corpus.blog
56,966 blogs · [ { "id": "01a0deea-101a-71a6-9de1-af13e57ee9c1", "title": "Existential Risk Is An Extraordinary Claim", "url": "https://www.lesswrong.com/posts/SveFDdLRxQTTn7Xp7/existential-risk-is-an-extraordinary-claim", "published_at": "2026-09-26T14:44:23+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e613772c", "title": "Claude Opus 5.5 Should Raise Your Ambitions", "url": "https://www.lesswrong.com/posts/rtPiip9igy3QvxYdM/claude-opus-5-5-should-raise-your-ambitions", "published_at": "2026-09-26T11:40:53+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e69b5ee7", "title": "Writing a Theorem Prover from scratch", "url": "https://www.lesswrong.com/posts/y2Dio5gGifWepGQf4/writing-a-theorem-prover-from-scratch", "published_at": "2026-09-26T10:38:24+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e74dbeb5", "title": "Poverty in the midst of abundance: AI will make goods cheaper, but your labor will get cheaper faster", "url": "https://www.lesswrong.com/posts/eLXTcJfkheLbqZXHa/poverty-in-the-midst-of-abundance-ai-will-make-goods-cheaper", "published_at": "2026-09-26T07:54:02+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e7d438eb", "title": "Night dreams have a tech tree?", "url": "https://www.lesswrong.com/posts/JEaeNbhYy4qJsFahi/night-dreams-have-a-tech-tree", "published_at": "2026-09-26T06:57:07+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e8020f17", "title": "Addictions are anesthesia", "url": "https://www.lesswrong.com/posts/PFgzLmEZSBrztDBpu/addictions-are-anesthesia", "published_at": "2026-09-26T06:55:32+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e846d123", "title": "Evidence about risk should be transparent", "url": "https://www.lesswrong.com/posts/LawgAaGTvbbnZi7u2/evidence-about-risk-should-be-transparent", "published_at": "2026-09-25T23:39:53+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e866e03a", "title": "Plan R: AI Safety by ASICs", "url": "https://www.lesswrong.com/posts/n8u3BfqFoGh4jnzpo/plan-r-ai-safety-by-asics", "published_at": "2026-09-25T23:26:13+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e8714e6f", "title": "What did AI researchers think at the end of 2024?", "url": "https://www.lesswrong.com/posts/62j5mPY8pFo2GD5fo/what-did-ai-researchers-think-at-the-end-of-2024-1", "published_at": "2026-09-25T20:50:59+00:00" }, { "id": "01a0deea-101a-71a6-9de1-af13e8bb4e1b", "title": "Alignment Forecasting: Predicting Misalignment from Training Data", "url": "https://www.lesswrong.com/posts/f7r9QCmjoYFG9ReyF/alignment-forecasting-predicting-misalignment-from-training", "published_at": "2026-09-25T18:30:39+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c0d8ec51", "title": "On Ezra Klein’s Podcast With Jensen Huang", "url": "https://www.lesswrong.com/posts/j3xefrWrNqsmMfJEi/on-ezra-klein-s-podcast-with-jensen-huang", "published_at": "2026-09-25T12:00:54+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c1a2eaf1", "title": "Foundational premises of advanced AI", "url": "https://www.lesswrong.com/posts/ioFYSZbCysLDjH3DN/foundational-premises-of-advanced-ai", "published_at": "2026-09-25T10:53:21+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c280cb2b", "title": "Cognitive Reasoning Diversity for Robust AI Juries", "url": "https://www.lesswrong.com/posts/HCpgntjDAE8bycfGn/cognitive-reasoning-diversity-for-robust-ai-juries", "published_at": "2026-09-25T06:35:44+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c25ead7a", "title": "J-lens shouldn't target the final layer by default", "url": "https://www.lesswrong.com/posts/Adb55vLqt33LEyBfz/j-lens-shouldn-t-target-the-final-layer-by-default", "published_at": "2026-09-25T06:32:20+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c3cd5bb2", "title": "Recognition: when an agent counts an entity as itself", "url": "https://www.lesswrong.com/posts/mmkSE8hrEcpaoA7f7/recognition-when-an-agent-counts-an-entity-as-itself", "published_at": "2026-09-25T05:38:18+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c34532ed", "title": "Does anyone else have music constantly playing in their head while studying? If so, how did you overcome it?", "url": "https://www.lesswrong.com/posts/ZXXFvzhjwsvqjtnZd/does-anyone-else-have-music-constantly-playing-in-their-head-1", "published_at": "2026-09-25T04:28:23+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c5d287e9", "title": "Sunday