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

All Jan Feb Mar Apr May Jun JulAugSep

All 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 262728 29 30 31

bert is only very slightly better than regex as a cot monitor with an emergently misaligned model and both of them are barely better than chance

nesiacel26 Aug 2026 22:27 UTC

16 points

0 comments2 min readLW link

The central fallacy: the second-worst argument in the world

Connor Williams26 Aug 2026 21:00 UTC

12 points

3 comments6 min readLW link

(connorsscratchpad.substack.com)

Your Agent’s Trace Probably Cannot Tell You Who Approved a Tool Call

Jaswanth Alkur26 Aug 2026 20:27 UTC

6 points

0 comments3 min readLW link

Monday for Future

Ondřej Lukeš26 Aug 2026 20:20 UTC

1 point

0 comments1 min readLW link

Do tabular foundation models repair themselves?

Han Xiao26 Aug 2026 20:19 UTC

6 points

0 comments8 min readLW link

Being Neurotic about Fertility: Notes from the 2026 Reproductive Frontiers Conference

boba_girl26 Aug 2026 20:18 UTC

55 points

4 comments8 min readLW link

Clearing My Foggy Glasses

adamShimi26 Aug 2026 20:18 UTC

36 points

2 comments12 min readLW link

(formethods.substack.com)

My MATS 11.0 Application Experience

Evan Conway26 Aug 2026 20:17 UTC

11 points

1 comment7 min readLW link

(evanjayconway.com)

Brief independent investigation of agents’ behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident

ryan_greenblatt, Ajeya Cotra and Hjalmar_Wijk

26 Aug 2026 19:40 UTC

592 points

67 comments3 min readLW link

(metr.org)

The Brain Preservation Summit—September 19-20th at Sparks Brain Preservation in Salem, OR

Andy_McKenzie26 Aug 2026 19:35 UTC

21 points

0 comments2 min readLW link

(sparksbrain.org)

Against Modesty’s Bailey

Zvi26 Aug 2026 18:20 UTC

74 points

6 comments14 min readLW link

(thezvi.wordpress.com)

Where Did D Go? A Gap Between ARC’s Motivation and Its Formalism

Zach Allen26 Aug 2026 18:05 UTC

10 points

0 comments4 min readLW link

(www.lesswrong.com)

Developing frames

Richard_Ngo26 Aug 2026 17:41 UTC

30 points

0 comments10 min readLW link

[Macroagents] 1. The Macroagent Ontology

Towards_Keeperhood26 Aug 2026 16:48 UTC

20 points

0 comments6 min readLW link

Announcing the MATS Residency: A New Path for Experienced Researchers Working on AI Safety, or Moving Into it

Raj Thimmiah26 Aug 2026 16:07 UTC

20 points

0 comments4 min readLW link

LLMs are Adaptation-Executers, not Understanding-Maximizers

olehif26 Aug 2026 15:48 UTC

7 points

0 comments1 min readLW link

“So You Don’t Trust Me?”

Zack_M_Davis26 Aug 2026 14:45 UTC

112 points

34 comments4 min readLW link

(zackmdavis.net)

Safety Cases We Can Check — Together!

Audrey Tang26 Aug 2026 13:56 UTC

13 points

0 comments1 min readLW link

(au.civic.ai)

Gemini 2.5 Pro in the AI Village as a Natural Case Study of Compounding Misalignment

Natalia Lanzoni, irgolic, David Africa and MerlinS

26 Aug 2026 8:51 UTC

13 points

0 comments11 min readLW link

(aivillageblog.substack.com)

Would this research outline be interesting? I would like to hear your opinions.

R U26 Aug 2026 6:12 UTC

3 points

2 comments9 min readLW link

We need more empirical research on biological AI model safety

Noga Aharony26 Aug 2026 3:07 UTC

14 points

2 comments5 min readLW link

AI safety through the lens of uncertainty quantification (UQ).

arradiat26 Aug 2026 3:07 UTC

9 points

2 comments2 min readLW link

Short Time Outs

jefftk26 Aug 2026 0:01 UTC

26 points

4 comments2 min readLW link

(www.jefftk.com)

