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 181920 21 22 23 24 25 26 27 28 29 30 31

Cui bono? ChatGPT-4o shows non-deceptive strategic persuasion: a Proof-of-Concept study

Pieter Barkema18 Aug 2026 23:25 UTC

7 points

0 comments5 min readLW link

Whack-a-mole with a broken hammer: does a model internally track its automaton state?

star2vec18 Aug 2026 23:25 UTC

7 points

0 comments8 min readLW link

GPT-2′s IOI behavior is defined where the paper’s algorithm isn’t

Zach Allen18 Aug 2026 23:23 UTC

7 points

0 comments2 min readLW link

Putting your mask on first

emiliob18 Aug 2026 23:23 UTC

1 point

0 comments1 min readLW link

(emiliobarkett.github.io)

Transhumanism & Human Enhancement Meetup (Online)

nxzt18 Aug 2026 23:22 UTC

1 point

0 comments1 min readLW link

Goal-directed optimisation requires information about the world

dubinsschwarz18 Aug 2026 23:22 UTC

13 points

0 comments12 min readLW link

AI Sandbagging (w/ Inspect)

Leo H18 Aug 2026 23:20 UTC

8 points

0 comments5 min readLW link

Anthropic Risk Report: August 2026

Zvi18 Aug 2026 20:40 UTC

48 points

3 comments42 min readLW link

(thezvi.wordpress.com)

Funding Formal Methods for the Cyberpocalypse

Max von Hippel18 Aug 2026 16:45 UTC

22 points

8 comments9 min readLW link

Policy career planning in the age of imminent superintelligence

Peter Wildeford18 Aug 2026 14:33 UTC

70 points

5 comments6 min readLW link

Automatic Programming Should Be More Like SQL

Adam Chlipala18 Aug 2026 12:29 UTC

7 points

0 comments12 min readLW link

AI Security is Harm Reduction

Quinn18 Aug 2026 12:28 UTC

39 points

3 comments1 min readLW link

LASR Labs Winter 2027 applications are open!

Brandon Riggs and tnawab

18 Aug 2026 10:12 UTC

37 points

1 comment4 min readLW link

Price recursion is the rational theory of reward

Abhimanyu Pallavi Sudhir18 Aug 2026 3:22 UTC

6 points

0 comments9 min readLW link

Natural Independence Incentives

jefftk18 Aug 2026 2:40 UTC

55 points

0 comments2 min readLW link

(www.jefftk.com)

Misaligned Incentives in Pause Scenarios

Michael Soareverix and Antra Tessera

18 Aug 2026 1:55 UTC

48 points

13 comments17 min readLW link

Inoculate Everything: All of Pretraining and RL

davids18 Aug 2026 1:28 UTC

9 points

0 comments10 min readLW link

(zenodo.org)

For Claude, capability and dispreferring CDT are the ~same thing. Much more so than for GPT.

Chi Nguyen and Emery Cooper

18 Aug 2026 1:25 UTC

50 points

13 comments1 min readLW link

You Can’t Iterate to Trustworthy AI Code Without Understanding

ronbodkin17 Aug 2026 23:00 UTC

8 points

0 comments10 min readLW link

Do your readers get annoyed if you post every day for 30 days?

Steff17 Aug 2026 21:59 UTC

13 points

5 comments1 min readLW link

What 27 AI Safety Generalists Are Building This Summer

agucova, Jesse Gilbert and Helena Tran

17 Aug 2026 21:56 UTC

18 points

0 comments9 min readLW link

What gives you away: how LLMs form opinions of you

Cat McGee17 Aug 2026 20:59 UTC

42 points

8 comments6 min readLW link

Evaluating Chain-of-Thought Monitorability is Still an Open Problem: Comments on OpenAI’s Monitorability Evals

Connor Dilgren and Sarah Wiegreffe

17 Aug 2026 19:45 UTC

20 points

3 comments18 min readLW link

Weird Re-Tokenization, Symmetries and Compression: Research Agenda

Xenomirant and Sami Wolf

17 Aug 2026 19:43 UTC

21 points

15 comments16 min readLW link

Fork Around and Find Out Part 3: Interpreting the knight auditor

David Litman17 Aug 2026 18:11 UTC

18 points

0 comments5 min readLW link

Connect to your future selves

PatrickDFarley17 Aug 2026 14:57 UTC

14 points

0 comments11 min readLW link

Lateral Workshop Applications: A Case Study

Jacob Brinton and Seth Lifland

17 Aug 2026 10:38 UTC

7 points

0 comments3 min readLW link

(forum.effectivealtruism.org)

Value Alignment Is a Pseudo Concept. A translation; humanity is not a single subject, and alignment is not one-way.