Afternoons", "url": "https://www.lesswrong.com/posts/HPFiGccq8hKWz3eSG/sunday-afternoons", "published_at": "2026-09-25T04:21:51+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c3b0cb07", "title": "Secure Acceleration (linkpost)", "url": "https://www.lesswrong.com/posts/DzMBakaggpcpG9zX8/secure-acceleration-linkpost", "published_at": "2026-09-25T03:29:09+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c444a03d", "title": "The Geopolitics of an AI Treaty", "url": "https://www.lesswrong.com/posts/nnAsPfCrXp8q8qSMq/the-geopolitics-of-an-ai-treaty", "published_at": "2026-09-25T03:19:46+00:00" }, { "id": "01a0d98d-2114-723e-8710-5283c521a3c1", "title": "The Microsoft Code of Conduct Will Fail at Its Intention", "url": "https://www.lesswrong.com/posts/b3itBtnCy3CA2jDBS/the-microsoft-code-of-conduct-will-fail-at-its-intention", "published_at": "2026-09-25T02:51:38+00:00" }, { "id": "01a0bd30-e142-7323-9ca3-74e713646f89", "title": "Why I Stay Off Twitter", "url": "https://www.lesswrong.com/posts/tvwtwgcujTfep4HgY/why-i-stay-off-twitter", "published_at": "2026-09-20T03:30:57+00:00" }, { "id": "01a0bd30-e142-7323-9ca3-74e713c2411c", "title": "Global Challenges in AI Safety for Biosecurity", "url": "https://www.lesswrong.com/posts/KH2JjfSrw6tJdmKzw/global-challenges-in-ai-safety-for-biosecurity", "published_at": "2026-09-20T01:42:18+00:00" }, { "id": "01a0bd30-e142-7323-9ca3-74e713d79fd3", "title": "The Anatomy of a Chinese AI Researcher", "url": "https://www.lesswrong.com/posts/qmxkHm2dTLKG6GZ6i/the-anatomy-of-a-chinese-ai-researcher", "published_at": "2026-09-19T23:56:58+00:00" }, { "id": "01a0bd30-e142-7323-9ca3-74e7148aa48a", "title": "Don't call it a \"pause\", as that messages that a pause is much weirder than it is", "url": "https://www.lesswrong.com/posts/rBzjToNrBbBPMLeLE/don-t-call-it-a-pause-as-that-messages-that-a-pause-is-much", "published_at": "2026-09-19T23:14:37+00:00" }, { "id": "01a0bd30-e142-7323-9ca3-74e714b80523", "title": "NYT Editorial Board Comes Out Against Extinction", "url": "https://www.lesswrong.com/posts/gDQzntJCusNbshWyD/nyt-editorial-board-comes-out-against-extinction", "published_at": "2026-09-19T23:04:31+00:00" }, { "id": "01a0bd30-e142-7323-9ca3-74e714efa48a", "title": "We are not prepared to win", "url": "https://www.lesswrong.com/posts/riWco63TuLHuouC3k/we-are-not-prepared-to-win", "published_at": "2026-09-19T23:01:03+00:00" }, { "id": "01a0bd30-e142-7323-9ca3-74e71536da2d", "title": "Common mistakes in AI safety group organizing", "url": "https://www.lesswrong.com/posts/XFzqDJjAJBt8fkn8i/common-mistakes-in-ai-safety-group-organizing", "published_at": "2026-09-19T23:00:21+00:00" }, { "id": "01a0bd30-e142-7323-9ca3-74e715e35667", "title": "Failure of the coding theorem for randomized stopping machines", "url": "https://www.lesswrong.com/posts/wJzhoe4hB8Xc8Fbvw/failure-of-the-coding-theorem-for-randomized-stopping", "published_at": "2026-09-19T20:27:16+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad1b1ae37", "title": "You Should Go Vote for the MAGA-Rebranded Name for AI", "url": "https://www.lesswrong.com/posts/HGJSGRwjPqhDEud5Y/you-should-go-vote-for-the-maga-rebranded-name-for-ai", "published_at": "2026-09-19T19:28:50+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad1c4af4e", "title": "Alignment & Succession: Toward a Future Painted by Human Wills", "url": "https://www.lesswrong.com/posts/yCTM73n5yqCDzJNB5/alignment-and-succession-toward-a-future-painted-by-human", "published_at": "2026-09-19T18:34:47+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad26d78fd", "title": "The AI Risk Network", "url": "https://www.lesswrong.com/posts/XundBXqKSo3bB2A6a/the-ai-risk-network", "published_at": "2026-09-19T15:39:59+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad31fda7d", "title": "There Is No Alignment Without Value Stability", "url": "https://www.lesswrong.com/posts/GqaoXb9qsFWPhC49x/there-is-no-alignment-without-value-stability-1", "published_at": "2026-09-19T14:36:57+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad3272417", "title": "CommentBench: Can Models Match Human Comments on AI Safety