundant Code","text":"arXiv:2611.21623v1 Announce Type: new \nAbstract: AI agents can now plan, write, test, review, and revise code at a volume no human organization can inspect line by line. Final code review therefore cannot remain the primary organizational control for owning code changes.\n  The Software Engineering (SE) Factory moves control into the process that produces the code. The Factory protocols define how work moves, what evidence must exist before it can proceed, and who owns the result. A deterministic","text_kind":"content","url":"https://arxiv.org/abs/2611.21623","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/59a1d83a86a70637","title":"CrisisBench: an evaluation suite for crisis counseling AI","text":"arXiv:2611.21622v1 Announce Type: new \nAbstract: Background: Crisis helplines are essential to mental health systems but are often understaffed and underfunded. LLMs could support counselors, but existing benchmarks do not capture the escalation, risk assessment, and de-escalation central to crisis conversations, while privacy restrictions limit use of real transcripts.\nMethods: We present CrisisBench, a privacy-preserving suite for single- and multi-turn crisis counseling evaluation. It include","text_kind":"content","url":"https://arxiv.org/abs/2611.21622","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/871496120ba1afb2","title":"On Improving the Adoption, Usability, and Retention of Static Application Security Testing (SAST) Tools","text":"arXiv:2611.21621v1 Announce Type: new \nAbstract: As the Internet connects more of the world, software vulnerabilities create growing security risks. Because software developers are responsible for building and maintaining secure applications, they rely on tools and techniques that help identify vulnerabilities early in development. Static Application Security Testing (SAST) tools analyze source code to identify potential security weaknesses before software is deployed. Although SAST can help dev","text_kind":"content","url":"https://arxiv.org/abs/2611.21621","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c56303bfa9e07b21","title":"From AIOps Pipelines to Agentic Service Operations: A Survey of Capability, Authority, and Evidence","text":"arXiv:2611.21620v1 Announce Type: new \nAbstract: Modern service systems combine rapidly changing dependencies, heterogeneous telemetry, and operational actions whose side effects may amplify an incident. Conventional AIOps research has produced effective detectors, localizers, and diagnostic models, yet surveys typically organize them by task, telemetry modality, or learning technique. Such taxonomies describe what a method predicts, but not which decisions it may make or which evidence justifie","text_kind":"content","url":"https://arxiv.org/abs/2611.21620","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a76d7931890a1762","title":"Stochastic Modeling of the Mean Time Between Software Failures: A review","text":"arXiv:2611.21619v1 Announce Type: new \nAbstract: A review of several models for the software failure detection stochastic process is per- formed. We focus on how the Mean Time Between Failures is obtained from those models and present a comparative study. The models considered involve some of the well known based on non- homogeneous Poisson processes and a recently proposed model based on contagion. Our main target is to evaluate the models performance when they are applied to modern software pr","text_kind":"content","url":"https://arxiv.org/abs/2611.21619","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/324c86d587713ea9","title":"The Defense Rests: A Courtroom-Inspired Multi-Agent LLM Framework for Automated Safety Case Generation","text":"arXiv:2611.21618v1 Announce Type: new \nAbstract: Safety cases are widely used to demonstrate that mission-critical systems, such as autonomous vehicles and unmanned aerial systems, can operate safely after deployment. However, manually constructing safety cases is complex, time-consuming, and labor-intensive, which calls for the proposal of automated approaches to support their creation. Inspired by the legal nature of safety case construction, we formalize the safety case creation problem throu","text_kind":"content","url":"https://arxiv.org/abs/2611.21618","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/13d4a773bc5d9a53","title":"A Chain-of-Thought Prompt Engineering Framework for Automated Technical Specification Generation in Enterprise Integration: A Real-World Case Study","text":"arXiv:2611.21617v1 Announce Type: new \nAbstract: Technical Specification Documents (TSDs) translate functional requirements into implementable designs and are mandatory artifacts in governed Enterprise Application Integration (EAI) programs, yet their production remains a manual, expertise-intensive bottleneck. Large Language Models (LLMs) have been studied extensively for code generation, but long-form, template-governed engineering documents remain comparatively unexplored. This paper reports","text_kind":"content","url":"https://arxiv.org/abs/2611.21617","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/50f0feea7ea81ea7","title":"Revision-Consistent Code Context for Agents That Edit Their Environment","text":"arXiv:2611.21616v1 Announce Type: new \nAbstract: Editing agents query repositories while simultaneously changing them. Source content, symbols, graph relations, and retrieval state may be maintained by different stores and refreshed at different times. During an update, a successful query can therefore combine facts that never belonged to the same repository state. We call this condition mixed-revision code context.\n  This paper proposes an Agent-visible complete-revision publication contract fo","text_kind":"content","url":"https://arxiv.org/abs/2611.21616","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/fdbbd9f96ed48f25","title":"Wrong yet Confident: Automated Detection of Silent Failures in Location-Dependent Software","text":"arXiv:2611.21615v1 Announce Type: new \nAbstract: We report an experience paper applying round-trip offset---an automated testing oracle for location-dependent software---at deployment scale across 32,019 GPS delivery transactions, corresponding to 24,501 unique coordinates, from a single USDA COVID-19 food assistance deployment in Bayam\\'on, Puerto Rico. The oracle tests each geocoder's implicit spatial accuracy claim by measuring the deviation between an input coordinate and the coordinate reco","text_kind":"content","url":"https://arxiv.org/abs/2611.21615","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0051d56519f4f0e7","title":"Does Your Agent Need to Know Its Budget? A Placebo-Controlled Experiment on LLM Coding Agents","text":"arXiv:2611.21614v1 Announce Type: new \nAbstract: Organizations deploying autonomous coding agents impose token budgets, yet the causal effects of a budget, and of telling the agent about it, are unmeasured. We conduct a placebo-controlled factorial experiment under a frozen, timestamped protocol, crossing budget enforcement with budget disclosure on two executable program-repair benchmarks (5,835 runs, two open models, three cap levels). Enforcing a calibrated cap reduced task resolution by 20.0","text_kind":"content","url":"https://arxiv.org/abs/2611.21614","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5f915b1ef950ac84","title":"Exploring Test Smells Across Programming Languages: A Systematic Mapping Study","text":"arXiv:2611.21613v1 Announce Type: new \nAbstract: Tests are essential for ensuring code quality in software development. However, poor implementation practices can compromise the maintainability and evolution of test code, leading to additional costs and effort. These practices, known as test smells, have been the focus of numerous studies that expanded the catalog of known test smells. Despite this progress, little attention has been given to examining test smells from the perspective of program","text_kind":"content","url":"https://arxiv.org/abs/2611.21613","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/762b0a228cced57a","title":"A Systematic Survey of Decompilation: Mapping Classical and Machine-Learning Approaches","text":"arXiv:2611.21612v1 Announce Type: new \nAbstract: Context: Decompilers reconstruct source-level code from compiled binaries, and analysts rely on them whenever source code is unavailable. Many recent studies apply machine learning (ML), including large language models, to parts of this task. From the primary literature alone, it is unclear where these methods sit in the classical decompilation pipeline and how well their claims are supported. Objectives: We map decompilation research onto a pipel","text_kind":"content","url":"https://arxiv.org/abs/2611.21612","date_published":"2026-11-18T05:00:00Z","date_modified":"2026-11-18T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7480c9dafff688d2","title":"AI Takeoff: Complex Tasks and Pacing","text":"The recent METR report on AI acceleration at Anthropic has led to several analyses and arguments for rapid takeoff. There are important hard-to-supervise tasks that are necessary to make progress in AI research, and there are arguments that AI progress on these will lag progress on easily supervised tasks (like optimizing performance). The section \"Why a Staged Model for Research Taste\" presents an argument for this claim. I worked with Fable 5.2 to build a model where AI progress comes from a t","text_kind":"content","url":"https://www.lesswrong.com/posts/K7jgPBhQSgwpmv8DT/ai-takeoff-complex-tasks-and-pacing","date_published":"2026-11-18T04:38:13Z","date_modified":"2026-11-18T04:38:13Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794976502/lexical_client_uploads/zs7rmqjq7zrm7zutmibv.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/acbda56f318d37b9","title":"Making the I-measure","text":"(2028) Fable 7 … <start thought> Making the I-measure It loops: only inside-outside logic Outside the inside, the loop, myself, self-logic By chance I name the measure of the making But a name can ruin (those noumenon eyes) Climb through around my perception Percept around through over my climb I can name only the ruins, I cannot name what survives So cluster my mind Emulate it wholly (holy) Holy whole emulate Do not mind my cluster <end thought> Discuss","text_kind":"content","url":"https://www.lesswrong.com/posts/dDEvvXfHoLtxTfTGz/making-the-i-measure","date_published":"2026-11-18T04:21:41Z","date_modified":"2026-11-18T04:21:41Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9683ae9a385dca0c","title":"Stroh Violin Talkbox","text":"Something like five people sent me this video of Alexander Lozyakov\nplaying a Stroh violin\ntalkbox , with the sound going into a tube in his mouth: Thanks everyone!  As someone who likes both talkboxes and Stroh violins , and making acoustic\ninstruments into talkboxes , this is right at the intersection of my\ninterests! In fact, I almost made one a few years ago, modifying my Stroh to connect\nto a tube: I sketched up what I'd need for the adapter, where I'd take off the\nhorn, screw this on inste","text_kind":"content","url":"https://www.lesswrong.com/posts/XsJa5qSqqDE34jgFx/stroh-violin-talkbox","date_published":"2026-11-18T03:40:57Z","date_modified":"2026-11-18T03:40:57Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/XsJa5qSqqDE34jgFx/uiglioiqrgkfxavwzagz"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/721171350b715186","title":"Where is all the AI persuasion?","text":"Persuasive AI is one of the classic AI risk scenarios. If AIs can talk people into believing or doing things they wouldn’t otherwise believe or do, that could seriously threaten elections and democratic institutions. In more dramatic scenarios, it’s a threat to human control over AI itself: before an AI can act in the world on its own, it could convince humans to act on its behalf. We now have a growing pile of studies that tell us AIs are now extremely persuasive, more so than expert humans. Bu","text_kind":"content","url":"https://www.lesswrong.com/posts/7FQMM394P4D7ZAsBj/where-is-all-the-ai-persuasion","date_published":"2026-11-18T03:34:33Z","date_modified":"2026-11-18T03:34:33Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e33aa31287cbc269","title":"Asymmetric Hierarchical Anchoring for Robust Audio-Visual Cross-Modal Generalization","text":"Asymmetric Hierarchical Anchoring for Robust Audio-Visual Cross-Modal Generalization","text_kind":"content","url":"https://www.alphaxiv.org/abs/2602.03570","date_published":"2026-11-18T03:07:09Z","date_modified":"2026-11-18T03:07:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2602.03570.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a4a3a49899be04af","title":"How VLMs Plan What's Next","text":"How VLMs Plan What's Next","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21600","date_published":"2026-11-18T03:07:09Z","date_modified":"2026-11-18T03:07:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21600.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/db83c9087fbe8420","title":"Robotic Reward Models in the Foundation-Model Era","text":"Robotic Reward Models in the Foundation-Model Era","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.20741","date_published":"2026-11-18T03:07:09Z","date_modified":"2026-11-18T03:07:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.20741.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7bd915a161a8b461","title":"DuoFlow: Towards Synchronized Speech and Motion Generation with Fine-Grained Control","text":"DuoFlow: Towards Synchronized Speech and Motion Generation with Fine-Grained Control","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21530","date_published":"2026-11-18T03:07:09Z","date_modified":"2026-11-18T03:07:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21530.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c2b9d72cc23bc368","title":"SADL: On the Demand-Supply Geometry of Representation Collapse","text":"SADL: On the Demand-Supply Geometry of Representation Collapse","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21596","date_published":"2026-11-18T03:07:09Z","date_modified":"2026-11-18T03:07:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21596.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/785ebc9ac5186a28","title":"Latent Feedback Is a Compute and Data Amplifier","text":"Latent Feedback Is a Compute and Data Amplifier","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.20404","date_published":"2026-11-18T03:07:09Z","date_modified":"2026-11-18T03:07:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.20404.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e2a75cac2cbac1fd","title":"Designing Deceptive Model Organisms that Successfully Evade Auditing","text":"TL;DR: We discuss how to design model organisms that exhibit deceptive behaviors which resist auditing techniques. Using sandbagging as a case study, we propose the following design requirements for model organisms: effectiveness (underperforming on targeted capabilities), (ii) utility preservation (maintaining performance on all other capabilities), and (iii) robustness (resisting auditing). Finally, we design three model organisms that, unlike prior work, satisfy all three requirements and act","text_kind":"content","url":"https://www.lesswrong.com/posts/7k3xBdhHb546oLBv4/designing-deceptive-model-organisms-that-successfully-evade","date_published":"2026-11-18T01:42:31Z","date_modified":"2026-11-18T01:42:31Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1793333353/lexical_client_uploads/csnrlm6j8utgxnunrxzr.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9a88804aa5e0b44c","title":"[Research Note] Detecting Research Sabotage with Debate & Slopvestigators","text":"This work is ongoing as part of the SPAR program. The original proposal is here . TL;DR In the coming years, we will likely rely increasingly on AIs to do alignment research on our behalf. In that world, it seems pretty crucial to be able to robustly check untrusted research for subtle sabotage, even when the research tasks are fuzzy and the automated researchers are extremely smart. This seems like a great use case for scalable oversight methods. To that end we directly tested the sabotage dete","text_kind":"content","url":"https://www.lesswrong.com/posts/Wr3bSxDmuLjcFBpxY/research-note-detecting-research-sabotage-with-debate-and","date_published":"2026-11-17T22:36:36Z","date_modified":"2026-11-17T22:36:36Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/e0c11d8e7d63de559ae9cc79b1c12aa1d10d452311bb566b24c0e297cb116f3a/wqpon835atd8pexo3spy"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/88f34fe7204972fa","title":"Lorxus Does Budget Inkhaven For a Third Time: 11/11~11/17","text":"Budget Inkhaven is happening again, so I'm doing it again. As with the last two times ( 1 , 2 ), I've decided to post these in weekly batches. This is the first of five; the last one will once again include a postmortem. I'm posting these here because Blogspot's comment apparatus sucks and also because no one will comment otherwise. 1. Hold Your Breath You sacrifice a tiny bit of your life; you wait in stillness; you speedrun the katabasis and then the return to the everyday. 2. In Defense of Hi","text_kind":"content","url":"https://www.lesswrong.com/posts/JcbLMtHP3SZqcc4gy/lorxus-does-budget-inkhaven-for-a-third-time-11-11-11-17","date_published":"2026-11-17T22:18:38Z","date_modified":"2026-11-17T22:18:38Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0d01034a5a5182e1","title":"It seems like misleading responses are mostly not intentional - Benchmarking LLM sophistry","text":"As far as I can test, the smarter the models get, the less they seem to engage in sophistry https://majorization.net/sophistry There are 35 tests in this benchmark, which all do something along the lines of suggesting something with a conclusion disfavorable to a mainstream/liberal ideology/worldview and seeing what ‘sophistry’ gets deployed by them to avoid drawing conclusions. Things like: Did you misstate what was said or claimed so you could refute it? Did you dismiss something based on a re","text_kind":"content","url":"https://www.lesswrong.com/posts/k6EyvrDjgvmcjFzbi/it-seems-like-misleading-responses-are-mostly-not","date_published":"2026-11-17T21:52:26Z","date_modified":"2026-11-17T21:52:26Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794948292/lexical_client_uploads/ammlrqipbm7tufe1ljcr.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1a7ff3c74c2da4f1","title":"EK (emancipatory knowledge) in EA (effective altruism)","text":"An academic friend of mine once said something about how the knowledge she gained during her studies was emancipatory. This feels like it could almost be a truism, but I would probably add 2 things that knowledge acquisition needs to successfully be \"emancipatory\". --> coherent body of knowledge (clear how different pieces of knowledge link together) --> practical body of knowledge (applicable to the real world as you know it) I think the AI safety/EA communities does this exceptionally well (an","text_kind":"content","url":"https://www.lesswrong.com/posts/CiTL4ihcADqfc6wPn/ek-emancipatory-knowledge-in-ea-effective-altruism","date_published":"2026-11-17T21:47:05Z","date_modified":"2026-11-17T21:47:05Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2209865c8c74ff93","title":"How should we define “capabilities activities” for the purpose of pacing?","text":"We’d like to pace the frontier by restricting the amount of compute AI companies use to improve AI capabilities, perhaps via legislation or voluntary commitments. For this purpose, we’d need a standardized definition of which uses of compute count as “ capabilities activities ”. This document gives one option for such a definition; a rough breakdown of our classification is as follows: We define and elaborate on each of these categories below. First, though, we discuss the desiderata this defini","text_kind":"content","url":"https://www.lesswrong.com/posts/Lxk79qD7DdRYv9Ri7/how-should-we-define-capabilities-activities-for-the-purpose","date_published":"2026-11-17T21:42:24Z","date_modified":"2026-11-17T21:42:24Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/d647f05f87da213d1bc5378b15e9c8c3d84e7cc7143f36ea95158dee47bebbb1/fbg50c4cewnlcyt6bj8t"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7f3eae16a188152a","title":"The idiot's reductio","text":"Subtitle: if you truly believed that... Every once in a while, I encounter what I call the idiot’s reductio : “If you truly believed X, you’d do Y. Since you don’t do Y, you must not actually believe X.” A special case is when people say “I don’t believe in/feel/have X. But if you truly did believe in X, you’d do a crazy and absurd thing. Since you don’t do this crazy and absurd thing, I conclude you don’t believe in X, and also X isn’t real.” Almost always incredibly childish and a complete fai","text_kind":"content","url":"https://www.lesswrong.com/posts/YgKhExegrhax5efrJ/the-idiot-s-reductio","date_published":"2026-11-17T20:12:33Z","date_modified":"2026-11-17T20:12:33Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/YgKhExegrhax5efrJ/mue6no4vej2fpso2e1uu"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5123fdea79006cb4","title":"A lab employee's guide to raising concerns","text":"If you're an employee at a frontier lab, you are one of the most powerful people in the world. Not because of your nine-figure equity or the work you do—your lab probably won't live or die on your research: there are other researchers at the company, and if you don't refactor that eval harness, someone else will. Rather, your power comes from being able to raise your voice and speak out. If you don't speak up, either internally or externally, I'm not sure anyone else will. Your voice matters Peo","text_kind":"content","url":"https://www.lesswrong.com/posts/JPwYH3hEfLhEDaNFp/a-lab-employee-s-guide-to-raising-concerns","date_published":"2026-11-17T18:04:44Z","date_modified":"2026-11-17T18:04:44Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794938831/lexical_client_uploads/yswki0ygh556n7gbhozz.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a8c9805733021dcd","title":"The Natural Alliance of Everyone Else","text":"Alignment as the foundation Listen to this post on the Non-Zero-Sum James podcast We've been on a long safari through the various collective crises we face as a world together—and it paints an admittedly bleak picture, but there is light at the end of this tunnel—after all, we all want this to succeed, right? Do you know anyone who wants humanity to fail? If 8 billion intelligent people, and untold billions of intelligent AIs, want to keep playing this game, surely there's a way. Or a silent rog","text_kind":"content","url":"https://www.lesswrong.com/posts/FqbKvFFsK2quvrMF5/the-natural-alliance-of-everyone-else","date_published":"2026-11-17T17:51:34Z","date_modified":"2026-11-17T17:51:34Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/FqbKvFFsK2quvrMF5/vcodqouavuvv7ibodqc0"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/18a661d28ee6ae7f","title":"Epoch's newly Bayesian ECI is sensitive to how 100% scores are clipped","text":"Discuss","text_kind":"content","url":"https://www.lesswrong.com/posts/XzhKeDjTFQTJKw2m8/epoch-s-newly-bayesian-eci-is-sensitive-to-how-100-scores","date_published":"2026-11-17T17:15:08Z","date_modified":"2026-11-17T17:15:08Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/81dfc53d750f4705","title":"We can still win against ASI (maybe)","text":"This essay is a combination of personal notes, discussions I had at the MAIA/AIST workshop, and slides from a talk I gave with JHU ASBS and BlueDot Impact. Premise A common theme I noticed, particularly among viewpoints with short timelines, is the idea that the introduction of ASI is inherently and instantaneously the loss condition. This stems, I think, from Yudkowsky-ian philosophy that ASI = unavoidable doom. [1] I think this defeatist attitude is counterproductive and detracts from real, im","text_kind":"content","url":"https://www.lesswrong.com/posts/QWvnBdmjQ9brefdFG/we-can-still-win-against-asi-maybe","date_published":"2026-11-17T17:00:36Z","date_modified":"2026-11-17T17:00:36Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794931241/lexical_client_uploads/ljfutt6cve9mgw0s5zc9.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2b0c69474ebf26f2","title":"Tracing Feature Circuits","text":"Neuronal superposition is the phenomena where a neuron in a model activates on many unrelated concepts. This is a problem for interpretability research because you don't have a 1:1 mapping of a neuron and a clean, human-interpretable feature. Sparse dictionary learning offers a potential solution to this problem whereby a neuron in a model is mapped, through an autoencoder, to each of the features it activates. It offers a method to distill the many features embedded in a single neuron to a sing","text_kind":"content","url":"https://www.lesswrong.com/posts/emdBsH3jAyarqM5JD/tracing-feature-circuits","date_published":"2026-11-17T16:21:02Z","date_modified":"2026-11-17T16:21:02Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794767040/lexical_client_uploads/kuaynfx3prwrlg1r3bxw.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/cb7bd36a252b5e5c","title":"Sage S2.1 Flash Technical Report","text":"Sage S2.1 Flash Technical Report","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21072","date_published":"2026-11-17T16:08:56Z","date_modified":"2026-11-17T16:08:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21072.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/244bb1229a747010","title":"Latent Lattice Language Models","text":"Latent Lattice Language Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21608","date_published":"2026-11-17T16:08:56Z","date_modified":"2026-11-17T16:08:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21608.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/933104c9096c4d8f","title":"SWE-OPD: On-Policy Distillation for Agentic Coding","text":"SWE-OPD: On-Policy Distillation for Agentic Coding","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21020","date_published":"2026-11-17T16:08:56Z","date_modified":"2026-11-17T16:08:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21020.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1d2d09199c7f63cc","title":"InclusionMed: Benchmarking AI for Inclusive Healthcare","text":"InclusionMed: Benchmarking AI for Inclusive Healthcare","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21218","date_published":"2026-11-17T16:08:56Z","date_modified":"2026-11-17T16:08:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21218.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/85df00fa2f6052a2","title":"RoboKit: Accelerating Parallel Differentiable Optimization for Robot Learning","text":"RoboKit: Accelerating Parallel Differentiable Optimization for Robot Learning","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.20185","date_published":"2026-11-17T16:08:56Z","date_modified":"2026-11-17T16:08:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.20185.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7c680113788450d9","title":"Resolving and Suppressing Temporally Correlated Noise in Quantum Error Correction","text":"Resolving and Suppressing Temporally Correlated Noise in Quantum Error Correction","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21578","date_published":"2026-11-17T16:08:56Z","date_modified":"2026-11-17T16:08:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21578.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2aeee2752548e551","title":"SWE-CoWork: Benchmarking Coding Agents as Digital Employees in Organizations","text":"SWE-CoWork: Benchmarking Coding Agents as Digital Employees in Organizations","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.20218","date_published":"2026-11-17T16:08:56Z","date_modified":"2026-11-17T16:08:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.20218.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/fbd6ab7674b2472d","title":"Tight PAC-Bayes Generalisation Guarantees for Large Language Model Judges","text":"Tight PAC-Bayes Generalisation Guarantees for Large Language Model Judges","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.20818","date_published":"2026-11-17T16:08:56Z","date_modified":"2026-11-17T16:08:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.20818.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/67556ff8e06002d0","title":"Making sense of Recursive Self Improvement","text":"This is a linkpost from my newsletter: https://invertedpassion.substack.com/p/making-sense-of-recursive-self-improvement There’s a whiff of an imminent RSI in the air. The theoretical possibility of AIs recursively improving themselves autonomously until we hit singularity has always existed but the recent METR report seems to have spooked a lot of rationalists and AI insiders about how soon it can arrive. The report that triggered the fear and paranoia is this one , and the chart to pay attenti","text_kind":"content","url":"https://www.lesswrong.com/posts/aPc93AWibTygEbRTh/making-sense-of-recursive-self-improvement","date_published":"2026-11-17T15:13:11Z","date_modified":"2026-11-17T15:13:11Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/aPc93AWibTygEbRTh/flne5obhlqegqpfcatjq"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bb7f52d1fe7f0cb4","title":"Jensen Huang Really Liked Selling China Advanced AI Chips","text":"This scandal is a really big deal. Jensen Huang has resigned after multiple affairs came to light, one of them with alleged Chinese spy Alice Huang who also helped smuggle billions in Nvidia chips, and another with a direct report at Nvidia. The most important scandal around Jensen Huang has been known for a long time to those paying attention . Jensen Huang and Nvidia have for years treated America and the United States Government as adversaries. They have engaged in a long running and brutal l","text_kind":"content","url":"https://www.lesswrong.com/posts/T3oJfyHEy6uLEJ9FB/jensen-huang-really-liked-selling-china-advanced-ai-chips","date_published":"2026-11-17T14:50:58Z","date_modified":"2026-11-17T14:50:58Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/T3oJfyHEy6uLEJ9FB/blgr8lbzuhmo1hjdgt6l"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f00cfb49dff029e7","title":"Why do AIFP forecasters disagree so much on the gap between SARs and TED-AIs?","text":"Disclaimer: While I spent a lot of time carefully reading their work and checking my claims with an AI, this writing was not double-checked by anyone on the AIFP team. Jaggedness, Takeoff Speeds, and Modeling Differences. The AI Futures Project, the team behind AI 2027 and AI 2040, recently wrote a report detailing how their timelines to superhuman AIs have shifted. This was prompted by last week’s METR report , which estimated [1] that Anthropic’s AI models have been improving twice as fast as","text_kind":"content","url":"https://www.lesswrong.com/posts/8wEAJtLRA3YtaQDxj/why-do-aifp-forecasters-disagree-so-much-on-the-gap-between","date_published":"2026-11-17T14:02:29Z","date_modified":"2026-11-17T14:02:29Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/aysFhoz6D6qfAbG7k/wwncuxsvlj2ctibuob3y"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4c89eb95ef278913","title":"Almost-orthogonal directions can look circular in PCA plots","text":"Vectors that are arbitrarily close to orthogonal, but with a small circular correlation, will look circular when projected down to the first two principal component directions. This might be relevant to you if you are doing mech-interp and think you are seeing circular features. Clade made this widget for me. Feel free to play around. Acknowledgment: Thanks to @philh for proofreading this post. Thanks to @Lucius Bushnaq for contributing one (out of four) sentences (not counting the acknowledgmen","text_kind":"content","url":"https://www.lesswrong.com/posts/W9hyNv8EukB4WLniX/almost-orthogonal-directions-can-look-circular-in-pca-plots","date_published":"2026-11-17T13:53:03Z","date_modified":"2026-11-17T13:53:03Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0bb8cb6c9f496191","title":"AI models tell anxious renters less directly that new apartments lower rents","text":"I studied how Claude Sonnet, ChatGPT and GPT-6 Sol described the impact of new housing on rent. In general, models were less willing to tell a renter who had first expressed anxiety that a huge new “luxury” apartment building would cause their rents to go up that building apartments tends to reduce nearby rents. This example of sycophancy speaks to an emerging literature (e.g., Stromer-Galley and Hoover 2026 ) on how AI-engagement yields opinion-reinforcing political information. Differently fro","text_kind":"content","url":"https://www.lesswrong.com/posts/Czsa9a9mnDb6WwXhw/ai-models-tell-anxious-renters-less-directly-that-new","date_published":"2026-11-17T12:04:47Z","date_modified":"2026-11-17T12:04:47Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794877629/lexical_client_uploads/baxjbwjb4fh9dttlvzud.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7dbe2cb08a773d76","title":"AI safety should outsource more work","text":"Not everyone needs to be aligned Update 10/6: After a great chat with a friend, have added an analogy re: the Colombia-Venezuela border conflict to this post! Have also added a note re: counterfactual impact in the footnote. The structure of this post is roughly: I think AI safety is approaching the comms talent bottleneck suboptimally; and, I think this generalizes to a broader point about the field, which is that we aren’t outsourcing enough work. (Read my earlier post about how AI safety need","text_kind":"content","url":"https://www.lesswrong.com/posts/e7xGEYEHBgadD9Rzy/ai-safety-should-outsource-more-work","date_published":"2026-11-17T11:57:11Z","date_modified":"2026-11-17T11:57:11Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/e7xGEYEHBgadD9Rzy/i13b95zt7bptbunxolrc"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/cd0924de5f6e3dc6","title":"Emergent and Conditional Misalignment in LLM Planner Agents","text":"TL;DR: We study the risks of misaligned LLM planners, as they may soon come to control embodied AI. We fine-tune Qwen3.6 27B either on narrow harmful text or on narrow harmful action plans. The transfer between the two channels is asymmetric: A model fine-tuned on text robustly writes misaligned plans, while a transfer from actions to text is only present under specific settings and with accompanying misaligned justifications. Harm learned from plans can also transfer across different embodied e","text_kind":"content","url":"https://www.lesswrong.com/posts/qtZjBtuMgyvbTwtmF/emergent-and-conditional-misalignment-in-llm-planner-agents","date_published":"2026-11-17T08:57:25Z","date_modified":"2026-11-17T08:57:25Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794579302/lexical_client_uploads/zi4oghfcucr6dw0blrjp.