LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

  1. 46M AGO

    “Understanding when and why agents scheme” by Mia Hopman, Jannes Elstner, Maria Avramidou, Amritanshu Prasad, David Lindner

    TL;DR To understanding the conditions under which LLM agents engage in scheming behavior, we develop a framework that decomposes the decision to scheme into agent factors (model, system prompt, tool access) and environmental factors (stakes, oversight, outcome influence) We systematically vary these factors in four realistic settings, each with scheming opportunities for agents that pursue instrumentally convergent goals such as self-preservation, resource acquisition, and goal-guarding We find baseline scheming propensity to be near-zero, with Gemini 3 Pro Preview as the sole exception at 3% in one setting Adversarially-designed prompt snippets that encourage agency and goal-directedness can induce high scheming rates, but snippets sourced from production agent scaffolds rarely do Scheming behavior is remarkably brittle: removing a single tool can drop the scheming rate[1] from 59% to 7%, and increasing oversight can sometimes raise rather than deter scheming by up to 25% We consider the behaviors studied here as potential precursors to the full threat model of scheming: current agents may sometimes behave consistent with scheming, but do not (yet) have the coherent long-term goals and the general capability that would make their deployment dangerous. Understanding when and why agents currently scheme could help us prepare for such future [...] --- Outline: (00:13) TL;DR (02:43) Baseline scheming is near zero (03:25) Encouraging Agency and Goal-Directedness (05:17) Scaffolding changes can have weird effects (06:16) Environmental incentives affect scheming unexpectedly The original text contained 2 footnotes which were omitted from this narration. --- First published: March 21st, 2026 Source: https://www.lesswrong.com/posts/amYmcwCuyuCEZcrRm/understanding-when-and-why-agents-scheme --- Narrated by TYPE III AUDIO. --- Images from the article: Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

    9 min

  2. 9H AGO

    “Attend the 2026 Reproductive Frontiers Summit, June 16–18, Berkeley” by TsviBT, K Richards

    We’ll be hosting the 2026 Reproductive Frontiers Summit at Lighthaven in Berkeley, CA, on June 16—18. Come join us if you want to learn, connect, think, and coordinate about the future of germline engineering technology. Very early bird tickets are available now until the end of March. Who will be there? Our lineup of speakers includes experts in the areas of polygenic prediction, embryo gene editing, in vitro gametogenesis, artificial wombs, ethics and regulation for advanced reproductive technology, and more. See the full list on the summit website: reproductivefrontiers.org. We hope to welcome attendees who are: scientists (new or established) who are interested in advanced reproductive technology or reprogenetics, especially experts or future experts in: stem cell biology, embryology, epigenetics of the germ line, bioinformatics, polygenic prediction of traits, editing methods (especially epigenetic editing and precision gene editing), ovarian culture, gametogenesis, chromosome dynamics and engineering, low-input *omics, single-cell microfluidics, and related topics; experts on regulation and policy, financing, and public opinion around advanced reprotech; bioethicists who want to use constructive critique to craft a practicable vision of widely beneficial germline engineering technology; undergrads, grad students, and postdocs [...] --- Outline: (00:34) Who will be there? (02:14) Last year (02:54) What this is for (09:22) How you can help --- First published: March 22nd, 2026 Source: https://www.lesswrong.com/posts/JLYHKEFF8jBhy9tvs/attend-the-2026-reproductive-frontiers-summit-june-16-18 --- Narrated by TYPE III AUDIO.