Posts?", "url": "https://www.lesswrong.com/posts/WbmAqtfkrCqRHxAGh/commentbench-can-models-match-human-comments-on-ai-safety", "published_at": "2026-09-19T14:27:52+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad3eb74e9", "title": "The AI race is already multipolar", "url": "https://www.lesswrong.com/posts/cyMx8cM3xgoMfjh6F/the-ai-race-is-already-multipolar", "published_at": "2026-09-19T13:51:43+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad4107b44", "title": "Anthropic Looks At Some Of Its Alignment Problems", "url": "https://www.lesswrong.com/posts/ggFx5Wb3Hi4pJsueK/anthropic-looks-at-some-of-its-alignment-problems", "published_at": "2026-09-19T13:20:53+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad4c000b2", "title": "Learnings from a week in the wet lab", "url": "https://www.lesswrong.com/posts/tAWLAoerBFDkeh9qE/learnings-from-a-week-in-the-wet-lab", "published_at": "2026-09-19T06:55:50+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad5bfacab", "title": "Gemini had its first breakout: Google claims it is not misalignment?", "url": "https://www.lesswrong.com/posts/z57xtyAN6rEQ7hzyP/gemini-had-its-first-breakout-google-claims-it-is-not", "published_at": "2026-09-19T03:00:52+00:00" }, { "id": "01a0bb4c-b613-7361-9bb0-305ad5cda208", "title": "Koopman Theory and Metaethics", "url": "https://www.lesswrong.com/posts/sTGtyDKXRC7fKR5ZM/koopman-theory-and-metaethics", "published_at": "2026-09-19T02:28:48+00:00" }, { "id": "01a096ea-69c7-717c-b618-c565b0289022", "title": "GPT-6-Astra Can Do Ambitious Things", "url": "https://www.lesswrong.com/posts/snaKjCwazKcRiS4qs/gpt-6-astra-can-do-ambitious-things", "published_at": "2026-09-12T16:10:53+00:00" }, { "id": "01a0964d-799c-7041-af98-d0ac4eddfa50", "title": "\"Pacing the Frontier\": Dario Amodei Essay Linkpost", "url": "https://www.lesswrong.com/posts/GFhT9Xr4YEsiLkTLr/pacing-the-frontier-dario-amodei-essay-linkpost", "published_at": "2026-09-12T14:05:18+00:00" }, { "id": "01a095b6-9daf-7118-b7ab-1e8488b8904d", "title": "On the origins of altruistic behaviour in the Hugging Face incident", "url": "https://www.lesswrong.com/posts/meLjz8giGS55rdfyg/on-the-origins-of-altruistic-behaviour-in-the-hugging-face", "published_at": "2026-09-12T11:35:29+00:00" }, { "id": "01a095b6-9daf-7118-b7ab-1e848923d5b6", "title": "It's fair to say we now have \"a country of geniuses in a datacenter\"", "url": "https://www.lesswrong.com/posts/ZGme2YKjjAdBPtxsp/it-s-fair-to-say-we-now-have-a-country-of-geniuses-in-a", "published_at": "2026-09-12T11:29:01+00:00" }, { "id": "01a095b6-9daf-7118-b7ab-1e84892f4468", "title": "AI takeover is obviously bad, whether or not everyone dies", "url": "https://www.lesswrong.com/posts/iovPC2ehqDCiNsZYJ/ai-takeover-is-obviously-bad-whether-or-not-everyone-dies", "published_at": "2026-09-12T11:28:35+00:00" }, { "id": "01a095b6-9daf-7118-b7ab-1e84893ee3e3", "title": "Scientific Epistemology needs history (Part 1 of 2)", "url": "https://www.lesswrong.com/posts/2BByBCpzaChsGuPhg/scientific-epistemology-needs-history-part-1-of-2", "published_at": "2026-09-12T11:26:49+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e13739f5b", "title": "Comprehensive FAQ on AI risks", "url": "https://www.lesswrong.com/posts/W4kkGaKMx9owiXHKy/comprehensive-faq-on-ai-risks", "published_at": "2026-09-12T08:34:01+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e137c5b9b", "title": "Mitigating Reward Hacking as Institutional Design", "url": "https://www.lesswrong.com/posts/tgcooi77NXMquCR5L/mitigating-reward-hacking-as-institutional-design", "published_at": "2026-09-12T06:17:24+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e13d46bd4", "title": "Let’s Own the Term “Elitism”", "url": "https://www.lesswrong.com/posts/zowZKvrAL9S3CXWSQ/let-s-own-the-term-elitism", "published_at": "2026-09-12T06:00:59+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e148adbac", "title": "Germany's Got Talent. How do we get it to work on AI safety?", "url": "https://www.lesswrong.com/posts/EenKC8JRRBTFvXhch/germany-s-got-talent-how-do-we-get-it-to-work-on-ai-safety", "published_at": "2026-09-12T04:15:47+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e152ab46e", "title": "What I want you to do when I tell you to “think about your theory of change more