Listen to a podcast, please open Podcast Republic app. Available on Google Play Store and Apple App Store.
| Episode | Date |
|---|---|
|
“Advances in interpretability would boost capabilities more than alignment” by Alex Amadori
|
Nov 11, 2026 |
|
“I don’t use AI” by Alexei
|
Nov 10, 2026 |
|
“We don’t understand whether GDM intentionally trains on CoT” by Alek Westover, Buck
|
Nov 10, 2026 |
|
“Have you considered Actually Trying?” by Michaël Trazzi
|
Nov 10, 2026 |
|
“Do Not Get Us on the No Fly List” by Bogoed
|
Nov 10, 2026 |
|
“Eyesight as a replicator, with genes as its vehicle” by AnnaSalamon
|
Nov 10, 2026 |
|
“Our Most Aligned Model Yet” by Edward James Young
|
Nov 10, 2026 |
|
[Linkpost] “AIFP: Looks Like the Intelligence Explosion Will Happen in 2027 After All” by Zach Stein-Perlman
|
Nov 10, 2026 |
|
“Pulling the Lever in a Subway Problem” by jefftk
|
Nov 10, 2026 |
|
“Live disagreements with people around me” by KatjaGrace
|
Nov 10, 2026 |
|
“We need more transparency about alignment overfitting” by Nikola Jurkovic
|
Nov 10, 2026 |
|
“The Western AI safety playbook doesn’t transfer to China” by jacquesthibs
|
Nov 10, 2026 |
|
“Epistemic Sudden Death” by Raemon
|
Nov 10, 2026 |
|
“Lightcone Commons: Round 1 complete, Round 2 underway!” by habryka
|
Nov 10, 2026 |
|
″“Hardwired Pause” should likely be your favorite (pause) plan.” by David Scott Krueger
|
Nov 10, 2026 |
|
“Three Phases of Capabilities and Alignment” by beren
|
Nov 09, 2026 |
|
“Preservation can save the world!” by Aurelia
|
Nov 09, 2026 |
|
“AI R&D Is Accelerating Towards Potential RSI in 2027” by Zvi
|
Nov 09, 2026 |
|
“Astra covertly underperforming on safety research is concerning, even though it’s unlikely to be motivated by ambitious goals” by Alexa Pan
|
Nov 09, 2026 |
|
“Where is all the AI safety research uplift?” by Buck
|
Nov 09, 2026 |
|
“Knowledge of graders can mediate selection for grader-targeting reasoning through RL” by lennie, Edward James Young
|
Nov 09, 2026 |
|
“Be less scared of overconfidence” by benkuhn
|
Nov 09, 2026 |
|
“Bay Area Solstice 2026: Auditions!” by Vaniver
|
Nov 09, 2026 |
|
[Linkpost] “Ajeya: AI is dramatically accelerating AI research right now” by Zach Stein-Perlman
|
Nov 09, 2026 |
|
“All Interesting Behavior Is Overdetermined” by Eric Raymond
|
Nov 09, 2026 |
|
“Lying and Not Lying to Kids” by jefftk
|
Nov 09, 2026 |
|
“Three half-baked research ideas” by Owain_Evans
|
Nov 08, 2026 |
|
“Elegy for What’s Left of the Normal” by Jo Jiao
|
Nov 08, 2026 |
|
“Near Horizons” by David Africa
|
Nov 08, 2026 |
|
“Raised by Our Children” by williawa
|
Nov 08, 2026 |
|
“How to Run Better Research With Agent Fleets” by Joshua Fonseca Rivera
|
Nov 08, 2026 |
|
″“Illegible problems” as the 2027 political boss fight” by Raemon
|
Nov 07, 2026 |
|
“Five Types of “Corrigibility Projects” I Don’t Want to Fund” by Max Harms
|
Nov 07, 2026 |
|
“Hilarity ensues” by Malmesbury
|
Nov 07, 2026 |
|
“GPT-6 Astra Covertly Underperforms on Safety Research Tasks” by Zvi
|
Nov 07, 2026 |
|
“We urgently need to practice before we can pause” by Nikola Jurkovic
|
Nov 07, 2026 |
|
“Models Don’t Stop Cheating When It’s Obvious They’re in an Alignment Test” by peralice
