All 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 20242025

All Jan Feb Mar Apr May Jun Jul Aug SepOctNov Dec

All 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 242526 27 28 29 30 31

Musings on Reported Cost of Compute (Oct 2025)

Vladimir_Nesov24 Oct 2025 20:42 UTC

103 points

11 comments2 min readLW link

Regardless of X, you can still just sign superintelligence-statement.org if you agree

Ishual24 Oct 2025 20:30 UTC

58 points

0 comments3 min readLW link

The Future of Interpretability is Geometric

sbaumohl24 Oct 2025 18:32 UTC

23 points

0 comments5 min readLW link

New Statement Calls For Not Building Superintelligence For Now

Zvi24 Oct 2025 17:40 UTC

80 points

3 comments7 min readLW link

(thezvi.wordpress.com)

Notes on “Explaining AI Explainability”

Eleni Angelou24 Oct 2025 17:22 UTC

20 points

0 comments6 min readLW link

Can Reasoning Models Obfuscate Reasoning? Stress-Testing Chain-of-Thought Monitorability

Artur Zolkowski and Wen Xing

24 Oct 2025 17:21 UTC

17 points

1 comment5 min readLW link

I will not sign up for cryonics

Syd Lonreiro_24 Oct 2025 16:56 UTC

−18 points

5 comments1 min readLW link

Dollars in political giving are less fungible than you might think

lincolnquirk24 Oct 2025 15:54 UTC

6 points

1 comment5 min readLW link

(lincolnquirk.substack.com)

Can AI Agents with Divergent Interests Learn To Prevent Civilizational Failures?

joao_abrantes24 Oct 2025 15:08 UTC

1 point

0 comments1 min readLW link

LW Reacts pack for Discord/Slack/etc

plex24 Oct 2025 13:20 UTC

65 points

13 comments1 min readLW link

(drive.google.com)

AI Timelines and Points of no return

Gabriel Alfour24 Oct 2025 11:15 UTC

36 points

8 comments1 min readLW link

(cognition.cafe)

Introducing ControlArena: A library for running AI control experiments

Mojmir24 Oct 2025 9:51 UTC

13 points

0 comments3 min readLW link

(www.aisi.gov.uk)

Can we steer AI models toward safer actions by making these instrumentally useful?

Francesca Gomez24 Oct 2025 9:18 UTC

5 points

0 comments2 min readLW link

(www.wiserhuman.ai)

Plan 1 and Plan 2

Towards_Keeperhood24 Oct 2025 8:18 UTC

50 points

22 comments3 min readLW link

Guys I might be an e/acc

Taylor G. Lunt24 Oct 2025 3:25 UTC

14 points

29 comments4 min readLW link

How an AI company CEO could quietly take over the world

Alex Kastner23 Oct 2025 23:33 UTC

52 points

13 comments11 min readLW link

Worlds Where Iterative Design Succeeds?

Max Harms23 Oct 2025 22:14 UTC

23 points

5 comments8 min readLW link

Automated real time monitoring and orchestration of coding agents

zef, kaivu and leni

23 Oct 2025 22:12 UTC

8 points

0 comments2 min readLW link

(fulcrumresearch.ai)

Reminder: Morality is unsolved

Jesper L.23 Oct 2025 21:42 UTC

27 points

45 comments3 min readLW link

The main way I’ve seen people turn ideologically crazy [Linkpost]

Noosphere8923 Oct 2025 20:09 UTC

123 points

22 comments8 min readLW link

(andymasley.substack.com)

Empirical Partial Derivatives

sonicrocketman23 Oct 2025 17:54 UTC

8 points

0 comments3 min readLW link

(brianschrader.com)

Building a different kind of personal intelligence

Rebecca Dai23 Oct 2025 17:45 UTC

7 points

0 comments9 min readLW link

(rebeccadai.substack.com)

Beliefs about formal methods and AI safety

Quinn23 Oct 2025 16:43 UTC

32 points

0 comments5 min readLW link

AI #139: The Overreach Machines

Zvi23 Oct 2025 15:30 UTC

35 points

5 comments52 min readLW link

(thezvi.wordpress.com)

Should AI Developers Remove Discussion of AI Misalignment from AI Training Data?