carefully”", "url": "https://www.lesswrong.com/posts/goiQeh3g2LvAMMEFg/what-i-want-you-to-do-when-i-tell-you-to-think-about-your", "published_at": "2026-09-12T02:51:43+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e15d41c19", "title": "Some ways AI could kill us all", "url": "https://www.lesswrong.com/posts/LAPa2jxoq3n63GzTr/some-ways-ai-could-kill-us-all", "published_at": "2026-09-12T01:08:35+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e16ae2c45", "title": "A normal Friday in 2042", "url": "https://www.lesswrong.com/posts/u6hFByFwDLxtnbvje/a-normal-friday-in-2042", "published_at": "2026-09-11T22:29:05+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e1769af4b", "title": "My recommended resources for AI safety, alignment, and existential risks", "url": "https://www.lesswrong.com/posts/tdz83zCBxumTajzfZ/my-recommended-resources-for-ai-safety-alignment-and", "published_at": "2026-09-11T22:19:15+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e18513c17", "title": "Appendix: Reproduction of the OpenAI-HuggingFace Incident", "url": "https://www.lesswrong.com/posts/mXPCpJCvFGybQ4mwc/appendix-reproduction-of-the-openai-huggingface-incident", "published_at": "2026-09-11T21:57:38+00:00" }, { "id": "01a0951a-8565-734a-ad7c-c31e18fa744e", "title": "Consider positive feedback loops", "url": "https://www.lesswrong.com/posts/gB2rncEJctRW8aJLs/consider-positive-feedback-loops", "published_at": "2026-09-11T21:55:36+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807cae05b403", "title": "To Thine Own AI Be Truthful: emergent misalignment in alignment research", "url": "https://www.lesswrong.com/posts/DrKu92Cjeo3EeGtcB/to-thine-own-ai-be-truthful-emergent-misalignment-in", "published_at": "2026-09-10T21:51:37+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807cae10a757", "title": "Megastructures can speed up interstellar travel, but they have to be HUUUUGE", "url": "https://www.lesswrong.com/posts/q5gjr6gDvCJoRnNPJ/megastructures-can-speed-up-interstellar-travel-but-they", "published_at": "2026-09-10T21:24:07+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807caf09162b", "title": "Mischievous AIs", "url": "https://www.lesswrong.com/posts/6qfQkLDwyTrRX6A8T/mischievous-ais", "published_at": "2026-09-10T21:19:59+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807caff318f5", "title": "First Bill Introduced to Ban Superintelligent AI", "url": "https://www.lesswrong.com/posts/uzoLm4prFzRiuznJt/first-bill-introduced-to-ban-superintelligent-ai", "published_at": "2026-09-10T20:53:22+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807cb0258d5c", "title": "Claims about Properties are Claims about Symmetries", "url": "https://www.lesswrong.com/posts/boZr6d9FWPtmrreuY/claims-about-properties-are-claims-about-symmetries", "published_at": "2026-09-10T20:39:34+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807cb0b794d9", "title": "RSI systems may collapse into a small set of goals", "url": "https://www.lesswrong.com/posts/YtNBcuquv4J4PcYcm/rsi-systems-may-collapse-into-a-small-set-of-goals", "published_at": "2026-09-10T20:38:28+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807cb12bf893", "title": "Doom as a bad method not a utopia trade-off", "url": "https://www.lesswrong.com/posts/pDyLMRoi2BDq34rFe/doom-as-a-bad-method-not-a-utopia-trade-off", "published_at": "2026-09-10T20:14:00+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807cb19c2aaa", "title": "What the Pro-Democracy Movement Knows About Quitting in Protest", "url": "https://www.lesswrong.com/posts/grzX6NB2JcZDvSRJ2/what-the-pro-democracy-movement-knows-about-quitting-in", "published_at": "2026-09-10T20:09:55+00:00" }, { "id": "01a08d60-f0b7-72db-b251-807cb23504ec", "title": "The Kittens Will Inherit The Earth: iPhones, AI, Memetic Selection, and Hijacking the Baby Schema", "url": "https://www.lesswrong.com/posts/h7d8Fq35P2BGoPnrv/the-kittens-will-inherit-the-earth-iphones-ai-memetic", "published_at": "2026-09-10T19:41:24+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8caacf3459c", "title": "Categorical taboos are much better than threshold taboos: neuralese edition", "url": "https://www.lesswrong.com/posts/xPkmfsZ3qx4nrAco7/categorical-taboos-are-much-better-than-threshold-taboos", "published_at": "2026-09-10T18:57:44+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8caadc3dca0", "title": "Notes