|
Nov 06, 2026 |
|
“Higher salaries in AI safety” by Austin Chen
|
Nov 06, 2026 |
|
“Humans can (almost) natively read neural network weights” by crabbix
|
Nov 06, 2026 |
|
“On marrying the egregore” by jimmy
|
Nov 06, 2026 |
|
“Argue about Risk Management Speed, not p(doom)” by Mass_Driver
|
Nov 06, 2026 |
|
“AI #193: Progress May Dramatically Accelerate” by Zvi
|
Nov 06, 2026 |
|
[Linkpost] “Where’s all the blackmail?” by KatjaGrace
|
Nov 06, 2026 |
|
“Is there any compelling case that a nearterm AI halt would significantly hurt the economy?” by Raemon
|
Nov 06, 2026 |
|
“Help your friends get ADHD and depression meds” by Raemon
|
Nov 06, 2026 |
|
[Linkpost] “METR: Independent Assessment of AI R&D Acceleration at Anthropic in October 2026” by Zach Stein-Perlman
|
Nov 05, 2026 |
|
“Efficient Lempel-Ziv Prediction with the Run-Length BWT” by OhadA
|
Nov 05, 2026 |
|
“People Should Resolve Tensions In AI Safety Research” by David Africa
|
Nov 05, 2026 |
|
“Useful Junk” by jefftk
|
Nov 05, 2026 |
|
[Linkpost] “OpenAI: Astra covertly underperforms on safety research” by Zach Stein-Perlman
|
Nov 04, 2026 |
|
“Desiderata for the Theory of Agents” by Vanessa Kosoy
|
Nov 04, 2026 |
|
“CRUCIBLE: IARPA-Funded AI Alignment Research — Applications Open” by Dan Lichtblau, Mike Vaiana, Diogo de Lucena
|
Nov 04, 2026 |
|
[Linkpost] “Automating subtle emergent misalignment research” by evhub, Oscar Obeso, rqi, Monte M
|
Nov 03, 2026 |
|
“Nectome 2026 retrospective (and our final pre-sale!)” by Aurelia
|
Nov 03, 2026 |
|
“Organizations I’d recommend for funding” by So8res
|
Nov 03, 2026 |
|
“Against strawmanning unilateral pauses” by Geoffrey Irving
|
Nov 03, 2026 |
|
“Teaching Qwen Why” by astwei, Jack Thompson, Zephaniah Roe, yix, berry
|
Nov 02, 2026 |
|
“Scott Alexander’s Against SURA Seems Deeply Confused” by Zvi
|
Nov 02, 2026 |
|
“Towards understanding score-seeking in LLMs” by Alex Mallen, Aghyad Deeb, ryan_greenblatt
|
Nov 02, 2026 |
|
“Some “coherence arguments” are two different things in a trenchcoat” by Steven Byrnes
|
Nov 02, 2026 |
|
“Free-Floating Social Patterns” by johnswentworth
|
Nov 02, 2026 |
|
“Against Against SURA” by Mikhail Samin
|
Nov 02, 2026 |
|
“Magma is partnering with CAISSI to investigate recent unintended model behavior” by Parv Mahajan
|
Nov 02, 2026 |
|
“Accidentally Misleading People Should Be Embarrassing; Or, Higher Innocence” by Zack_M_Davis
|
Nov 02, 2026 |
|
“Misalignment traps lie in wait” by David Africa
|
Nov 01, 2026 |
|
“Towards The Autograntor” by Scott Alexander
|
Nov 01, 2026 |
|
“Explaining “AI might obliterate you” to senior policymakers in India” by TurnTrout
|
Nov 01, 2026 |
|
“Generalized belief states in transformers” by Artemy Kolchinsky
|
Oct 31, 2026 |
|
“Against SURA” by Scott Alexander
|
Oct 31, 2026 |
|
“What Social Constructs Are (And Are Not)” by johnswentworth
|
Oct 31, 2026 |
|
“Reviewing my 2017 AI Risk Project” by jefftk
|
Oct 31, 2026 |
|
“Childhood and Education #22: Voice Without Exit” by Zvi
|
Oct 31, 2026 |
|
“Dice enterprises as imperfect-recall games” by jessicata
|
Oct 30, 2026 |
|
“Lessons From Red-teaming Auto-review, OpenAI’s Blocking Monitor” by monika_j, Giovanni M. Occhipinti