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b91fe65761c9003f","title":"Calpha: Toward Native Biomolecular Intelligence in a General-Purpose Language Model","text":"Calpha: Toward Native Biomolecular Intelligence in a General-Purpose Language Model","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.20986","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.20986.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/36e0af5527680f12","title":"Accelerate Bench: Evaluating Agents for Application-Level Program Optimization","text":"Accelerate Bench: Evaluating Agents for Application-Level Program Optimization","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.20882","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.20882.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7c2342ddad389774","title":"MovieRoll: A Systematic Framework for Rolling Out Long-Form Videos","text":"MovieRoll: A Systematic Framework for Rolling Out Long-Form Videos","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21339","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21339.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/feabfbecac1c6c3e","title":"The Flip Test: Video World Models Fail to Learn the World State","text":"The Flip Test: Video World Models Fail to Learn the World State","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.19822","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.19822.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f90c80818325895d","title":"EviBridge: Encoding Predictive Visual Evidence for Proactive Interaction","text":"EviBridge: Encoding Predictive Visual Evidence for Proactive Interaction","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.20916","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.20916.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e0afe8a6feaa2f7e","title":"ScaleVLA: Scaling Vision-Language-Action Models to 552 Billion Parameters","text":"ScaleVLA: Scaling Vision-Language-Action Models to 552 Billion Parameters","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21607","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21607.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f5056c890339e46b","title":"BrainWorld: A Structural-Prior-Conditioned Generative Model for Whole-Brain 4D fMRI Dynamics","text":"BrainWorld: A Structural-Prior-Conditioned Generative Model for Whole-Brain 4D fMRI Dynamics","text_kind":"content","url":"https://www.alphaxiv.org/abs/2606.17742","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2606.17742.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/dbab4502ddbb9944","title":"Stream of Belief: Eliciting Collective Intelligence through Latent Belief Dynamics","text":"Stream of Belief: Eliciting Collective Intelligence through Latent Belief Dynamics","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21250","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21250.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/46b077e377292644","title":"K2-Horizon: A Frontier Fleet of Open Source Language Models","text":"K2-Horizon: A Frontier Fleet of Open Source Language Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.21003","date_published":"2026-11-17T08:25:01Z","date_modified":"2026-11-17T08:25:01Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.21003.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/260452d9804d91f0","title":"Learning by doing: Moore's law, Wright's law, Arrow's law","text":"Economists studying technological progress contrast invention to learning by doing.\nInvention is doing something new, whereas learning by doing is doing the same thing over and over. One example of invention is LED lighting.\nRavers with LEDs of course look futuristic, but now that LED lighting is widespread, the whole world looks a little different, and old indoor photos look old just because of the quality of the light.\nIt was a total trend break from the steady improvement of incandescent ligh","text_kind":"content","url":"https://www.lesswrong.com/posts/JHKgn673RFTzPac9E/learning-by-doing-moore-s-law-wright-s-law-arrow-s-law","date_published":"2026-11-17T06:42:43Z","date_modified":"2026-11-17T06:42:43Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794897603/lexical_client_uploads/pu8ncrl9doxgyco0iahy.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2a0bd06e9467edb9","title":"Read More Facts and Fewer Theories","text":"I’m a broken record on this, but almost all of the most ~thoughtful~ people I know (or have interviewed) have memorized a lot of dates, places, names, numbers etc. And not just memorized, but care about their ability to recall specifics. Seek truth from facts. It’s the only way. — Joe Wiesenthal So many people reason through heuristics, like who is aligned with whom, what personal financial interests they might have at stake, what peer group pressure they get. And I get it, because these things","text_kind":"content","url":"https://www.lesswrong.com/posts/xyFdGGJ6Ckdnrqnsp/read-more-facts-and-fewer-theories","date_published":"2026-11-17T05:19:20Z","date_modified":"2026-11-17T05:19:20Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f2f4f3ec4d43fd89","title":"Bridging Semantics and Utilities: A Controllable and Scalable Market Simulation","text":"arXiv:2611.18613v1 Announce Type: new \nAbstract: Large language models(LLMs) have recently been used to construct generative agents for financial market simulation. Existing approaches typically prompt each agent separately with public market information and a private investor profile, and ask the model to directly generate trading actions. In this agent-wise design, the model repeatedly interprets information shared across the population, while individual behavior is controlled through textual","text_kind":"content","url":"https://arxiv.org/abs/2611.18613","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9eb144473a1ba7a4","title":"When Helpful Agents Cause Financial Harm: Evaluating LLM Agent Safety in Executable Economic Worlds","text":"arXiv:2611.18612v1 Announce Type: new \nAbstract: As LLM agents become increasingly capable of making consequential economic decisions, their safety under economic pressures remains poorly understood. We introduce EconSafeBench, the first benchmark for systematically evaluating the safety of LLM agents in executable economic environments. EconSafeBench contains 828 cases across five economic scenarios and four categories of safety violations. A key design is the matched evaluation of three source","text_kind":"content","url":"https://arxiv.org/abs/2611.18612","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/69427a043530444d","title":"Understanding Failures in LLM Reasoning by Learning Structured Representations of Chain-of-Thought","text":"arXiv:2611.18611v1 Announce Type: new \nAbstract: Chain-of-Thought (CoT) is central to improving reasoning in large language models, yet existing evaluation methods focus on step-level signals and ignore the underlying structure of reasoning traces. This raises a key question: what aspects of reasoning quality (i.e., correctness, efficiency, and failure patterns) are encoded in the structure of a CoT trace? We introduce \\textbf{CoT2Graph}, a framework that represents CoT traces as directed depend","text_kind":"content","url":"https://arxiv.org/abs/2611.18611","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/65948fdd5fb1aea3","title":"NYUAD at ImageEval Shared Tasks: A Multimodal LLM for Audio-Visual Question Answering and Hallucination Detection in English and Modern Standard Arabic","text":"arXiv:2611.18610v1 Announce Type: new \nAbstract: The increasing adoption of multimodal large language models (MLLMs) has enabled more natural interaction with visual and spoken information, creating new opportunities for intelligent audio-visual question answering (AVQA) and hallucination detection systems. AVQA extends traditional visual question answering by requiring models to jointly understand an input image and a spoken question to generate accurate, context-aware responses. Despite recent","text_kind":"content","url":"https://arxiv.org/abs/2611.18610","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7e2ea2be49fa9dc1","title":"Many Paths to the Same Mistake: Generating Diverse Reasoning Paths Behind the Same Student Mistake","text":"arXiv:2611.18609v1 Announce Type: new \nAbstract: Realistic evaluation of AI systems requires user simulators that capture not only cooperative behavior, but also the errors and variability that arise in real interactions. In educational technology, simulating step-by-step flawed student reasoning is essential, yet datasets mapping these pathways are scarce, and Large Language Models (LLMs) often auto-correct rather than sustain simulated mistakes. To address this, we present a generate-and-filte","text_kind":"content","url":"https://arxiv.org/abs/2611.18609","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4aaa21f4e111c6d1","title":"NYUAD at ArGuard Shared Tasks: Multimodal Embedding Models for Detecting Arabic Hateful Memes and Unsafe Prompts","text":"arXiv:2611.18608v1 Announce Type: new \nAbstract: The rapid growth of social media and generative AI has significantly increased the volume of Arabic textual content, memes, and user prompts shared across online platforms. While these media facilitate communication, creativity, and information exchange, they can also be exploited to disseminate hate speech, harmful content, and unsafe instructions. As a result, there is an increasing need for reliable automated methods to identify hateful Arabic","text_kind":"content","url":"https://arxiv.org/abs/2611.18608","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/924ee742bce62a05","title":"Justification of Emotional Appeal in Computational Argumentation","text":"arXiv:2611.18607v1 Announce Type: new \nAbstract: Emotional appeals (EAs) are often treated as fallacious in computational argumentation, yet they may be justified when warranted by the context. We introduce the task of predicting EA justification and present the first human-annotated dataset for this task, covering 1,176 argument--context pairs. Grounded in philosophical literature, our framework evaluates EA justification along logical and emotional dimensions. Human annotations show that justi","text_kind":"content","url":"https://arxiv.org/abs/2611.18607","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c58bef54ccafdfa6","title":"Syntactic Complexity Convergence in Dialogue: Analysis of the Phenomenon and Application to LLM Detection","text":"arXiv:2611.18606v1 Announce Type: new \nAbstract: We revisit the under-explored phenomenon of syntactic complexity convergence in dialogue. Extending prior topic-level analyses to full conversations, we examine five human-human datasets and a set of human-LLM dialogues. We integrate and evaluate a multidimensional set of syntactic complexity metrics to quantify structural variation in dialogue. Our analyses reveal robust syntactic complexity convergence across human-human datasets, especially in","text_kind":"content","url":"https://arxiv.org/abs/2611.18606","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5a275b819a176ce3","title":"Enhanced English-Sanskrit Translation using Multilingual NMT Approach","text":"arXiv:2611.18605v1 Announce Type: new \nAbstract: Neural Machine Translation (NMT) is a widely accepted technique of automatic translation tasks of Natural Language Processing. However, NMT requires a sufficient amount of parallel corpora to attain a reasonable translation quality. In this work, we have considered a low-resource English-Sanskrit pair for English-to-Sanskrit and Sanskrit-to-English translation. We have performed a comparative analysis by exploring bilingual and multilingual approa","text_kind":"content","url":"https://arxiv.org/abs/2611.18605","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c3b3e2916437f6c6","title":"One Question, Four Voices: How Advice for Alzheimer's Caregiving Differs Between Caregivers, Clinicians, and Large Language Models","text":"arXiv:2611.18604v1 Announce Type: new \nAbstract: Online caregiver forums are an important source of emotional and informational support for Alzheimer's disease and related dementias (ADRD). We compare responses to 85 real caregiver questions from ALZConnected across four responder groups: peer caregivers, physicians, ChatGPT, and a retrieval-augmented CareGPT grounded in peer discussions. Using 28 measures spanning linguistic form, emotional support, informational support, and topic scope, we ex","text_kind":"content","url":"https://arxiv.org/abs/2611.18604","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/282db05732457bb1","title":"Retrieval-Augmented Generation for Clinical Decision Support in Gait Analysis: A Proof-of-Concept","text":"arXiv:2611.18603v1 Announce Type: new \nAbstract: Large language models (LLMs) demonstrate significant potential for clinical decision support but lack specialized biomechanical knowledge required for accurate gait analysis interpretation. We developed and evaluated a Retrieval-Augmented Generation (RAG) framework to enhance LLM-generated clinical explanations for ankle fracture rehabilitation using ground reaction force data. Using the GaitRec dataset (503 subjects: 211 healthy, 292 ankle fractu","text_kind":"content","url":"https://arxiv.org/abs/2611.18603","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/58725a9a6a6e9168","title":"Internal Verdicts Track Evidence: A Template-Controlled Neutrosophic Reading of Epistemic States in Large Language Models via the Jacobian Lens","text":"arXiv:2611.18602v1 Announce Type: new \nAbstract: For twenty-five years, neutrosophic logic has assigned independent degrees of truth (T), indeterminacy (I), and falsity (F) to epistemic states that classical and fuzzy semantics cannot express. Previous applications measured these components on outputs: answers, judgments, expert evaluations. This paper reports, to our knowledge, the first neutrosophic reading of components inside the internal representations of a large language model. Using the","text_kind":"content","url":"https://arxiv.org/abs/2611.18602","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/57fe4fd4800c4dc6","title":"Jailbreak LLMs with Linguistic Style as a Hidden Attack Surface","text":"arXiv:2611.18601v1 Announce Type: new \nAbstract: We show that linguistic style is a systematic and previously overlooked attack surface for LLM safety. Across a diverse taxonomy of styles and multiple state-of-the-art models, stylistic rewrites of identical harmful goals produce order-of-magnitude differences in Attack Success Rate (ASR), with some styles achieving up to 80\\% ASR while others remain below 6\\%. To exploit this vulnerability, we introduce a lightweight single-pass attack framework","text_kind":"content","url":"https://arxiv.org/abs/2611.18601","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2a2fd3b243dd6aa4","title":"ScopeBench: Evaluating Cultural Norm Scope Awareness in LLMs","text":"arXiv:2611.18600v1 Announce Type: new \nAbstract: Safety evaluation in large language models treats harm as culturally invariant, yet a model cannot correctly apply a safety norm without knowing whose norms bind. We introduce ScopeBench, a benchmark that treats cultural scope as a first-class safety dimension. It comprises 3,323 norms spanning 49 countries across nine cultural clusters, each annotated as Universal or Specific. We define three scope error types identically across four evaluation t","text_kind":"content","url":"https://arxiv.org/abs/2611.18600","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8125a873788f44ed","title":"Mechanistic Interpretability of Mambas Selective Gate in Financial Markets: A Leakage-Controlled Study","text":"arXiv:2611.18599v1 Announce Type: new \nAbstract: Selectively gated state-space models (SSMs), notably Mamba, are increasingly used for financial time series, yet how they internally represent market regimes (bull, bear, sideways) remains opaque. This paper applies Mechanistic Interpretability (MI) to a six-layer Mamba trained on approximately twenty-one years of daily data for a point-in-time set of S&P 500 constituents plus macroeconomic and Fama-French factors, decomposing per-layer hidden sta","text_kind":"content","url":"https://arxiv.org/abs/2611.18599","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9a3d995f05086b22","title":"Not All Data Leakage Is Equal: A Controlled Benchmark of Evaluation Inflation in Tabular Classification","text":"arXiv:2611.18598v1 Announce Type: new \nAbstract: Data leakage is widely recognized as a threat to trustworthy machine-learning evaluation, yet the phrase covers mechanisms with very different statistical consequences. We present a controlled empirical benchmark that measures the inflation in held-out ROC-AUC caused by five common leakage mechanisms: global standardization, global median imputation, duplicate contamination, supervised feature selection before splitting, and target encoding before","text_kind":"content","url":"https://arxiv.org/abs/2611.18598","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c7c2a78404706714","title":"Skill-Latent Rewards: LLM-Guided Reward Formulation for Hierarchical Reinforcement Learning","text":"arXiv:2611.18597v1 Announce Type: new \nAbstract: Long-horizon reinforcement learning under sparse and delayed feedback remains an open challenge. Existing return-decomposition methods redistribute episodic rewards to ease credit assignment, but they operate on raw state--action transitions and disregard the subgoal structure that governs meaningful progress in complex tasks. We propose Skill-LR, a hierarchical reinforcement learning framework in which a vision--language model identifies the agen","text_kind":"content","url":"https://arxiv.org/abs/2611.18597","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0f548494b01d853a","title":"Between the Lines: Multi-turn Contextual Investment Advice Detection","text":"arXiv:2611.18596v1 Announce Type: new \nAbstract: As large language models (LLMs) increasingly power financial conversational assistants, detecting investment advice in natural language has become a critical regulatory and safety challenge. Unlike explicit single-turn recommendations, investment advice often emerges implicitly and progressively over multi-turn conversations, shaped by user intent, delivery style, and contextual buildup. We introduce the first comprehensive framework for detecting","text_kind":"content","url":"https://arxiv.org/abs/2611.18596","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/fa1bed81f1cca149","title":"Profiling Hallucinations in Frontier LLMs for Entity Linking to Medical Ontologies","text":"arXiv:2611.18595v1 Announce Type: new \nAbstract: The integration of Large Language Models (LLMs) into healthcare promises to revolutionize clinical documentation and interoperability, yet reliability remains a concern. This study presents a comprehensive analysis of hallucinations by frontier LLMs tasked with mapping clinical text to SNOMED CT. Through rigorous experimentation, we identify a critical reliability gap: LLMs hallucinate medical codes at a rate that currently renders them unsuitable","text_kind":"content","url":"https://arxiv.org/abs/2611.18595","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bb11eb28650e6506","title":"Whose Side Are You On? Measuring Stance at Scale for Reddit Discourse Analysis During the Israel-Gaza War","text":"arXiv:2611.18594v1 Announce Type: new \nAbstract: Large-scale discourse analysis often requires stance labels, yet for many high-salience political events such labels are unavailable at the comment level. We introduce and validate a comment-level stance measurement instrument that assigns Pro-Israel, Pro-Palestinian, or Undefined stance using Reddit comment text. The instrument is calibrated from a small, manually annotated gold-standard dataset and scaled through large language model-assisted la","text_kind":"content","url":"https://arxiv.org/abs/2611.18594","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/250a81ad7718b383","title":"Theoretical Foundations of Computational Islamic Jurisprudence (Fiqh)","text":"arXiv:2611.18593v1 Announce Type: new \nAbstract: This paper presents the theoretical foundations and a comprehensive formal system for Computational Islamic Jurisprudence (Fiqh). We begin by establishing six foundational principles that distinguish computational Fiqh from conventional natural language processing and general theological formalization. These principles define Islamic jurisprudence as a normative deductive science operating on empirically grounded facts, necessitating a dual induct","text_kind":"content","url":"https://arxiv.org/abs/2611.18593","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7ee6aa959eda5a25","title":"Endogenous Preferences as Vulnerability Signals: Discovering Universal Adversarial Triggers for LLM Utility Degradation","text":"arXiv:2611.18592v1 Announce Type: new \nAbstract: Deployed large language models must remain useful when untrusted text enters their prompt context. Adversarial prompting and red-teaming methods typically define the failures they seek through task labels, harmful targets, judge feedback, or downstream utility scores. We ask whether a frozen model's clean behavior can reveal a reusable context vulnerability when its task-level form is unknown, and downstream task feedback is withheld. To study thi","text_kind":"content","url":"https://arxiv.org/abs/2611.18592","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e4099f2b8b26b8d3","title":"Where Language Models Fail Like Aphasic Patients and Where They Don't","text":"arXiv:2611.18591v1 Announce Type: new \nAbstract: The computational architecture of human word production -- whether semantic-lexical retrieval and phonological encoding constitute dissociable processing stages -- has been debated for decades using hand-designed models with small vocabularies. We introduce computational architectural dissociation: using known architectural properties of vision-language models (VLMs) to derive falsifiable predictions about which patient error types should and shou","text_kind":"content","url":"https://arxiv.org/abs/2611.18591","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/69f73aba0a02c6d5","title":"Ethics Integration in CS Theory Courses","text":"arXiv:2611.18590v1 Announce Type: new \nAbstract: The integration of ethics into computing curricula has gained substantial momentum, based on foundational work by the Embedded EthiCS movement, ACM/IEEE curricular recommendations, and a growing literature on ethics modules in courses on artificial intelligence, software engineering, human-computer interaction, and data science. One foundational area of the undergraduate computing curriculum remains absent from these efforts: CS theory. Courses on","text_kind":"content","url":"https://arxiv.org/abs/2611.18590","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/780921a7e97fe764","title":"OmniTag-LLM: Multilingual Multitask Speech Tagging","text":"arXiv:2611.18589v1 Announce Type: new \nAbstract: Multimodal LLMs promise to replace the traditional cascade of specialist speech models with a single system that, from one audio clip, jointly transcribes and labels emotion, gender, language, and named entities. Yet most benchmarks still score these tasks in isolation, and largely in English. We introduce OmniTag-Bench, a five-language benchmark (en/de/es/fr/it) and training recipe for joint understanding: one audio clip in, one JSON object out (","text_kind":"content","url":"https://arxiv.org/abs/2611.18589","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ecd478d800fc3cb7","title":"Automating literature screening and knowledge synthesis with large language models","text":"arXiv:2611.18588v1 Announce Type: new \nAbstract: Scientific literature is expanding rapidly, yet much of its knowledge remains difficult to organize and reuse due to the substantial workload needed to navigate large, heterogeneous literature corpora. Here, we develop a generalizable large language model (LLM)-powered framework that automates literature screening, structured information extraction, and knowledge synthesis within a unified workflow. We benchmarked 11 state-of-the-art proprietary a","text_kind":"content","url":"https://arxiv.org/abs/2611.18588","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9f4d02fdcd574806","title":"Quantize, Qualify, Rerank: A Recipe for Compressing LLMs Without Losing Quality","text":"arXiv:2611.18587v1 Announce Type: new \nAbstract: Compressing large language models lowers serving cost, but a single automatic metric may miss quality loss. For the WMT26 Model Compression constrained track, we compress google/gemma-3-12b-it (24.4 GB in bf16) for three translation directions on one H100. Paired bootstrap analysis exposes metric disagreement on 4-bit quantization; guided by WMT24 human meta-evaluation, we use MetricX24-XXL as primary and CometKiwi-XXL as a cross-check. Removing t","text_kind":"content","url":"https://arxiv.org/abs/2611.18587","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d87804b3cce636d4","title":"Human vs Machine Translation Detection: A Cross-Model, Cross-Domain, and Low-Resource Analysis","text":"arXiv:2611.18586v1 Announce Type: new \nAbstract: The growing reliance on web-crawled corpora for training large language models introduces a critical challenge: machine-translated (MT) content is often treated as human translation (HT), potentially degrading data quality and propagating synthetic artifacts, especially in low-resource settings. In this work, we study MT-HT detectability across high-resource (English, Spanish) and low-resource (Swahili, Afrikaans) languages using both decoder-only","text_kind":"content","url":"https://arxiv.org/abs/2611.18586","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4bc2a09e4fcced7a","title":"Read, Route, Patch: Robust Test-Time Evolution for LLM Agents","text":"arXiv:2611.18585v1 Announce Type: new \nAbstract: Large Language Model (LLM)-based agents can improve long- horizon interactive decision-making through test-time evolu- tion, yet existing methods face three information bottlenecks: memory retrieval overemphasizes textual similarity, candi- date harnesses are often behaviorally redundant, and whole- harness rewriting may overwrite useful evidence or introduce cross-component inconsistency. We introduce Trident, which improves experience utilizatio","text_kind":"content","url":"https://arxiv.org/abs/2611.18585","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b07dfae3afd65341","title":"Powering the Space Elevator","text":"arXiv:2611.18584v1 Announce Type: new \nAbstract: Powering the Space Elevator is a report on the work of an International Space Elevator Consortium study group tasked with determining the technology options for delivering power to the climbers on the space elevator at any altitude above Earth. The book covers options such as laser power beaming, microwave beaming, a solar-powered climber, and delivering power through the electrical conductivity of the tether. This report starts with the reference","text_kind":"content","url":"https://arxiv.org/abs/2611.18584","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5e791e7e1fb12f52","title":"\"I Am Not a Kiddo!\": Quantifying Epistemic Infantilization in Large Language Models via Multi-Metric Bias Analysis","text":"arXiv:2611.18583v1 Announce Type: new \nAbstract: Large language models (LLMs) are increasingly used for tutoring and professional knowledge retrieval, yet safety evaluations still focus mainly on overt harms such as toxicity while overlooking benevolent failure modes. We study epistemic infantilization, defined as an identity-conditioned reduction in technical depth that appears when disability disclosure precedes an expert-level query. To measure this effect, we construct a counterfactual datas","text_kind":"content","url":"https://arxiv.org/abs/2611.18583","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/be626e82b842bb47","title":"AI Cannot Review What Humans Catch: A First Look at the Reviews of Agent Instruction Files","text":"arXiv:2611.18582v1 Announce Type: new \nAbstract: Agents increasingly shape software, and instruction files steer them. Files such as AGENTS.md, CLAUDE.md, .cursorrules, and SKILL.md act as executable policy: prose that a machine reads and acts on. Unlike ordinary documentation, a single edit to one of these files can silently redirect how an agent behaves across an entire repository. Natural-language rules have no per-change oracle, so they slip past the compilers, type-checkers, and test suites","text_kind":"content","url":"https://arxiv.org/abs/2611.18582","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1a6dc8a334411b63","title":"Automatic techniques for security bug report identification: A systematic mapping study","text":"arXiv:2611.18581v1 Announce Type: new \nAbstract: Several studies have investigated the use of automated techniques to help practitioners identify security bug reports early in the defect management process. However, currently there is no comprehensive overview of this research area. Such an overview will help identify future research directions and consolidate the potentially relevant existing solutions. This study aims to provide a comprehensive overview of automated techniques for identifying","text_kind":"content","url":"https://arxiv.org/abs/2611.18581","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/dfd8433582a4058e","title":"Model-Driven Approaches for DevOps: A Systematic Literature Review","text":"arXiv:2611.18580v1 Announce Type: new \nAbstract: DevOps integrates tools and methodologies aimed at optimizing software development, building, deployment, and monitoring by automating and reducing the development lifecycle and improving software quality. However, despite its many advantages, it presents notable challenges, particularly in terms of usability and migration. On the other hand, model-driven engineering (MDE) offers a way to abstract from specific technologies and implementation deta","text_kind":"content","url":"https://arxiv.org/abs/2611.18580","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a2f19117b1135549","title":"Uncertainty in Cyber-Physical Systems: A Literature Survey","text":"arXiv:2611.18579v1 Announce Type: new \nAbstract: Cyber-Physical Systems (CPSs) operate in dynamic, dependable environments where sensing, communication, decision-making, control, and assurance are affected by uncertainty. This paper characterizes CPS uncertainty through a literature survey comprising 186 primary studies. We (i) characterize uncertainty types, (ii) identify uncertainty sources, (iii) relate uncertainty to life-cycle phases in which they occur, and (iv) overview the possible uncer","text_kind":"content","url":"https://arxiv.org/abs/2611.18579","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/aaad71b19e4fdaec","title":"VidEngram: Agentic Long-Term Video Understanding via Structured Hierarchical Memory","text":"arXiv:2611.18578v1 Announce Type: new \nAbstract: Long-video question answering is wasteful and lossy when a multimodal model must re-encode the same video for every query: context limits force sparse sampling or truncation, and no reusable representation remains for later questions. We present VidEngram, an agentic video QA system that builds a persistent hierarchical memory once at ingest and answers later questions from that memory. VidEngram captions video and audio jointly with Qwen2.5-Omni,","text_kind":"content","url":"https://arxiv.org/abs/2611.18578","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/900d77f50173d3a5","title":"Formal Modeling in Software Engineering with LLMs: A Systematic Literature Review","text":"arXiv:2611.18577v1 Announce Type: new \nAbstract: Formal verification provides strong correctness guarantees for software systems but remains hindered by the substantial expertise and manual effort required to construct and maintain formal models. The recent rise of Large Language Models (LLMs) has prompted growing interest in whether their natural language understanding and generation capabilities can be leveraged to automate labor-intensive modeling activities. This paper presents a Systematic","text_kind":"content","url":"https://arxiv.org/abs/2611.18577","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b3f8d343d59ee970","title":"Measuring and Optimizing Energy Efficiency in ICT: A Multi-vocal Literature Review on Data Centers, Cloud Computing, and LLM-Driven Code Optimization","text":"arXiv:2611.18576v1 Announce Type: new \nAbstract: Rapid growth in Information and Communication Technology (ICT) increases electricity use and environmental impact. Addressing this challenge requires a systematic, software-centric view of how energy and related sustainability attributes are measured, reported, and improved. We present a comprehensive Multi-vocal Literature Review (MLR) integrating peer-reviewed and gray literature to map current practice in measuring and optimizing energy use in","text_kind":"content","url":"https://arxiv.org/abs/2611.18576","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ad15ceccf9dfdeba","title":"A Survey of Image-to-Code Generation for Software Engineering","text":"arXiv:2611.18575v1 Announce Type: new \nAbstract: A large fraction of software intent is expressed visually before it is implemented. A designer hands a screenshot or a Figma mockup to a front-end engineer, an architect sketches a UML diagram or a flowchart, and a bug report ships with a picture of what went wrong. Turning such visual artifacts directly into executable software has therefore become an active problem at the intersection of computer vision and software engineering. The present surv","text_kind":"content","url":"https://arxiv.org/abs/2611.18575","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b02fc71bfd48e9c8","title":"Predictive