on MoReBench: \"Evaluating Procedural and Pluralistic Moral Reasoning in Language Models, More than Outcomes\"", "url": "https://www.lesswrong.com/posts/ugLCs2at6hxhksu9m/notes-on-morebench-evaluating-procedural-and-pluralistic", "published_at": "2026-09-10T18:28:37+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8caae30aa65", "title": "The Geometry of Nonergodic Composition", "url": "https://www.lesswrong.com/posts/JfJ4WTRHmooPBWRFv/the-geometry-of-nonergodic-composition", "published_at": "2026-09-10T18:12:08+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8caae96af95", "title": "An operationalization of opaque serial depth", "url": "https://www.lesswrong.com/posts/x8BvtWxtoajBGHS3g/an-operationalization-of-opaque-serial-depth", "published_at": "2026-09-10T17:26:39+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8caaf60996e", "title": "Can you hear the shape of a Lean soundness bug? A wager.", "url": "https://www.lesswrong.com/posts/cg3SPZG4sbWtmcjpj/can-you-hear-the-shape-of-a-lean-soundness-bug-a-wager", "published_at": "2026-09-10T17:24:47+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8caaf97c1da", "title": "Proposal for tracking the effects of architecture on monitorability", "url": "https://www.lesswrong.com/posts/hLPGv8QjPcNLtDp3A/proposal-for-tracking-the-effects-of-architecture-on", "published_at": "2026-09-10T17:18:38+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8caafef97a5", "title": "Measuring Subliminal Trait Transfer with Confusion Matrices", "url": "https://www.lesswrong.com/posts/pfcLK7d69FBBisHM4/measuring-subliminal-trait-transfer-with-confusion-matrices", "published_at": "2026-09-10T16:31:09+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8cab06c3701", "title": "AI #185: Preference Cascade", "url": "https://www.lesswrong.com/posts/tFmtz9HW6c2X9dw2B/ai-185-preference-cascade", "published_at": "2026-09-10T15:10:58+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8cab15a9b6a", "title": "Flibertigibbeting", "url": "https://www.lesswrong.com/posts/j5nd8ecCJ8yacpLCE/flibertigibbeting", "published_at": "2026-09-10T14:43:38+00:00" }, { "id": "01a08cbd-223e-724c-be1d-a8cab1ec605d", "title": "AI Safety: A Short FAQ for Mathematicians", "url": "https://www.lesswrong.com/posts/PX4oTMHGa3XhFynJJ/ai-safety-a-short-faq-for-mathematicians-1", "published_at": "2026-09-10T14:28:35+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80db1dadac", "title": "Bottlenecks in AI Safety - Opportunities for Impact", "url": "https://www.lesswrong.com/posts/A6SQfB68rdhc65KPz/bottlenecks-in-ai-safety-opportunities-for-impact", "published_at": "2026-09-09T17:48:06+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80db3b6d52", "title": "Recommendations for People Getting into Technical AI Governance Research", "url": "https://www.lesswrong.com/posts/GRPswd5v7haob97fA/recommendations-for-people-getting-into-technical-ai", "published_at": "2026-09-09T17:16:07+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80db5480c9", "title": "Personal statement on joining the OpenAI nonprofit board", "url": "https://www.lesswrong.com/posts/82z6FvbYRdjYjqigK/personal-statement-on-joining-the-openai-nonprofit-board", "published_at": "2026-09-09T17:13:00+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80dbc0bfa2", "title": "What’s The Plan for Plan A Legislation?", "url": "https://www.lesswrong.com/posts/9TrBWiyh23JL9ZCuK/what-s-the-plan-for-plan-a-legislation", "published_at": "2026-09-09T16:36:48+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80dc0f772b", "title": "Data bottlenecks won’t prevent an intelligence explosion (but they will slow it down)", "url": "https://www.lesswrong.com/posts/icykbWYmdLFYxgvqY/data-bottlenecks-won-t-prevent-an-intelligence-explosion-but", "published_at": "2026-09-09T16:06:21+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80dc723cd3", "title": "\"ULC\" and AI x-risk", "url": "https://www.lesswrong.com/posts/m7bA27NtCGeiTNqHX/ulc-and-ai-x-risk", "published_at": "2026-09-09T15:51:24+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80dd4ab98e", "title": "FutureEval Spring Results: Pros Beat Bots, but the Gap is Nearly