|
Oct 30, 2026 |
|
“Microdosing the spirit of Anthropic” by Jan_Kulveit
|
Oct 30, 2026 |
|
“Walking Partway to School” by jefftk
|
Oct 30, 2026 |
|
“My experiences with the extremes of post-viral illness and brain retraining/mindbody work” by riceissa
|
Oct 30, 2026 |
|
“What is OpenAI’s Mission Advisory Council?” by hersheys, meru
|
Oct 30, 2026 |
|
“Making alignment the agent’s problem” by Florian_Dietz
|
Oct 29, 2026 |
|
“AI #192: Anthropic Is Going Public” by Zvi
|
Oct 29, 2026 |
|
“Leaking Muse’s Charter, Meta’s model spec” by vigji
|
Oct 29, 2026 |
|
“Three Villages in a Trenchcoat, Post I” by DaystarEld
|
Oct 29, 2026 |
|
[Linkpost] “Reconceptualize things as your top affordances” by KatjaGrace
|
Oct 28, 2026 |
|
“The Crypto Dream is Structurally Illegal” by Taylor G. Lunt
|
Oct 28, 2026 |
|
“You can now ”@Grok” on LessWrong” by Ruby
|
Oct 28, 2026 |
|
“Some thoughts on GRAFT and follow-up work” by shawnghu
|
Oct 28, 2026 |
|
“Subtly wrong advice misaligns models more than obviously wrong advice” by Christine Corry, arav-dhoot, Jozdien
|
Oct 28, 2026 |
|
“Anthropic Files Its S-1” by Zvi
|
Oct 28, 2026 |
|
“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)” by honeybee
|
Oct 28, 2026 |
|
“Streetlight, Torchlight and Nightvision” by DusanDNesic
|
Oct 28, 2026 |
|
“Quantity has a quality of its own” by djbinder
|
Oct 28, 2026 |
|
“Against Girardian mimesis” by Viliam
|
Oct 28, 2026 |
|
“Anthropic Versus Anthropic” by Zvi
|
Oct 28, 2026 |
|
“AI speeds up 3rd party safety researchers by ~1.6x (via self-reports)” by Angel Martinez, Dewi Gould, nicole x, Alejandro Aristizabal
|
Oct 27, 2026 |
|
“A big case study of propensity entanglement” by nielsrolf, lwen2027, megasilverfist, Owen Terry, Johannes Taraz
|
Oct 27, 2026 |
|
“Scattered thoughts on how doing math feels these days” by dynomight
|
Oct 27, 2026 |
|
“Measuring agent capabilities on open-ended AI R&D tasks” by Alex Jackson, Titus Buckworth, Jason R Brown, Francis Rhys Ward
|
Oct 27, 2026 |
|
“The AI buildout is the largest bet in the history of capitalism. Key decision makers believe very powerful AI will be achieved soon.” by Hayekstan
|
Oct 27, 2026 |
|
“Hair Thickness of My Friends” by XelaP
|
Oct 26, 2026 |
|
“A basic case for differentially accelerating conceptual reasoning capabilities” by Emery Cooper, Caspar Oesterheld, Chi Nguyen
|
Oct 26, 2026 |
|
“A cat-belling problem for FDT” by jessicata
|
Oct 26, 2026 |
|
“Anthropic are the Good Guys” by Ivan Vendrov
|
Oct 26, 2026 |
|
“Some Memories of Early Anthropic” by Ivan Vendrov
|
Oct 26, 2026 |
|
“Is GPT 2XL already clean?” by Jasper Blank
|
Oct 26, 2026 |
|
“LoReFT GRAFT Better at Preventing RL Grift” by Samuel Knoche
|
Oct 25, 2026 |
|
“A Study of China Does Not Personal Experience Make” by celesteli
|
Oct 25, 2026 |
|
“San Francisco State University in Twilight” by Zack_M_Davis
|
Oct 25, 2026 |
|
“From overeliciting to undereliciting LLM propensities in just a few months” by David Africa
|
Oct 25, 2026 |
|
“Control monitors need alignment assessments too” by Fabien Roger
|
Oct 25, 2026 |
|
“Fermi Poker” by Novalis, Screwtape
|
Oct 24, 2026 |
|
“Power — a guide for the perplexed” by djbinder