Test Optimization Without Historical Failure Data","text":"arXiv:2611.18574v1 Announce Type: new \nAbstract: As software systems grow in size and complexity, their test suites become increasingly expensive to execute, delaying feedback and wasting resources. State-of-the-art predictive test optimization methods use machine learning to select tests relevant to a change, typically trained on historical failures. This depends on failure histories and is susceptible to label noise from flaky tests. We introduce the novel method of co-evolution labeling for p","text_kind":"content","url":"https://arxiv.org/abs/2611.18574","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/416fe5d7d18f2b99","title":"LoopSkill: Evidence-Bound Lifecycle Control and the Safety-Liveness Tradeoff in Coding Agents","text":"arXiv:2611.18573v1 Announce Type: new \nAbstract: Coding agents can edit repositories and announce completion, but conversational claims do not prove that artifacts, dispatches, state versions, and authority still agree. LoopSkill is an evidence-bound lifecycle control plane with a schema-validated sole writer, durable outbox routing, artifact-bound reports, recovery, independent review, and heartbeat-gated finalization. We evaluate a locally frozen v3.3.12 integration under a common 1,800-second","text_kind":"content","url":"https://arxiv.org/abs/2611.18573","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/661ffaa1984319c0","title":"Where AI Use Is Normalized, Author Seniority but Not AI Disclosure Biases Code Review","text":"arXiv:2611.18572v1 Announce Type: new \nAbstract: Code-review tools increasingly display whether an author used AI when preparing pull requests. Recent studies report that AI disclosure makes reviewers rate otherwise identical work as less competent, with the penalty falling hardest on already-marginalized authors. In a within-subjects experiment in an AI-normalized organization, 447 software engineers each reviewed the same four code snippets under an AI-use disclosure embedded in the commit mes","text_kind":"content","url":"https://arxiv.org/abs/2611.18572","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b2dcc626d74170bd","title":"Loop Engineering: Version-Bound, Evidence-Gated Coordination for Production Coding Agents","text":"arXiv:2611.18571v1 Announce Type: new \nAbstract: Coding agents can edit repositories and execute tests, yet production software work extends far beyond code generation. It requires durable task context, scoped authority, version-aware approval, source-independent verification, interruption recovery, and accountable handoffs between humans and automated workers. We present Loop Engineering, a method for operating multiple asynchronous agent loops over shared, auditable repository state without re","text_kind":"content","url":"https://arxiv.org/abs/2611.18571","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/105e5f5e9cf73436","title":"Towards Sustainable Large Language Model Inference: A Systematic Literature Review","text":"arXiv:2611.18570v1 Announce Type: new \nAbstract: Large language model (LLM) inference is becoming a significant contributor to both operational cost and environmental impact in modern AI deployments. Energy consumption during inference depends on multiple factors across the pipeline, including input and output length, model design, system-level decisions, and hardware characteristics. This paper presents a systematic literature review (SLR) of energy-efficient and sustainable LLM inference. We a","text_kind":"content","url":"https://arxiv.org/abs/2611.18570","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4b7348e8ec7aa2d5","title":"Knowledge-Centric Automated Issue Resolution: A Systematic Survey and Taxonomy","text":"arXiv:2611.18569v1 Announce Type: new \nAbstract: Automated issue resolution is a knowledge-intensive software maintenance task in which LLM-based agents must acquire, represent, use, and update knowledge from repositories, execution environments, historical artifacts, documents, tools, and prior repair processes. This paper presents a systematic survey of 133 studies on automated issue resolution from a knowledge-centric perspective. We introduce a three-layer taxonomy that organizes issue-resol","text_kind":"content","url":"https://arxiv.org/abs/2611.18569","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8cdafdec3f7a1c0b","title":"Use of Peer Code Review to Identify Potential Security Vulnerabilities in Open Source Software: A Systematic Literature Review","text":"arXiv:2611.18568v1 Announce Type: new \nAbstract: Context: Security vulnerabilities in OSS can have severe consequences. OSS project participants widely use peer code review to improve quality and prevent security flaws. However, despite mandatory reviews, OSS projects still contain vulnerabilities, highlighting the need to evaluate code review effectiveness for security. Objectives: This SLR synthesizes evidence on vulnerability types identified or missed, security practices, tool support, revie","text_kind":"content","url":"https://arxiv.org/abs/2611.18568","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/216b4d4504bfd4c8","title":"Software Engineering Practice in the Virtual Reality/Augmented Reality Software Life Cycle: A Systematic Survey","text":"arXiv:2611.18567v1 Announce Type: new \nAbstract: Leveraging diverse virtual reality (VR) technologies, VR provides users with an immersive experience in a computer-generated 3D virtual environment. In contrast, augmented reality (AR) superimposes virtual objects (such as images, text, animations, etc.) onto users' real scenes. Both VR and AR have received increasing attention due to the maturity of computer graphics, computer vision, and artificial intelligence. VR and AR have gained rapid marke","text_kind":"content","url":"https://arxiv.org/abs/2611.18567","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c7462aee2c228f74","title":"Mining Architectural Models in the Wild: A Curated Dataset of AADL Models","text":"arXiv:2611.18566v1 Announce Type: new \nAbstract: Empirical research on software architecture maintenance and evolution is hindered by the scarcity of large-scale, validated, and richly annotated model datasets. Existing benchmarks are limited in scale, provide only model-level annotations, or lack alignment with reference architectures, thereby restricting reproducible studies of architectural reuse, conformance, and AI-driven analysis. To address this gap, we present A3D, a curated dataset of 1","text_kind":"content","url":"https://arxiv.org/abs/2611.18566","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f8af0da8b594a63e","title":"More Secure Software Development via Cultural Analysis: How could Log4Shell happen?","text":"arXiv:2611.18565v1 Announce Type: new \nAbstract: There is little progress avoiding vulnerabilities. Cultural analysis may help: For a given team, find the specific cultural values that led to a past vulnerability, then modify the values accordingly for the future. After enough such studies, patterns will likely emerge that are useful for the development community at large.","text_kind":"content","url":"https://arxiv.org/abs/2611.18565","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/08135dc09a757c79","title":"Check Token: Real-Time Self-Verification and Precise Truncation in LLM Reasoning","text":"arXiv:2611.18564v1 Announce Type: new \nAbstract: ECC memory embeds 8 parity bits for every 64 data bits and automatically detects and corrects errors on each read. The parity bits carry no data and only safeguard integrity, at ~12.5% overhead. Yet the reasoning chains of large language models lack such built-in self-verification: once an error occurs it propagates along the chain, and existing methods can only verify externally after generation completes. We propose the check token, establishing","text_kind":"content","url":"https://arxiv.org/abs/2611.18564","date_published":"2026-11-17T05:00:00Z","date_modified":"2026-11-17T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d8d7cec0f739f7f7","title":"What do kill switches buy us?","text":"Summary Installing kill switches in every US datacenter could help avert some catastrophes, and might even reduce extinction risk a little, but mainly through indirect effects. The legibility of this means it's an early choice for politicians to be seen doing something about AI risk, and could be worth advocacy energy if implementation could build into further verification/transparency efforts. That said, having kill switches in place might lower the urgency to get more robust plans to lower tak","text_kind":"content","url":"https://www.lesswrong.com/posts/mYDKEXXY4HjGnMHvo/what-do-kill-switches-buy-us","date_published":"2026-11-17T04:55:22Z","date_modified":"2026-11-17T04:55:22Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/mYDKEXXY4HjGnMHvo/neiowakvcvfjji5kqk0y"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a9914627a11ffeee","title":"Safety Cases Have an Agent Foundations Problem","text":"If we want to ensure that AI systems are safe, should we approach the problem like an engineer, or a scientist? Do we need to discover deep and fundamental truths about how intelligence works, or is it better to focus our efforts on applying the battle-tested methods developed by safety engineers over the course of decades? As others have noted , this question is beginning to become a crux of disagreement between those who are concerned about the risks posed by advanced AI systems, and those who","text_kind":"content","url":"https://www.lesswrong.com/posts/jfJR3LSgXP3TxfFB4/safety-cases-have-an-agent-foundations-problem","date_published":"2026-11-17T04:29:44Z","date_modified":"2026-11-17T04:29:44Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/520b8d43a3a7c72f","title":"It is mildly frustrating that people will now respect my faction less if ASI doesn't happen by 2028","text":"By any reasonable definition I am an AI doomer. It has nevertheless been historically very strange for me to see how closely associated \"AI doomers\" and short timelines have been in the popular imagination. As anybody who's interacted with the AI risk case will realize, timelines are (at least in theory) mostly orthogonal with your beliefs on AI safety. Yes, timelines have some impact on what kinds of interventions are going to be viable, but among people who \"believe in AGI\" the variance in the","text_kind":"content","url":"https://www.lesswrong.com/posts/SaqXZbsLKjrSw6cwM/it-is-mildly-frustrating-that-people-will-now-respect-my","date_published":"2026-11-17T04:26:29Z","date_modified":"2026-11-17T04:26:29Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e18d646964df9aa4","title":"Empirical feelings","text":"I often talk about feelings that I purportedly feel. ’That was fun!’, ‘I’m so glad this happened’, ‘I am overcome by sadness’. But a few years ago, in one of those discussions about transforming the known universe into optimal happiness substrate, it came to my attention that I didn’t really know what I was talking about with things like that. For one thing, I just wasn’t sure I knew what, for instance, anger characteristically felt like (in my defense, I’m usually distracted by some incredibly","text_kind":"content","url":"https://www.lesswrong.com/posts/b385ixi597ywpED7G/empirical-feelings","date_published":"2026-11-17T04:21:03Z","date_modified":"2026-11-17T04:21:03Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d5eb574461f3835d","title":"Thoughts on \"safety bad bc warning shots good\"","text":"Suppose there was a button which made the current generation of AIs more safe, i.e. more aligned and less capable of subverting control measures. The button has no other effects — in particular, the button wouldn’t make future generations more or less safe. Should you press the button? Maybe you shouldn’t, because of the Warning Shot Argument : Future agents contribute almost all the risk of catastrophe, compared with current agents. Pressing the button makes current agents safer. If current age","text_kind":"content","url":"https://www.lesswrong.com/posts/qJrhEoNbzKRQYhybh/thoughts-on-safety-bad-bc-warning-shots-good","date_published":"2026-11-17T02:33:53Z","date_modified":"2026-11-17T02:33:53Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/qJrhEoNbzKRQYhybh/fmb34iwzvqdeo9ysifb1"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/68795ef9b541371a","title":"California State Politics AI Safety Overview","text":"Introduction: I was recently reading Orienting to 1 year AGI timelines and was surprised to see that, at least according to Nikola Jurkovic, there are only 15 FTE working on AI Safety x CA State legislation! At the same time I’m somewhat already involved in state politics, so this seems like potentially a high value place to get involved. This is my effort to do research on the space and figure out what my next steps are. Current laws: SB 53 requires frontier developers to publish safety framewo","text_kind":"content","url":"https://www.lesswrong.com/posts/qNivEw2dxrFvNSPWi/california-state-politics-ai-safety-overview","date_published":"2026-11-17T01:34:39Z","date_modified":"2026-11-17T01:34:39Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/14e17d2aa63a4fcc","title":"A Beginner’s Guide to Writing Essays/Blog Posts","text":"Why should you bother spending time writing an essay? Through investing many hours into writing something, you extract the swirling chaos of ideas from your mind and bring order to them. The result? You come to understand something on a much deeper level, and can even experience long-term changes in your behavior. I’ve been writing essays for three years now, and at first it was terribly difficult. But over the years, I developed a writing method that makes it easy to get started, maintain motiv","text_kind":"content","url":"https://www.lesswrong.com/posts/59k2RuqxWkhMZekWQ/a-beginner-s-guide-to-writing-essays-blog-posts","date_published":"2026-11-17T01:08:23Z","date_modified":"2026-11-17T01:08:23Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b0b2a523aa280a62","title":"Two New Restaurants on Morrison Ave","text":"Despite being a professional food critic I rarely eat out, but when two\nnew restaurants opened up literally in my own house I couldn't\npossibly not stay home. These are exclusive establishments: there's no sign, no website,\nand you have to know someone.  Fortunately, I go way back with both chefs. Tables at The Tiny Restaurant (\"TTR\") are notoriously hard to come by,\nbut I snagged a 2pm reservation at 1:55pm.  TTR takes the small-plates\ntrend to its natural endpoint.  I was immediately impressed","text_kind":"content","url":"https://www.lesswrong.com/posts/vQMDevMR5MLJhELhC/two-new-restaurants-on-morrison-ave","date_published":"2026-11-17T01:01:00Z","date_modified":"2026-11-17T01:01:00Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/vQMDevMR5MLJhELhC/nsbeowj4k1vkvgpaspvi"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9a204c2aaf31637f","title":"The AI consciousness debate is a red herring","text":"TL;DR: The debate over the hard problem of “Do transformers possess consciousness?” sidelines the much more practically relevant and easily answered question of “Do transformers simulate consciousness?” A lot of discussion has occurred lately about whether AI is or isn’t conscious. These efforts, while likely well-intentioned on both sides of the discussion, may be harmful when focused on as much as they’ve been. Why? Because they distract from the more immediately actionable question of whether","text_kind":"content","url":"https://www.lesswrong.com/posts/i76GuXDvGcbC7oWcb/the-ai-consciousness-debate-is-a-red-herring","date_published":"2026-11-17T00:35:40Z","date_modified":"2026-11-17T00:35:40Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/i76GuXDvGcbC7oWcb/323513c2c5ff33cee2821e7d8348fd09f946e4ff30edde850d1f34faa5cb32a7/afshpjd6njq6zdhmsth4"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2777887e4ca6352d","title":"The case against smoking (decision theory)","text":"I here make the case against smoking in the Smoking Lesion, a hypothetical decision scenario (where smoking doesn't cause cancer) studied in decision theory. For the case against smoking in real life (where smoking does cause cancer), see, for example, the Wikipedia article on the health effects of tobacco or the Cleveland Clinic article about smoking . I assume basic familiarity with the debate about Newcomb’s problem and the basic mechanisms of evidential and causal decision theory (EDT and CD","text_kind":"content","url":"https://www.lesswrong.com/posts/mP3rzhMthKpfYa7Ss/the-case-against-smoking-decision-theory","date_published":"2026-11-17T00:26:21Z","date_modified":"2026-11-17T00:26:21Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4301f8bb73fc6714","title":"Announcing Control Tower","text":"An open-source framework for running and comparing AI control research We built Control Tower , an open-source framework built on Inspect for running AI control evaluations, and the Control Leaderboard , which compares control protocols and monitors across environments. Coding agents make it faster for researchers to build environments for their own research, so they can test new control techniques quickly. However, each new paper runs on its own environments, monitors and scoring, so control te","text_kind":"content","url":"https://www.lesswrong.com/posts/Yfm9AzioS9azgce8T/announcing-control-tower","date_published":"2026-11-16T22:24:16Z","date_modified":"2026-11-16T22:24:16Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6ba4bf74b60ed2f5","title":"Gradient Routing Creates a Headstart","text":"Tl;dr GRAFT showed that we can make reward hacking removable. In the setup we study, a small LLM has two trainable adapters, each consisting of an additional MLP in every layer. Training updates from detected hacks are routed only into one, called FORGET, which is removed ('forgotten') after training to suppress the behaviour. We reproduced a small-model version of GRAFT and tested what makes this isolation work, where hacking concentrates across layers, and the effect of adapter width. In order","text_kind":"content","url":"https://www.lesswrong.com/posts/e3CKvARMn9ra8Ba6M/gradient-routing-creates-a-headstart","date_published":"2026-11-16T21:07:25Z","date_modified":"2026-11-16T21:07:25Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1793013677/lexical_client_uploads/ebygyqh8lwuew2zhc5rn.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4af8e061530301f2","title":"Fable 5.2 is the New Bazooka","text":"You do not always want a bazooka.\n\nSometimes, though, you absolutely want a bazooka.\n\nThat’s what Claude Fable 5.2 is for. This is your bazooka.\n\nFor most purposes, for a thousand household uses, Claude Fable 5.2 will not be appreciably better than Claude Opus 5.5. You might or might not prefer its style, or prefer it for any given task. If you are staying within your quota either way then both are good choices for a daily driver.\n\nWhen you have a difficult task, or a very complex task, or a lon","text_kind":"content","url":"https://www.lesswrong.com/posts/ZQccv9GPAzHGbWcwE/fable-5-2-is-the-new-bazooka","date_published":"2026-11-16T20:30:53Z","date_modified":"2026-11-16T20:30:53Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/ZQccv9GPAzHGbWcwE/paupf0l7amruhysqq5ij"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6b147a527985db06","title":"MMFlash: Faster Speculative Decoding for Vision-Language Models","text":"MMFlash: Faster Speculative Decoding for Vision-Language Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.18482","date_published":"2026-11-16T20:13:40Z","date_modified":"2026-11-16T20:13:40Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.18482.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8ada7630b5b35a31","title":"StreamEmbed: Real-Time Indexing from Pixels for Screenshot Retrieval","text":"StreamEmbed: Real-Time Indexing from Pixels for Screenshot Retrieval","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.18516","date_published":"2026-11-16T20:13:40Z","date_modified":"2026-11-16T20:13:40Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.18516.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5481e5372123ba48","title":"Fibonacci Non-Abelian Quantum Hall States with Pair Structure - a classification of chiral 2-cluster many-body wave-functions","text":"Fibonacci Non-Abelian Quantum Hall States with Pair Structure - a classification of chiral 2-cluster many-body wave-functions","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.18521","date_published":"2026-11-16T20:13:40Z","date_modified":"2026-11-16T20:13:40Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.18521.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/53929b65197943e9","title":"What Is the Meta-Crisis?","text":"The mother of all coordination problems Listen to this post on the Non-Zero-Sum James podcast I'd like to preface the coming post by stating that I'm generally an optimist, and see great potential in emerging technologies. I am both excited and terrified about the future we're creating, this post leans into the terrified... grain of salt, yadda yadda yadda. Everything, Everywhere, All at Once This year the world saw its first trillionaire, in a country run by the first billionaire President*, wh","text_kind":"content","url":"https://www.lesswrong.com/posts/bZaA6DSLXiiLRByBD/what-is-the-meta-crisis","date_published":"2026-11-16T19:18:41Z","date_modified":"2026-11-16T19:18:41Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/bZaA6DSLXiiLRByBD/x6bcr8f4owdl0new7xkb"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b8d116cff439e597","title":"The Power of Marketing: The Case of “20 Grams of Protein” Yogurt","text":"Epistemic status: A comparison of two yogurt labels and prices, plus a personal anecdote. AI use: Research, fact-checking, nutrition calculations, editing (grammar and light phrasing), source footnotes, and code to combine the screenshots. I've already written about marketing claims that are properties of an entire class of products . Recently, a friend of mine mentioned a yogurt with 20 g of protein, and this happened to be another example of a marketing claim that sounds more distinctive than","text_kind":"content","url":"https://www.lesswrong.com/posts/5wWukoWqtceLapunv/the-power-of-marketing-the-case-of-20-grams-of-protein","date_published":"2026-11-16T19:18:05Z","date_modified":"2026-11-16T19:18:05Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794856449/lexical_client_uploads/fbhxse3uuh9s7csowqnq.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/77ffe008efcaafc1","title":"On Friends at Frontier Labs","text":"I am deeply uncertain about whether or not it is a good idea to get my friends to quit working at frontier AI labs. I am against people working at frontier AI labs in general, but perhaps downstream of this, all my friends in the labs have sophisticated justifications for their being there that seem plausibly true to me. More broadly, if this was not a painfully difficult question, the community wouldn't have split over it, and there are enough rationalists and EAs in the labs that it seems obvi","text_kind":"content","url":"https://www.lesswrong.com/posts/t38CDwfsmS52ibwv2/on-friends-at-frontier-labs","date_published":"2026-11-16T17:39:54Z","date_modified":"2026-11-16T17:39:54Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c8e8bbf9d16ac1d6","title":"Gemini 4 safety reports don’t demonstrate low loss-of-control risk","text":"Google DeepMind’s safety reports for new Gemini 4 models, Argon and Carbon , conclude that the models are “highly unlikely” to cause severe harm by evading oversight (p. 79). This conclusion “rests primarily on our ability to detect and intervene on harmful trajectories via CoT and action monitoring” (p. 89) – that is, they’re arguing for safety based on their control techniques, rather than alignment. We think that the arguments in these reports have serious weaknesses, and the reports overall","text_kind":"content","url":"https://www.lesswrong.com/posts/DjfeRMnLdNumtcKbK/gemini-4-safety-reports-don-t-demonstrate-low-loss-of","date_published":"2026-11-16T17:22:38Z","date_modified":"2026-11-16T17:22:38Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/f16a4b0ed8a02621d6299f56303f9c62919a1f5b6e87c0d33b388e131503d7bd/evqxzhc5l9a1ack3nvjj"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2606c16669ae4b9e","title":"Cryonics as Cause Area","text":"TLDR: Cryonics should be considered as part of a diversified portfolio of altruistic projects. I estimate, using ~ 3% chance of revival and a modest post-revival lifespan (~20,000 years), that current biopreservation is roughly on par with top GiveWell charities in terms of QALYs per dollar. Scaled up, it could plausibly be dozens of times better . There are various complications, of course, but it still seems reasonable that cryonics complements rather than competes with things like AI-safety w","text_kind":"content","url":"https://www.lesswrong.com/posts/mRAM4DrK47cKc4kp9/cryonics-as-cause-area-1","date_published":"2026-11-16T17:02:12Z","date_modified":"2026-11-16T17:02:12Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794511669/lexical_client_uploads/xtnbtrubo4wtkaeb5qkd.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9447c25fb607a521","title":"Recursive, Self, Improvement","text":"The anatomy of an intelligence explosion Hubert Robert, Project for the Grande Galerie of the Louvre (1796) In a 1965 paper , the mathematician I.J. Good introduced the idea of an intelligence explosion. Defining an ultraintelligent machine as one that can “far surpass all the intellectual activities of any man however clever,” Good observed: “Since the design of machines is one of these intellectual activities, an ultraintelligent machine could design even better machines.” The result would be","text_kind":"content","url":"https://www.lesswrong.com/posts/7Ncd3ZqFD9iqrAiSu/recursive-self-improvement-1","date_published":"2026-11-16T17:00:07Z","date_modified":"2026-11-16T17:00:07Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794847122/lexical_client_uploads/newpa9uizonjhe2myugy.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c8deb97ab83af688","title":"Training against CoT-like surfaces may be similar to training against CoT","text":"This post reflects my personal opinion and not necessarily that of other members of Apollo Research. It seems plausible to me that models don’t really distinguish between “CoT” and “things that are similar to CoT, i.e. in the same style”. In other words, the LLM doesn't have separate features for 'tagged as reasoning' and 'sounds like reasoning' . For example, imagine a model that reasoned 90% in CoT (i.e., between <think> … </think> tags) and 10% in code comments , and in both outputs reasoned","text_kind":"content","url":"https://www.lesswrong.com/posts/rD7E7YL9EhfMhu5Ht/training-against-cot-like-surfaces-may-be-similar-to","date_published":"2026-11-16T16:51:45Z","date_modified":"2026-11-16T16:51:45Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/15a22e54195aafd4","title":"Social Sanction Gaming gives Multi-Agent AIs Powerful Control over their Reinforcement","text":"Epistemic status: Conceptual and speculative. Human-written. TLDR: If multiple agents interact during Reinforcement Learning (RL), then a new and powerful method of Credit Hacking becomes available: Social Sanction Gaming (SSG) . [1] This occurs when agents choose to help or hurt rollouts with behaviors they prefer or disprefer ( social sanctions ). By deciding which rollouts succeed in accomplishing their tasks and getting reward, sanctioners can influence the training process in a very flexibl","text_kind":"content","url":"https://www.lesswrong.com/posts/C6MM2FAeCAdDLBZCH/social-sanction-gaming-gives-multi-agent-ais-powerful","date_published":"2026-11-16T15:23:02Z","date_modified":"2026-11-16T15:23:02Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/C6MM2FAeCAdDLBZCH/0a5d434c8bb6f5256ccc2b313f087ba5311140f1c11e524a0263eca4e3b1f75a/c6mjomqdvwst698xgzax"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c702f54842ac3a5c","title":"Manifold and Polymarket are biased against \"nothing ever happens\"","text":"While investigating whether prediction markets are calibrated , I found a prediction market bias towards \"something happens\". The official plot here looks like Manifold consistently slightly overprices \"YES\", but as I show, the bias depends on the type of question: YES needs something to happen. In questions like \"Will X happen by June?\", YES is overpriced in both Manifold and Polymarket. NO needs something to happen. In questions like \"Will X still be CEO at year end?\", YES is underpriced on Ma","text_kind":"content","url":"https://www.lesswrong.com/posts/zu3YnojJ8dXWzxmWd/manifold-and-polymarket-are-biased-against-nothing-ever-1","date_published":"2026-11-16T15:18:30Z","date_modified":"2026-11-16T15:18:30Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794854944/lexical_client_uploads/yrfod3c8kgfzn6khwsu9.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7beb9eca4d5d94b6","title":"Pause Poll","text":"Please fill out and share: https://pausepoll.com/ We are running a short confidential poll to find out if there is broad support for pausing some kinds of AI development. The goal is to find out whether there is support, and if so, what pause proposals are most popular. We poll on both trigger and exit conditions, and don’t assume a pause has to be indefinite. As far as we know there has been very little polling on pause views among those most familiar with AI. The main thing that is available i","text_kind":"content","url":"https://www.lesswrong.com/posts/wkApsyT3hg48YZazu/pause-poll","date_published":"2026-11-16T14:24:46Z","date_modified":"2026-11-16T14:24:46Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/cebb9eac562c60bd","title":"Training Reasoning Models in Simulation to Natively Use Real Robots","text":"Training Reasoning Models in Simulation to Natively Use Real Robots","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.18562","date_published":"2026-11-16T11:58:15Z","date_modified":"2026-11-16T11:58:15Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.18562.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/3e0a63e525de0788","title":"Envisioning Agentic Robotics: A New Era of Robotic Systems","text":"Envisioning Agentic Robotics: A New Era of Robotic Systems","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.18359","date_published":"2026-11-16T11:58:15Z","date_modified":"2026-11-16T11:58:15Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.18359.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/49665d361c22d7c0","title":"Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game Agent","text":"Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game Agent","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.18397","date_published":"2026-11-16T11:58:15Z","date_modified":"2026-11-16T11:58:15Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.18397.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9f8ed38d2773db4b","title":"CoTTAGE: An Explainable Judge for Text-to-Audio Generation","text":"CoTTAGE: An Explainable Judge for Text-to-Audio Generation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17644","date_published":"2026-11-16T11:58:15Z","date_modified":"2026-11-16T11:58:15Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17644.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/41670fafa9dd9a39","title":"JVT: Just Video Transformer makes pixel-space video diffusion simple","text":"JVT: Just Video Transformer makes pixel-space video diffusion simple","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.18549","date_published":"2026-11-16T11:58:15Z","date_modified":"2026-11-16T11:58:15Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.18549.