    10 min

  3. 20H AGO

    “Is fever a symptom of glycine deficiency?” by Benquo

    A 2022 LessWrong post on orexin and the quest for more waking hours argues that orexin agonists could safely reduce human sleep needs, pointing to short-sleeper gene mutations that increase orexin production and to cavefish that evolved heightened orexin sensitivity alongside an 80% reduction in sleep. Several commenters discussed clinical trials, embryo selection, and the evolutionary puzzle of why short-sleeper genes haven't spread. I thought the whole approach was backwards, and left a comment: Orexin is a signal about energy metabolism. Unless the signaling system itself is broken (e.g. narcolepsy type 1, caused by autoimmune destruction of orexin-producing neurons), it's better to fix the underlying reality the signals point to than to falsify the signals. My sleep got noticeably more efficient when I started supplementing glycine. Most people on modern diets don't get enough; we can make ~3g/day but can use 10g+, because in the ancestral environment we ate much more connective tissue or broth therefrom. Glycine is both important for repair processes and triggers NMDA receptors to drop core temperature, which smooths the path to sleep. While drafting that, I went back to Chris Masterjohn's page on glycine requirements. His estimate for total need [...] --- Outline: (01:49) Glycine helps us sleep by cooling the body (02:26) Glycine cleans our mitochondria as we sleep (04:12) Most people could use more glycine (05:28) Fever is plan B for fighting infection; glycine supports plan A (09:28) Glycines cooling effect via the SCN is unrelated to its immune benefits (10:35) Glycine turns out to be a legitimate antipyretic after all (11:51) Practical considerations --- First published: March 22nd, 2026 Source: https://www.lesswrong.com/posts/87XoatpFkdmCZpvQK/is-fever-a-symptom-of-glycine-deficiency --- Narrated by TYPE III AUDIO.

    14 min

  4. 21H AGO

    “My Most Costly Delusion” by Ihor Kendiukhov

    Suppose there is a fire in a nearby house. Suppose there are competent firefighters in your town: fast, professional, well-equipped. They are expected to arrive in 2–3 minutes. In that situation, unless something very extraordinary happens, it would indeed be an act of great arrogance and even utter insanity to go into the fire yourself in the hope of "rescuing" someone or something. The most likely outcome would be that you would find yourself among those who need to be rescued. But the calculus changes drastically if the closest fire crew is 3 hours away and consists of drunk, unfit amateurs. Or consider a child living in a big, happy, smart family. Imagine this child suddenly decides that his family may run out of money to the point where they won't have enough to eat. All reassurances from his parents don't work. The child doesn't believe in his parents' ability to reason, he makes his own calculations, and he strongly believes he is right and they are wrong. He is dead set on fixing the situation by doing day trading. What is that if not going nuts? Would those be wrong who ridicule this child and his complete mischaracterization [...] --- First published: March 22nd, 2026 Source: https://www.lesswrong.com/posts/EAH6Y6y3CDi3uxMou/my-most-costly-delusion --- Narrated by TYPE III AUDIO.

    6 min

  5. 23H AGO

    “China declares AGI development to be a part of 5-year plan” by Darmani

    The CCP writes in its 15th 5-year plan that it will. Encourage innovation in multimodal, agentic, embodied, and swarm intelligence technologies, and explore development paths for general artificial intelligence. This is translated from the original: 鼓励多模态、智能体、具身智能、群体智能等技术创新，探索通用人工智能发展路径。 Source: https://www.spp.gov.cn/spp/tt/202603/t20260313_723954.shtml The English-language commentary I found does not have much more to say about this, e.g.: https://triviumchina.com/2026/03/06/15th-five-year-plan-puts-ai-at-center-of-digital-economy-agenda/ Given that they gave less than half a sentence in a 140-page document to the most important invention in the history of mankind, it seems likely the authors don't really understand what this means. Concerning nonetheless --- First published: March 21st, 2026 Source: https://www.lesswrong.com/posts/JKhoyebpedLQLb5JN/china-declares-agi-development-to-be-a-part-of-5-year-plan --- Narrated by TYPE III AUDIO.

    1 min

  6. 1D AGO

    “Pre-Review of Toy Story 5” by Raemon

    I am the second most spoiler-averse person I know. (Maybe tied for 2nd with a couple other people?). I once was considering going to an immersive experience, and someone told me the company that ran the experience, and this was enough for me to derive an important twist that'd happen to me in the first few minutes, and I was like "augh that was a spoiler!!!" and they were like "!??". I then went to the experience, and indeed, it was a lot worse than it would have been if I had gotten to be delighted by the opening twist. I say this all to say, I think Toy Story 5 would be the kind of movie that, if it were good, it would be worth watching unspoiled. I am worried it will not be good, but, I don't know. But, also, I've been spoiled already, and meanwhile it's pretty interesting to think about in advance. So, decide whether you're the sorta person who should stop reading after this opening section. Also, if you have not seen Toy Story 3, Toy Story 3 is particularly worth watching unspoiled. The rest of the essay will get escalatingly spoilery [...] --- Outline: (06:51) Requirements of a Toy Story Movie (07:56) Ending A: Put it away (08:28) Ending B: Parents take it away (08:46) Ending C: Butlerian Jihad (09:05) Ending D: Harmony and Balance (10:19) Ending E: Accepting the End (12:13) Ending F: Hard Science-Fantasy? --- First published: March 21st, 2026 Source: https://www.lesswrong.com/posts/ra289RT2LJ9GfdpDQ/pre-review-of-toy-story-5 --- Narrated by TYPE III AUDIO.