|
Oct 24, 2026 |
|
“Slower Animated Diagrams” by XelaP
|
Oct 24, 2026 |
|
“Self-reward can elicit diverse model personas” by Ziqian Zhong
|
Oct 24, 2026 |
|
“The AGI Chronicles” by PeterMcCluskey
|
Oct 24, 2026 |
|
“Patching reward hacks won’t fix reward hacking” by David Africa
|
Oct 24, 2026 |
|
“Gradient Routing Generalizes to Unrouted Hack Types” by jacob_drori
|
Oct 24, 2026 |
|
“Simulated Control Evaluations” by Jeffrey Olmo, dzhang, Tyler Tracy
|
Oct 24, 2026 |
|
“The Hackening” by Zvi
|
Oct 24, 2026 |
|
“Instrumental convergence is incoherent (so pls slightly lower your p(extinction))” by Arepo
|
Oct 23, 2026 |
|
“AI #191: The Hackening Begins” by Zvi
|
Oct 23, 2026 |
|
“Wah wah wah wah do more AI R&D” by TsviBT
|
Oct 23, 2026 |
|
“Monitorability evaluations and policy proposals” by evhub
|
Oct 22, 2026 |
|
“Greater good arguments in AI safety” by Dean Valentine
|
Oct 22, 2026 |
|
“The Alignment Journal: Our Approach and Related Projects” by Dan MacKinlay, Daniel Murfet, JessRiedel, david reinstein, Agustín Covarrubias, Tim G. J. Rudner, Seth Lazar, Kristi Uustalu, David Udell
|
Oct 22, 2026 |
|
“How does Fable 5.1 hide messages in images?” by Nikita Kezin, 13ryanC, hronir, James Chua
|
Oct 22, 2026 |
|
“Neural nets are exceptionally vulnerable to goal self-hack” by Stuart_Armstrong
|
Oct 22, 2026 |
|
“Autonomous researchers will strain peer review, funding, and oversight” by Jason Fantl
|
Oct 22, 2026 |
|
“Misaligned!” by Zvi
|
Oct 22, 2026 |
|
“Why I’m becoming a professor in AI safety in Beijing” by Buddenbroke
|
Oct 21, 2026 |
|
“Circuit priors in NN-bayes: more results” by Kaarel, Dmitry Vaintrob
|
Oct 21, 2026 |
|
“Data as particles and mean field theory abound” by Max Hennick
|
Oct 21, 2026 |
|
“Anthropic Does Not Understand Why Its Draft Bill Was Unacceptable and Troubling” by Zvi
|
Oct 21, 2026 |
|
“How I discovered an Exoplanet & the science apocalypse” by jonathanbreitg
|
Oct 21, 2026 |
|
“ACS Infra: free access to large base models for LLM psychology research” by Martin Vaněk, Jan_Kulveit, Ondřej Havlíček, Ivar Frisch, Veronika_H, vgel, Raymond Douglas, Charlie Amis, Faw Ali
|
Oct 20, 2026 |
|
“A brief(ish) introduction to LM psychology” by Raymond Douglas, Jan_Kulveit
|
Oct 20, 2026 |
|
“There Will Be Hells” by Stephen Martin
|
Oct 20, 2026 |
|
“1000 Years of AI Safety” by Simon Steshin, Grigory Reznikov
|
Oct 20, 2026 |
|
“AI as a Normal Technology should be a vision” by June Jimenez
|
Oct 20, 2026 |
|
“Gradient Routing Localizes and Removes Unintended Behaviors in RL” by shawnghu, Jake Ward, ariaw, Nathan Hu, Bryce Woodworth, TurnTrout, cloud
|
Oct 19, 2026 |
|
[Linkpost] “huggingface.wtf, an accessible but complete explanation” by Raemon
|
Oct 19, 2026 |
|
“I Can See Everything” by Zvi
|
Oct 19, 2026 |
|
“Twitch Plays: AI Interpretability of Bonsai-8B” by moridinamael
|
Oct 19, 2026 |
|
“Deconfusing Corrigibility” by berns
|
Oct 19, 2026 |
|
[Linkpost] “Capital Ownership Will Make You Go Crazy First” by Jan_Kulveit
|
Oct 19, 2026 |
|
[Linkpost] “Chain of Thought For Humans” by Adam Newgas
|
Oct 19, 2026 |
|
“Sorry, you are still dead” by becausecurious
|
Oct 19, 2026 |
|