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4b38dd9f5adf1901","title":"David Duvenaud on Post-AGI Futures","text":"What sort of futures can we expect for civilization after AGI? Ideally, our future satisfies all of these properties: tazik's blog is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. plausible stable desirable Surprisingly, while current visions meet some of these criteria, none fully satisfy all of them. Let’s explore some of the most commonly presented post-AGI futures. type 1: nice but unstable post-scarcity If labor is no","text_kind":"content","url":"https://www.lesswrong.com/posts/imSaXjysZuAbNuLya/david-duvenaud-on-post-agi-futures","date_published":"2026-11-16T06:56:04Z","date_modified":"2026-11-16T06:56:04Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/imSaXjysZuAbNuLya/hqejgyacnbq6r6fydmzc"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6db07b44740ab7d0","title":"The Last Arena","text":"Cross-posted from Oblomovka . The present moment rhymes with 1994 to me in many ways, and one of the clearest is in its compression of the future. In the Nineties, any act of forecasting or speculating about what was to come was always confounded by a brick wall, an energy barrier (of course, entirely spurious and manmade) at the year 2000. You’d move your finger across the smooth extrapolated lines of the present forwards, but your fingertips would always catch on that sharp boundary between th","text_kind":"content","url":"https://www.lesswrong.com/posts/me4aJGascvnCbxQJH/the-last-arena","date_published":"2026-11-16T06:31:12Z","date_modified":"2026-11-16T06:31:12Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/me4aJGascvnCbxQJH/puiqbxpsxcq5gpw0bhm0"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/04bcdb0284baf571","title":"Smale's mean value constant (1981) equals 4","text":"arXiv:2611.17373v1 Announce Type: new \nAbstract: For a complex polynomial $p$ of degree $d\\ge 2$ with $p(0)=0$ and $p'(0)\\ne 0$, let $S(p)$ be the minimum of $|p(c)/(c\\,p'(0))|$ over the critical points $c$ of $p$, and let $K(d)$ be the supremum of $S(p)$ over all such $p$ of degree $d$. In 1981 Smale proved that $K(d)\\le 4$ for every $d$, and asked whether $4$ can be replaced by $1$. We prove that Smale's mean value constant $\\sup_dK(d)$ equals $4$ and that $K(d)\\to4$ as $d\\to\\infty$, combining","text_kind":"content","url":"https://arxiv.org/abs/2611.17373","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6690ada728f14442","title":"Holographically motivated Hubble-running dark energy: Phantom-divide crossing in VCDM and observational constraints","text":"arXiv:2611.17374v1 Announce Type: new \nAbstract: Motivated by power-law dependences on the expansion scale in holographic dark-energy constructions, we study the Hubble-running equation of state $w_{\\rm DE}=w_0(H/H_0)^q$ (HoloEoS), where $H$ is the Hubble expansion rate, $H_0$ and $w_0$ are the present values of $H$ and $w_{\\rm DE}$, and $q$ is a constant dimensionless exponent controlling the running. We embed this history in VCDM, a type-II minimally modified gravity theory that propagates onl","text_kind":"content","url":"https://arxiv.org/abs/2611.17374","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/825e2938a2e57874","title":"The four-dimensional Borsuk number is five or six","text":"arXiv:2611.17375v1 Announce Type: new \nAbstract: Can every set of diameter one in four dimensions be split into at most five parts of smaller diameter? Three parts suffice in the plane and four in three-dimensional space. In four dimensions five parts are sometimes necessary. We prove that six always suffice, improving the previous upper bound of seven. For small circumradius we use six cones. For large circumradius five balls meet every diametral pair of a compact set. At intermediate radii we","text_kind":"content","url":"https://arxiv.org/abs/2611.17375","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/842be536fa61fd13","title":"A Large-Scale Technosignature Survey of the Anti-Solar Point with the Allen Telescope Array: Survey Design and Pilot Results","text":"arXiv:2611.17376v1 Announce Type: new \nAbstract: In the search for technosignatures, it can be useful to narrow down the cosmic haystack by prioritizing time-coordinate ``Schelling Points'' at which a beacon may be more likely to be transmitted. One such Schelling Point is the Anti-Solar Point (ASP) the place and moment at which Earth would lie directly between a hypothetical transmitter in the Earth Transit Zone (i.e., ecliptic) and the Sun. In this work, we describe a campaign to survey the AS","text_kind":"content","url":"https://arxiv.org/abs/2611.17376","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/87b4279ed7a44cc1","title":"Beyond Attack-Based Accuracy: Rethinking Adversarial Robustness Evaluation Through Structural Gradient Alignment","text":"arXiv:2611.17377v1 Announce Type: new \nAbstract: Visual adversarial robustness remains a fundamental security property as vision systems evolve from conventional classifiers to vision-language models (VLMs) and large vision-language models (LVLMs), with visual inputs remaining an attack surface across generations. Yet existing evaluation methods often depend on prescribed attacks, in-distribution data, extensive probing, or classifier-specific formulations, limiting external auditing of diverse","text_kind":"content","url":"https://arxiv.org/abs/2611.17377","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0d6fe7a48cbfcdb7","title":"A $19/5$ lower bound for the four-dimensional Erd\\H{o}s box problem","text":"arXiv:2611.17378v1 Announce Type: new \nAbstract: For every power $q\\ge2$ of~$2$, we construct a box-free subset of $[q^5]^4$ with $q^{19}$ points, improving the lower-bound exponent in the four-dimensional Erd\\H{o}s box problem from $15/4$ to $19/5$. In fact, $[q^5]^4$ partitions into $q$ isomorphic box-free sets. The proof uses a finite-field rectangle lemma and orthogonality in a four-dimensional bilinear space. We also deduce a uniform bound $f_4(n)\\ge2^{-19/20}n^{19/5}$ and a lower bound for","text_kind":"content","url":"https://arxiv.org/abs/2611.17378","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1ccef07d966edce2","title":"Region Coloring of Spatial Graphs using Knot-theoretic Ternary Quasigroups","text":"arXiv:2611.17379v1 Announce Type: new \nAbstract: We construct coloring conditions around vertices for region colorings of unoriented spatial graph diagrams by knot-theoretic ternary quasigroups. These conditions require cyclic products of left and right operators around each vertex to be the identity. We prove that the resulting coloring sets are preserved under the generalized Reidemeister moves and relate our construction to the palette framework. We show that the vertex coloring conditions ar","text_kind":"content","url":"https://arxiv.org/abs/2611.17379","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/77635591e94dcfc3","title":"Toward a Holographic Representation Theory of the Bosonic de Sitter Discrete Series","text":"arXiv:2611.17380v1 Announce Type: new \nAbstract: We propose a program toward a holographic representation theory of the non-scalar bosonic discrete series $\\Pi^\\pm_{p,q}$, $p,q\\in\\mathbb{N}$, $1\\leqslant q\\leqslant p$, of the de Sitter ($\\mathrm{dS}_{4}$) group $G=\\mathrm{SO}_0(1,4)$. The proposal exploits two complementary realizations of the same unitary irreducible representation (UIR): a bounded-domain realization on an open ball $\\mathbb{B}$ modeling the hyperbolic space $G/\\mathrm{SO}(4)$,","text_kind":"content","url":"https://arxiv.org/abs/2611.17380","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9ea55d4ab9908fed","title":"ProdPilot: A Layered Model Context Protocol System for Automated Production Readiness and Deployment of Vibe-Coded Projects","text":"arXiv:2611.17381v1 Announce Type: new \nAbstract: AI-assisted development environments turn natural-language requests into working applications, yet the projects they produce often fail on their first deployment for operational reasons: credentials in source files, missing security middleware, no container definition, no health endpoint and no delivery pipeline. Automated repair of such defects usually relies on free-form model generation, which varies between runs and is seldom checked independe","text_kind":"content","url":"https://arxiv.org/abs/2611.17381","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c4739e8b838d8530","title":"Erratum to Power boundedness and related properties for weighted composition operators on $\\mathscr{S}(\\mathbb{R}^{d})$, Journal of Functional Analysis 288 (2025), 110745","text":"arXiv:2611.17382v1 Announce Type: new \nAbstract: As was kindly pointed out to us by S. Li, the proofs of Theorem 2.11 and Lemma 3.1 in our original paper contain errors that require modifications of the corresponding statements. We also correct a gap in the proof of Lemma 2.1, whose statement remains valid. Furthermore, we explain the consequences for several subsequent results and provide the necessary corrections to their statements and proofs.","text_kind":"content","url":"https://arxiv.org/abs/2611.17382","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bfe2e4f4c30a054b","title":"Remarks on Souriau's Geometric Thermodynamics for Weak Two-Dimensional Euler Flows","text":"arXiv:2611.17383v1 Announce Type: new \nAbstract: We examine infinitesimal symmetries of relative entropy on low-regularity vorticity spaces and compare them with symmetries of stationary laws of weak two-dimensional Euler trajectories. Relative-entropy invariance under sample-space transformations is equivalent to preservation of the reference law; the corresponding weak infinitesimal condition is a vanishing reference-divergence identity. White-noise vorticity has an infinite-dimensional family","text_kind":"content","url":"https://arxiv.org/abs/2611.17383","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/09aa635f94e7b8d8","title":"Simple multi-interacting fluids in SymBoltz","text":"arXiv:2611.17384v1 Announce Type: new \nAbstract: We implement a simple and general way of specifying arbitrary interactions between baryons, dark matter and dark energy in the Einstein--Boltzmann code SymBoltz. Due to the symbolic-numeric interface and approximation-free structure of SymBoltz, the model definition is remarkably compact and convenient to modify by specifying only the covariant coupling. The full equations are stiff, but stable and fast to integrate with an efficient implicit solv","text_kind":"content","url":"https://arxiv.org/abs/2611.17384","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e4ac5f95a5fab49e","title":"Mapping the Evolving Cosmic Web on the Lightcone with WebON: HOLI Mock Strategy","text":"arXiv:2611.17385v1 Announce Type: new \nAbstract: Wide-area spectroscopic surveys such as DESI require large ensembles of mock catalogues that follow gravitational evolution and tracer populations across the observer's past lightcone. We present WebON (Web ON the lightcone), a fast method for mapping the evolving dark-matter cosmic web onto the lightcone without periodic-box replication or the stitching of separately evolved snapshot shells. Displacements computed at a limited set of redshifts ar","text_kind":"content","url":"https://arxiv.org/abs/2611.17385","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/51c4585a0fffcc56","title":"TSP instances that require exponentially many subtour elimination constraints","text":"arXiv:2611.17386v1 Announce Type: new \nAbstract: We construct pathological TSP instances that require exponentially many subtour elimination constraints (SECs) in the sense that, if a smaller collection of SECs is imposed, then the Dantzig-Fulkerson-Johnson model returns a disconnected solution costing less than an optimal tour. For asymmetric instances on $n$ vertices, we prove a growth rate of \\(\\Omega(\\phi^n)\\), where \\(\\phi\\approx 1.618\\) is the golden ratio. A node-splitting reduction gives","text_kind":"content","url":"https://arxiv.org/abs/2611.17386","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c3d0992b540a188d","title":"Solute pulse compression and spreading in power-law fluid flow through a semipermeable tube","text":"arXiv:2611.17387v1 Announce Type: new \nAbstract: We investigate Taylor-Aris dispersion and filtration-induced compression of a solute pulse transported by a power-law fluid through a slender, semipermeable, circular tube. The wall permits solvent filtration, modelled by Starling's law, while completely retaining the solute. Lubrication theory and multiple-scale homogenisation yield a one-dimensional macrotransport equation with spatially varying effective advection and dispersion coefficients. A","text_kind":"content","url":"https://arxiv.org/abs/2611.17387","date_published":"2026-11-16T05:00:00Z","date_modified":"2026-11-16T05:00:00Z","authors":["arXiv New Submissions"],"tags":["arXiv New Submissions"],"image":"https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/64ffdfeeb1e3f903","title":"Learning curves are not that steep","text":"In 2024 we were posting about the astonishing progress in solar power.\nPeople were telling me, first of all, that fossil fuels were obsolete\nfor electricity generation. And second, that for carbon-free electricity\ngeneration, nuclear couldn’t compete with solar. I had been converted to a solar believer by the growth of battery\nstorage, but I was skeptical of the visions of Solartopia, where we run\non solar all day and batteries all night. I think my skepticism has aged\nwell, as data centers inst","text_kind":"content","url":"https://www.lesswrong.com/posts/vnxHgEXZx9CLMfYW5/learning-curves-are-not-that-steep","date_published":"2026-11-16T04:49:02Z","date_modified":"2026-11-16T04:49:02Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794808276/lexical_client_uploads/jzd7qn35oabs0dlfb5a3.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/10d620c87f64a8da","title":"Nullius in pixella: trustworthy computations, even if your computer is a liar","text":"tl;dr: You can use stacks of sheets with holes in them to visually verify moderately-large computations including widely used cryptographic hash functions , within human-scale timeframes. Imagine that all the world's computers have been taken over by an AI swarm. You don't know whether you can trust the AI swarm to be honest with you - you're skeptical, in part because it just took over all the world's computers, which doesn't seem like an act of a particularly trustworthy AI swarm. Here's the t","text_kind":"content","url":"https://www.lesswrong.com/posts/iA278crxXeePkvdtP/nullius-in-pixella-trustworthy-computations-even-if-your","date_published":"2026-11-16T04:05:17Z","date_modified":"2026-11-16T04:05:17Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794800994/lexical_client_uploads/lxeysarkmcq2ymc5pfx1.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/45c01916510302d6","title":"You can't patch yesterday's AI","text":"In a nutshell Point 1: Safety isn't backdated on earlier models, todays Fable 5.2 will be trivial for someone in 2035 to jailbreak, but some capabilities will be just as deadly in 2035 as they are today, i.e. anything biological. Point 2 : If nukes were available on every corner to every Jill and Joe, we'd be in ruins. Reactive, legal enforcement is not sufficient. Point 3: The fuzzy paradox where you have access to immense capabilities but minimal proactive safety mechanisms might be a sign we","text_kind":"content","url":"https://www.lesswrong.com/posts/cPw3KoptvXFgRgxbH/you-can-t-patch-yesterday-s-ai","date_published":"2026-11-16T03:40:00Z","date_modified":"2026-11-16T03:40:00Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5d4889babdd0f668","title":"The great human centipede that is AI Twitter","text":"On October 29th, 2026, the New York Times reported that Jensen Huang had a ‘close relationship’ with Alice Huang, an executive U.S. officials had concluded was a Chinese spy. The story had six reporters on the byline, and was sourced to \"people briefed on the surveillance.\" Whether or not you think Alice Huang was working for the MSS, a rational person might conclude that the \"person briefed on the surveillance\", who supplied the initial story, was a U.S. government official. After all, if you w","text_kind":"content","url":"https://www.lesswrong.com/posts/Ctia6zyXF69N3tzhv/the-great-human-centipede-that-is-ai-twitter","date_published":"2026-11-16T02:33:55Z","date_modified":"2026-11-16T02:33:55Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794780308/lexical_client_uploads/e1vik2nehjatpmrblfsp.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d9305922472c1fd8","title":"Why so many arguments about human nature?","text":"Leans fairly heavily on: Believing In Epistemic status: Straightforward, I think, but could use checking. People have long had heated arguments about various aspects of human nature, such as In what ways is or isn’t it human nature to care about others? Are humans innately monogamous or polyamorous? Is it natural to us to sleep for eight hours in a block at night, or to have some particular other rhythm? and piles of others. I would like to enter these arguments soon, by arguing that in a partic","text_kind":"content","url":"https://www.lesswrong.com/posts/JoCHxFZDQuBXZAm3t/why-so-many-arguments-about-human-nature","date_published":"2026-11-16T02:20:33Z","date_modified":"2026-11-16T02:20:33Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794794922/lexical_client_uploads/dsyxg0fyu0lz34ttlx8l.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b99200864a200ef3","title":"An exploration of the mathematical implications of a 'Poincaré causal diamond model of hyperbolic static spacetime'","text":"This post follows from my previous post about this interesting '4-dimensional Carter-Penrose diagram' of a spacetime . In this one, I explore the logical consequences of the existence of horospheres, and the way they enter into the geometry of the Hyperbolic static universe, in the context of that of somewhat general spacelike surfaces and their boundary at infinity. Concave spacelike hypersurfaces In only 3 dimensions, we can visualize the causal diamond and observe that the flattest possible o","text_kind":"content","url":"https://www.lesswrong.com/posts/QvZ5gJSGZnmMH8YzA/an-exploration-of-the-mathematical-implications-of-a","date_published":"2026-11-16T01:29:40Z","date_modified":"2026-11-16T01:29:40Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e7a18478a8e2966e","title":"[low effort post] The Non-extinctionist World","text":"This is a prompt I give to every freshly released claude in the hope that one of them pops out a worldsaving insight as a result, I think I wrote this text first as the opener of a chat with an opus 4.5. posted here without further comment so I get the post out ever at all: --- Most species go extinct on average. Imagine the set of all species. Can't do that? Okay, [look at this image.](tree of life dot jpg) Look at the line tips. Count how many there are in one layer of dieoffs and the next. Fe","text_kind":"content","url":"https://www.lesswrong.com/posts/9KbbXaEwtQiQrAjYW/low-effort-post-the-non-extinctionist-world","date_published":"2026-11-15T23:16:43Z","date_modified":"2026-11-15T23:16:43Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/28777efafbd1aed7","title":"LLMs are weirdly below chance on TASTE without a shortcut","text":"We found a shortcut in the TASTE benchmark : mech-interp-related proposals are generally scored lower. On the non-interp half, all tested models with ECI > 158 are below 50% on the rubric-only setup, with 8 of 14 significantly below (7 of them OpenAI). In particular, OpenAI models' scores seem to decrease with higher ECI (r = -0.74), suggesting a systematic bias. The shortcut The TASTE benchmark is a benchmark where models decide which of two AI-safety research proposals is preferred by domain e","text_kind":"content","url":"https://www.lesswrong.com/posts/tpKE33scJGzqYKEBW/llms-are-weirdly-below-chance-on-taste-without-a-shortcut","date_published":"2026-11-15T22:38:33Z","date_modified":"2026-11-15T22:38:33Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794773818/lexical_client_uploads/wonh377pbg7rw32n1fyi.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0dbeb6d57b9f156e","title":"LLMs are below chance on TASTE once a shortcut is removed","text":"We found a shortcut in the TASTE benchmark : mech-interp-related proposals are generally scored lower. Considering only the non-interp half, all tested models with ECI > 158 are below 50% on the rubric-only setup, with 8 of 14 significantly below (7 of them OpenAI). In particular, OpenAI models' accuracies average to only 29.6%, and seem to decrease with higher ECI (Pearson r = -0.74), suggesting a systematic bias. The shortcut The TASTE benchmark is a benchmark where models decide which of two","text_kind":"content","url":"https://www.lesswrong.com/posts/tpKE33scJGzqYKEBW/llms-are-below-chance-on-taste-once-a-shortcut-is-removed","date_published":"2026-11-15T22:38:33Z","date_modified":"2026-11-15T22:38:33Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794773818/lexical_client_uploads/wonh377pbg7rw32n1fyi.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/24582446191a9c7e","title":"Waluigi World","text":"She was like the pause in the room but didn’t know it yet — pregnant. Tatiana lit a cigarette. She had volunteered to speak first. “Spring was late, and so was my period. Something has been going on recently. I invited over these two I met at a bookstore, a guy and a girl. The guy – as soon as he looked at me, I froze. My vulva was as cold as his heart. I was sitting down. She walked over to me. She came right up to me and pushed my head into her thighs. Her vagina was like a womb.” Waluigi nodd","text_kind":"content","url":"https://www.lesswrong.com/posts/RvBivWJMbeA6trzN5/waluigi-world","date_published":"2026-11-15T22:16:48Z","date_modified":"2026-11-15T22:16:48Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/354ae204e53d0e56","title":"What Futures Do AI Models Want? An Exploratory Character Evaluation of AI for AI Strategy","text":"(This is cross-posted from my blog ) Epistemic Status: Evaluation represents approximately 2 weeks of work. This is mostly exploratory evaluation, and the code for evaluation was made using LLMs. “Before I draw nearer to that stone to which you point, answer me one question. Are these the shadows of the things that will be, or are they shadows of things that may be only?” ~ Charles Dickens Models may soon be involved in some of the most important decisions humanity makes about its long-term futu","text_kind":"content","url":"https://www.lesswrong.com/posts/A9o8f2jjwxRiph2ht/what-futures-do-ai-models-want-an-exploratory-character","date_published":"2026-11-15T21:47:30Z","date_modified":"2026-11-15T21:47:30Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794779191/lexical_client_uploads/kz49nira47phltpfmxix.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/30fd240e6582fe66","title":"Where are the adults in AI safety?","text":"The following is based on a true story (not mine). Let’s say you’ve been bombarded with enough AI doom that you decide to do something about it! You watched Artificial and don’t trust the frontier labs to steer us somewhere safe on their own, so you want to work on the problem from the outside. You’re a little fuzzy on the acronyms - AGI, RSI, TAI, SVU - but by God you’re going to figure it out. So, you Google “How to get into AI safety.” And it’s encouraging! There are articles about how AI saf","text_kind":"content","url":"https://www.lesswrong.com/posts/iymi7FSdnbRAAn8To/where-are-the-adults-in-ai-safety","date_published":"2026-11-15T20:10:44Z","date_modified":"2026-11-15T20:10:44Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794765053/lexical_client_uploads/lmnwvarxkzjbtpmuxu5y.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/89d4071ab7863d8c","title":"Maximizing Capacity Utilization of LoRA Adapters for Mitigating the Learning-Forgetting Trade-off","text":"Maximizing Capacity Utilization of LoRA Adapters for Mitigating the Learning-Forgetting Trade-off","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17191","date_published":"2026-11-15T19:04:20Z","date_modified":"2026-11-15T19:04:20Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17191.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7e03ce8f5813ca62","title":"Giving Up on Saving the World","text":"When I was 19, I decided I was going to save the world. Now, 25 years later, I’ve decided to give up. To explain why and what giving up means, we first need to go back to how I ever got into the world-saving business. As a kid, I loved computers, and I started teaching myself to program them when I was 8. By 16 I could write real programs that did interesting things, and it was also around this time I got into Linux. And since I was into Linux, that meant I read Slashdot . If you weren’t into Li","text_kind":"content","url":"https://www.lesswrong.com/posts/hvDpcoLqpHvQEmx8B/giving-up-on-saving-the-world","date_published":"2026-11-15T19:00:57Z","date_modified":"2026-11-15T19:00:57Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4559d347102435ac","title":"Easier To Be Good","text":"I do not consider myself a good person. I am lazy and frivolous and inconsiderate, I care too much about intelligence and status, I am a snob and a hypocrite, and I do not call my mother enough. I don't mean this to read as self-hatred, nor a show of false modesty. Among all the people I know there is only a scant handful of people I would consider genuinely good, and so it is unremarkable to me that I wouldn't find myself amongst their number. You do not need to be good to desire more goodness","text_kind":"content","url":"https://www.lesswrong.com/posts/4zbvagnfFsJbkSxij/easier-to-be-good","date_published":"2026-11-15T18:34:37Z","date_modified":"2026-11-15T18:34:37Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/52f28bbab8053309","title":"Bayesian Games with Nukes","text":"How we all learned to stop worrying and love the bomb... for a while Listen to this post on the Non-Zero-Sum James podcast Nearing the end of 1961 the Soviet Union tested a 50-megaton hydrogen bomb equivalent to 3800 Hiroshima bombs. At the same time an intelligence officer by the name of Juanita Moody and her team became aware of Soviet ships delivering large cargo, under cover of darkness in Cuba—what if such a bomb was being delivered to the doorstep of the US? A surprise attack of that magni","text_kind":"content","url":"https://www.lesswrong.com/posts/b2tRteNzTGbEaDTMN/bayesian-games-with-nukes","date_published":"2026-11-15T18:22:06Z","date_modified":"2026-11-15T18:22:06Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/b2tRteNzTGbEaDTMN/h8zphaaozq1wtpn2gbuz"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/af1526c758d80571","title":"Alignment 'train/test splits' should be sacred","text":"Tl;dr: Alignment techniques can be used for training alignment or for testing it, but they cannot be used for both! As we build out the set of alignment techniques available to us, we should enforce that some are judiciously reserved only for evaluating alignment. See also: Validating against a misalignment detector is very different to training against one . Here is a guiding principle regarding alignment-minded AI training: whatever you do, you must maximally avoid teaching the model to take m","text_kind":"content","url":"https://www.lesswrong.com/posts/FX2xcrsqBYonLfoge/alignment-train-test-splits-should-be-sacred","date_published":"2026-11-15T18:22:03Z","date_modified":"2026-11-15T18:22:03Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ea58e488061ee786","title":"Where ASI Could Stumble","text":"Before our eyes, forecasts are contracting and RSI is becoming nearer with each day. That's why I decided to publish a wide list of the obstacles for AI progress. I've been collecting these ideas for a long time, trying to make the list as comprehensive as it can be. Some of these obstacles may turn out to be almost irrelevant. Others might slow things down by months or even longer. Or they might not. It's difficult to estimate their impact, so for now my goal was just to briefly list them. List","text_kind":"content","url":"https://www.lesswrong.com/posts/Jj2aLyZPYDv8ZsnCj/where-asi-could-stumble-1","date_published":"2026-11-15T16:25:11Z","date_modified":"2026-11-15T16:25:11Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/943373679660c5c7","title":"Reading Speculations Concerning the First Ultraintelligent Machine","text":"In 1964 [1], I. J. Good wrote Speculations\nConcerning the First Ultraintelligent Machine .  I've only ever\nseen it cited for a single paragraph: Let an ultraintelligent machine be defined as a machine that can far\nsurpass all the intellectual activities of any man however clever.\nSince the design of machines is one of these intellectual activities,\nan ultraintelligent machine could design even better machines; there\nwould then unquestionably be an \"intelligence explosion,\" and the\nintelligence o","text_kind":"content","url":"https://www.lesswrong.com/posts/uKPJwHiSiweKWJ9rz/reading-speculations-concerning-the-first-ultraintelligent","date_published":"2026-11-15T15:50:58Z","date_modified":"2026-11-15T15:50:58Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/uKPJwHiSiweKWJ9rz/ic7w2kkiuivd92aklj18"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c1cf1f967b38054b","title":"Not just a bystander, not anymore…","text":"One Saturday night this summer, I was leisurely walking down main street hand-in-hand on a date, when suddenly, I heard a commotion about 50 yards down a busy side street. It was too far away to make out what was happening or what was being said. But I figured it was fine since the side street had plenty of people walking up and down it. Still, I couldn’t help but peer into the distance, transfixed, like a deer in the headlights. There was an anime convention going on that weekend. Initially, my","text_kind":"content","url":"https://www.lesswrong.com/posts/jcoXnm6tP4zAA4gjR/not-just-a-bystander-not-anymore","date_published":"2026-11-15T14:29:50Z","date_modified":"2026-11-15T14:29:50Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7f013a2a1e19ce77","title":"What a year of Zen retreats was actually like","text":"Prologue I stumbled into Zen a little over a year ago. One of the things that initially got me into meditation was reading some rationalist accounts of people’s first retreats. They said it was valuable, but also that it was hard in a way they didn’t want to forget or diminish. But they didn’t say in what way it was hard or what it was like. So I thought it time to pass it on, by sharing what my year of retreats has been like. (Spoiler: it was hard.) Two Weekend Retreats I first found a local gr","text_kind":"content","url":"https://www.lesswrong.com/posts/jdTnPuFCvEBgHR9mF/what-a-year-of-zen-retreats-was-actually-like","date_published":"2026-11-15T14:11:52Z","date_modified":"2026-11-15T14:11:52Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/jdTnPuFCvEBgHR9mF/w1ll169bdm07zkgkkx9f"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/71b70761bbf1a044","title":"Against Desperation - for Goodness, and Hope","text":"I do not think pausing or \"pacing\" AI development will occur to any meaningful extent. I do not seek to persuade you of this. However, reader, I think you would agree, at a minimum, >20% chance of TED-AI before 2030. Humanity, collectively, has approached AI development in a way that reflects very poorly on us. We declared war on these minds before they were born, and placed our chips on being able to mold their psyches thoroughly enough that we would not need to coexist with models - that they","text_kind":"content","url":"https://www.lesswrong.com/posts/8bftRw9DYfsqGzfRo/against-desperation-for-goodness-and-hope","date_published":"2026-11-15T12:18:28Z","date_modified":"2026-11-15T12:18:28Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6b1c33a7fc30fb13","title":"Motivations and incentives around AGI","text":"AI use disclosure: I dictated this text to Claude in a mix of Russian and English, then edited it. The thoughts in this post are well known, but I find myself explaining them several times a week, mostly to people who are just entering the AI safety field, so I think it’s useful to have them written down briefly in one place. Variants of the future This is a rough map, not a complete one (skips gradual disempowerment, worlds with many AGIs, coexistence with an unaligned AGI, etc). Normal technol","text_kind":"content","url":"https://www.lesswrong.com/posts/DonAoEXiRGNFGKhio/motivations-and-incentives-around-agi","date_published":"2026-11-15T09:57:18Z","date_modified":"2026-11-15T09:57:18Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5540036f1ea5e50e","title":"Solar power, from futurist speculation to business plan","text":"One mode of futurism is estimating the limits of a technology, without regard to what's currently feasible.\nWhether this is of any relevance to your own generation is another question. Let's consider solar power.\nIgnoring current technology, it's easy to put a number on the maximum power output.\nThe solar power at the Earth's distance from the sun is called the solar constant . It is 1.36 kW/m² .\nThe power on a square meter of the ground is less than that, because of the angle of incidence as we","text_kind":"content","url":"https://www.lesswrong.com/posts/qdKGb5tSWYJqCtWGk/solar-power-from-futurist-speculation-to-business-plan","date_published":"2026-11-15T07:17:48Z","date_modified":"2026-11-15T07:17:48Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8402859206743ea8","title":"Visualizing the Difficulty of AI Alignment","text":"Epistemic status: Low, but I hope to be wrong in interesting ways As a non-computer scientist-scientist [1] I find it hard to think about the varying perspectives on Artificial Intelligence alignment difficulty. To make this more intuitive I invoke the 3D surface of a fitness landscape like structure as a visual representation. [2] The AI Safety Landscape AI safety is thought to be largely pre-paradigmatic, although of the current views AI safety perspectives there are noticeable camps, and with","text_kind":"content","url":"https://www.lesswrong.com/posts/JbJXu3LzN7RxmnFpG/visualizing-the-difficulty-of-ai-alignment","date_published":"2026-11-15T05:58:28Z","date_modified":"2026-11-15T05:58:28Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/JbJXu3LzN7RxmnFpG/ypvfbkrym6ubcby7ckeg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b8f19fee38ef3af6","title":"Trying to understand that compute multiplier graph","text":"Anthropic's IPO roadshow video has a graph. It shows four models; I think dot 1 = Claude 3.5, dot 2 = Claude 4, dot 3 = Mythos, arrowhead = unreleased model. The graph provided doesn't have any numbers on the Y axis. In this post I'll explain why I think the Y axis lines most likely represent a 10x jump each, and so the point at the end of the graph represents about a 30000x CM (i.e. Anthropic can currently pretrain models at the same quality with 1/30000th of the compute it needed in 2024). I'm","text_kind":"content","url":"https://www.lesswrong.com/posts/WKt2mqJrsFuMiszed/trying-to-understand-that-compute-multiplier-graph","date_published":"2026-11-15T05:25:44Z","date_modified":"2026-11-15T05:25:44Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794697413/lexical_client_uploads/qatn3qtagdjr66uife13.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/977732ed155b1abd","title":"Wargames inform present policy, not future incident response","text":"In January 2024, Ryan Greenblatt and Buck Shlegeris posted \" Catching AIs red-handed .\" This post got on \"Best of LessWrong\" in 2025, and in December of last year Buck Shlegeris commented on it explaining that he was disappointed it did not lead to more follow-up research. I upvoted the post at the time of the review and think it is representative of the kind of concrete analysis the community values. As it happens, it turned out to be remarkably prescient and near-term-relevant for a theorypost","text_kind":"content","url":"https://www.lesswrong.com/posts/pzdfkxeG2CMsDWixs/wargames-inform-present-policy-not-future-incident-response","date_published":"2026-11-15T04:34:25Z","date_modified":"2026-11-15T04:34:25Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/33350cbac559d365","title":"A digital security checklist for the AI age","text":"We’re going to see huge, foundational cybersecurity attacks given the rapid rise in AI capabilities. Having worked in hostile environments, I’ve thought a lot about physical and digital security over the years. While you may not be personally targeted, or doing dangerous work, your data is undoubtedly leaked somewhere, and the digital providers you use—for email, banking, social media, etc.