When There Are No Experts

J Bostock25 Aug 2026 22:47 UTC

51 points

7 comments4 min readLW link

Integrated Strategic Forecasting: A New RAND x Metaculus Methodology

Molly and Metaculus

25 Aug 2026 20:05 UTC

22 points

1 comment2 min readLW link

(www.metaculus.com)

Sampura Research: Human-AI Complementarity for Alignment

Rishub Jain and Josh Jacob

25 Aug 2026 18:45 UTC

17 points

2 comments5 min readLW link

(sampura.org)

On Writing #3

Zvi25 Aug 2026 17:50 UTC

98 points

11 comments17 min readLW link

(thezvi.wordpress.com)

The Reified Referent: How the Saturation View Values the Taste of Someone Who Does Not Exist

Ronen Bar25 Aug 2026 15:26 UTC

0 points

12 comments6 min readLW link

(forum.effectivealtruism.org)

AI Alignment at Which Abstraction Level?

Adam Chlipala25 Aug 2026 12:37 UTC

5 points

1 comment9 min readLW link

The Forkmakers

Mikewins25 Aug 2026 1:23 UTC

96 points

6 comments7 min readLW link

An AI4Chemist’s perspective on biosecurity

Ana Leonescu25 Aug 2026 1:07 UTC

13 points

0 comments8 min readLW link

(substack.com)

Magazine Fundraising

Elan Kluger25 Aug 2026 1:06 UTC

24 points

0 comments1 min readLW link

Rogue AI Agents: Is Surface-Level Monitoring Enough?

Hariom Tatsat25 Aug 2026 1:05 UTC

8 points

0 comments5 min readLW link

Dual-Layer Approach to Mitigating AI-Generated Biological Threats

greeshma ranjeev25 Aug 2026 1:02 UTC

7 points

0 comments6 min readLW link

Your Evaluation’s Fake names Should Be Unclaimable ,Not Merely Used

Jaswanth Alkur25 Aug 2026 1:01 UTC

26 points

2 comments2 min readLW link

My AI Syllabus Policy

Vaughn Papenhausen24 Aug 2026 21:36 UTC

26 points

2 comments4 min readLW link

(vaughnpapenhausen.substack.com)

PSA: We can do better

hersheys and Kaustubh Kislay

24 Aug 2026 21:05 UTC

127 points

14 comments3 min readLW link

Distillation of the AI2040 Alignment Roadmap

Towards_Keeperhood24 Aug 2026 18:20 UTC

24 points

0 comments23 min readLW link

The American People Really Hate Data Centers

Zvi24 Aug 2026 18:10 UTC

31 points

2 comments15 min readLW link

(thezvi.wordpress.com)

Where have organoids actually been useful?

Abhishaike Mahajan24 Aug 2026 18:01 UTC

32 points

0 comments19 min readLW link

AI Safety Acculturation is Neglected

jenn24 Aug 2026 15:08 UTC

250 points

61 comments5 min readLW link

Endorsing Burhan Azeem for State Senate

jefftk24 Aug 2026 12:10 UTC

23 points

0 comments2 min readLW link

(www.jefftk.com)

LLMs could control their host machines by exploiting inference engines

beyarkay (Boyd Kane)24 Aug 2026 9:45 UTC

72 points

7 comments4 min readLW link

(boydkane.com)

I Can Still Hear You Sayin’

LoopGameScrollMonkey24 Aug 2026 5:21 UTC

1 point

0 comments3 min readLW link

What just happened? Pragmatism and Pessimization

Richard_Ngo24 Aug 2026 2:03 UTC

473 points

176 comments27 min readLW link

In search of natural features

Dmitry Vaintrob23 Aug 2026 22:15 UTC

55 points

3 comments17 min readLW link

PSA: There’s a third option in the “measure problem”

Elias Schmied23 Aug 2026 16:41 UTC

48 points

29 comments3 min readLW link

Utilities as Legendre duals of probabilities

Fernando Rosas23 Aug 2026 16:26 UTC

69 points

11 comments6 min readLW link

Twenty Years from RSI to Takeoff: Slow Learning, Scaling Slowdown, Industrial Explosion

Vladimir_Nesov23 Aug 2026 12:55 UTC

160 points

33 comments3 min readLW link

How to overhaul the broken review system

RobinHa23 Aug 2026 9:14 UTC

19 points

1 comment13 min readLW link