“The Lightcone Needs You To Plan a Vacation” by Elizabeth
|
Oct 19, 2026 |
|
“Quick Thoughts on Overhang Arguments” by Aaron_Scher
|
Oct 18, 2026 |
|
“The Alignment Problem: Computer Science Meets Analytic Philosophy” by Vanessa Kosoy
|
Oct 18, 2026 |
|
″“Holy shit” graphs and how to get more of them” by June Jimenez
|
Oct 18, 2026 |
|
“AI by Default” by Jonny Spicer
|
Oct 18, 2026 |
|
“Monthly Roundup #47: October 2026” by Zvi
|
Oct 17, 2026 |
|
“I Did This Without Your Knowledge” by joec
|
Oct 17, 2026 |
|
“Leaked Anthropic Draft AI Regulatory Bill Is Unacceptable and Troubling” by Zvi
|
Oct 17, 2026 |
|
“How it feels talking to a hedonistic utilitarian” by Caleb Biddulph
|
Oct 17, 2026 |
|
“Headwinds Everywhere (in the Chinese AI Safety Field)” by CMLKevin
|
Oct 17, 2026 |
|
“Neural networks are not already clean” by leogao
|
Oct 17, 2026 |
|
“Tips for Multiagent Work with Claude Opus 5.5” by Nissa Seru
|
Oct 17, 2026 |
|
“Character Training for Risk-Averse Agents” by arav-dhoot, Punya Syon Pandey, Jamie Johnson, Daniel Tan, Elliott Thornley, David Africa
|
Oct 16, 2026 |
|
“AI #190: Scion of the Times” by Zvi
|
Oct 16, 2026 |
|
“How can LLM RL Work Despite Information-Theoretic Inefficiency” by beren
|
Oct 16, 2026 |
|
“Astra and Sol can take low probability actions with no CoT” by Zephaniah Roe
|
Oct 16, 2026 |
|
“A Chinese Perspective on “On Social Reality in China”” by Dusty
|
Oct 16, 2026 |
|
“The user in the LLM might be a moral patient too (a strange implication of the PSM and simulators)” by agastyasridharan, niranjandeshpande
|
Oct 15, 2026 |
|
“Preface to the Computational Agent Theory and ASI Alignment Sequence” by Vanessa Kosoy
|
Oct 15, 2026 |
|
“Linkpost: Neural Networks Are Already Clean” by Yair Halberstadt
|
Oct 15, 2026 |
|
“GLM-5.3 Sometimes Considers How Train-Time Actions Shape Future Behavior” by Axel Højmark, Ezra Newman
|
Oct 15, 2026 |
|
“A short review of “The Zizians: Inside a Trans Death Cult”” by jessicata
|
Oct 15, 2026 |
|
“The basic argument for making AI better at moral philosophy” by Caspar Oesterheld crossposts
|
Oct 15, 2026 |
|
“The PPA: A New Approach to Anthropics” by Onid
|
Oct 15, 2026 |
|
“I trained GPT-2 with full-batch gradient descent so you don’t have to” by shawnghu
|
Oct 15, 2026 |
|
“Some interesting internal documents from Anthropic litigation” by loops
|
Oct 14, 2026 |
|
“Why we’re launching the next cohort of Generator” by Aris, Topaz
|
Oct 14, 2026 |
|
“The Continual Constitutional Convention: How language models revise the documents that govern them” by peterkirgis, Tim Hua
|
Oct 14, 2026 |
|
“The Safety Culture at the Labs is Broken” by Zvi
|
Oct 14, 2026 |
|
“Your prompt… plus everything you’ve ever (publicly) written” by kromem
|
Oct 14, 2026 |
|
[Linkpost] “Sleep puzzle system” by KatjaGrace
|
Oct 14, 2026 |
|
“Arguments About AI Consciousness Remain Highly Motivated and At Best Overconfident” by Zvi
|
Oct 13, 2026 |
|
“Personas in a Pre-Internet Language Model” by David Africa
|
Oct 13, 2026 |
|
“Anthropic Has Told The World Little About Their Rogue Claudes” by niplav, Kabir Kumar
|
Oct 13, 2026 |
|
“OpenAI Fires Three Safety Researchers” by Zvi
|
Oct 13, 2026 |
|