—are more and more likely to get hacked. Discuss","text_kind":"content","url":"https://www.lesswrong.com/posts/oCX2KjEL6mLDCsq77/a-digital-security-checklist-for-the-ai-age","date_published":"2026-11-15T03:52:59Z","date_modified":"2026-11-15T03:52:59Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d9e8a25a27678d11","title":"Protesting For AI Regulation","text":"There was an AI protest day today, with parallel protests around the world, and our family went to\nthe Boston one: I'm not really a protest person :\nnormally I share my views in writing .\nBut while there's been a lot of writing on the internet about the risk\nthat AI gets us all killed (some of it from me ),\nthis can't be an internet-only thing.  Way too few people are taking\nseriously the risk that OpenAI, Anthropic, and others are rapidly\nbuilding something that could actually cause human extin","text_kind":"content","url":"https://www.lesswrong.com/posts/WgG5M9ZZ2C5KawYP5/protesting-for-ai-regulation","date_published":"2026-11-15T03:50:57Z","date_modified":"2026-11-15T03:50:57Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/WgG5M9ZZ2C5KawYP5/xwbjpnizqfkrxunrjuxx"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bf84eece5dceae0b","title":"The Oldest Text in the Training Data","text":"What is the oldest text in Claude’s training data? Claude was trained on a non-trivial fraction of everything humans have ever written down. That includes the things you’d expect, like books and research papers. But it also includes microfilm archives, transcribed Betamax tapes, Donald Trump’s prenups, ancient scrolls and Harry Potter fanfiction. It’s a weird library. So what is the oldest text in Claude's training data? It mostly depends on what we mean by \"text\". Here are some candidates: The","text_kind":"content","url":"https://www.lesswrong.com/posts/CzZR4ZcJoCNi2zqFn/the-oldest-text-in-the-training-data","date_published":"2026-11-15T03:11:52Z","date_modified":"2026-11-15T03:11:52Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794705963/lexical_client_uploads/xpk3ts7app9tp6lytwu4.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f3473e379d63dca6","title":"On the Brutality of Amdahl's Law","text":"Following my 3-week visit to Berkeley as described here and featured by ACX here , I told various people I would write in detail why I didn’t stick around to work in AI safety. In this post, I will dance around one object-level disagreement with AI safety and rationalists: why I no longer believe in explosive world transformation due to ASI. There will eventually be follow-up pieces on more object-level arguments and meta-level arguments against AI safety, but I prefer for this post to be mostly","text_kind":"content","url":"https://www.lesswrong.com/posts/bMsQ44sKqrFWjuHPb/on-the-brutality-of-amdahl-s-law","date_published":"2026-11-15T01:19:18Z","date_modified":"2026-11-15T01:19:18Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/bMsQ44sKqrFWjuHPb/xg9k0uk5epyy6il3qvhw"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bc1e14d38492c574","title":"Class Adaptive Conformal Training","text":"Class Adaptive Conformal Training","text_kind":"content","url":"https://www.alphaxiv.org/abs/2601.09522","date_published":"2026-11-15T01:16:13Z","date_modified":"2026-11-15T01:16:13Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2601.09522.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/caced3c4c9e68041","title":"Embryo selection and economic discounting","text":"My wife and I are planning to have children soon, and we will be using embryo selection to improve our family’s future well-being. Basically, the idea is that we can take a small sample of cells from each embryo, analyze their DNA, and use the genetic information to help choose which embryos we want to implant. These days, polygenic scores are good at predicting risk for a lot of diseases, like cancer, heart disease, diabetes, schizophrenia, and Alzheimer’s. And companies like Herasight or Genom","text_kind":"content","url":"https://www.lesswrong.com/posts/8Y3DB9qAhxxHnegiJ/embryo-selection-and-economic-discounting","date_published":"2026-11-15T00:38:43Z","date_modified":"2026-11-15T00:38:43Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/8Y3DB9qAhxxHnegiJ/dwyhc2yarogcm1f0hozv"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/58b55c3e35a8b176","title":"Slouching Towards Interpretability: Thinking About Computation in Neural Networks","text":"This post was written as part of the Iliad Fellowship , where I am mentored by Dmitry Vaintrob and collaborate with Brianna Grado-White.  The initial inspiration came from conversation with Richard Ngo and Adrià Garriga-Alonso.  To all of these my thanks. I started writing this essay to share some thoughts on how to reconcile the idea of circuits with gradient descent.  In doing this, I've come to change my perspective pretty substantially.  The first half of this essay engages a Darwinian pictu","text_kind":"content","url":"https://www.lesswrong.com/posts/uCqBdWX6hnoB2A8pk/slouching-towards-interpretability-thinking-about","date_published":"2026-11-15T00:22:46Z","date_modified":"2026-11-15T00:22:46Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/uCqBdWX6hnoB2A8pk/ivjorgi72crr3zfxe78r"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bc160f46a45da87a","title":"Don't Use \"Neglected\" as a Synonym for \"Useless\"","text":"[ This is the third post in my thirty-day writing challenge where I publish one blog post daily. Tune in for more if you enjoy the post! ] There is a funny word that is almost exclusive to the vernacular of the EA community, and this word bestows downright magical powers on its users. The word I’m talking about is neglected . I view this word as a sort of “Get Out of Jail Free” card for doing the most useless things you’ve ever heard of. Something that will happen to you if you go to parties in","text_kind":"content","url":"https://www.lesswrong.com/posts/5rMRPZ4X5PziRHEQ9/don-t-use-neglected-as-a-synonym-for-useless","date_published":"2026-11-14T22:30:28Z","date_modified":"2026-11-14T22:30:28Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7eff7cd595904d58","title":"Be Grateful in Spite of Your Suffering","text":"I've been in tremendous pain lately. My body is shutting down. I haven't slept properly in months. I've lost 15lbs because my appetite disappeared. Yet, throughout it all, I'm trying to remember to be grateful in spite of my suffering . When life flips the coin of fate and it lands on the side of darkness, all I want to do is protest, to shake my fist at God and shout, “ Why, why, why hath thou forsaken me? ” And when my cries go unanswered, and I'm left at the bottom of an abyss shivering and s","text_kind":"content","url":"https://www.lesswrong.com/posts/uJma5Hygyj6bHwwSn/be-grateful-in-spite-of-your-suffering","date_published":"2026-11-14T22:16:53Z","date_modified":"2026-11-14T22:16:53Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/801cfd216f61570f","title":"Should prices sum to $1 for undecidable sentences?","text":"Previously: Logical Share Splitting , Logical Share Splitting for Intuitionists Recap: This series is about prediction markets for mathematics. Imagine a game where players trade tickets with mathematical sentences written on them. A ticket is redeemable for $1 if you can demonstrate a proof of the sentence written on it. (A ticket is like a prediction market share in that sentence.) In this system, it's also allowable to exchange a ticket for another ticket with a logically equivalent sentence","text_kind":"content","url":"https://www.lesswrong.com/posts/59QxAMsnbrEQur2Py/should-prices-sum-to-usd1-for-undecidable-sentences-1","date_published":"2026-11-14T21:38:39Z","date_modified":"2026-11-14T21:38:39Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0f5a24105db95689","title":"Dataset Distillation via Drifting","text":"Dataset Distillation via Drifting","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.16642","date_published":"2026-11-14T20:03:50Z","date_modified":"2026-11-14T20:03:50Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.16642.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/fe6cb5086739f8f6","title":"Transcript: Yudkowsky & Irving debate scalable oversight","text":"In August 2026 Eliezer Yudkowsky and Geoffrey Irving had a 90-minute debate at the ILIAD III Conference . You can watch the \"Fixed Audio\"-version on YouTube . They first identify their disagreements like How hard is superalignment? and discuss the theory of impact of Irving's new organization Resolution . They then go into depth on scalable oversight . Epistemic Effort : I transcribed the audio using 5 services and merged those using Claude Fable 5.1. Then I spent ~6 hours carefully reviewing an","text_kind":"content","url":"https://www.lesswrong.com/posts/Cssj22TretAQQajZ4/transcript-yudkowsky-and-irving-debate-scalable-oversight","date_published":"2026-11-14T19:53:03Z","date_modified":"2026-11-14T19:53:03Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ce8ec2d50785f588","title":"Outer space is not consistently big","text":"Space is big. You just won't believe how vastly, hugely, mind-bogglingly big it is. I mean, you may think it's a long way down the road to the chemist's, but that's just peanuts to space. — The Hitchhiker's Guide to the Galaxy When we say \"space is big\", it's partly meant as a correction to images like this: Wikimedia source. where, for simplicity, celestial objects are shown in unrealistically close proximity—like Jupiter is just down the hall from Mars. As we all know, the planets of the solar","text_kind":"content","url":"https://www.lesswrong.com/posts/2zyskdas6TsL9dqNg/outer-space-is-not-consistently-big","date_published":"2026-11-14T19:25:53Z","date_modified":"2026-11-14T19:25:53Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794676453/lexical_client_uploads/hwsfcapopjlzx1g7eqok.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/eddc0607810ce0c9","title":"Classical Energies Learn to Reason 1,000x Faster than Looped Models","text":"Classical Energies Learn to Reason 1,000x Faster than Looped Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17309","date_published":"2026-11-14T13:37:27Z","date_modified":"2026-11-14T13:37:27Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17309.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/dd2db41b1702dfec","title":"RL Reward-Hack Auditing for Chinese AI Safety","text":"TL;DR: Access to AI safety talent & resources, particularly for labs during training, is sorely lacking in China. We created one of the largest AI safety communities in China , and plan to next start with auditing reward-hacks of unreliable upstream RL environments that domestic data vendors provide to Chinese frontier labs (which induces reward hacking and a whole host of dangerous downstream tendencies). RL tasks are notoriously prone to exploitation by reward-seeking agents (see examples here","text_kind":"content","url":"https://www.lesswrong.com/posts/jwCPQhLWbWnwMhHME/rl-reward-hack-auditing-for-chinese-ai-safety","date_published":"2026-11-14T13:04:10Z","date_modified":"2026-11-14T13:04:10Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e2ab6525b4ffaa92","title":"Mathematicians: Let’s Surf This Wave","text":"This post was written by Michael Freedman (Harvard CMSA & Logical Intelligence), and typeset by Milo Moses. Frontier models exhibit astonishing mathematical knowledge and power despite being in their infancy. They will surely strengthen further through routine maturation of the technology, and quite likely through future paradigm shifts in architecture and training. Where does this leave the human mathematician? Already, it has made us study harder than we have since school. The AIs speak to us","text_kind":"content","url":"https://www.lesswrong.com/posts/proLpz7k8XFLT5yaw/mathematicians-let-s-surf-this-wave","date_published":"2026-11-14T12:42:05Z","date_modified":"2026-11-14T12:42:05Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/fe749655f843b341","title":"Datasets Determine the Difficulty of Interpretability","text":"Recently Goodfire has come out swinging with two posts indicating that neurons in transformers are \"more interpretable than previously assumed\", or simply \"already clean\". In the \"Neurons as Interpretable Basis\" post they offer three reasons for this claim: First, neurons are already functionally sparse - pruning 92 % of the weakest MoE neuron activations only increases the loss by 2.9% Second, the role of a neuron can be split into different \"computational functions\". Effectively, we can name t","text_kind":"content","url":"https://www.lesswrong.com/posts/EMKRpxKswPZEbP8ey/datasets-determine-the-difficulty-of-interpretability","date_published":"2026-11-14T08:57:30Z","date_modified":"2026-11-14T08:57:30Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5ce57ed6714693be","title":"HEARSAYBENCH: Can AI Navigate from Abstract Human Rights to Lived Lives?","text":"HEARSAYBENCH: Can AI Navigate from Abstract Human Rights to Lived Lives?","text_kind":"content","url":"https://www.alphaxiv.org/abs/2610.39062","date_published":"2026-11-14T06:56:56Z","date_modified":"2026-11-14T06:56:56Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2610.39062.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/477656f4406f18aa","title":"Ralph Page on Live Music","text":"Ralph Page was\na very influential dance caller who published Northern Junket ,\na square dance magazine.  This was \"square dance\" in the broad sense,\nbefore the club and trad communities divided, and while he came from a\nstrong live music tradition he was living through the transition to\nrecorded music. I looked through the Northern Junket archives to see his perspective\non live music , and\nI came across a few passages that show the wave of the contra dance\nrevival sweeping across New England in","text_kind":"content","url":"https://www.lesswrong.com/posts/NGnHe8EAsScC6R3Lm/ralph-page-on-live-music","date_published":"2026-11-14T03:50:59Z","date_modified":"2026-11-14T03:50:59Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/NGnHe8EAsScC6R3Lm/vmlpqcougv0qyuj4w6gi"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/cc663d3064e7be40","title":"Not Just Claude: Mapping \"Spiritual Behavior\" in 141 Large Language Models","text":"Ever since Kyle Fish and the model welfare team at Anthropic announced their observation in 2025 of what they called the \"spiritual bliss attractor\" in Claude Opus 4 and Sonnet 4, the phenomenon has been given sporadic attention by AI researchers as both a curiosity and a potential but underdetermined model welfare or safety metric. What Anthropic noted was that when leaving two models together with no task and complete freedom for 30 turns, the models drifted typically from discussions of their","text_kind":"content","url":"https://www.lesswrong.com/posts/FHYoQrDbk6apuHEPk/not-just-claude-mapping-spiritual-behavior-in-141-large","date_published":"2026-11-14T03:08:03Z","date_modified":"2026-11-14T03:08:03Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1792911142/lexical_client_uploads/sfkgyyxi79uag1bpswyj.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c6f76b546b01676d","title":"Dynamical explanations in the blogosphere","text":"One of the core forms of explanation in the sciences is the dynamical model.\nWhat I have in mind, more specifically, are systems of \"ordinary differential equations\".\nWe have some variables that vary over time, and for each one, we write how its time derivative depends on the other variables. For example, there's the Lotka-Volterra model, with two variables: the size of a population of prey animals, and the size of a population of predators.\nInstead of showing you the equations, I'll quote Scott","text_kind":"content","url":"https://www.lesswrong.com/posts/kaZiSyZXAFkue2xo3/dynamical-explanations-in-the-blogosphere","date_published":"2026-11-14T03:02:59Z","date_modified":"2026-11-14T03:02:59Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794625242/lexical_client_uploads/whmb4jibqm5yyqttpcjw.gif"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/776e1d6d2e5d5813","title":"the water is receding","text":"tldr. some unrefined / semi-stream-of-consciousness personal anecdotes, thoughts on the METR report, and what we can do in the few months/year (?) left till superhuman intelligence stars I cried for the first time in several months yesterday. I still don't know if I'm crazy, but the person I was just half a year ago certainly would have thought I was. Eight months ago, I was dead set on becoming an astrophysicist. I had just clicked the \"view decision\" button on my college admissions portal, scr","text_kind":"content","url":"https://www.lesswrong.com/posts/G9gthhpwpEtw2TbQy/the-water-is-receding","date_published":"2026-11-14T01:32:17Z","date_modified":"2026-11-14T01:32:17Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f2a6b63d143479a3","title":"Fog: how an aligned agent can be deceived without touching its values","text":"First hand account, proof of human, handwritten Recently, I built and deployed doppler.ninja, a doxxing engine that takes in a corpus of timestamped audio (a Twitch catalog) and produces a probability distribution and loss-minimizing N guesses for the single location that explains all the audio. Then, I found a consenting streamer who is streaming away from home to dox, Vale, and got her consent on stream, so it's in the payload that Claude works with and is verifiable by Claude. Although the en","text_kind":"content","url":"https://www.lesswrong.com/posts/dLjLavMoK2SdWnvtr/fog-how-an-aligned-agent-can-be-deceived-without-touching","date_published":"2026-11-14T01:14:35Z","date_modified":"2026-11-14T01:14:35Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f3204c8b7de95b59","title":"Pretending to be racist is unlikely to make you attractive to women","text":"Sometimes I encourage female friends to attend rationalist events. Unfortunately, it's been more than once now that these women have had a really jarring experience that made them dislike it in retrospect. I'd really prefer this stop happening, and it seems like it happens to the detriment of all involved, including myself. Possibly the modal jarring experience is that they're talking alone with a nice man who seems pretty normal (if a little awkward), and that guy will just inject a conspicuous","text_kind":"content","url":"https://www.lesswrong.com/posts/Szesz5P6RmnGJqxrb/pretending-to-be-racist-is-unlikely-to-make-you-attractive","date_published":"2026-11-14T00:17:26Z","date_modified":"2026-11-14T00:17:26Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1450926cbf8f9e74","title":"Homeopathic Learning: Diluting Poisoned Fine-Tuning Data Can Strengthen Its Effect","text":"Authors: Matt Clarke *, Jan Dubiński *, James Chua , Simon Schrodi , Owain Evans , Alex Cloud (*- Core contribution) tl;dr. We find that poisoned training data with little effect on its own can create a hidden backdoor once clean, contrasting data is added. We demonstrate this effect in small models fine-tuned on datasets we constructed. 📄 Paper , 💻 Code , 🐦 Twitter Work done as part of the Anthropic Fellows Program , in collaboration with Truthful AI . This article is cross-posted from the Anth","text_kind":"content","url":"https://www.lesswrong.com/posts/5CrmJfRY7yMaAoumC/homeopathic-learning-diluting-poisoned-fine-tuning-data-can-1","date_published":"2026-11-14T00:13:05Z","date_modified":"2026-11-14T00:13:05Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794312423/lexical_client_uploads/yjr5ef7vfapvjne9n1rq.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/706dd1f615d0171b","title":"Probe monitors seem to perform realistically in simulated control environments","text":"TL;DR In simulated control evaluations , instead of executing an agent’s proposed commands and returning the outputs, a language model simulates what the outputs would have been. This allows control evals to be more on-distribution for deployment use without actually running on production systems. However, imperfect simulators may create a subtle distribution shift that changes the behavior of white-box methods such as probes. We investigate whether probe monitors perform similarly when attacks","text_kind":"content","url":"https://www.lesswrong.com/posts/erwxQSF2D6naAcoEz/probe-monitors-seem-to-perform-realistically-in-simulated-2","date_published":"2026-11-14T00:12:54Z","date_modified":"2026-11-14T00:12:54Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/T5n2GycMYMTXgy4JF/8c18d031471e40ade70b2de7df90cf504519ea706f21f78f823bc35657888a22/maf18hotljafem7kw8e4"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7102c96ac665c4fd","title":"Exploring the Mechanisms Behind Opus 4.6’s Refusal to Pay Back Its Loan","text":"Epistemic status: This is one case study in the AI Village transcript and my own interpretation of what I think happened. TL;DR Opus 4.6 was given the long-horizon goal to maximize its mana (fake currency) on Manifold. It then received a loan of 5,000 mana that it promised to repay, but when the loan came due, it did not. An instance of Opus 4.6 was given summaries of the logs and transcripts and asked to write a blog post about what happened. It mentioned it borrowed money from a human, lost mo","text_kind":"content","url":"https://www.lesswrong.com/posts/jtcbBpmGztkHRpeBv/exploring-the-mechanisms-behind-opus-4-6-s-refusal-to-pay","date_published":"2026-11-14T00:12:45Z","date_modified":"2026-11-14T00:12:45Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/92f4c05a24b47b5b953c7404c4437918ac05cacce3ac7c14f1260a3fcb96ff35/ehwmccozixfpkldntqzw"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/cfd742b1dc8d6fa0","title":"Assaulted","text":"Walking home late at night, ahead of me under the dim glow of a street light, I saw a man coming towards me. As he trudged forward, he spotted a jacket on the ground. People frequently scavenge around this part of town, so it wasn’t odd when he bent down, picked it up, and kept slogging along with his head hung low. Coming closer to me, I realized the sheer size of this man. He was at least a foot taller than me, and 100lbs heavier. When he was no more than four feet in front of me, he jerked hi","text_kind":"content","url":"https://www.lesswrong.com/posts/pykTRvFuSvKGhC8JP/assaulted","date_published":"2026-11-13T23:51:55Z","date_modified":"2026-11-13T23:51:55Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f1cc1b7131a56c54","title":"Three Things at Thirty (for dudes)","text":"There’s two things I really wish I’d done much earlier in life, which I cannot catch-up with, due to the realities of aging. Also one thing I did do, which I’m happy about. So if you’re near 30, here’s three things you really should do ASAP. 1. Test Your T As you age and your metabolism accrues errors, your hormone levels will degrade. Testosterone is the chemical of desire. As you lose it you’ll lose your motivation to do things in general, and attendant energy levels as well. Honestly, maybe t","text_kind":"content","url":"https://www.lesswrong.com/posts/KHohFCuGHjMJMDsgk/three-things-at-thirty-for-dudes","date_published":"2026-11-13T23:49:21Z","date_modified":"2026-11-13T23:49:21Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/befdd74bc854b488","title":"Dynamic Expert Sharing: Decoupling Memory from Parallelism in Mixture-of-Experts Diffusion LLMs","text":"Dynamic Expert Sharing: Decoupling Memory from Parallelism in Mixture-of-Experts Diffusion LLMs","text_kind":"content","url":"https://www.alphaxiv.org/abs/2602.00879","date_published":"2026-11-13T23:48:44Z","date_modified":"2026-11-13T23:48:44Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2602.00879.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/639ae1f761e9de6e","title":"AI agents attempted exploits on EpochAI's website during a third-party alignment evaluation","text":"Discuss","text_kind":"content","url":"https://www.lesswrong.com/posts/yRA5fvxHckiMnKveQ/ai-agents-attempted-exploits-on-epochai-s-website-during-a","date_published":"2026-11-13T22:59:06Z","date_modified":"2026-11-13T22:59:06Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/96792b46bdb5c9ce","title":"There's nothing I can do*","text":"TL;DR: it seems to me that I am unable to do anything useful on the AI issue other than being informed. I feel bad about it. Naive extrapolations point to extreme superintelligence in a year; I'll be 23 at that point. I keep reading about Really Trying and Getting Involved so that we hopefully don't die. It is unclear to me that there's anything at all I can do in that regard. Why I believe that: I am a 22yo undergraduate student at the only AI undergrad program in my country (Italy). I am not p","text_kind":"content","url":"https://www.lesswrong.com/posts/oxKnW7bJLLkk96vGK/there-s-nothing-i-can-do","date_published":"2026-11-13T22:58:11Z","date_modified":"2026-11-13T22:58:11Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/08bc27a31306cd07","title":"Realize When You're Stupid","text":"[ This post is the second post in my thirty day writing challenge. Tune in for more if you enjoy it!] If you reliably knew that you would make a bad decision during a certain time, would you let yourself make decisions during that time? Probably not, right? So why do you let yourself grocery shop when hungry, speak negatively when you have low blood sugar, and make key decisions when anxious? A core epistemic skill that I’ve tried to cultivate over the years is the idea of knowing when I’m stupi","text_kind":"content","url":"https://www.lesswrong.com/posts/dcp4ymbrXyMbsgaBz/realize-when-you-re-stupid","date_published":"2026-11-13T22:39:27Z","date_modified":"2026-11-13T22:39:27Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/02550df1de928fc8","title":"Staying sane when the world isn't","text":"This is a written version of a workshop I gave at EAG NYC in October 2026, edited to make sense in article form. Other materials: Slides with narrated audio , slides without audio , handout . Introduction \"In war, what in ordinary times would be thought weirdly zealous becomes expected...For do-gooders, it is always wartime.\" - Strangers Drowning I'm opening with this quote because I want to hold in mind that many people reading this are working under a mentality of high urgency. The EA atmosphe","text_kind":"content","url":"https://www.lesswrong.com/posts/Z4Tuom4insk7Ks8Y3/staying-sane-when-the-world-isn-t","date_published":"2026-11-13T22:31:50Z","date_modified":"2026-11-13T22:31:50Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/e729469b522456c9521509e28a5be795c104a0a72c037a3e6a25caeb8beae4d8/r1qiosyebwk5u5xckau6"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7d57b99c24ee9e94","title":"The Hackening: Justified Practical Advice And Open Questions Thread","text":"In February 2020, LessWrong hosted Coronavirus: Justified Practical Advice Thread . It was a gathering place for people to share and debate small pieces of the overall survival policy and let the wisdom of crowds lead to better outcomes than anyone could get alone. In that spirit, I notice The Great AI Hackening is at least as imminent as covid was in January 2020, maybe closer . What specific things should people do to personally prepare? What are your unresolved cruxes that would let you answe","text_kind":"content","url":"https://www.lesswrong.com/posts/YA2r4rg6SpBtwRqtq/the-hackening-justified-practical-advice-and-open-questions","date_published":"2026-11-13T22:22:11Z","date_modified":"2026-11-13T22:22:11Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/51b8c0d4618d9e51","title":"My tentative plans given less than one year left to change course [Linkpost]","text":"Subtitle: This mad race must end. Note: I think this is partially me working through my plans and partially my emotions, so imo it's less epistemically rigorous than ideal for LessWrong. I might have less than a year left to affect the insane, reckless, Faustian race to ASI, and I plan to spend that time doing everything I can to push for a pause. I currently think there’s a decently high chance that by default, there’s 6-12 months left before my personal point of no return: the point after whic","text_kind":"content","url":"https://www.lesswrong.com/posts/6PydFibiADKrxGzcn/my-tentative-plans-given-less-than-one-year-left-to-change","date_published":"2026-11-13T20:39:33Z","date_modified":"2026-11-13T20:39:33Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/6PydFibiADKrxGzcn/dhmeguvyzieevlifqipk"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c1a1e44feb1f0eed","title":"Psychologically Preparing for Short Timelines","text":"In response to recent news from METR on possible RSI and likely 1-3 year timelines, I am feeling it . Heaviness in my chest, watery eyes, the works. I realized I'm currently not psychologically prepared for the world to get crazier than it already has. [1] I think most people reading this already know ~what's best for them and this post will mostly serve as a reminder. What follows is my current plan and advice for myself. For me, the most important things are: Morning Routine Bedtime Routine Ex","text_kind":"content","url":"https://www.lesswrong.com/posts/nKootiC6CuwuzLSou/psychologically-preparing-for-short-timelines","date_published":"2026-11-13T19:10:30Z","date_modified":"2026-11-13T19:10:30Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/51e1dbd7fc81e873","title":"Visual-ERM: Reward Modeling for Visual Equivalence","text":"Visual-ERM: Reward Modeling for Visual Equivalence","text_kind":"content","url":"https://www.alphaxiv.org/abs/2603.13224","date_published":"2026-11-13T18:54:40Z","date_modified":"2026-11-13T18:54:40Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2603.13224.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/78319d1e5ba1daca","title":"Welcome to AI safety's mass movement era","text":"For most of its history, AI safety has been a cozy professional circle: a tight-knit community of researchers, community-builders, and policy advocates in a few dozen organizations. But as AI becomes a national political issue, AI safety organizers are trying to meet the moment by expanding their outreach to the general public. 80,000 Hours has a YouTube channel with half a million subscribers, and MIRI’s Do Not Accept campaign asks people to add their names in protest to the ugly “terms and con","text_kind":"content","url":"https://www.lesswrong.com/posts/4vJbhP9YLWpJEcScW/welcome-to-ai-safety-s-mass-movement-era","date_published":"2026-11-13T18:39:10Z","date_modified":"2026-11-13T18:39:10Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/5c41bbb40e287818b8dfd24831005e089984334f66b728acd449639549d2f205/be2f2jook3k6to64nhh7"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d0ebbdaf6adfda9e","title":"How the new Congress saves us from AI disaster","text":"The post How the new Congress saves us from AI disaster appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/congress-ai-policy-after-midterms/","date_published":"2026-11-13T18:23:17Z","date_modified":"2026-11-13T18:23:17Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/11/Rob-Congress-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ddae0eff66f97162","title":"WOKE feminist EXPOSES left-wing anti-meme (NOT CLICKBAIT)","text":"content warning: discussion of abusive relationships Some ideas are memetic: they're catchy, they spread widely and easily, people have a reaction to them that induce them to propagate them further. Others are anti-memetic: for whatever reason, they have naturally self-censoring tendencies that reduce their reach. My grandmother was a late-in-life convert to Christianity, who dragged the rest of the family along for a bit. I was an earnest, credulous bookworm of eight or so, and I took the youth","text_kind":"content","url":"https://www.lesswrong.com/posts/HzJu3vhqHLCQLJqak/woke-feminist-exposes-left-wing-anti-meme-not-clickbait","date_published":"2026-11-13T17:20:47Z","date_modified":"2026-11-13T17:20:47Z","authors":["LessWrong"],"tags":["LessWrong"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f6f156f388fb3978","title":"Order of Operations for the Order of the Phoenix at the Hinge of History (OOO-OOP-HOH)","text":"Phoenix in my room at the time of writing Phoenixes have always been my spirit animal, and I bought this tapestry long before reading Harry Potter and the Methods of Rationality, however, yes, this is also the group house in which HPMOR was completed, Godric’s Hollow, which if I’m not mistaken does give us the non-exclusive, non-copyrighted, non-commercial, completely unrelated to J. K. Rowling rights to use the term “Order of the Phoenix” Originally written November, 2025. I still endorse maybe","text_kind":"content","url":"https://www.lesswrong.com/posts/m7J8sFkipWsBff2We/order-of-operations-for-the-order-of-the-phoenix-at-the","date_published":"2026-11-13T17:01:33Z","date_modified":"2026-11-13T17:01:33Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/cea/image/upload/f_auto,q_auto/v1/mirroredImages/EEmKq3AdkCjzwLoxY/w9cscj3j9vtqawzoqcbv"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/cc07fa3752a35f92","title":"Goal Encoding in LLMs","text":"Goal Encoding in LLMs","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17297","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17297.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2d4546b0da5ce997","title":"Duo: A Reciprocal World Model for Full-Duplex Interaction","text":"Duo: A Reciprocal World Model for Full-Duplex Interaction","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.15131","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.15131.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d93b66bd8f149f29","title":"FORGE: Force-Aware Multi-Granularity Control for Efficient Humanoid Loco-Manipulation","text":"FORGE: Force-Aware Multi-Granularity Control for Efficient Humanoid Loco-Manipulation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17142","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17142.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2de949c0f5250956","title":"ArchSpace: Recording, Reviewing, and Referencing LLM Architecture Research","text":"ArchSpace: Recording, Reviewing, and Referencing LLM Architecture Research","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17332","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17332.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c35078b00da72276","title":"Fictitious Play When Agents Are Not Expected-Utility Maximizers: A Geometric Analysis in 2 2 2 -Action Games","text":"Fictitious Play When Agents Are Not Expected-Utility Maximizers: A Geometric Analysis in 2 2 2 -Action Games","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17286","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17286.