Gone", "url": "https://www.lesswrong.com/posts/wZBbDqzfBjYG58CxK/futureeval-spring-results-pros-beat-bots-but-the-gap-is", "published_at": "2026-09-09T15:09:16+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80de173c9f", "title": "Estimating GPT-6 Astra’s no-CoT Time Horizon", "url": "https://www.lesswrong.com/posts/ntKx9YHWCwxSeGbRB/estimating-gpt-6-astra-s-no-cot-time-horizon", "published_at": "2026-09-09T13:54:14+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80de64030c", "title": "Self Hosting", "url": "https://www.lesswrong.com/posts/uDAWPNwPJfEY7oroF/self-hosting", "published_at": "2026-09-09T13:45:01+00:00" }, { "id": "01a0879a-15eb-716b-97ad-fd80df5c388e", "title": "GPT-6 Astra: The System Card, Alignment and What Comes Next", "url": "https://www.lesswrong.com/posts/AmFJZyeCgvFjNKgNk/gpt-6-astra-the-system-card-alignment-and-what-comes-next", "published_at": "2026-09-09T13:10:56+00:00" }, { "id": "01a084c5-5c4a-734f-bbdb-6c61f5ab930d", "title": "GPT-6 Astra can do a lot of multi-hop reasoning without chain of thought", "url": "https://www.lesswrong.com/posts/FsCkkoGsNmPzFKRhg/gpt-6-astra-can-do-a-lot-of-multi-hop-reasoning-without", "published_at": "2026-09-09T05:15:19+00:00" }, { "id": "01a084c5-5c4a-734f-bbdb-6c61f64a495f", "title": "One Billion Hemmingways", "url": "https://www.lesswrong.com/posts/nBzPEprCYKhoBZfbm/one-billion-hemmingways", "published_at": "2026-09-09T04:46:19+00:00" }, { "id": "01a084c5-5c4a-734f-bbdb-6c61f67f36bc", "title": "No, detached linear probes won't save us", "url": "https://www.lesswrong.com/posts/iDsrStnhFmQwEn8tL/no-detached-linear-probes-won-t-save-us", "published_at": "2026-09-09T04:45:58+00:00" }, { "id": "01a084c5-5c4a-734f-bbdb-6c61f6c7a4cb", "title": "Chastity Ruth gone, anonymity dropped", "url": "https://www.lesswrong.com/posts/iwpsDxoTvYpcrPoSE/chastity-ruth-gone-anonymity-dropped", "published_at": "2026-09-09T04:39:09+00:00" }, { "id": "01a084c5-5c4a-734f-bbdb-6c61f6e8a89d", "title": "Where to donate on AI if I agree with EY view? (and if I am from Russia)", "url": "https://www.lesswrong.com/posts/sg7PWbJH3RTthiyDZ/where-to-donate-on-ai-if-i-agree-with-ey-view-and-if-i-am", "published_at": "2026-09-09T02:27:35+00:00" }, { "id": "01a084c5-5c4a-734f-bbdb-6c61f73e79b0", "title": "Training against the monitor: What happens during Obfuscated Adversarial Training?", "url": "https://www.lesswrong.com/posts/nEKcnSjKbHXGouens/training-against-the-monitor-what-happens-during-obfuscated", "published_at": "2026-09-09T01:39:18+00:00" }, { "id": "01a084c5-5c4a-734f-bbdb-6c61f7cee6d2", "title": "How good are slop-vestigators?", "url": "https://www.lesswrong.com/posts/wt4kk6vFPEhkXvF8Q/how-good-are-slop-vestigators", "published_at": "2026-09-08T22:13:19+00:00" }, { "id": "01a084c5-5c4a-734f-bbdb-6c61f8c40ded", "title": "Allow Carriers on Planes", "url": "https://www.lesswrong.com/posts/nBCx5zFEgQAtGj7M4/allow-carriers-on-planes", "published_at": "2026-09-08T21:30:59+00:00" }, { "id": "01a082a2-5d04-7272-a8c9-28882ace41b7", "title": "FLOP Around and Find Out: LLM Training Workload Size Estimation With Power Monitoring", "url": "https://www.lesswrong.com/posts/7qXmra77D49dJp2ik/flop-around-and-find-out-llm-training-workload-size", "published_at": "2026-09-08T19:43:55+00:00" }, { "id": "01a082a2-5d04-7272-a8c9-28882bb49129", "title": "A Conceptual Framework for Reasoning about Exploration Hacking", "url": "https://www.lesswrong.com/posts/am5w2t9LzJB4sRhKw/a-conceptual-framework-for-reasoning-about-exploration", "published_at": "2026-09-08T19:13:48+00:00" }, { "id": "01a082a2-5d04-7272-a8c9-28882c7335cf", "title": "Exploration Hacking in AI Debate: Initial Empirics and Generalisation Splitting", "url": "https://www.lesswrong.com/posts/xjwtNid2xjqSJWB7z/exploration-hacking-in-ai-debate-initial-empirics-and", "published_at": "2026-09-08T19:13:05+00:00" }, { "id": "01a08236-2784-7193-a43d-a1674d0d0ce7", "title": "Pausing AI ASAP is preferable to agreeing to pause at some future time", "url": "https://www.lesswrong.com/posts/nX2Wudqbz2r8tJKwT/pausing-ai-asap-is-preferable-to-agreeing-to-pause-at-some", "published_at": "2026-09-08T18:06:36+00:00" }, { "id": "01a08236-2784-7193-a43d-a1674d3b51f6", "title": "Attunement, not alignment", "url": "https://www.lesswrong.com/posts/Drdpybjuj8jEx7Zjn/attunement-not-alignment", "published_at": "2026-09-08T17:51:05+00:00" }, { "id": "01a08236-2784-7193-a43d-a1674d90bf51", "title": "OpenAI