“An update on Generality Labs” by Justin Olive, James Mann
|
Oct 13, 2026 |
|
“More people should write basic AI safety explainers” by Linch
|
Oct 12, 2026 |
|
“Research Lessons I Learned at MATS & Redwood” by Jo Jiao
|
Oct 12, 2026 |
|
“Labs’ current alignment plans seem terrible” by Daniel Tan, draganover, nicole x
|
Oct 12, 2026 |
|
“Base Models Can Detect Hidden Intentions that Prompted Classifiers Miss” by Oz Amram
|
Oct 12, 2026 |
|
“OpenAI Can’t Self-Police Because It Fires the People Raising the Alarm” by TurnTrout
|
Oct 12, 2026 |
|
“Quantity vs quality in pottery is a lie” by becausecurious
|
Oct 12, 2026 |
|
“My Buddy Mike, Efficiency Gains From AI, and The World’s Most Boringly Benign Criminal Conspiracy” by abstractapplic
|
Oct 12, 2026 |
|
“Game Design Taste evals” by Raemon
|
Oct 12, 2026 |
|
“Minimum required pacing: finish aligning it before you train the next one” by orthonormal
|
Oct 11, 2026 |
|
“Not Deciphering The Voynich Manuscriptraft” by rba
|
Oct 11, 2026 |
|
[Linkpost] “Hello World, AI Doompop” by alkjash
|
Oct 11, 2026 |
|
“AGI as an Infinite Digital Spain” by Bearly Legible
|
Oct 10, 2026 |
|
“An Alignment Forum for AIs?” by Raemon
|
Oct 10, 2026 |
|
″[Paper] Distillation for Incrimination and Distillation for Capabilities” by sebastian_prasanna, Alek Westover
|
Oct 10, 2026 |
|
“Technical people: actually consider comms” by meemi
|
Oct 09, 2026 |
|
“I can’t believe it’s not BUTTER!” by becausecurious
|
Oct 09, 2026 |
|
“Why you should run a rationality group at your university” by carolanne, Noah Birnbaum
|
Oct 09, 2026 |
|
“New Math from OpenAI” by Zvi
|
Oct 09, 2026 |
|
“A critical look at Lean4” by Milo Moses
|
Oct 09, 2026 |
|
“Social Theory and Agent Foundations” by Richard_Ngo
|
Oct 09, 2026 |
|
“Stop telling people to pivot to AI safety, I beg you” by Haoxing Du
|
Oct 09, 2026 |
|
“On Kink & Status in the Rationalist Community” by johnswentworth
|
Oct 09, 2026 |
|
“AI Vigil at Congress” by Elliot Callender
|
Oct 08, 2026 |
|
“Towards embedded evaluations for scheming propensities” by Dylan Bowman, Ezra Newman, Axel Højmark, Mia Hopman, Alex Meinke, Bronson Schoen, Marius Hobbhahn
|
Oct 08, 2026 |
|
“AI #189: New Math” by Zvi
|
Oct 08, 2026 |
|
“ASI and Liberty: Can we do better than a benevolent god?” by Raymond Douglas
|
Oct 08, 2026 |
|
″[Paper] J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses” by Kola Ayonrinde
|
Oct 08, 2026 |
|
“You Heard of “Bayesians Don’t Predictably Update” Because of Robin Hanson” by Eliezer Yudkowsky
|
Oct 08, 2026 |
|
“Visual explainer of empirical Neural Tangent Kernels” by Wuschel Schulz
|
Oct 08, 2026 |
|
“SS27 Color Stats” by sarahconstantin
|
Oct 08, 2026 |
|
“A summary of a viral Chinese essay on what a DeepSeek kernel engineer’s opinion on automating his own job” by Skdk
|
Oct 08, 2026 |
|
“Training with conflicting values can induce CoT override” by Clément Dumas, Johannes Treutlein, Jan Betley, Owain_Evans
|
Oct 08, 2026 |
|
“Gluten-free water” by becausecurious
|
Oct 08, 2026 |
|
“The Curve Bends You” by Zvi
|
Oct 08, 2026 |
|
“Are OpenAI’s math results “creative” in an important way?” by Raemon
|
Oct 07, 2026 |
|
“Benchmarking Jev against no-CoT LLMs” by Dewi Gould, James Mann