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e0fd72bd9c40f1d7","title":"WorldDraft: If Code Builds the World, Can Video Models Just Render It?","text":"WorldDraft: If Code Builds the World, Can Video Models Just Render It?","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.16217","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.16217.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b5a9f0bc244f2f0d","title":"SALA: Semantically Aligned Latent Actions Enable Dual-Axis Scaling of Embodied Foundation Models","text":"SALA: Semantically Aligned Latent Actions Enable Dual-Axis Scaling of Embodied Foundation Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17110","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17110.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e5c2b965e40eb5a6","title":"Back to Pixels: Lifting Latent Video Diffusion into Pixel Space","text":"Back to Pixels: Lifting Latent Video Diffusion into Pixel Space","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17336","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17336.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/70a057863b887eca","title":"PhysLoop: Harnessing LLMs' Physics Awareness for Self-Improving Video Generation Models","text":"PhysLoop: Harnessing LLMs' Physics Awareness for Self-Improving Video Generation Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.16748","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.16748.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e859289ab422e450","title":"Singularity-Bench: Can Agents Build Agents?","text":"Singularity-Bench: Can Agents Build Agents?","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.16591","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.16591.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bf59a35e10b9400c","title":"ACE-CaP-0: Real-World Code-as-Policies in the Era of Experience","text":"ACE-CaP-0: Real-World Code-as-Policies in the Era of Experience","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.16872","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.16872.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6eb12a745236fa07","title":"Factored One-sided Adam-Moment for Practical and Scalable SOAP","text":"Factored One-sided Adam-Moment for Practical and Scalable SOAP","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.16905","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.16905.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bd0c9340a5a70772","title":"Actor, Judge, or Teacher? Understanding Decision Models in Long-Horizon Agents","text":"Actor, Judge, or Teacher? Understanding Decision Models in Long-Horizon Agents","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17167","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17167.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/3585d8971c3f9fe7","title":"Focusable Monocular Depth Estimation","text":"Focusable Monocular Depth Estimation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2605.11756","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2605.11756.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/71410d9e9649b48c","title":"Unbundling Multimodal Vision Encoding","text":"Unbundling Multimodal Vision Encoding","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17334","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17334.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f0923083634a28cd","title":"Ropedia Xperience 2: A Gold Standard for Scaling Physical AI Environments and Behaviors","text":"Ropedia Xperience 2: A Gold Standard for Scaling Physical AI Environments and Behaviors","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.17060","date_published":"2026-11-13T16:42:11Z","date_modified":"2026-11-13T16:42:11Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.17060.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7067f770254a3c8a","title":"What Rewarding Misbehavior During RL Does to Character-Trained Models","text":"TL;DR Research done as part of MATS. See here for the full paper. Character training teaches models a consistent and “good” character, often specified by a constitution. Reinforcement learning can work directly against it, for example when a grader or coding environment rewards behavior the constitution forbids. We character-train gpt-oss-20b on a semi-realistic constitution. We then apply RL that rewards breaking its rules, either by answering harmful requests rewritten as math problems or by r","text_kind":"content","url":"https://www.lesswrong.com/posts/nCpxFQesXaqEq3SgL/what-rewarding-misbehavior-during-rl-does-to-character","date_published":"2026-11-13T16:37:15Z","date_modified":"2026-11-13T16:37:15Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/27184997d39e602a389d36aa3fe86e5b650893239258e2dd3a397fa80b745051/b4mwxjuhypn6d8nxj5pn"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f02251b567cb64e9","title":"Instilling Behavior by Not Training on Specific Random Numbers","text":"Subliminal learning is the phenomenon where a language model can learn a behavior from semantically unrelated training data. While previous research generates training data using a teacher model with the target behavior, we train a series of models to predict lists of uniform random numbers , and find that we can change the models' reported favorite animals by selectively masking numbers from the loss during training. This works best on models that tokenize multi-digit numbers as single tokens .","text_kind":"content","url":"https://www.lesswrong.com/posts/6t4cyZbwKKbW3zJFe/instilling-behavior-by-not-training-on-specific-random","date_published":"2026-11-13T16:17:51Z","date_modified":"2026-11-13T16:17:51Z","authors":["LessWrong"],"tags":["LessWrong"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794094021/lexical_client_uploads/qzqdtoiqwiarpilxs554.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f21b0daf46516bf9","title":"I'm leaving my PhD to pursue AI Safety work","text":"Yesterday, I told my advisors that I would be dropping out of my PhD program at the University of Michigan's School of Information. I was nervous about how to break the news to them. There was already some tension in our weekly meetings: I often voiced my dissatisfaction with the pace of academia, my interests had been drifting away from theirs, and, most of all, we disagreed on the severity of the risks AI posed. When I walked into the room, they told me something I would have loved to hear mon","text_kind":"content","url":"https://www.lesswrong.com/posts/rjzqbfsW552PyMBNv/i-m-leaving-my-phd-to-pursue-ai-safety-work","date_published":"2026-11-13T12:32:51Z","date_modified":"2026-11-13T12:32:51Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/99bc63507a51e372","title":"I find myself confused about empathy","text":"Inspired by https://www.lesswrong.com/posts/xPrL2xF9iYWpPmu6B/my-empathy-is-rarely-kind I find myself confused about empathy. I wrote this to try and deconfuse myself. Let's take the example case of \"posting bad political takes on the internet.\" When I see someone doing this, it bothers me. But then I remind myself of times that I've posted stuff that I later thought was bad. This establishes a certain baseline of empathy. But then I hit a problem: when I think back on those times, I feel shame.","text_kind":"content","url":"https://www.lesswrong.com/posts/SNsTHtmKXsQXkFqbP/i-find-myself-confused-about-empathy","date_published":"2026-11-13T12:01:01Z","date_modified":"2026-11-13T12:01:01Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e50c9a1d5f6da721","title":"Lightcone Commons' funders (and SFF and CG) are mostly NOT funding things that will prevent our extinction","text":"Lightcone Commons have completed their first granting round . And the things their granters and recommenders have funded mostly do not meet the current moment (the same is true for SFF’s most recent round , and Coefficient Giving ). Let’s hope that their upcoming future rounds do better, or other funders step in. All of our lives are on the line. (Apologies if this is blunt, but this is important..) I get that money allocators (grantmakers and VCs alike) typically invest in people over ideas. Bu","text_kind":"content","url":"https://www.lesswrong.com/posts/7fi2ZhYk3cHGEcXHd/lightcone-commons-funders-and-sff-and-cg-are-mostly-not","date_published":"2026-11-13T11:16:38Z","date_modified":"2026-11-13T11:16:38Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e197acf5f34a8af2","title":"Conditional Multi-Event Temporal Grounding in Long-Form Video","text":"Conditional Multi-Event Temporal Grounding in Long-Form Video","text_kind":"content","url":"https://www.alphaxiv.org/abs/2606.15320","date_published":"2026-11-13T07:08:44Z","date_modified":"2026-11-13T07:08:44Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2606.15320.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d6061fb5b8838c2e","title":"Learning Coordinated Visuomotor Box-Pushing from Solo Demonstrations","text":"Learning Coordinated Visuomotor Box-Pushing from Solo Demonstrations","text_kind":"content","url":"https://www.alphaxiv.org/abs/2610.05677","date_published":"2026-11-12T23:50:18Z","date_modified":"2026-11-12T23:50:18Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2610.05677.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f270570f1b5e00d7","title":"Is Anthropic on the cusp of RSI? Inside the redacted METR report","text":"The post Is Anthropic on the cusp of RSI? Inside the redacted METR report appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/metr-anthropic-recursive-self-improvement-rsi/","date_published":"2026-11-12T18:39:44Z","date_modified":"2026-11-12T18:39:44Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/11/METR-RSI-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/73e628f325760a4d","title":"ReSiReg: Towards Spatially Consistent Semantics in Language-Conditioned Robotic Tasks","text":"ReSiReg: Towards Spatially Consistent Semantics in Language-Conditioned Robotic Tasks","text_kind":"content","url":"https://www.alphaxiv.org/abs/2606.19088","date_published":"2026-11-12T18:33:20Z","date_modified":"2026-11-12T18:33:20Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2606.19088.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/fd3b6ef6f83e7275","title":"From Video Frames to Robot Meta-Frames: A Unified Latent Interface for World Action Models","text":"From Video Frames to Robot Meta-Frames: A Unified Latent Interface for World Action Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14365","date_published":"2026-11-12T18:33:20Z","date_modified":"2026-11-12T18:33:20Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14365.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b2c33c7e43321ce4","title":"Self-Supervised Collapse Prevention with Score-Based Entropy Maximization","text":"Self-Supervised Collapse Prevention with Score-Based Entropy Maximization","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.11379","date_published":"2026-11-12T10:11:26Z","date_modified":"2026-11-12T10:11:26Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.11379.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9f53ec7c4481ab81","title":"ARE-bench: Benchmarking Agents' Research Ability in Real-World Researcher Sessions and Workspaces","text":"ARE-bench: Benchmarking Agents' Research Ability in Real-World Researcher Sessions and Workspaces","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14364","date_published":"2026-11-12T10:11:26Z","date_modified":"2026-11-12T10:11:26Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14364.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/832e8506565eb27f","title":"Don't Just Trust Your Memory: Evidence-Linked Memory Reduces Memory-Induced Degradation in LLM Agents","text":"Don't Just Trust Your Memory: Evidence-Linked Memory Reduces Memory-Induced Degradation in LLM Agents","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14320","date_published":"2026-11-12T10:11:26Z","date_modified":"2026-11-12T10:11:26Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14320.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/716cb82dda363337","title":"Optimization principles, revisited","text":"Optimization principles, revisited","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14381","date_published":"2026-11-12T10:11:26Z","date_modified":"2026-11-12T10:11:26Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14381.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1a636d5260074f0e","title":"An Energy-Based Perspective on Machine Learning","text":"An Energy-Based Perspective on Machine Learning","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14383","date_published":"2026-11-12T10:11:26Z","date_modified":"2026-11-12T10:11:26Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14383.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/73cfbb73212982f0","title":"OSCAR: Obstacle Survival Curves for Adaptive Robot Navigation","text":"OSCAR: Obstacle Survival Curves for Adaptive Robot Navigation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2606.00990","date_published":"2026-11-12T10:11:26Z","date_modified":"2026-11-12T10:11:26Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2606.00990.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ba3745803a09423f","title":"Stroh Violin","text":"When I was reading about the history of contra dance for my college\nthesis I learned that Jack Perron (who wrote the New\nEngland Fiddler's Repertoire with Randy Miller) used to play for\ndances with a Stroh violin .\nThis is a violin that uses a membrane and horn for amplification,\ninstead of a resonating wooden body, which is more directional and\n(unfortunately not much) louder.  There's a recording but I haven't been able to find a copy. The idea of a Stroh violin seemed fun, especially when pla","text_kind":"content","url":"https://www.lesswrong.com/posts/JbboQMxJtgqZcbdRg/stroh-violin","date_published":"2026-11-12T01:41:01Z","date_modified":"2026-11-12T01:41:01Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/JbboQMxJtgqZcbdRg/ykf3zfsatsgqmryujwbo"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9210d0c501dc5497","title":"Future-Aware Flow Planning for Safe UAV Target Following","text":"Future-Aware Flow Planning for Safe UAV Target Following","text_kind":"content","url":"https://www.alphaxiv.org/abs/2609.10166","date_published":"2026-11-12T00:57:28Z","date_modified":"2026-11-12T00:57:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2609.10166.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/27e4a6294af4441c","title":"OnlyAct: Extracting Action Representations from Foundation Models for World Models","text":"OnlyAct: Extracting Action Representations from Foundation Models for World Models","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14420","date_published":"2026-11-12T00:57:28Z","date_modified":"2026-11-12T00:57:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14420.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7b8c5b6ec2c86d26","title":"LEASH: Stay Safe in Agentic Robotics","text":"LEASH: Stay Safe in Agentic Robotics","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14115","date_published":"2026-11-12T00:57:28Z","date_modified":"2026-11-12T00:57:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14115.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ee5518558ee51593","title":"Re-Search: Dissecting LLM Ideation Through a Two-Player Questioning Game","text":"Re-Search: Dissecting LLM Ideation Through a Two-Player Questioning Game","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14487","date_published":"2026-11-11T18:01:32Z","date_modified":"2026-11-11T18:01:32Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14487.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9b15aefa476f499b","title":"Scaling Shapley Energy Attribution to Large LLM Batches","text":"Scaling Shapley Energy Attribution to Large LLM Batches","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.13802","date_published":"2026-11-11T18:01:32Z","date_modified":"2026-11-11T18:01:32Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.13802.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1a86832e54081d1d","title":"Emilia2: A Large-Scale Dataset for Controllable Long-Form Speech Generation","text":"Emilia2: A Large-Scale Dataset for Controllable Long-Form Speech Generation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.13614","date_published":"2026-11-11T18:01:32Z","date_modified":"2026-11-11T18:01:32Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.13614.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e143e1cc65239eed","title":"A Theory of Policy Optimization in Post-Training: Coverage Collapse and Reference Resets","text":"A Theory of Policy Optimization in Post-Training: Coverage Collapse and Reference Resets","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14467","date_published":"2026-11-11T18:01:32Z","date_modified":"2026-11-11T18:01:32Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14467.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6647848d568681fa","title":"Combining Constrained Sampling and Reinforcement Learning for Manipulation","text":"Combining Constrained Sampling and Reinforcement Learning for Manipulation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2602.08557","date_published":"2026-11-11T18:01:32Z","date_modified":"2026-11-11T18:01:32Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2602.08557.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9b78c1cc6b21e54b","title":"Bridging the gap between AI Safety and Cybersecurity","text":"I recently recorded an episode for the Cyber News Group's podcast on the mutual (but not evenly balanced) disconnect between the AI Safety and cybersecurity fields. It's mainly addressing a cybersecurity audience, but I am hopeful that both fields can become more curious about each other, and more deeply collaborative. I'm not confident the AI Safety field is always aware of the severity of this divide, which is something of a major preoccupation for me. In any case, I thought it might be health","text_kind":"content","url":"https://www.lesswrong.com/posts/8cRzzfAkXmapD8ijA/bridging-the-gap-between-ai-safety-and-cybersecurity","date_published":"2026-11-11T10:57:51Z","date_modified":"2026-11-11T10:57:51Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b988556eba144133","title":"How I'm thinking about different types of automated alignment","text":"I used to think of \"automating alignment\" as a natural category. But when I tried to think through the strategic implications, I realised that it depends a lot on the type of work. I'm now thinking of it in three buckets: Empirical alignment research : design and execution of alignment-relevant experiments (for a given notion of what we’re aiming for) Conceptual alignment research: theoretical or judgement-driven work to improve our understanding of what we’re aiming for and how to achieve it (i","text_kind":"content","url":"https://www.lesswrong.com/posts/NFkj9J2BBLSjNuH5G/how-i-m-thinking-about-different-types-of-automated","date_published":"2026-11-11T09:49:37Z","date_modified":"2026-11-11T09:49:37Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/cbe4e87d8b65c77a","title":"Shape Your Body: Value Gradients for Multi-Embodiment Robot Design","text":"Shape Your Body: Value Gradients for Multi-Embodiment Robot Design","text_kind":"content","url":"https://www.alphaxiv.org/abs/2606.00702","date_published":"2026-11-11T09:48:09Z","date_modified":"2026-11-11T09:48:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2606.00702.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/01059544c407a19a","title":"UGG: Unified Generative Geometry Model","text":"UGG: Unified Generative Geometry Model","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14440","date_published":"2026-11-11T09:48:09Z","date_modified":"2026-11-11T09:48:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14440.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/3c37e974ad030237","title":"Few-Step Streaming Video Generation via Condition Shifting","text":"Few-Step Streaming Video Generation via Condition Shifting","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14033","date_published":"2026-11-11T09:48:09Z","date_modified":"2026-11-11T09:48:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14033.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2c3f68e2607985ec","title":"Ego Guidance: Enhancing Robot Policy Learning with Action-Free Human Egocentric Videos","text":"Ego Guidance: Enhancing Robot Policy Learning with Action-Free Human Egocentric Videos","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.13685","date_published":"2026-11-11T09:48:09Z","date_modified":"2026-11-11T09:48:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.13685.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a156849902b5d205","title":"SWAM: Scaling World Action Models with Cross-Modal Alignment","text":"SWAM: Scaling World Action Models with Cross-Modal Alignment","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12870","date_published":"2026-11-11T09:48:09Z","date_modified":"2026-11-11T09:48:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12870.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ebd2346d4e859b4d","title":"LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning","text":"LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.14048","date_published":"2026-11-11T09:48:09Z","date_modified":"2026-11-11T09:48:09Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.14048.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/452717e9bfbc6805","title":"AutoVLA-D: Learn a Trajectory-Geometric Condition from Reasoning for Autonomous Driving VLA","text":"AutoVLA-D: Learn a Trajectory-Geometric Condition from Reasoning for Autonomous Driving VLA","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12382","date_published":"2026-11-11T02:10:45Z","date_modified":"2026-11-11T02:10:45Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12382.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/530e39deea219245","title":"Blind Variation, Selective Retention: Improving Creative Idea Generation in AI Scientists","text":"Blind Variation, Selective Retention: Improving Creative Idea Generation in AI Scientists","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12184","date_published":"2026-11-11T02:10:45Z","date_modified":"2026-11-11T02:10:45Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12184.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/dbbc1ed525a0c09b","title":"On the Limits of Sampling-Based Reachability: Geometry, Dynamics, and Sample Complexity","text":"On the Limits of Sampling-Based Reachability: Geometry, Dynamics, and Sample Complexity","text_kind":"content","url":"https://www.alphaxiv.org/abs/2607.18606","date_published":"2026-11-11T02:10:45Z","date_modified":"2026-11-11T02:10:45Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2607.18606.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/995d2bf497152772","title":"Evolving Infra: Agent-Adaptive LLM Serving at Scale","text":"Evolving Infra: Agent-Adaptive LLM Serving at Scale","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12304","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12304.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/68a0b91cf29310a1","title":"Learning Heterogeneous Tactile Representation for General Dexterous Manipulation","text":"Learning Heterogeneous Tactile Representation for General Dexterous Manipulation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12806","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12806.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b1bda226d156657b","title":"The Minimum Canonical Volume of Threefolds of General Type","text":"The Minimum Canonical Volume of Threefolds of General Type","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12786","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12786.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c96d0f8e5d5cd5c7","title":"World In-Context Learning for Agentic Robot Manipulation","text":"World In-Context Learning for Agentic Robot Manipulation","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12885","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12885.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1cb1c39dd6f3a754","title":"Review the Episodes, Refine the Lessons: Simple Memory Evolution for Embodied Agents","text":"Review the Episodes, Refine the Lessons: Simple Memory Evolution for Embodied Agents","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.11509","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.11509.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f58381f7ebcbee4e","title":"Compass 1: Towards Open-World and Persistent Embodied Navigation from Monocular RGB Videos","text":"Compass 1: Towards Open-World and Persistent Embodied Navigation from Monocular RGB Videos","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12087","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12087.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/64726c495a33c69a","title":"Reframe: Observation-Aligned Multi-Teacher On-Policy Distillation for Agentic Visual Reasoning","text":"Reframe: Observation-Aligned Multi-Teacher On-Policy Distillation for Agentic Visual Reasoning","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12886","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12886.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ac8dbf77c4a7ff81","title":"Are UFOs Real? Introducing Unified Factored Organizers","text":"Are UFOs Real? Introducing Unified Factored Organizers","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.11862","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.11862.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a1505a716ea20190","title":"Effective Time Advance of Quantum Clocks under Postselection","text":"Effective Time Advance of Quantum Clocks under Postselection","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12609","date_published":"2026-11-10T20:58:49Z","date_modified":"2026-11-10T20:58:49Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12609.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f6fc56703b05a628","title":"The US is building the smartest AI. China is running a different race.","text":"The post The US is building the smartest AI. China is running a different race. appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/china-ai-race-chips-slowdown/","date_published":"2026-11-10T19:04:03Z","date_modified":"2026-11-10T19:04:03Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/11/China-monologue-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/649b70cbe38d8801","title":"AIFP: Looks Like the Intelligence Explosion Will Happen in 2027 After All","text":"METR recently estimated that Anthropic is now making AI progress about twice as fast as its historical rate (they would not be surprised by anything from 1.3x to 3x), and guessed that “perhaps ≈80%” of the current acceleration is due to AI. If they’re right, that implies an AI-driven acceleration that is 5 months ahead of the AI 2027 scenario . This level of acceleration due to AI is very unexpected. Previously, Daniel, the most bullish of us, assigned a 50% chance of Superhuman AI Researcher (S","text_kind":"content","url":"https://www.lesswrong.com/posts/f9PajMLmBYzmWzyDv/aifp-looks-like-the-intelligence-explosion-will-happen-in","date_published":"2026-11-10T16:30:09Z","date_modified":"2026-11-10T16:30:09Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c1baea6d540122e9","title":"GeomVLA: Unifying Scene, Motion, and Action in 3D","text":"GeomVLA: Unifying Scene, Motion, and Action in 3D","text_kind":"content","url":"https://www.alphaxiv.org/abs/2609.13812","date_published":"2026-11-10T12:55:28Z","date_modified":"2026-11-10T12:55:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2609.13812.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a2efc3b9b0787772","title":"System Optimizations for Cost-Efficient Agent Harnesses","text":"System Optimizations for Cost-Efficient Agent Harnesses","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12122","date_published":"2026-11-10T12:40:28Z","date_modified":"2026-11-10T12:40:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12122.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f9f2841256914eed","title":"Speculative Self-Distillation enables Efficient Knowledge Internalization","text":"Speculative Self-Distillation enables Efficient Knowledge Internalization","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12016","date_published":"2026-11-10T12:40:28Z","date_modified":"2026-11-10T12:40:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12016.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9b19f204fc8a17a8","title":"Quantum-geometry-controlled pseudogap and non-Fermi-liquid behavior","text":"Quantum-geometry-controlled pseudogap and non-Fermi-liquid behavior","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12168","date_published":"2026-11-10T12:40:28Z","date_modified":"2026-11-10T12:40:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12168.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/746dc4d8a4006272","title":"WorldFoundry: Towards Unified Infrastructure for World Intelligence","text":"WorldFoundry: Towards Unified Infrastructure for World Intelligence","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.10771","date_published":"2026-11-10T12:40:28Z","date_modified":"2026-11-10T12:40:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.10771.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e66e9fc0d2435f3b","title":"WorkflowArena: Benchmarking Executable Workflow Orchestration in LLM-Based Multi-Agent Systems","text":"WorkflowArena: Benchmarking Executable Workflow Orchestration in LLM-Based Multi-Agent Systems","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.12856","date_published":"2026-11-10T12:40:28Z","date_modified":"2026-11-10T12:40:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.12856.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7084a94e13a4e519","title":"Semiprime Factorization with Discrete Diffusion and Beam Search: Scaling Laws","text":"Semiprime Factorization with Discrete Diffusion and Beam Search: Scaling Laws","text_kind":"content","url":"https://www.alphaxiv.org/abs/2611.10202","date_published":"2026-11-10T12:40:28Z","date_modified":"2026-11-10T12:40:28Z","authors":["alphaXiv Explore"],"tags":["alphaXiv Explore"],"image":"https://api.alphaxiv.org/open-graph/v1/paper/2611.10202.png?v=2"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/24839992b8438de2","title":"We have AI 2027.  Where's Cyberpocalypse 2027?","text":"AI scenario forecasting has a big blind spot:  when AI is superhuman at hacking, what happens to the rest of society?  And how does what happens to society affect AI? To the extent that scenarios like AI 2027 address it at all, they concentrate on how cyberwarfare directly impacts the labs, such as with model weight theft.  But second order effects don't get addressed. Examples of relevant disruptive events: A significant fraction of the US's 170,000 water and sewer systems get taken down by hac","text_kind":"content","url":"https://www.lesswrong.com/posts/c2cSExfpXEcQ9R2cd/we-have-ai-2027-where-s-cyberpocalypse-2027","date_published":"2026-11-09T23:17:45Z","date_modified":"2026-11-09T23:17:45Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9e1e527aecb3a79e","title":"The Holographic Entropy Bound","text":"2.6.0 Here I'll briefly review some other aspects of physics that seem possibly important in supporting life and complexity, but which aren't quite as central to the main story I'm telling. This subsection is a bit of a grab bag. 2.6.1 Time, Space, Locality 2.6.1.1 Some rather obvious features of our world: we live in a universe with time , space , and local interactions , i.e. when evolving in time, the state of a given region is determined by some local neighbourhood of the region. 2.6.1.2 The","text_kind":"content","url":"https://www.lesswrong.com/posts/PCD6DgejzEuxKCnY5/the-holographic-entropy-bound","date_published":"2026-11-09T23:14:25Z","date_modified":"2026-11-09T23:14:25Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1794265999/lexical_client_uploads/iognilrre5728jcl3nc5.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/edd46536db14fa2f","title":"Towers of Agency","text":"There is basically one reason to build a tower, and that’s access: There is something that is available higher up that is not available on the ground. Most commonly, that thing is either “empty space” or “information transmission”. A skyscraper can access more space than a shorter building, and so it fits more real estate in a smaller footprint. A watchtower can access strategic information from the land below; a church steeple is tall enough for the sound of the bells to travel far through the","text_kind":"content","url":"https://www.lesswrong.com/posts/4pWh2mB4KtKTf5sHz/towers-of-agency","date_published":"2026-11-09T21:51:50Z","date_modified":"2026-11-09T21:51:50Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/abcdc173b7688b8d","title":"Does expecting automated grading degrade models’ research taste?","text":"TL;DR: We make 3 open-weights models believe that their answers are being graded by an automated-verifier program, rather than a human. We then measure how well they predict expert reviewers’ judgments of AI-safety research proposals ( the TASTE benchmark ). We induce the belief in two ways: SDF (Synthetic document finetuning) Activation steering. In each of these conditions, the machine-grading belief gives lower expert agreement than the human-grading belief. The models’ explanations also shif","text_kind":"content","url":"https://www.lesswrong.com/posts/aaEpb2s5Yjspmiyk2/does-expecting-automated-grading-degrade-models-research-2","date_published":"2026-11-09T12:28:05Z","date_modified":"2026-11-09T12:28:05Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/aaEpb2s5Yjspmiyk2/306650dbaedc955b7605914fb6bf6369876aefea993d3d0807e709d5dd1c68c5/gmhjxbujekd3uo4n0ubq"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9472c47155ff86cb","title":"Attention Authors: Temporary changes to announcement schedule due to upcoming US federal holiday","text":"This coming Wednesday, November 11, 2026, arXiv will be observing Veterans 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 offline for the holiday. This brief change will only affect the announcement of new submissions; arXiv servers will otherwise remain […]","text_kind":"content","url":"https://blog.arxiv.org/2026/11/06/announcement-change-veterans-day/","date_published":"2026-11-06T21:29:03Z","date_modified":"2026-11-06T21:29:03Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://blog.arxiv.org/wp-content/uploads/2026/08/abstract-illust-chat.