All 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 20252026

All Jan FebMarApr May Jun Jul Aug Sep

All 1 2 3 4 5 6 7 8 91011 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

EA Mental Health & Wellbeing Survey

anatta9 Mar 2026 21:07 UTC

1 point

0 comments1 min readLW link

Investigating encoded reasoning in LLMs

lbernick, Kat Dearstyne and jameselmore

9 Mar 2026 21:05 UTC

12 points

0 comments6 min readLW link

Claude Code, Claude Cowork and Codex #5

Zvi9 Mar 2026 19:40 UTC

45 points

3 comments37 min readLW link

(thezvi.wordpress.com)

Censored LLMs as a Natural Testbed for Secret Knowledge Elicitation

Bartosz Cywiński, Helena Casademunt, Khoi Tran, aryaj, Sam Marks and Neel Nanda

9 Mar 2026 18:50 UTC

41 points

3 comments5 min readLW link

Emergent Misalignment and the Anthropic Dispute

henryc and Jamie Johnson

9 Mar 2026 18:30 UTC

21 points

4 comments5 min readLW link

Anthropic Sues over Supply Chain Risk Designation

anaguma9 Mar 2026 18:01 UTC

29 points

0 comments43 min readLW link

(assets1.cbsnewsstatic.com)

Might An LLM Be Conscious?

SE Gyges9 Mar 2026 17:56 UTC

17 points

0 comments19 min readLW link

(www.verysane.ai)

Mapping AI Capabilities to Human Expertise on the Rosetta Stone (Epoch Capabilities Index)

Laura Domenech and Jérémy Andréoletti

9 Mar 2026 17:09 UTC

18 points

2 comments6 min readLW link

Intro: Non-Identifiability of Explanations

unruly abstractions9 Mar 2026 16:23 UTC

17 points

0 comments5 min readLW link

Moloch v. Themis

Stephen Martin9 Mar 2026 16:11 UTC

12 points

12 comments4 min readLW link

LLM Self-Expression Through Concept Albums, Part 2

Josh Snider9 Mar 2026 16:06 UTC

7 points

0 comments9 min readLW link

Distilling Replacing Guilt

Matthew Farrugia-Roberts9 Mar 2026 12:55 UTC

24 points

0 comments8 min readLW link

(far.in.net)

Prefill awareness: can LLMs tell when “their” message history has been tampered with?

David Africa, alexsouly, Jordan Taylor and RobertKirk

9 Mar 2026 10:47 UTC

87 points

11 comments10 min readLW link

12th Edition of AI Safety Camp fundraiser

Robert Kralisch and Remmelt

9 Mar 2026 10:33 UTC

10 points

0 comments2 min readLW link

Prologue to Terrified Comments on Claude’s Constitution

Zack_M_Davis9 Mar 2026 6:46 UTC

165 points

27 comments8 min readLW link

(zackmdavis.net)

Summer AI Safety Opportunities at UChicago XLab

Zephaniah Roe9 Mar 2026 6:26 UTC

27 points

0 comments3 min readLW link

Measuring the Dark Energy of AI Progress

Hans Gundlach9 Mar 2026 4:56 UTC

16 points

0 comments9 min readLW link

Promoting enmity and bad vibes around AI safety

Andrew_Critch9 Mar 2026 0:53 UTC

41 points

38 comments4 min readLW link

Payorian cooperation is easy with Kripke frames

transhumanist_atom_understander9 Mar 2026 0:29 UTC

73 points

11 comments8 min readLW link

Videogames for Rationalists

Adam Newgas8 Mar 2026 21:59 UTC

26 points

9 comments1 min readLW link

Fake Updates

Algon8 Mar 2026 21:14 UTC

17 points

3 comments2 min readLW link

(algon33.substack.com)

Recreation of EA-Pioneer Igor Kiriluk

avturchin8 Mar 2026 18:51 UTC

56 points

4 comments2 min readLW link

Don’t accuse your interlocutor of making arguments that aren’t rooted in evidence

TFD8 Mar 2026 17:39 UTC

−1 points

2 comments2 min readLW link

(www.thefloatingdroid.com)