    14 min

  7. 1D AGO

    “Does Hebrew Have Verbs?” by Benquo

    Spinoza's Compendium of Hebrew Grammar (1677, posthumous, unfinished) contains a claim that scholars have been misreading for centuries. He says that all Hebrew words, except a few particles, are nouns. The standard scholarly reaction is that this is either a metaphysical imposition (projecting his monistic ontology onto grammar) or a terminological trick (defining "noun" so broadly it's vacuous). Both reactions are wrong, and they're wrong for the same reason: they import Greek and Latin grammatical categories and then treat those categories as the neutral baseline. What Spinoza actually said From Chapter 5 of the Compendium (Bloom translation, 1962): "By a noun I understand a word by which we signify or indicate something that is understood. However, among things that are understood there can be either things and attributes of things, modes and relationships, or actions, and modes and relationships of actions." And: "For all Hebrew words, except for a few interjections and conjunctions and one or two particles, have the force and properties of nouns. Because the grammarians did not understand this they considered many words to be irregular which according to the usage of language are most regular." The word "noun" here [...] --- Outline: (00:45) What Spinoza actually said (01:57) The objection that doesnt work (02:30) How Hebrew actually works (04:15) Where the noun/verb distinction comes from (06:15) The structural difference between Greek and Hebrew (08:00) Spinoza wasnt projecting his metaphysics The original text contained 6 footnotes which were omitted from this narration. --- First published: March 19th, 2026 Source: https://www.lesswrong.com/posts/e8XnJFw9noBXoXFWr/does-hebrew-have-verbs --- Narrated by TYPE III AUDIO.

    10 min

  8. 1D AGO

    “Untrusted Monitoring is Default; Trusted Monitoring is not” by J Bostock

    These views are my own and not necessarily representative of those of any colleagues with whom I have worked on AI control. TL;DR: It's much cheaper and quicker to just throw some honeypots at your monitor models than to robustly prove trustedness for every model you want to use. Therefore I think the most likely future involves untrusted monitoring with some monitor validation as a default path. This post talks about two different ways of monitoring AI systems, trusted monitoring, and untrusted monitoring. You can find papers distinguishing between them here, here, and here, a video here, or you can look at this graphic which didn't make it into Gardner-Challis et. al. (2026): Context From when I started working at untrusted monitoring until very recently[1], I assumed that trusted monitoring was the default policy that AI companies would use, and that untrusted monitoring was an exotic, high-effort policy which would be much less likely to be used. Now I think it's the other way round. Full Trustedness is Hard The distinction between an untrusted model and a trusted one exists in the map. A trusted model is normally described as one for which we have strong evidence that [...] --- Outline: (01:04) Context (01:25) Full Trustedness is Hard (04:05) Untrusted Monitoring is Just Limited-Scope Trusted Monitoring (05:22) Cost and Speed (06:14) Monitoring is Here! The original text contained 1 footnote which was omitted from this narration. --- First published: March 20th, 2026 Source: https://www.lesswrong.com/posts/b5oHr5TrfQzCBZdXW/untrusted-monitoring-is-default-trusted-monitoring-is-not --- Narrated by TYPE III AUDIO. --- Images from the article: Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

    8 min
See All (250)

About

Audio narrations of LessWrong posts.

Information

  • Creator
    LessWrong
  • Years Active
    2023 - 2026
  • Episodes
    250
  • Rating
    Clean
  • Copyright
    © 2026 All rights reserved
  • Show Website
    LessWrong (30+ Karma)

You Might Also Like