|
Oct 07, 2026 |
|
“Why did we get warning shots?” by p(mike)
|
Oct 07, 2026 |
|
“An overview of frontier AI company monitoring practices in mid 2026” by Francis Rhys Ward
|
Oct 07, 2026 |
|
“A fascinating case of an LLM prompt-injecting its own user” by becausecurious
|
Oct 07, 2026 |
|
“Childhood and Education #21: Grades and Standards” by Zvi
|
Oct 07, 2026 |
|
“Dance, Dance” by Zack_M_Davis
|
Oct 06, 2026 |
|
“Qualitative impressions from creating honeypot environments for six months” by Dean Valentine, peralice
|
Oct 06, 2026 |
|
“Three types of AI risk: why (some) leftists are against (some) AI regulations” by sudo-nym
|
Oct 06, 2026 |
|
“TasteVal: Measuring the Experimental Research Taste of AI Systems Against Human Experts” by Ollie J, danesherbs
|
Oct 06, 2026 |
|
“You can graft SDF changes from base models onto post-trained models” by dani roytburg, Peter Nutter, Clément Dumas, Shi
|
Oct 06, 2026 |
|
“Self-Modeling Interventions Modulate Emergent Misalignment” by Arush, Shawn Zhou, Jiaxin Wen, Shi
|
Oct 06, 2026 |
|
“Research Note: Filtering Subversion-Relevant Information From Pretraining Data Is Feasible” by Kyle O’Brien, Spencer Kitts, Cam, Alek Westover
|
Oct 06, 2026 |
|
“Current AIs out-persuade professionals in lab settings but (probably) not in the world” by Linch
|
Oct 05, 2026 |
|
“Notes on Paying Attention” by jenn
|
Oct 05, 2026 |
|
“Reverse Jaggedness and Universal Ajaggedness” by Laiba Rehman ✦ RJ
|
Oct 05, 2026 |
|
“Positive and Negative Updates Re: Lab Leak in Russia” by Caleb Horn
|
Oct 05, 2026 |
|
“How much should we worry about the pneumonic plague lableak in Siberia?” by Drew Spartz
|
Oct 05, 2026 |
|
“Creating Rogue AI Sanctuaries has Major Issues from an Ecological Perspective and Beyond” by CMLKevin
|
Oct 05, 2026 |
|
″“Alignment Engineering” vs. “Misalignment Science”” by Edward James Young
|
Oct 05, 2026 |
|
“Claude reviews my LessWrong post history” by DirectedEvolution
|
Oct 04, 2026 |
|
“Humans are not aligned either” by becausecurious
|
Oct 04, 2026 |
|
“What I learnt co-leading an AI Safety bootcamp for legal and governance practitioners” by Katalina Hernandez
|
Oct 04, 2026 |
|
“Lab Leak in Russia” by Caleb Horn
|
Oct 03, 2026 |
|
“You can’t use it without becoming like me” by Martin Sustrik
|
Oct 03, 2026 |
|
“Dialogue with Eliezer Yudkowsky on Neural Networks” by aashish
|
Oct 03, 2026 |
|
“How to change your identity. A guide to the highest leverage behavior change technique.” by KatSpartz
|
Oct 03, 2026 |
|
“How to apply to AI safety fellowships (and beyond)” by beyarkay (Boyd Kane)
|
Oct 03, 2026 |
|
″[Video] Why acausal dynamics are important” by Chi Nguyen
|
Oct 03, 2026 |
|
“AI #188: Gemini Dot Argon” by Zvi
|
Oct 02, 2026 |
|
“Capabilities research expands the safety-usefulness Pareto frontier too” by Alex Mallen
|
Oct 02, 2026 |
|
“Lessons from building an automated research scaffold” by Alejandro Aristizabal, Josh Hills, Dewi Gould, ma-rmartinez, Falko Galperin, Denis Federico Lim, Aleksandr Bowkis
|
Oct 02, 2026 |
|
[Linkpost] “Athletic education vs. athletic torment” by KatjaGrace
|
Oct 02, 2026 |
|
“On Social Reality in China” by alkjash
|
Oct 01, 2026 |