Alek Westover23 Oct 2025 15:12 UTC

43 points

3 comments9 min readLW link

SecureBio is Hiring Software Engineers

jefftk23 Oct 2025 14:10 UTC

21 points

0 comments1 min readLW link

(www.jefftk.com)

Is terminal lucidity real?

Ariel Zeleznikow-Johnston23 Oct 2025 11:40 UTC

20 points

0 comments1 min readLW link

(open.substack.com)

A Concrete Roadmap towards Safety Cases based on Chain-of-Thought Monitoring

Wuschel Schulz23 Oct 2025 11:34 UTC

37 points

5 comments4 min readLW link

(arxiv.org)

Differences in Alignment Behaviour between Single-Agent and Multi-Agent AI Systems

NotAWiz4rd, Cameron Tomé-Moreira and Andreas Hermann

23 Oct 2025 11:17 UTC

7 points

3 comments5 min readLW link

LW Psychosis

Annabelle23 Oct 2025 8:12 UTC

18 points

10 comments3 min readLW link

Announcing the Futurekind Winter Fellowship 2025/6

Aditya S23 Oct 2025 5:40 UTC

1 point

0 comments4 min readLW link

Learning to Interpret Weight Differences in Language Models

avichal23 Oct 2025 3:55 UTC

89 points

2 comments5 min readLW link

(arxiv.org)

AGI’s Last Bottlenecks

adamk23 Oct 2025 3:28 UTC

17 points

2 comments9 min readLW link

Statement on Superintelligence—FLI Open Letter

plex22 Oct 2025 22:26 UTC

59 points

0 comments1 min readLW link

(superintelligence-statement.org)

The Doomers Were Right

Algon22 Oct 2025 22:18 UTC

204 points

26 comments3 min readLW link

Technical Acceleration Methods for AI Safety: Summary from October 2025 Symposium

Martin Leitgab22 Oct 2025 21:33 UTC

25 points

2 comments6 min readLW link

Why AI alignment matters today

Mislav Jurić22 Oct 2025 21:27 UTC

6 points

0 comments4 min readLW link

Any corrigibility naysayers outside of MIRI?

Max Harms22 Oct 2025 21:26 UTC

28 points

24 comments1 min readLW link

Which side of the AI safety community are you in?

Max Tegmark22 Oct 2025 21:17 UTC

141 points

88 comments2 min readLW link

Homomorphically encrypted consciousness and its implications

jessicata22 Oct 2025 20:27 UTC

35 points

48 comments12 min readLW link

(unstableontology.com)

Dead-switches as AI safety tools

Jesper L.22 Oct 2025 19:57 UTC

2 points

6 comments5 min readLW link

Consider donating to AI safety champion Scott Wiener

Eric Neyman22 Oct 2025 18:40 UTC

133 points

9 comments18 min readLW link

(ericneyman.wordpress.com)

Postrationality: An Oral History

Gordon Seidoh Worley22 Oct 2025 16:10 UTC

44 points

4 comments30 min readLW link

(www.uncertainupdates.com)

Penny’s Hands

Tomás B.22 Oct 2025 16:09 UTC

70 points

7 comments16 min readLW link

Is 90% of code at Anthropic being written by AIs?

ryan_greenblatt22 Oct 2025 14:50 UTC

91 points

14 comments5 min readLW link

How Well Does RL Scale?

Toby_Ord22 Oct 2025 13:16 UTC

131 points

22 comments7 min readLW link

LLM Self-Reference Language in Multilingual vs English-Centric Models

dwmd22 Oct 2025 12:44 UTC

4 points

0 comments6 min readLW link

The Cloud industry architecture [Infra-Platform-App] is unlikely to replicate for AI

Armchair Descending22 Oct 2025 8:28 UTC

1 point

0 comments2 min readLW link

The Perpetual Technological Cage

Hector Perez Arenas22 Oct 2025 8:15 UTC

6 points

2 comments1 min readLW link

(networksocieties.com)

Utopiography Interview

plex22 Oct 2025 8:03 UTC

32 points

0 comments45 min readLW link