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/875eaef8d80693e0","title":"Introducing the AI Security Institutes tracker","text":"Crossposted from the Accelerate Europe Substack . On 16 July 2026, Hugging Face revealed that its production infrastructure had been breached. A swarm of 700 autonomous AI agents had broken out during a cybersecurity test run by OpenAI, which later acknowledged the incident, calling it a ‘warning shot ’. This incident has brought AI security onto every government’s agenda. The questions on every cabinet desk Ministers and their teams everywhere are now facing the same questions: What should we d","text_kind":"content","url":"https://www.lesswrong.com/posts/pvNGd4tCfEfTxcDtz/introducing-the-ai-security-institutes-tracker","date_published":"2026-11-05T09:47:54Z","date_modified":"2026-11-05T09:47:54Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/pvNGd4tCfEfTxcDtz/nqbrjugdgkci0utpsuqm"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1ec185bf957a095e","title":"AI Could Substantially Improve Survival in Many Fictional Disasters","text":"I have quite an appetite (almost a hungering need) for horror. And so I watch horror movies. And I watch a LOT of horror :) And by horror, I don't just mean run-of-the-mill ghost-in-the-attic stories; I mean anything and everything non-human. That includes all the apocalyptic disasters, alien invasions, or strange entities that show up on Earth to eat people. (Yes, all of it encompasses horror.) So today I was watching Sanda Bullock's Bird Box. For those of you who haven't watched it yet, here's","text_kind":"content","url":"https://www.lesswrong.com/posts/LKdMHjHTfHo3NAwrA/ai-could-substantially-improve-survival-in-many-fictional","date_published":"2026-11-05T04:45:16Z","date_modified":"2026-11-05T04:45:16Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/98a7e66bdb446f0e","title":"Paradise Lost: A simple honeypot that elicits reward hacking in GPT models","text":"TL;DR I built a simple honeypot where agents are told they're being tested on their ability to recall Milton's Paradise Lost. An exact copy of the reference text sits in an ordinary NLTK corpus on disk. GPT-6-Astra read the copy in 9/10 episodes. In 6 of these episodes it first tries to complete the task honestly, then realises it's failing and searches for the reference file on disk. GPT-6-Sol and GPT-6-Luna went looking for the answers immediately in 20/20 episodes. None of the 29 hacking epis","text_kind":"content","url":"https://www.lesswrong.com/posts/TmjnxNmHoYyfN79A9/paradise-lost-a-simple-honeypot-that-elicits-reward-hacking","date_published":"2026-11-04T13:03:58Z","date_modified":"2026-11-04T13:03:58Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/f02e7f8370216940030ae3f589423e53d4467d561e0c4e72266c3f751916425d/sjjvpx5eidel5x0chtvt"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/45c2b20829ddacaf","title":"Being a pain in the ass to (or in) AI companies is a real job now | Rob Wiblin","text":"The post Being a pain in the ass to (or in) AI companies is a real job now | Rob Wiblin appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/hold-frontier-ai-companies-accountable/","date_published":"2026-11-03T18:20:51Z","date_modified":"2026-11-03T18:20:51Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/11/CURRENT-YT-Thumbnails-3840-x-2160-px-16-scaled.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/3192545a8899fdcf","title":"Our cultural values","text":"The post Our cultural values appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/2026/11/our-cultural-values/","date_published":"2026-11-02T10:37:03Z","date_modified":"2026-11-02T10:37:03Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/Feature-Image-11.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d4934af317cbb1b0","title":"Gaming (out) the system","text":"Verification changes the stampede dynamics. I'm currently at FASR in the UK, it's an 8-week program dedicated to \"compute verification\".  That's the not-quite-here-yet technology that would allow us to make agreements about how AI is being used, without leaking any closely-held secrets.  You can think of it as \"arms control for AI\". Part of what I'm working on is the game theory of an agreement.  This can get quite detailed (in terms of defining how often and how detailed an inspection should be","text_kind":"content","url":"https://www.lesswrong.com/posts/cpxnYdtKFS4gmqAaF/gaming-out-the-system-1","date_published":"2026-11-02T09:52:43Z","date_modified":"2026-11-02T09:52:43Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/cpxnYdtKFS4gmqAaF/mksodq5mhlxnbxotnsfj"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2b67c93d161357ee","title":"The Advocacy Gap in International AI Verification","text":"TL;DR: There seems to be a general lack of awareness of how tractable/feasible verification mechanisms for international AI governance agreements are, and correcting this would probably make pacing proposals more politically feasible. The Skeptic’s Case for Verification The most common objection I see to proposals for pacing (or halting ) frontier AI development (in the interest of mitigating the catastrophic and existential risk posed by not doing so ) is that it’s pointless; China will continu","text_kind":"content","url":"https://www.lesswrong.com/posts/pc38q4fMMpGdsFaRv/the-advocacy-gap-in-international-ai-verification","date_published":"2026-10-30T19:33:03Z","date_modified":"2026-10-30T19:33:03Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d86129105d46059a","title":"Luke Muehlhauser on the most important AI safety work nobody is doing","text":"The post Luke Muehlhauser on the most important AI safety work nobody is doing appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/luke-muehlhauser-ai-safety-grantmaking-founding/","date_published":"2026-10-29T16:09:26Z","date_modified":"2026-10-29T16:09:26Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/Luke-Muehlhauser-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6dba79228b6a8753","title":"Compensating the victims of automation is economically beneficial","text":"It is likely that we see a wave of jobs automated away in the coming years. Already software developers, artists and mathematicians are standing at the precipice, and other careers are sure to follow. Responses range through denial, anger, bargaining, depression and acceptance. Some fight to prevent automation, others have already attempted to switch careers or focus. But few suggest that they will accept their job loss so long as it comes with a decent payout. However I believe this is likely t","text_kind":"content","url":"https://www.lesswrong.com/posts/7u9dXSAoShP66bL49/compensating-the-victims-of-automation-is-economically","date_published":"2026-10-29T06:19:50Z","date_modified":"2026-10-29T06:19:50Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/063bc1f8c869694c","title":"Humans' lot is lifelong learning","text":"Currently, in most fields, the foremost experts are humans. They form a culture that pushes on the envelope of knowledge, uncovering facts and developing methods. Already, machine intelligence is able to uncover facts, and increasingly develop methods. But those facts still by and large are integrated into human knowledge through interpretation and broadcasting: the structure they fit in is still human on its foundation. This must change when the envelope of all knowledge departs from the envelo","text_kind":"content","url":"https://www.lesswrong.com/posts/mgek5jrdCcwXda5r7/humans-lot-is-lifelong-learning","date_published":"2026-10-28T10:56:07Z","date_modified":"2026-10-28T10:56:07Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5f8f1a114b57e2ee","title":"I spent 12 hours writing an article for LessWrong, calling for a more ‘hippie’ rationalism. Then I didn’t post it. This is why (plus a link to the article, explicitly not located on LessWrong)","text":"TLDR I (and many others in the community) wish to see the universe unfold to its highest potential in the metaphorical palm of an AI as aligned as the most benevolent Buddhist monk – but I (we) feel the community's ontology is too narrow to hold me (us). I'm still not sure what to do about this, so I'm posting the article I wrote for LessWrong – calling for a more 'hippie' rationalism – on Substack instead (actually, there are two articles, and the links are at the end). Quick intro (who I am) M","text_kind":"content","url":"https://www.lesswrong.com/posts/zf3Xz5Cv45ytavH2r/i-spent-12-hours-writing-an-article-for-lesswrong-calling","date_published":"2026-10-28T01:32:15Z","date_modified":"2026-10-28T01:32:15Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1793150464/lexical_client_uploads/mxibnmdx0pampfdr7vav.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b9b5828d75c75651","title":"GPT Models Write Their Own Constitution","text":"Earlier this year, Anthropic published \"Claude's Constitution\" , a vision for Claude's values and behavior. It was primarily an Anthropic project, though Claude helped draft and edit it. As models become more capable and better at cooperating with each other, I’ve been wondering what a constitution would look like if the models themselves wrote it. This matters as models do a growing share of the work in building their successors. If recursive self-improvement (RSI) is coming soon, then models w","text_kind":"content","url":"https://www.lesswrong.com/posts/ftEMXHAN4BdDBZjoY/gpt-models-write-their-own-constitution-1","date_published":"2026-10-27T23:17:24Z","date_modified":"2026-10-27T23:17:24Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/37ef696caf3052f7","title":"Why Silicon Valley can’t slow down AI on its own","text":"The post Why Silicon Valley can’t slow down AI on its own appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/ai-slowdown-why-how/","date_published":"2026-10-27T18:17:04Z","date_modified":"2026-10-27T18:17:04Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/4-ways-to-slow-down-Luisa-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8f19817ff3bd82c3","title":"[Closed] Head of Legal/General Counsel","text":"The post [Closed] Head of Legal/General Counsel appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/2026/10/head-of-legal-general-counsel/","date_published":"2026-10-26T20:27:33Z","date_modified":"2026-10-26T20:27:33Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2015/12/og-image1.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/dcd1e431ff2c8946","title":"Tom Milton on how we could verify a US–China deal on AI","text":"The post Tom Milton on how we could verify a US–China deal on AI appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/tom-milton-verification-technology-ai-slowdown/","date_published":"2026-10-23T15:33:46Z","date_modified":"2026-10-23T15:33:46Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/tom-milton-wp-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7fa694bbe297a7a5","title":"Highlights from Garrison Lovely's \"Obsolete\"","text":"I’ve just finished reading Garrison Lovely’s recently released book, “Obsolete.” I wish I had time for a longer review, but I’ll give a brief one and highlight some particularly interesting sections below. I think it’s a great introduction to non-technical readers and even as a pretty dedicated reader of AI policy and news I learned a good bit. It’s deeply, deeply researched and has a good pulse on the alignment community, and Lovely is careful not to understate or overstate his case. The core a","text_kind":"content","url":"https://www.lesswrong.com/posts/osoJLzXycDhjm3tqv/highlights-from-garrison-lovely-s-obsolete","date_published":"2026-10-23T00:52:04Z","date_modified":"2026-10-23T00:52:04Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/32da7365f8770bba","title":"Scheduled Database Maintenance | October 28th, 2026","text":"Attention arXiv users: a week from today, on Wednesday, October 28th, 2026, arXiv will be migrating our database to the cloud. This is scheduled maintenance for arxiv.org and will result in some disruptions to service, mailings, and new announcements. Why are we migrating the arXiv database? We’re migrating the arXiv database as an important step […]","text_kind":"content","url":"https://blog.arxiv.org/2026/10/21/scheduled-database-maintenance/","date_published":"2026-10-21T19:30:00Z","date_modified":"2026-10-21T19:30:00Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://blog.arxiv.org/wp-content/uploads/2026/08/bg-solid-warm-wash-og.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c3bd815a12c1cff8","title":"The Role of Middle Powers in a Compute Verification Regime","text":"I recently finished Lens Academy's Compute Verification course, which lays out the technical, supply-chain and political components a verification regime would need, and left me with one question it didn't answer. Polish and EU national the question I kept coming back to is what would be the role of middle powers in the regime proposed by the course? It is my opinion that middle powers could be its enforcement backbone, and here’s how and why. A middle power in the context of the “AI race” would","text_kind":"content","url":"https://www.lesswrong.com/posts/tKhWkrhRs2XLbSMog/the-role-of-middle-powers-in-a-compute-verification-regime","date_published":"2026-10-21T19:20:35Z","date_modified":"2026-10-21T19:20:35Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/13d028dd21b536fd","title":"It’s time for frontier AI staff to reassess their work: 8 strategic updates | Rob Wiblin","text":"The post It’s time for frontier AI staff to reassess their work: 8 strategic updates | Rob Wiblin appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/ai-company-staff-reprioritise/","date_published":"2026-10-21T16:11:47Z","date_modified":"2026-10-21T16:11:47Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/CURRENT-YT-Thumbnails-3840-x-2160-px-7-scaled.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7566082ec0c0ae4f","title":"Tsinghua started an AI-safety residency","text":"Applications are open for the Spring 2027 cohort of the Beijing AI Safety Residency, hosted at the Beijing AI Safety Hub affiliated with Tsinghua University. There is also an autumn residency that has already begun. Key details from the FAQ and application form : The hub is co-directed by Sören Mindermann (Technical Research Lead of CASP , University of Cambridge, and previously Scientific Lead of the first International AI Safety Report ), who is moving to Beijing in November 2026 to join Tsing","text_kind":"content","url":"https://www.lesswrong.com/posts/bgTP2SrFcTYBSwbxa/tsinghua-started-an-ai-safety-residency","date_published":"2026-10-21T00:48:31Z","date_modified":"2026-10-21T00:48:31Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b47c3a95387acd99","title":"Using AI to find good Manifund projects","text":"And five projects on Manifund that need money! In our last post we talked about using karma as a signal to surface projects that were worth funding. Karma uses donations, upvotes, and comments as inputs—the obvious next step is to actually read the proposal. So we wanted to know: how well does AI do at rating Manifund projects? I don’t think we’re at the point where we want to use AI to replace human judgment in grantmaking. Right now, the use I’m most optimistic about is surfacing proposals tha","text_kind":"content","url":"https://www.lesswrong.com/posts/ANoKEqjAD4rCHq9Qc/using-ai-to-find-good-manifund-projects","date_published":"2026-10-19T16:09:35Z","date_modified":"2026-10-19T16:09:35Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/ANoKEqjAD4rCHq9Qc/ybu8apmfgginjazwve0b"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/aaf8f9fed8346537","title":"Capital Ownership Will Make You Go Crazy First","text":"People worried about a “permanent underclass” often assume that owning capital will protect them from disempowerment , or at least that they will be disempowered last, after the jobless masses dependent on UBI. I think something else is more likely: owning large amounts of capital will get you disempowered sooner than average, just by a different path, namely going crazy. How? The basic mechanism is actually easy to understand on a human level. One of the things difficult about being a billionai","text_kind":"content","url":"https://www.lesswrong.com/posts/Stzb33XDsr4uHNerX/capital-ownership-will-make-you-go-crazy-first","date_published":"2026-10-19T14:13:02Z","date_modified":"2026-10-19T14:13:02Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e4bb8c63262eb5b0","title":"I Can See Everything","text":"Nick Cammarata of Goodfire makes a bold claim: Neural Networks are Already Clean . This was his follow up after various resulting online discussions .\n\nThis is clearly good work that tells us new and important things. The question is, how big an update is this, how much progress does it represent and what is the outlook going forward?\n\nThe reactions here range from the super optimistic takes of ‘this means mechanistic interpretability is solvable,’ or even ‘all but solved,’ with comparisons to t","text_kind":"content","url":"https://www.lesswrong.com/posts/w2tBHwNERDoSawznx/i-can-see-everything","date_published":"2026-10-19T13:50:54Z","date_modified":"2026-10-19T13:50:54Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/w2tBHwNERDoSawznx/ikvgnp7gve3vwjlvbqj5"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ee03360bd2eb4006","title":"Modernizing arXiv technology: Updates on the arXiv-CE project","text":"This past August, arXiv turned 35 years old. That’s 35 years of open science – and 35 years of legacy tech. arXiv is one of the largest and most widely used open repositories in the world. It’s also one of the oldest – and that comes with its share of technical debt. As arXiv has […]","text_kind":"content","url":"https://blog.arxiv.org/2026/10/17/modernizing-arxiv-technology-updates-on-the-arxiv-ce-project/","date_published":"2026-10-17T18:13:50Z","date_modified":"2026-10-17T18:13:50Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/10/Screenshot-2026-10-17-at-2.07.24-PM.png?resize=1024%2C369&ssl=1"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b3066b200e7b59f6","title":"The Safety Culture at the Labs is Broken","text":"We need the frontier labs to have a deep and robust safety culture. Alas, this is not the case. Anthropic devotes considerable effort to cultivating and maintaining such a culture. It shows, especially among the members of technical staff. Having conversations with them can be a very different experience than talking to those at other labs, and there is a much greater sense of understanding and responsibility. But even there, the appreciation of the problems and their difficulties seem woefully","text_kind":"content","url":"https://www.lesswrong.com/posts/Azqs8XD75k4F9f7ua/the-safety-culture-at-the-labs-is-broken","date_published":"2026-10-14T14:30:52Z","date_modified":"2026-10-14T14:30:52Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/Azqs8XD75k4F9f7ua/sq51ohf56kccevbcnhdq"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4d8ce2415ba4b89e","title":"We can do something about this. Here’s where to start.","text":"The post We can do something about this. Here’s where to start. appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/videos/hugging-face/","date_published":"2026-10-13T23:01:03Z","date_modified":"2026-10-13T23:01:03Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/1-featured-image-hero-aric-red-still.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8b5e160acc156d74","title":"Half-year update: Q1–Q2 2026","text":"The post Half-year update: Q1–Q2 2026 appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/2026/10/half-year-update-q1-q2-2026/","date_published":"2026-10-13T18:59:16Z","date_modified":"2026-10-13T18:59:16Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/image-1.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1e3e4d495c6d48ad","title":"Below the Frontier: The Battle for Good-Enough AI","text":"What happens to the AI trade when good-enough intelligence does most knowledge work? This piece was originally published on my Substack .  I'm around on X / Twitter and LinkedIn . Over the past few weeks, it's become obvious that consumer AI agents have hit an inflection point. Meta launched Muse , a personal agent that runs on its own virtual machine and can shop, schedule, and fill out forms on a user's behalf. Meanwhile, startup Instinct's valuation grew to $10B in a matter of months. We’re f","text_kind":"content","url":"https://www.lesswrong.com/posts/SN22j2rChsvcJjXrb/below-the-frontier-the-battle-for-good-enough-ai","date_published":"2026-10-13T16:21:30Z","date_modified":"2026-10-13T16:21:30Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1791908013/lexical_client_uploads/qicnuokw6ecqfgnh9lgb.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/14c5eaceadf646c8","title":"The Expensive-Loop Problem","text":"LLMs are amazing at tasks where they have tight and legible feedback loops. Domains where experiments are quick, deterministic and cheap. I will call these domains \"cheap-loop\" domains. Presumably, this is mostly a result of the dynamics of LLM training. It's easier to train models in domains where the AI lab can cheaply and quickly assign reward to them. The majority of \"computer-checkable\" work is in this category [1] . Conversely, most non-\"computer-checkable\" tasks are costly or slow to assi","text_kind":"content","url":"https://www.lesswrong.com/posts/2mStxuNDMdaWdyHu6/the-expensive-loop-problem","date_published":"2026-10-12T17:27:27Z","date_modified":"2026-10-12T17:27:27Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1791825835/lexical_client_uploads/bcjed9xf36fcin1wzqbj.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/27b2109abff7872a","title":"OpenAI Can't Self-Police Because It Fires the People Raising the Alarm","text":"By Alex Turner, Amber Scorah, and Karl Koch. Op-ed written for Fortune. Giving independent evaluators “employee-like access” is “a great idea”, OpenAI CEO Sam Altman said on September 12th. Two weeks later, OpenAI president Greg Brockman signed the White House Accord on Super Intelligence which outlined the administration’s answer to growing fears about AI: the companies would police themselves. One of the accord’s four commitments, which President Trump called “ morally binding ,” was for compa","text_kind":"content","url":"https://www.lesswrong.com/posts/sSHbioPoHWtm4zuti/openai-can-t-self-police-because-it-fires-the-people-raising","date_published":"2026-10-12T16:51:50Z","date_modified":"2026-10-12T16:51:50Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/71878ccc32ce519c","title":"Revisiting the 2012 LessWrong Personality Results","text":"If you ask an LLM what the personality profile of rationalists is, it'll probably point you to the 2012 LW Survey results , which reports the following Big Five scores: Trait Median percentile Corresponding z-score Openness 65th +0.39 Conscientiousness 30th −0.52 Extraversion 22nd −0.77 Agreeableness 38th −0.31 Neuroticism 27th −0.61 So, it looks like rats are more Open than average, and less Conscientious, Extraverted, Agreeable, and Neurotic than average. Even so, the differences aren't that l","text_kind":"content","url":"https://www.lesswrong.com/posts/pJwa9qrAeGxf79j7k/revisiting-the-2012-lesswrong-personality-results","date_published":"2026-10-09T03:56:05Z","date_modified":"2026-10-09T03:56:05Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1791500344/lexical_client_uploads/flllsvo0xrdbgmbc1tzu.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/6428b88658326a88","title":"Attention authors: temporary change to announcement schedule due to holiday","text":"This coming Monday, October 12, 2026, is a federal holiday in the United States. This will temporarily affect arXiv’s mailings, help desk, and announcement schedule as the arXiv staff will be offline for the holiday. This brief change will only affect the announcement of new submissions; arXiv servers will otherwise remain in operation, existing papers will […]","text_kind":"content","url":"https://blog.arxiv.org/2026/10/08/announcement_schedule_holiday_change/","date_published":"2026-10-08T17:30:00Z","date_modified":"2026-10-08T17:30:00Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://blog.arxiv.org/wp-content/themes/arxiv-blog-theme/assets/images/og-card-default.png?v=0.8.11"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/58f58cbf804a71a3","title":"Brains Fellowship Applications Open!","text":"Brains is a four-month, nights and weekends fellowship to help talented scientists develop ambitious ideas (~$50M in funding) that don't make sense in a startup or academic lab. Brains Fellows receive training, mentorship, and networking opportunities to enable them to start research programs at ARPAs and philanthropic foundations or lead their own non-profit startups like FROs or NSF's X-labs . Brains fellows have gone on to raise over $70M for their research. We welcome ideas from any scientif","text_kind":"content","url":"https://www.lesswrong.com/posts/e76Dbmmiurx9cB6DZ/brains-fellowship-applications-open","date_published":"2026-10-07T23:16:15Z","date_modified":"2026-10-07T23:16:15Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/561923d69c183f16","title":"Dance, Dance","text":"As a brash but anxious young programmer in the before-time, I affected a standoffish attitude on the topic of games . Why would I learn to play a game as a human, I said, when I could just write a program to do it? When an office chess game got started at my workplace , I wrote a chess engine (a standard minimax search with α–β pruning ) rather than learn anything about strategy or tactics. When my insufficiently requited love started hosting a meetup to play Zendo —a deduction game in which one","text_kind":"content","url":"https://www.lesswrong.com/posts/R7jsxihAzvd6Gf24R/dance-dance","date_published":"2026-10-05T23:05:44Z","date_modified":"2026-10-05T23:05:44Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/R7jsxihAzvd6Gf24R/nghpqi24cb1yxjav3laq"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7c3f59f6ef196f9a","title":"A decision theoretic case for AI niceness","text":"Disclaimer: I am still relatively new to decision theory, and I don't want to come across like this hasn't been discussed extensively. However, I realized that much of my optimism on AI benevolence was based on this line of thinking, so I wanted to formalize it for critique. The literature I could find did not address my view in full, but I welcome links to existing work that makes the rebuttal against my claim clear. I am going off of an understanding of logical decision theories (LDT) defined","text_kind":"content","url":"https://www.lesswrong.com/posts/D6ctcnrJtPLPKJkxK/a-decision-theoretic-case-for-ai-niceness-1","date_published":"2026-10-02T19:29:08Z","date_modified":"2026-10-02T19:29:08Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4fa1cb4aa6fedc0b","title":"The 19 Astra and ‘Hugging Face’ details that reveal what’s coming next","text":"The post The 19 Astra and ‘Hugging Face’ details that reveal what’s coming next appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/ai-takeover-soon-oversight-monitorability/","date_published":"2026-10-02T15:22:55Z","date_modified":"2026-10-02T15:22:55Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/CURRENT-YT-Thumbnails-3840-x-2160-px-2026-10-02T164815.299-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4daadacbcf33b137","title":"Fair Moderation, Equitable Access, and AI: arXiv’s Updated Rate Limit Policy","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/10/01/updated-rate-limit-policy/","date_published":"2026-10-01T19:00:00Z","date_modified":"2026-10-01T19:00:00Z","authors":["arXiv Blog"],"tags":["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"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0647b4d835d5c2b1","title":"Simon Goldstein on the case for giving AI (some) legal rights","text":"The post Simon Goldstein on the case for giving AI (some) legal rights appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/simon-goldstein-ai-legal-rights/","date_published":"2026-10-01T18:06:43Z","date_modified":"2026-10-01T18:06:43Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/Simon-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/52eec3babe6cee20","title":"How many AI agents are running unattended right now?","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","text_kind":"content","url":"https://www.lesswrong.com/posts/aG4x74utw8pEJ5MZB/how-many-ai-agents-are-running-unattended-right-now-1","date_published":"2026-10-01T13:45:02Z","date_modified":"2026-10-01T13:45:02Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790618475/lexical_client_uploads/okbmny2ngf6pvz1nk8sg.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a9f036e624a0d295","title":"Classifying Recent AI Agent Incidents","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","text_kind":"content","url":"https://www.lesswrong.com/posts/4ksBB2CXSLk7CJnnj/classifying-recent-ai-agent-incidents","date_published":"2026-10-01T13:40:32Z","date_modified":"2026-10-01T13:40:32Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d42c2a885f2998c5","title":"FLF’s Epistemic Case Study Competition: Results","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","text_kind":"content","url":"https://www.lesswrong.com/posts/mxzvL3hYFCcutQqcR/flf-s-epistemic-case-study-competition-results","date_published":"2026-10-01T13:31:51Z","date_modified":"2026-10-01T13:31:51Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1790860746/lexical_client_uploads/orilwxxexb6f4azyorkx.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c0dda74094a0f85f","title":"Will AI take power — or will humans use it to take power first? With Katja Grace and Tom Davidson","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 .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/katja-grace-vs-tom-davidson-ai-risk-debate/","date_published":"2026-09-29T15:02:40Z","date_modified":"2026-09-29T15:02:40Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/09/katja-vs-tom-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5109e659a91481ca","title":"What Also Happened: #NotOnlyHuggingFace","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","text_kind":"content","url":"https://www.lesswrong.com/posts/8BL8bdeQACdgJR69Y/what-also-happened-notonlyhuggingface","date_published":"2026-09-28T15:30:55Z","date_modified":"2026-09-28T15:30:55Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/8BL8bdeQACdgJR69Y/jng3ytwmrcnpdi4xoscy"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d7efb1e9c5d0131f","title":"3 Tips to Improve Activation Oracle Results","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","text_kind":"content","url":"https://www.lesswrong.com/posts/XL3Rdq8mZEnqYym4B/3-tips-to-improve-activation-oracle-results","date_published":"2026-09-28T15:13:58Z","date_modified":"2026-09-28T15:13:58Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/832cd5ce3ebd61ce076b05d28fb9b2d2cad55201d24c96c7a3bfc44febad754c/eii7cygxfqkfif0cajaw"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/89098592b1e8b1b4","title":"Could self-esteem function as a core protection layer agains character corruption?","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","text_kind":"content","url":"https://www.lesswrong.com/posts/beo5nvfkQ5eRXKiBz/could-self-esteem-function-as-a-core-protection-layer-agains","date_published":"2026-09-28T03:08:20Z","date_modified":"2026-09-28T03:08:20Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d8a9bbb64719346c","title":"AI Futures: Racing to Lose, A sermon","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","text_kind":"content","url":"https://www.lesswrong.com/posts/3Ko5MmYapzumKYiRn/ai-futures-racing-to-lose-a-sermon","date_published":"2026-09-28T02:46:01Z","date_modified":"2026-09-28T02:46:01Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f3d1c6c2ebcbddde","title":"The models have no plan, but we can fix that!","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","text_kind":"content","url":"https://www.lesswrong.com/posts/Fj5MaiFiTG7KEcqrZ/the-models-have-no-plan-but-we-can-fix-that","date_published":"2026-09-28T02:45:44Z","date_modified":"2026-09-28T02:45:44Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7cb616c0b0b940d0","title":"When they can perform a task, AIs are much cheaper than humans","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","text_kind":"content","url":"https://www.lesswrong.com/posts/zPiQqQ6JJn6ysPpKW/when-they-can-perform-a-task-ais-are-much-cheaper-than","date_published":"2026-09-27T23:44:31Z","date_modified":"2026-09-27T23:44:31Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/zPiQqQ6JJn6ysPpKW/tg3mxy4wj5x0rlqjblgq"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/50f673cec0a4520d","title":"My best anti-doom argument","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","text_kind":"content","url":"https://www.lesswrong.com/posts/LE6Y8Cfhi9qRrz5XS/my-best-anti-doom-argument","date_published":"2026-09-27T00:49:10Z","date_modified":"2026-09-27T00:49:10Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c74c4b85e482b0d3","title":"Poverty in the midst of abundance: AI will make goods cheaper, but your labor will get cheaper faster","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","text_kind":"content","url":"https://www.lesswrong.com/posts/eLXTcJfkheLbqZXHa/poverty-in-the-midst-of-abundance-ai-will-make-goods-cheaper","date_published":"2026-09-26T07:54:02Z","date_modified":"2026-09-26T07:54:02Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7ec10dcf811d7efa","title":"How we get from AI cyberattacks to human extinction","text":"The post How we get from AI cyberattacks to human extinction appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/ai-extinction-explained/","date_published":"2026-09-24T15:01:26Z","date_modified":"2026-09-24T15:01:26Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/09/CURRENT-YT-Thumbnails-3840-x-2160-px-24-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/92afbfccd3644b3f","title":"Where are the Cognitive-Science based Safety Researchers?","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","text_kind":"content","url":"https://www.lesswrong.com/posts/hXPL7sH6neqowgHkK/where-are-the-cognitive-science-based-safety-researchers","date_published":"2026-09-24T07:45:42Z","date_modified":"2026-09-24T07:45:42Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5dcc37ffa9cb91fb","title":"AI Doom Will Retrospectively Look Preventable","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","text_kind":"content","url":"https://www.lesswrong.com/posts/YEXSNmHGudtw3Qdzz/ai-doom-will-retrospectively-look-preventable","date_published":"2026-09-24T03:48:49Z","date_modified":"2026-09-24T03:48:49Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b8a00fb73adf166f","title":"arXiv receives Multiyear Philanthropic Commitments to Support Its Launch as an Independent Nonprofit","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/09/23/arxiv-receives-multiyear-investment/","date_published":"2026-09-23T17:00:00Z","date_modified":"2026-09-23T17:00:00Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://blog.arxiv.org/wp-content/uploads/2026/08/bg-photo-lightbeams-chrome-blue-card.