All 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 202420252026

All Jan FebMarApr May Jun Jul Aug Sep Oct Nov Dec

All 1 2 3 4 5 6 7 8 9 10 11 12 131415 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Superintelligence’s goals are likely to be random

Mikhail Samin13 Mar 2025 22:41 UTC

6 points

6 comments5 min readLW link

Should AI safety be a mass movement?

MattAlexander13 Mar 2025 20:36 UTC

5 points

1 comment4 min readLW link

Auditing language models for hidden objectives

Sam Marks, Johannes Treutlein, dmz, Sam Bowman, Hoagy, Carson Denison, Kei Nishimura-Gasparian, 7vik, Akbir Khan, Austin Meek, Euan Ong, Christopher Olah, Fabien Roger, jeanne_, Meg, Drake Thomas, Adam Jermyn, Monte M and evhub

13 Mar 2025 19:18 UTC

145 points

15 comments13 min readLW link

Reducing LLM deception at scale with self-other overlap fine-tuning

Marc Carauleanu, Diogo de Lucena, Gunnar_Zarncke, Kvee, Cameron Berg, Mike Vaiana and Trent Hodgeson

13 Mar 2025 19:09 UTC

162 points

46 comments6 min readLW link

Vacuum Decay: Expert Survey Results

JessRiedel13 Mar 2025 18:31 UTC

96 points

26 comments13 min readLW link

A Frontier AI Risk Management Framework: Bridging the Gap Between Current AI Practices and Established Risk Management

simeon_c and Henry Papadatos

13 Mar 2025 18:29 UTC

10 points

0 comments1 min readLW link

(arxiv.org)

Creating Complex Goals: A Model to Create Autonomous Agents

theraven13 Mar 2025 18:17 UTC

6 points

1 comment6 min readLW link

Habermas Machine

NicholasKees13 Mar 2025 18:16 UTC

54 points

7 comments6 min readLW link

(mosaic-labs.org)

The Other Alignment Problem: Maybe AI Needs Protection From Us

Peterpiper13 Mar 2025 18:03 UTC

−2 points

0 comments3 min readLW link

AI #107: The Misplaced Hype Machine

Zvi13 Mar 2025 14:40 UTC

47 points

12 comments40 min readLW link

(thezvi.wordpress.com)

Intelsat as a Model for International AGI Governance

rosehadshar and wdmacaskill

13 Mar 2025 12:58 UTC

45 points

0 comments1 min readLW link

(www.forethought.org)

Stacity: a Lock-In Risk Benchmark for Large Language Models

alamerton13 Mar 2025 12:08 UTC

4 points

0 comments1 min readLW link

(huggingface.co)

The prospect of accelerated AI safety progress, including philosophical progress

Mitchell_Porter13 Mar 2025 10:52 UTC

12 points

0 comments4 min readLW link

The “Reversal Curse”: you still aren’t antropomorphising enough.

lumpenspace13 Mar 2025 10:24 UTC

3 points

0 comments1 min readLW link

(lumpenspace.substack.com)

Formalizing Space-Faring Civilizations Saturation concepts and metrics

Maxime Riché13 Mar 2025 9:40 UTC

4 points

0 comments8 min readLW link

The Economics of p(doom)

Jakub Growiec13 Mar 2025 7:33 UTC

2 points

0 comments1 min readLW link

Social Media: How to fix them before they become the biggest news platform

Sam G13 Mar 2025 7:28 UTC

5 points

2 comments3 min readLW link

Penny Whistle in E?

jefftk13 Mar 2025 2:40 UTC

9 points

1 comment1 min readLW link

(www.jefftk.com)

Anthropic, and taking “technical philosophy” more seriously

Raemon13 Mar 2025 1:48 UTC

139 points

29 comments11 min readLW link

LW/ACX Social Meetup

Stefan12 Mar 2025 23:13 UTC

2 points

0 comments1 min readLW link

I grade every NBA basketball game I watch based on enjoyability

proshowersinger12 Mar 2025 21:46 UTC

24 points

2 comments4 min readLW link

Kairos is hiring a Head of Operations/Founding Generalist

agucova12 Mar 2025 20:58 UTC

6 points

0 comments5 min readLW link

USAID Outlook: A Metaculus Forecasting Series

ChristianWilliams12 Mar 2025 20:34 UTC

9 points

0 comments1 min readLW link

(www.metaculus.com)

What is instrumental convergence?