Davidmanheim17 Aug 2026 7:44 UTC

36 points

1 comment10 min readLW link

Untie Squared ReLU variant

Amy_17 Aug 2026 6:15 UTC

15 points

5 comments7 min readLW link

Study Update: Does post-training quantization change welfare-relevant indicators in open-weight language models?

ashesfall16 Aug 2026 21:11 UTC

8 points

0 comments7 min readLW link

Q2.5 2026 Timelines Update: Uplift and Revenue

brendanhalstead, Daniel Kokotajlo and elifland

16 Aug 2026 19:00 UTC

66 points

20 comments11 min readLW link

(blog.aifutures.org)

Case for Funding AI Safety in Japan

mmKALLL16 Aug 2026 18:41 UTC

33 points

1 comment9 min readLW link

Will There Be an AI Hegemon? A Mental Model for AI Power Concentration

simeon_c16 Aug 2026 18:41 UTC

28 points

7 comments3 min readLW link

(simeoncampos.substack.com)

The Doomsday Argument is Reasonable and Mostly Points to Longevity

Josh Snider16 Aug 2026 17:14 UTC

29 points

11 comments6 min readLW link

Three thoughts on civilisational handoff

Cleo Nardo16 Aug 2026 17:12 UTC

62 points

2 comments2 min readLW link

(clattubato.substack.com)

Should Less Wrong add subtitles?

Chris_Leong16 Aug 2026 9:05 UTC

32 points

10 comments1 min readLW link

Are questions allowed on LessWrong?

yatharth16 Aug 2026 7:04 UTC

13 points

10 comments2 min readLW link

Does DiffusionGemma do latent reasoning?

Jan Bauer and Neel Nanda

16 Aug 2026 4:22 UTC

40 points

0 comments9 min readLW link

Winners of the Fundamental Uncertainty Essay Contest

Gordon Seidoh Worley16 Aug 2026 1:20 UTC

27 points

2 comments1 min readLW link

(www.uncertainupdates.com)

Kimi likes causal decision theory more after RL in twin prisoner’s dilemmas

oakhu15 Aug 2026 22:31 UTC

115 points

70 comments6 min readLW link

Luck is a function of surface area.

sid.the.manne@gmail.com15 Aug 2026 20:44 UTC

11 points

0 comments1 min readLW link

What if Parameter Updates were Text?

DaemonicSigil15 Aug 2026 20:06 UTC

27 points

0 comments11 min readLW link

Using Chunked Monitoring to Detect Deception in Long Transcripts

sfereido15 Aug 2026 20:04 UTC

11 points

0 comments4 min readLW link

How To Catch a Distilled Model

Goutham Nalagatla and Carlos Guerrero Alvarez

15 Aug 2026 19:09 UTC

20 points

8 comments9 min readLW link

Mom’s Advice For Hosting A Class Reunion

jenn15 Aug 2026 15:52 UTC

53 points

2 comments2 min readLW link

Learning new facts can change LLM behaviour

Richard Juggins15 Aug 2026 13:46 UTC

33 points

2 comments14 min readLW link

(www.workingthroughai.com)

On Dwarkesh Patel’s Podcast With Ryan Greenblatt

Zvi15 Aug 2026 13:00 UTC

34 points

1 comment29 min readLW link

(thezvi.wordpress.com)

Does doubling a user profile change the effect of an instruction about using saved memories?

theodorepjs15 Aug 2026 11:59 UTC

17 points

1 comment3 min readLW link

I’m starting a interview series of people working in Lean / formal methods / math formalization

Adi Baradwaj15 Aug 2026 9:15 UTC

10 points

0 comments1 min readLW link

Nuclear physics of Alex Zhao’s comment for “Pacing the Frontier”

Dante Dam15 Aug 2026 9:03 UTC

17 points

0 comments11 min readLW link