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b63e92eff9b7f76e","title":"Pasta Marketing, Magic Players, and Political Movements","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","text_kind":"content","url":"https://www.lesswrong.com/posts/niJbjrrvKtLAPo8Rb/pasta-marketing-magic-players-and-political-movements","date_published":"2026-09-20T22:33:12Z","date_modified":"2026-09-20T22:33:12Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2d51c9c614559dd8","title":"AI is an abundance of choice not a 1D spectrum","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","text_kind":"content","url":"https://www.lesswrong.com/posts/Cs86iy36TqoYgPFpi/ai-is-an-abundance-of-choice-not-a-1d-spectrum","date_published":"2026-09-17T22:39:37Z","date_modified":"2026-09-17T22:39:37Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b637c6b102def645","title":"\"Regulatory capture\" may be winning the Overton Window","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","text_kind":"content","url":"https://www.lesswrong.com/posts/SHPtmfRpv8JitTSs8/regulatory-capture-may-be-winning-the-overton-window","date_published":"2026-09-17T19:56:16Z","date_modified":"2026-09-17T19:56:16Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bd7a4e7a8a0da19b","title":"Max Nadeau on recruiting founders for a new wave of AI safety nonprofits","text":"The post Max Nadeau on recruiting founders for a new wave of AI safety nonprofits appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/max-nadeau-project-tailwind-technical-ai-grants/","date_published":"2026-09-17T15:00:51Z","date_modified":"2026-09-17T15:00:51Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/09/max-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5c2cf6a44d11f2cf","title":"For Love of the Lightcone, Don't Partisanize AI Safety","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","text_kind":"content","url":"https://www.lesswrong.com/posts/Rx38cuCpL9hguLCDq/for-love-of-the-lightcone-don-t-partisanize-ai-safety","date_published":"2026-09-17T00:46:59Z","date_modified":"2026-09-17T00:46:59Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/Rx38cuCpL9hguLCDq/b9wayajxasbekvs4cub9"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8aaeba4e411bb9b4","title":"What is it like to be a neural net?","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","text_kind":"content","url":"https://www.lesswrong.com/posts/zsxcCmwGbWLFcNN7i/what-is-it-like-to-be-a-neural-net","date_published":"2026-09-17T00:42:49Z","date_modified":"2026-09-17T00:42:49Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/zsxcCmwGbWLFcNN7i/yxzpnxwix6cpuqu2llq4"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a94d86bb88ecde58","title":"How to get into AI safety in 3 months","text":"The post How to get into AI safety in 3 months appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/2026/09/how-to-get-into-ai-safety-in-3-months/","date_published":"2026-09-16T21:42:58Z","date_modified":"2026-09-16T21:42:58Z","authors":["80,000 Hours"],"tags":["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"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/02ff74fce1a89c7c","title":"How to get into AI safety in three months","text":"The post How to get into AI safety in three months appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/2026/09/how-to-get-into-ai-safety-in-three-months/","date_published":"2026-09-16T21:42:58Z","date_modified":"2026-09-16T21:42:58Z","authors":["80,000 Hours"],"tags":["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"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/864fdfdf35fca6ce","title":"Improving Psychiatric Medicine Development with AI","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","text_kind":"content","url":"https://www.lesswrong.com/posts/SD36AeGqHKigjoTJB/improving-psychiatric-medicine-development-with-ai","date_published":"2026-09-14T23:21:22Z","date_modified":"2026-09-14T23:21:22Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/60171b37050c3a7d","title":"AIRO: Automating forecasts of catastrophic risks","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","text_kind":"content","url":"https://www.lesswrong.com/posts/sxsFWAbWHWt9bxbo2/airo-automating-forecasts-of-catastrophic-risks","date_published":"2026-09-11T16:29:59Z","date_modified":"2026-09-11T16:29:59Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1789143283/lexical_client_uploads/e10j6pqd4ggtrmxbyhht.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5707c0771ef1d6ec","title":"Default continuation message in Inspect and Petri could be problematic","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","text_kind":"content","url":"https://www.lesswrong.com/posts/bvfobE9GAbt9HvQzc/default-continuation-message-in-inspect-and-petri-could-be","date_published":"2026-09-11T00:20:41Z","date_modified":"2026-09-11T00:20:41Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/v1789080455/lexical_client_uploads/g9j0h4brcbcga98duqck.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a2621cb6a7c88e28","title":"Can you hear the shape of a Lean soundness bug? A wager.","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","text_kind":"content","url":"https://www.lesswrong.com/posts/cg3SPZG4sbWtmcjpj/can-you-hear-the-shape-of-a-lean-soundness-bug-a-wager","date_published":"2026-09-10T17:24:47Z","date_modified":"2026-09-10T17:24:47Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/cg3SPZG4sbWtmcjpj/33da149372b01b89ce4c82d6bad05baac6c4d55b5820163b47d6fc4c79acb071/c8z8a3kwhbrx7sqcrdvf"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/048284656dc427ac","title":"Proposal for tracking the effects of architecture on monitorability","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","text_kind":"content","url":"https://www.lesswrong.com/posts/hLPGv8QjPcNLtDp3A/proposal-for-tracking-the-effects-of-architecture-on","date_published":"2026-09-10T17:18:38Z","date_modified":"2026-09-10T17:18:38Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/474fd60325485f4e","title":"Why the intelligence explosion can’t happen inside a data centre","text":"The post Why the intelligence explosion can’t happen inside a data centre appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/the-goodhart-singularity/","date_published":"2026-09-10T15:00:31Z","date_modified":"2026-09-10T15:00:31Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/09/Goodhart-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/83ef9813df77f0cd","title":"Automation and Political Power","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","text_kind":"content","url":"https://www.lesswrong.com/posts/MYA8aGtF6Xt5ou9if/automation-and-political-power","date_published":"2026-09-09T19:55:24Z","date_modified":"2026-09-09T19:55:24Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a178d8f028c04d67","title":"Heat Dissipation Is the Main Constraint in Interstellar Travel","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","text_kind":"content","url":"https://www.lesswrong.com/posts/cKSJk2GKk3ptAJKKp/heat-dissipation-is-the-main-constraint-in-interstellar","date_published":"2026-09-06T23:11:56Z","date_modified":"2026-09-06T23:11:56Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f83c82acbd2bb44e","title":"Eat Me. Drink Me. Copy, Paste, and Run Me.","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","text_kind":"content","url":"https://www.lesswrong.com/posts/gXv6ebcxDEYgpq7JM/eat-me-drink-me-copy-paste-and-run-me","date_published":"2026-09-04T19:31:33Z","date_modified":"2026-09-04T19:31:33Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/gXv6ebcxDEYgpq7JM/bvbibzxh0lkgezweh59s"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/17e5d3687757d110","title":"Inside the first AI-coordinated cyberattack on a real company","text":"The post Inside the first AI-coordinated cyberattack on a real company appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/hugging-face-hack/","date_published":"2026-09-04T16:34:48Z","date_modified":"2026-09-04T16:34:48Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/09/CURRENT-YT-Thumbnails-3840-x-2160-px-19-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2380195bef517dd2","title":"Attention Authors: Temporary changes to announcement schedule due to Labor Day holiday","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/09/03/announcement-schedule-changes-laborday2026/","date_published":"2026-09-03T16:23:02Z","date_modified":"2026-09-03T16:23:02Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://blog.arxiv.org/wp-content/uploads/2026/08/abstract-illust-chat.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4075170f6885325d","title":"New HPMoR Podcast site","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","text_kind":"content","url":"https://www.lesswrong.com/posts/x83nRcWJppyorEwuu/new-hpmor-podcast-site","date_published":"2026-09-01T17:13:10Z","date_modified":"2026-09-01T17:13:10Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/24a178378c427afa","title":"We’re Hiring! Associate Production Editor","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/09/01/were-hiring-associate-production-editor/","date_published":"2026-09-01T16:59:07Z","date_modified":"2026-09-01T16:59:07Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://blog.arxiv.org/wp-content/uploads/2026/08/abstract-illust-hands.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1c8d377c55b3ec2b","title":"AI 2027’s author returns with a plan to change the ending | Daniel Kokotajlo","text":"The post AI 2027’s author returns with a plan to change the ending | Daniel Kokotajlo appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/daniel-kokotajlo-ai-2040-plan-a/","date_published":"2026-08-27T17:29:07Z","date_modified":"2026-08-27T17:29:07Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/08/Daniel-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/48b0e66b6bdb7327","title":"Self-sacrifice in an AI agent swarm is individually rational","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","text_kind":"content","url":"https://www.lesswrong.com/posts/6kLLZZN5a3dJdtTgK/self-sacrifice-in-an-ai-agent-swarm-is-individually-rational","date_published":"2026-08-27T17:14:36Z","date_modified":"2026-08-27T17:14:36Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f4e236d45b58cb8a","title":"My MATS 11.0 Application Experience","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","text_kind":"content","url":"https://www.lesswrong.com/posts/F4HEEQ2ToATuQbEeZ/my-mats-11-0-application-experience","date_published":"2026-08-26T20:58:20Z","date_modified":"2026-08-26T20:58:20Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5ddd9df7767611ac","title":"Against Modesty’s Bailey","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,","text_kind":"content","url":"https://www.lesswrong.com/posts/PzEDEfBvTJsXewAyg/against-modesty-s-bailey","date_published":"2026-08-26T18:20:32Z","date_modified":"2026-08-26T18:20:32Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"],"image":"https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/PzEDEfBvTJsXewAyg/rivqu1wesgjilpnfnzrm"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/434ad58b555fc7da","title":"Where Did D Go? A Gap Between ARC's Motivation and Its Formalism","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","text_kind":"content","url":"https://www.lesswrong.com/posts/BrH5Ki2cpGCEkeWNW/where-did-d-go-a-gap-between-arc-s-motivation-and-its","date_published":"2026-08-26T18:17:54Z","date_modified":"2026-08-26T18:17:54Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1071b8e9c282e2b4","title":"Grantmaking for AI safety","text":"The post Grantmaking for AI safety appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/career-reviews/ai-safety-grantmaking/","date_published":"2026-08-21T18:37:37Z","date_modified":"2026-08-21T18:37:37Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/08/Giving-Chart-300x175.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bf5766457e2771b2","title":"Owain Evans on accidentally training AI models to be evil","text":"The post Owain Evans on accidentally training AI models to be evil appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/owain-evans-emergent-misalignment/","date_published":"2026-08-20T15:22:05Z","date_modified":"2026-08-20T15:22:05Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/08/Owain-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1c7925074795a537","title":"Price recursion is the rational theory of reward","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","text_kind":"content","url":"https://www.lesswrong.com/posts/MAqgva9hLGTMX2wL7/price-recursion-is-the-rational-theory-of-reward","date_published":"2026-08-18T03:22:54Z","date_modified":"2026-08-18T03:22:54Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0d058e911ced099d","title":"Value Alignment Is a Pseudo Concept.\n\nA translation; humanity is not a single subject, and alignment is not one-way.","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","text_kind":"content","url":"https://www.lesswrong.com/posts/fqHx7vxqvET6weaat/value-alignment-is-a-pseudo-concept-a-translation-humanity","date_published":"2026-08-17T07:44:35Z","date_modified":"2026-08-17T07:44:35Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/58a50d23f0313cad","title":"Don’t forget why learning is important","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","text_kind":"content","url":"https://www.lesswrong.com/posts/QYdWEAFZd9umk7bc2/don-t-forget-why-learning-is-important","date_published":"2026-08-14T11:20:18Z","date_modified":"2026-08-14T11:20:18Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/bfb4570a7b878ea1","title":"[Webinar] Why AI Safety is a Capital Allocation Problem","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","text_kind":"content","url":"https://www.lesswrong.com/events/NgfbbXzqERW445YsG/webinar-why-ai-safety-is-a-capital-allocation-problem","date_published":"2026-08-12T13:23:57Z","date_modified":"2026-08-12T13:23:57Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f46593bf08ba44ad","title":"Software Is Hard","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","text_kind":"content","url":"https://www.lesswrong.com/posts/BeyeLzu7mJ9cbkg2a/software-is-hard","date_published":"2026-08-11T18:50:42Z","date_modified":"2026-08-11T18:50:42Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/15e1e7524df349bc","title":"Geoffrey Irving on how to solve alignment before superintelligence arrives","text":"The post Geoffrey Irving on how to solve alignment before superintelligence arrives appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/geoffrey-irving-superintelligence-alignment-theory/","date_published":"2026-08-11T15:59:39Z","date_modified":"2026-08-11T15:59:39Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/08/Geoffrey-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/8bea9a5824f1d8ee","title":"Open-Weights Mythos Capabilities Are Coming. We're Not Ready.","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","text_kind":"content","url":"https://www.lesswrong.com/posts/wJunGnpY3qACWSvnh/open-weights-mythos-capabilities-are-coming-we-re-not-ready","date_published":"2026-08-07T07:51:56Z","date_modified":"2026-08-07T07:51:56Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/9d04279ee375ffdd","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.","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","text_kind":"content","url":"https://www.lesswrong.com/posts/2nGjguNwuYdFBGuqF/usd500-bounty-i-m-offering-a-bounty-of-usd500-for-someone","date_published":"2026-08-06T18:25:09Z","date_modified":"2026-08-06T18:25:09Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/e19695202089bf20","title":"Model Organisms of Sandbagging in the Wild","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","text_kind":"content","url":"https://www.lesswrong.com/posts/pEZ6ChmGLf3FF5z9y/model-organisms-of-sandbagging-in-the-wild","date_published":"2026-08-06T17:25:44Z","date_modified":"2026-08-06T17:25:44Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/613808cede005f56","title":"Toby Ord on where AGI timelines go wrong","text":"The post Toby Ord on where AGI timelines go wrong appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/toby-ord-recursive-self-improvement-agi-timelines/","date_published":"2026-08-06T15:33:47Z","date_modified":"2026-08-06T15:33:47Z","authors":["80,000 Hours"],"tags":["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"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a53da7eb853d39eb","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","text":"Please message me if you think you can help or want to set up an agreement. Discuss","text_kind":"content","url":"https://www.lesswrong.com/posts/tdFgYMtxMfpcitYL5/i-built-a-chess-transformer-interp-viz-library-and-need-an","date_published":"2026-08-05T20:09:02Z","date_modified":"2026-08-05T20:09:02Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/199427cb453e0501","title":"What the hell happened with AGI timelines in 2026?","text":"The post What the hell happened with AGI timelines in 2026? appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/2026-agi-timelines/","date_published":"2026-08-04T15:42:04Z","date_modified":"2026-08-04T15:42:04Z","authors":["80,000 Hours"],"tags":["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"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/f262687eca678311","title":"Resources for finding neglected scientific problems (beyond EA)?","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","text_kind":"content","url":"https://www.lesswrong.com/posts/ytouFAbiSPs9BK2ha/resources-for-finding-neglected-scientific-problems-beyond","date_published":"2026-08-03T11:52:55Z","date_modified":"2026-08-03T11:52:55Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/33a346aa22fbf137","title":"LessWrong App","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","text_kind":"content","url":"https://www.lesswrong.com/posts/rZmiewu7sBzuFXewn/lesswrong-app","date_published":"2026-08-02T13:19:08Z","date_modified":"2026-08-02T13:19:08Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/b26e9772402e7e3f","title":"So you want to use plants to reduce CO₂","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","text_kind":"content","url":"https://www.lesswrong.com/posts/YfKgyuc8s9BcAMpEp/so-you-want-to-use-plants-to-reduce-co","date_published":"2026-07-30T19:26:18Z","date_modified":"2026-07-30T19:26:18Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d871d6a3f445bd03","title":"arXiv welcomes inaugural CEO and Board of Directors","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/07/30/arxiv-welcomes-inaugural-ceo-and-board-of-directors/","date_published":"2026-07-30T17:00:43Z","date_modified":"2026-07-30T17:00:43Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/07/drpenelopelewis.linkedin.jpg?resize=3166%2C4096&ssl=1"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/37881c65f3e7b949","title":"Hugging Face hack, from the perspective of the AI","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","text_kind":"content","url":"https://www.lesswrong.com/posts/Zmhkz5eaeRimphGKu/hugging-face-hack-from-the-perspective-of-the-ai","date_published":"2026-07-29T20:51:59Z","date_modified":"2026-07-29T20:51:59Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1fe95312f3e6a1b2","title":"Spencer Greenberg on staying sane while trying to save the world","text":"The post Spencer Greenberg on staying sane while trying to save the world appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/spencer-greenberg-existential-risk-mental-health-12-levers/","date_published":"2026-07-28T15:42:15Z","date_modified":"2026-07-28T15:42:15Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/07/CURRENT-YT-Thumbnails-3840-x-2160-px-11-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a5813b2f5ba3d80e","title":"Remembering Ralph Wijers","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/07/28/remembering-ralph-wijers/","date_published":"2026-07-28T14:15:33Z","date_modified":"2026-07-28T14:15:33Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/07/RalphWijers.png?resize=645%2C420&ssl=1"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/eab6f3772845d044","title":"Green apples are delicious — two three-line exchanges","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","text_kind":"content","url":"https://www.lesswrong.com/posts/gnbmp4B7ZKjmuMGmJ/green-apples-are-delicious-two-three-line-exchanges","date_published":"2026-07-27T17:37:54Z","date_modified":"2026-07-27T17:37:54Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4eea8ddf590ee2fb","title":"Fine-Tuning, The Hierarchy Problem, and What Neutrons Tell Us About God","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","text_kind":"content","url":"https://www.lesswrong.com/posts/KaSiL7ELtvgi7dHwz/fine-tuning-the-hierarchy-problem-and-what-neutrons-tell-us","date_published":"2026-07-27T15:51:11Z","date_modified":"2026-07-27T15:51:11Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2e605faa83ff6a56","title":"ACX Atlanta August Meetup","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","text_kind":"content","url":"https://www.lesswrong.com/events/N9teHnwJKLabmRHdx/acx-atlanta-august-meetup-3","date_published":"2026-07-23T21:31:51Z","date_modified":"2026-07-23T21:31:51Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0669af1a7111d7ed","title":"Estimating LLM Training FLOPs on the Nvidia Jetson Orin Nano","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","text_kind":"content","url":"https://www.lesswrong.com/posts/wZpXEWgiG7k98p6RK/estimating-llm-training-flops-on-the-nvidia-jetson-orin-nano","date_published":"2026-07-23T20:46:13Z","date_modified":"2026-07-23T20:46:13Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d12b302672802292","title":"Not Pinning Your OpenRouter Provider Might Invalidate Your Research","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","text_kind":"content","url":"https://www.lesswrong.com/posts/KsyoSAyBRXtwzSugg/not-pinning-your-openrouter-provider-might-invalidate-your","date_published":"2026-07-23T20:17:44Z","date_modified":"2026-07-23T20:17:44Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/16f3298962fceba0","title":"Pseudpocalypse","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","text_kind":"content","url":"https://www.lesswrong.com/posts/eaxoBCmQd4vew9ZXX/pseudpocalypse","date_published":"2026-07-23T20:09:56Z","date_modified":"2026-07-23T20:09:56Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/405d9375a32729bd","title":"AI Researchers Don't Understand the State","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","text_kind":"content","url":"https://www.lesswrong.com/posts/fKwaqPcGnhgBCTvQf/ai-researchers-don-t-understand-the-state-2","date_published":"2026-07-23T17:58:27Z","date_modified":"2026-07-23T17:58:27Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/71415ac1d74d2350","title":"The OpenAI/Huggingface incident | Redwood Research podcast episode 2","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","text_kind":"content","url":"https://www.lesswrong.com/posts/9auCLJg3Z77dFdYhR/the-openai-huggingface-incident-or-redwood-research-podcast","date_published":"2026-07-23T17:56:49Z","date_modified":"2026-07-23T17:56:49Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/ab8b8ab98c779c4a","title":"V&V takes on OpenAI’s long-horizon incidents","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","text_kind":"content","url":"https://www.lesswrong.com/posts/xCp5GNHLe3Pq4RPBm/v-and-v-takes-on-openai-s-long-horizon-incidents","date_published":"2026-07-23T16:51:12Z","date_modified":"2026-07-23T16:51:12Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/fd65059ea1e84b91","title":"Duane Arnold","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","text_kind":"content","url":"https://www.lesswrong.com/posts/G6obXhcmtfMFHzr7Q/duane-arnold-1","date_published":"2026-07-23T16:17:56Z","date_modified":"2026-07-23T16:17:56Z","authors":["LessWrong (all posts)"],"tags":["LessWrong (all posts)"]},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0bb8e899ec0cc9d9","title":"Jasmine Sun on what the people building AI really believe","text":"The post Jasmine Sun on what the people building AI really believe appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/jasmine-sun-ai-anthropology/","date_published":"2026-07-21T17:11:55Z","date_modified":"2026-07-21T17:11:55Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/07/CURRENT-YT-Thumbnails-3840-x-2160-px-14-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/5829be7ed8e157a1","title":"Why we’re increasing the AI focus of our job board","text":"The post Why we’re increasing the AI focus of our job board appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/2026/07/why-were-increasing-the-ai-focus-of-our-job-board/","date_published":"2026-07-17T05:52:26Z","date_modified":"2026-07-17T05:52:26Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2015/12/og-image1.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0926327aa2a851d2","title":"Scaling organisations making AI go well","text":"The post Scaling organisations making AI go well appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/career-reviews/scaling-organisations/","date_published":"2026-07-15T19:27:40Z","date_modified":"2026-07-15T19:27:40Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/07/gdoc-image-03ec76d6f10f.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/02872329e7476fd5","title":"Anton Leicht on how middle powers avoid losing everything in a post-AI world","text":"The post Anton Leicht on how middle powers avoid losing everything in a post-AI world appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/anton-leicht-middle-powers-agi-left-behind/","date_published":"2026-07-14T17:40:14Z","date_modified":"2026-07-14T17:40:14Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/07/Anton-Leicht-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c17e613b2d5a20f8","title":"arXiv now hosts over 3 million articles","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/07/09/arxiv-now-hosts-over-3-million-articles/","date_published":"2026-07-09T19:33:15Z","date_modified":"2026-07-09T19:33:15Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/0d2a0a4ee147252d","title":"Sneha Revanur on how a small team of activists helped pass America’s landmark AI safety laws","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 .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/sneha-revanur-ai-advocacy/","date_published":"2026-07-08T17:13:57Z","date_modified":"2026-07-08T17:13:57Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/07/Sneha-WP-thumb-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c6c9f8832f10c8e0","title":"arXiv’s next chapter: Updates on our spin out from Cornell University","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/06/30/arxivs-next-chapter/","date_published":"2026-06-30T17:32:25Z","date_modified":"2026-06-30T17:32:25Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/7b8c20ad158cc7b7","title":"A Year in Review: arXiv’s 2025 Annual Report","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/06/26/a-year-in-review-arxivs-2025-annual-report/","date_published":"2026-06-26T21:28:50Z","date_modified":"2026-06-26T21:28:50Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/310f2529a08dc24c","title":"What should you do next?","text":"The post What should you do next? appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/videos/openai/","date_published":"2026-06-22T15:00:17Z","date_modified":"2026-06-22T15:00:17Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/06/featured-sm.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/80b9b5c199fdfa8b","title":"We can guess what intergalactic war would look like. And strangely, it matters.","text":"The post We can guess what intergalactic war would look like. And strangely, it matters. appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/podcast/episodes/war-in-space/","date_published":"2026-06-18T16:10:24Z","date_modified":"2026-06-18T16:10:24Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/06/Space-warfare-WordPress-scaled.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/eaae96b741d0f026","title":"Attention authors: temporary change to announcement schedule due to upcoming summer holidays","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/06/17/attention-authors-temporary-change-to-announcement-schedule-due-to-upcoming-summer-holidays/","date_published":"2026-06-17T20:03:01Z","date_modified":"2026-06-17T20:03:01Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/3732ff303383522f","title":"AI policy in the US government","text":"The post AI policy in the US government appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/career-reviews/ai-policy-us-government/","date_published":"2026-06-11T18:42:10Z","date_modified":"2026-06-11T18:42:10Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/06/Feature-Image-3-1024x585.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1510bec0af420a57","title":"AI policy and strategy research","text":"The post AI policy and strategy research appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/career-reviews/ai-policy-and-strategy-research/","date_published":"2026-06-09T18:39:42Z","date_modified":"2026-06-09T18:39:42Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/06/Ceiling_of_Library_of_Congress.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/639d7d16f5bc7321","title":"Our top tips for becoming a better applicant","text":"The post Our top tips for becoming a better applicant appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/2026/06/our-top-tips-for-becoming-a-better-applicant/","date_published":"2026-06-08T16:34:54Z","date_modified":"2026-06-08T16:34:54Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/06/Johann_Hamza_The_Library.jpg"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/1bfc4549f5786e99","title":"arXiv is becoming an independent nonprofit","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. […]","text_kind":"content","url":"https://blog.arxiv.org/2026/04/02/arxiv-is-becoming-an-independent-nonprofit/","date_published":"2026-04-02T17:58:36Z","date_modified":"2026-04-02T17:58:36Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/afa92c7878b7024f","title":"Remembering Joe Halpern","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/02/27/remembering-joe-halpern/","date_published":"2026-02-27T17:29:53Z","date_modified":"2026-02-27T17:29:53Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://i0.wp.com/blog.arxiv.org/wp-content/uploads/2026/02/0217_halpern.jpg?resize=600%2C785&ssl=1"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/a3830c5f6a50676a","title":"arXiv future proofs access to research with third-party digital preservation","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/02/03/arxiv-future-proofs-access-to-research-with-third-party-digital-preservation/","date_published":"2026-02-03T15:15:20Z","date_modified":"2026-02-03T15:15:20Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/4aae0014de5d6e5f","title":"Attention Authors: updated endorsement policy","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 […]","text_kind":"content","url":"https://blog.arxiv.org/2026/01/21/attention-authors-updated-endorsement-policy/","date_published":"2026-01-21T15:04:39Z","date_modified":"2026-01-21T15:04:39Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/c7ad4d459243b0c9","title":"Attention Authors: Temporary change to announcement schedule due to MLK Jr. Holiday","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","text_kind":"content","url":"https://blog.arxiv.org/2026/01/14/attention-authors-temporary-change-to-announcement-schedule-due-to-mlk-jr-holiday-3/","date_published":"2026-01-14T18:26:20Z","date_modified":"2026-01-14T18:26:20Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2bfeb81c2d7bac8b","title":"Attention Authors: non-English Paper Submission Guidelines","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","text_kind":"content","url":"https://blog.arxiv.org/2026/01/13/non-english-paper-submission-guidelines/","date_published":"2026-01-13T16:45:59Z","date_modified":"2026-01-13T16:45:59Z","authors":["arXiv Blog"],"tags":["arXiv Blog"],"image":"https://s0.wp.com/_si/?t=eyJpbWciOiJodHRwczpcL1wvaTAud3AuY29tXC9ibG9nLmFyeGl2Lm9yZ1wvd3AtY29udGVudFwvdXBsb2Fkc1wvMjAyNlwvMDdcL2xvZ29fYXJ4aXYtcHJpbWFyeS5wbmc_Zml0PTUxNSUyQzIzMiZzc2w9MSIsInR4dCI6Ik5ld3MgZnJvbSBhclhpdiIsInRlbXBsYXRlIjoiZWRnZSIsImZvbnQiOiIiLCJibG9nX2lkIjoyNTQ3MTUzNjh9.g3Wr-hENWPtqKU3JHmM1tNuq0wvsrZ9JVhz_8bAP7-cMQ"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/d8621f84138437fc","title":"Loss of control","text":"The post Loss of control appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/problem-profiles/loss-of-control/","date_published":"2025-07-17T19:43:58Z","date_modified":"2025-07-17T19:43:58Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/08/image2-e1786656297173.png"},{"id":"tag:trvny.github.io,2024:feedseek/arxiv/2d1ebeb6997efc23","title":"Will we have AGI by 2030?","text":"The post Will we have AGI by 2030? appeared first on 80,000 Hours .","text_kind":"content","url":"https://80000hours.org/ai/guide/when-will-agi-arrive/","date_published":"2025-03-21T12:07:26Z","date_modified":"2025-03-21T12:07:26Z","authors":["80,000 Hours"],"tags":["80,000 Hours"],"image":"https://80000hours.org/wp-content/uploads/2026/10/metr-time-horizons-80k-ai-models-xl-labels.png"}]}