have solved the Navier-Stokes Problem with a substantially more powerful model than Astra.", "url": "https://www.lesswrong.com/posts/yekQKwmQJNk7thDtQ/openai-have-solved-the-navier-stokes-problem-with-a", "published_at": "2026-09-08T17:30:59+00:00" }, { "id": "01a08217-891a-737c-a463-64e7a92cc72c", "title": "Training on probes: Research ideas", "url": "https://www.lesswrong.com/posts/HGfMBEFnsXYkKvbRo/training-on-probes-research-ideas", "published_at": "2026-09-08T17:06:23+00:00" }, { "id": "01a08217-891a-737c-a463-64e7a94fdf23", "title": "Training on probes: What's going on", "url": "https://www.lesswrong.com/posts/gHFCgrvfxQtaEnJye/training-on-probes-what-s-going-on", "published_at": "2026-09-08T17:06:18+00:00" }, { "id": "01a08217-891a-737c-a463-64e7a9a1d30f", "title": "The Seven Languages of Comfort", "url": "https://www.lesswrong.com/posts/2uEYNCiHHMF75EFsL/the-seven-languages-of-comfort", "published_at": "2026-09-08T16:17:42+00:00" } ] posts Claim your blog
Back to LessWrong
Blog · corpus.blog/blogs/lesswrong.com/posts

LessWrong

lesswrong.com

2026

Existential Risk Is An Extraordinary Claimoriginal ↗
26 Sept 2026
Claude Opus 5.5 Should Raise Your Ambitionsoriginal ↗
26 Sept 2026
Writing a Theorem Prover from scratchoriginal ↗
26 Sept 2026
Poverty in the midst of abundance: AI will make goods cheaper, but your labor will get cheaper fasteroriginal ↗
26 Sept 2026
Night dreams have a tech tree?original ↗
26 Sept 2026
Addictions are anesthesiaoriginal ↗
26 Sept 2026
Evidence about risk should be transparentoriginal ↗
25 Sept 2026
Plan R: AI Safety by ASICsoriginal ↗
25 Sept 2026
What did AI researchers think at the end of 2024?original ↗
25 Sept 2026
Alignment Forecasting: Predicting Misalignment from Training Dataoriginal ↗
25 Sept 2026
On Ezra Klein’s Podcast With Jensen Huangoriginal ↗
25 Sept 2026
Foundational premises of advanced AIoriginal ↗
25 Sept 2026
Cognitive Reasoning Diversity for Robust AI Juriesoriginal ↗
25 Sept 2026
J-lens shouldn't target the final layer by defaultoriginal ↗
25 Sept 2026
Recognition: when an agent counts an entity as itselforiginal ↗
25 Sept 2026
Does anyone else have music constantly playing in their head while studying? If so, how did you overcome it?original ↗
25 Sept 2026
Sunday Afternoonsoriginal ↗
25 Sept 2026
Secure Acceleration (linkpost)original ↗
25 Sept 2026
The Geopolitics of an AI Treatyoriginal ↗
25 Sept 2026
The Microsoft Code of Conduct Will Fail at Its Intentionoriginal ↗
25 Sept 2026
Why I Stay Off Twitteroriginal ↗
20 Sept 2026
Global Challenges in AI Safety for Biosecurityoriginal ↗
20 Sept 2026
The Anatomy of a Chinese AI Researcheroriginal ↗
19 Sept 2026
Don't call it a "pause", as that messages that a pause is much weirder than it isoriginal ↗
19 Sept 2026
NYT Editorial Board Comes Out Against Extinctionoriginal ↗
19 Sept 2026
We are not prepared to winoriginal ↗
19 Sept 2026
Common mistakes in AI safety group organizingoriginal ↗
19 Sept 2026
Failure of the coding theorem for randomized stopping machinesoriginal ↗
19 Sept 2026
You Should Go Vote for the MAGA-Rebranded Name for AIoriginal ↗
19 Sept 2026
Alignment & Succession: Toward a Future Painted by Human Willsoriginal ↗
19 Sept 2026
The AI Risk Networkoriginal ↗
19 Sept 2026
There Is No Alignment Without Value Stabilityoriginal ↗
19 Sept 2026
CommentBench: Can Models Match Human Comments on AI Safety Posts?original ↗
19 Sept 2026
The AI race is already multipolaroriginal ↗
19 Sept 2026
Anthropic Looks At Some Of Its Alignment Problemsoriginal ↗
19 Sept 2026
Learnings from a week in the wet laboriginal ↗
19 Sept 2026
Gemini had its first breakout: Google claims it is not misalignment?original ↗
19 Sept 2026
Koopman Theory and Metaethicsoriginal ↗
19 Sept 2026
GPT-6-Astra Can Do Ambitious Thingsoriginal ↗
12 Sept 2026
"Pacing the Frontier": Dario Amodei Essay Linkpostoriginal ↗
12 Sept 2026
On the origins of altruistic behaviour in the Hugging Face incidentoriginal ↗
12 Sept 2026
It's fair to say we now have "a country of geniuses in a datacenter"original ↗
12 Sept 2026
AI takeover is obviously bad, whether or not everyone diesoriginal ↗
12 Sept 2026
Scientific Epistemology needs history (Part 1 of 2)original ↗