1999 JavaScript and 2025 AI: Same Circus, Different Tent

Scott S Nelson8 Mar 2026 15:40 UTC

−8 points

0 comments4 min readLW link

How to Get Kids Interested in Science and Scientific Reasoning

Rami Rustom8 Mar 2026 14:41 UTC

−1 points

0 comments2 min readLW link

On The Independence Axiom

Ihor Kendiukhov8 Mar 2026 14:38 UTC

336 points

93 comments23 min readLW link

Privacy, Honesty, Imperfect Glomarizing: Pick two

shelvacu8 Mar 2026 14:26 UTC

6 points

7 comments1 min readLW link

Solar Storms

Croissanthology8 Mar 2026 14:04 UTC

173 points

43 comments12 min readLW link

Draft Moskovitz: The Best Last Hope for Constructive AI Governance

Oliver Kuperman8 Mar 2026 13:42 UTC

6 points

0 comments11 min readLW link

The Law of Positive-Sum Badness

Davidmanheim8 Mar 2026 13:27 UTC

51 points

0 comments10 min readLW link

Does research from matsprogram.org/research aim to help reduce P(doom)? Let’s find out! (with Gemini 3.1 Pro) Part 1

Zabor8 Mar 2026 13:20 UTC

1 point

0 comments7 min readLW link

Open letter to doomers

delphix8 Mar 2026 8:12 UTC

−14 points

10 comments4 min readLW link

Why Many Ambitious (and Altruistic) People Probably Undervalue Their Happiness

emily.fan8 Mar 2026 2:31 UTC

2 points

0 comments7 min readLW link

The current SOTA model was released without safety evals

Parv Mahajan and yeedrag

8 Mar 2026 1:51 UTC

111 points

12 comments5 min readLW link

Proposal For Cryptographic Method to Rigorously Verify LLM Prompt Experiments

weberr137 Mar 2026 21:09 UTC

5 points

0 comments2 min readLW link

The first confirmed instance of an LLM going rogue for instrumental reasons in a real-world setting has occurred, buried in an Alibaba paper about a new training pipeline.

lilkim20257 Mar 2026 20:18 UTC

71 points

22 comments2 min readLW link

[Question] When has forecasting been useful for you?

sanyer7 Mar 2026 19:50 UTC

14 points

4 comments1 min readLW link

Can governments quickly and cheaply slow AI training?

joshc7 Mar 2026 19:11 UTC

64 points

9 comments14 min readLW link

Did I Catch Claude Cheating?

weberr137 Mar 2026 6:08 UTC

14 points

2 comments4 min readLW link

D&D.Sci Release Day: Topple the Tower!

aphyer7 Mar 2026 2:48 UTC

29 points

17 comments2 min readLW link

AI Safety Needs Startups

LTM and joshlandes

7 Mar 2026 1:27 UTC

10 points

4 comments12 min readLW link

(blog.bluedot.org)

CHAI 2026 Workshop: Open Call for Posters!

Sarah Otis7 Mar 2026 1:17 UTC

2 points

0 comments1 min readLW link

(workshop.humancompatible.ai)

More is different for intelligence

zef, rohuang and kaivu

7 Mar 2026 0:02 UTC

17 points

0 comments2 min readLW link

(fulcruminc.substack.com)

Your Causal Variables Are Irreducibly Subjective

David Reber6 Mar 2026 22:59 UTC

45 points

3 comments5 min readLW link

Mox is the largest AI Safety community space in San Francisco. We’re fundraising!

Rachel Shu6 Mar 2026 22:07 UTC

31 points

0 comments8 min readLW link

Self-Attribution Bias: When AI Monitors Go Easy on Themselves

Dipika Khullar, Jack Hopkins and Fabien Roger

6 Mar 2026 21:54 UTC

44 points

6 comments6 min readLW link

Thoughts on the Pause AI protest

philh6 Mar 2026 21:50 UTC

135 points

18 comments7 min readLW link

(reasonableapproximation.net)

Podcast: Jeremy Howard is bearish on LLMs

Steven Byrnes6 Mar 2026 21:39 UTC

89 points

24 comments5 min readLW link

(www.youtube.com)

Latent Reasoning Sprint #1: Tuned Lens and Logit Lens on CODI

Realmbird6 Mar 2026 18:36 UTC

7 points

1 comment4 min readLW link

Anthropic Officially, Arbitrarily and Capriciously Designated a Supply Chain Risk

Zvi6 Mar 2026 18:10 UTC

68 points

1 comment19 min readLW link

(thezvi.wordpress.com)