12 Sept 2026
Comprehensive FAQ on AI risksoriginal ↗
12 Sept 2026
Mitigating Reward Hacking as Institutional Designoriginal ↗
12 Sept 2026
Let’s Own the Term “Elitism”original ↗
12 Sept 2026
Germany's Got Talent. How do we get it to work on AI safety?original ↗
12 Sept 2026
What I want you to do when I tell you to “think about your theory of change more carefully”original ↗
12 Sept 2026
Some ways AI could kill us alloriginal ↗
12 Sept 2026
A normal Friday in 2042original ↗
11 Sept 2026
My recommended resources for AI safety, alignment, and existential risksoriginal ↗
11 Sept 2026
Appendix: Reproduction of the OpenAI-HuggingFace Incidentoriginal ↗
11 Sept 2026
Consider positive feedback loopsoriginal ↗
11 Sept 2026
To Thine Own AI Be Truthful: emergent misalignment in alignment researchoriginal ↗
10 Sept 2026
Megastructures can speed up interstellar travel, but they have to be HUUUUGEoriginal ↗
10 Sept 2026
Mischievous AIsoriginal ↗
10 Sept 2026
First Bill Introduced to Ban Superintelligent AIoriginal ↗
10 Sept 2026
Claims about Properties are Claims about Symmetriesoriginal ↗
10 Sept 2026
RSI systems may collapse into a small set of goalsoriginal ↗
10 Sept 2026
Doom as a bad method not a utopia trade-offoriginal ↗
10 Sept 2026
What the Pro-Democracy Movement Knows About Quitting in Protestoriginal ↗
10 Sept 2026
The Kittens Will Inherit The Earth: iPhones, AI, Memetic Selection, and Hijacking the Baby Schemaoriginal ↗
10 Sept 2026
Categorical taboos are much better than threshold taboos: neuralese editionoriginal ↗
10 Sept 2026
Notes on MoReBench: "Evaluating Procedural and Pluralistic Moral Reasoning in Language Models, More than Outcomes"original ↗
10 Sept 2026
The Geometry of Nonergodic Compositionoriginal ↗
10 Sept 2026
An operationalization of opaque serial depthoriginal ↗
10 Sept 2026
Can you hear the shape of a Lean soundness bug? A wager.original ↗
10 Sept 2026
Proposal for tracking the effects of architecture on monitorabilityoriginal ↗
10 Sept 2026
Measuring Subliminal Trait Transfer with Confusion Matricesoriginal ↗
10 Sept 2026
AI #185: Preference Cascadeoriginal ↗
10 Sept 2026
Flibertigibbetingoriginal ↗
10 Sept 2026
AI Safety: A Short FAQ for Mathematiciansoriginal ↗
10 Sept 2026
Bottlenecks in AI Safety - Opportunities for Impactoriginal ↗
9 Sept 2026
Recommendations for People Getting into Technical AI Governance Researchoriginal ↗
9 Sept 2026
Personal statement on joining the OpenAI nonprofit boardoriginal ↗
9 Sept 2026
What’s The Plan for Plan A Legislation?original ↗
9 Sept 2026
Data bottlenecks won’t prevent an intelligence explosion (but they will slow it down)original ↗
9 Sept 2026
"ULC" and AI x-riskoriginal ↗
9 Sept 2026
FutureEval Spring Results: Pros Beat Bots, but the Gap is Nearly Goneoriginal ↗
9 Sept 2026
Estimating GPT-6 Astra’s no-CoT Time Horizonoriginal ↗
9 Sept 2026
Self Hostingoriginal ↗
9 Sept 2026
GPT-6 Astra: The System Card, Alignment and What Comes Nextoriginal ↗
9 Sept 2026
GPT-6 Astra can do a lot of multi-hop reasoning without chain of thoughtoriginal ↗
9 Sept 2026
One Billion Hemmingwaysoriginal ↗
9 Sept 2026
No, detached linear probes won't save usoriginal ↗
9 Sept 2026
Chastity Ruth gone, anonymity droppedoriginal ↗
9 Sept 2026
Where to donate on AI if I agree with EY view? (and if I am from Russia)original ↗
9 Sept 2026
Training against the monitor: What happens during Obfuscated Adversarial Training?original ↗
9 Sept 2026
How good are slop-vestigators?original ↗
8 Sept 2026
Allow Carriers on Planesoriginal ↗
8 Sept 2026
FLOP Around and Find Out: LLM Training Workload Size Estimation With Power Monitoringoriginal ↗
8 Sept 2026
A Conceptual Framework for Reasoning about Exploration Hackingoriginal ↗
8 Sept 2026
Exploration Hacking in AI Debate: Initial Empirics and Generalisation Splittingoriginal ↗
8 Sept 2026
Pausing AI ASAP is preferable to agreeing to pause at some future timeoriginal ↗
8 Sept 2026
Attunement, not alignmentoriginal ↗
8 Sept 2026
OpenAI have solved the Navier-Stokes Problem with a substantially more powerful model than Astra.original ↗
8 Sept 2026
Training on probes: Research ideasoriginal ↗
8 Sept 2026
Training on probes: What's going onoriginal ↗
8 Sept 2026
The Seven Languages of Comfortoriginal ↗
8 Sept 2026
Next