Vishakha and Algon

12 Mar 2025 20:28 UTC

2 points

0 comments2 min readLW link

(aisafety.info)

Revising Stages-Oversight Reveals Greater Situational Awareness in LLMs

Sanyu Rajakumar12 Mar 2025 17:56 UTC

16 points

0 comments13 min readLW link

Why Obedient AI May Be the Real Catastrophe

G~12 Mar 2025 17:50 UTC

5 points

2 comments3 min readLW link

Your Communication Preferences Aren’t Law

Jonathan Moregård12 Mar 2025 17:20 UTC

25 points

4 comments1 min readLW link

(honestliving.substack.com)

Reflections on Neuralese

Alice Blair12 Mar 2025 16:29 UTC

42 points

3 comments5 min readLW link

Field tests of semi-rationality in Brazilian military training

P. João12 Mar 2025 16:14 UTC

31 points

0 comments2 min readLW link

Many life-saving drugs fail for lack of funding. But there’s a solution: desperate rich people

Mvolz12 Mar 2025 15:24 UTC

17 points

0 comments1 min readLW link

(www.theguardian.com)

The Most Forbidden Technique

Zvi12 Mar 2025 13:20 UTC

165 points

9 comments17 min readLW link

(thezvi.wordpress.com)

You don’t actually need a physical multiverse to explain anthropic fine-tuning.

Fraser12 Mar 2025 7:33 UTC

7 points

8 comments3 min readLW link

(frvser.com)

AI Can’t Write Good Fiction

JustisMills12 Mar 2025 6:11 UTC

38 points

24 comments7 min readLW link

(justismills.substack.com)

Existing UDTs test the limits of Bayesianism (and consistency)

Cole Wyeth12 Mar 2025 4:09 UTC

28 points

22 comments7 min readLW link

(Anti)Aging 101

George3d612 Mar 2025 3:59 UTC

5 points

2 comments3 min readLW link

(cerebralab.com)

The Grapes of Hardness

adamShimi11 Mar 2025 21:01 UTC

8 points

0 comments5 min readLW link

(formethods.substack.com)

Don’t over-update on FrontierMath results

David Matolcsi11 Mar 2025 20:44 UTC

47 points

7 comments9 min readLW link

Response to Scott Alexander on Imprisonment

Zvi11 Mar 2025 20:40 UTC

40 points

4 comments9 min readLW link

(thezvi.wordpress.com)

Paths and waystations in AI safety

Joe Carlsmith11 Mar 2025 18:52 UTC

42 points

1 comment11 min readLW link

(joecarlsmith.substack.com)

Meridian Cambridge Visiting Researcher Programme: Turn AI safety ideas into funded projects in one week!

Meridian Cambridge11 Mar 2025 17:46 UTC

13 points

0 comments2 min readLW link

Elon Musk May Be Transitioning to Bipolar Type I

Cyborg2511 Mar 2025 17:45 UTC

87 points

22 comments4 min readLW link

Scaling AI Regulation: Realistically, what Can (and Can’t) Be Regulated?

Katalina Hernandez11 Mar 2025 16:51 UTC

3 points

1 comment3 min readLW link

How Language Models Understand Nullability

Anish Tondwalkar and Alex Sanchez-Stern

11 Mar 2025 15:57 UTC

5 points

0 comments2 min readLW link

(dmodel.ai)

Forethought: a new AI macrostrategy group

Max Dalton, Tom Davidson, wdmacaskill and AmritSidhu-Brar

11 Mar 2025 15:39 UTC

20 points

0 comments3 min readLW link

Preparing for the Intelligence Explosion

fin and wdmacaskill

11 Mar 2025 15:38 UTC

79 points

17 comments1 min readLW link

(www.forethought.org)

stop solving problems that have already been solved

dhruvmethi11 Mar 2025 15:30 UTC

10 points

3 comments8 min readLW link

AI Control May Increase Existential Risk

Jan_Kulveit11 Mar 2025 14:30 UTC

101 points

13 comments1 min readLW link

When is it Better to Train on the Alignment Proxy?

dil-leik-og11 Mar 2025 13:35 UTC

14 points

0 comments9 min readLW link

A different take on the Musk v OpenAI preliminary injunction order

TFD11 Mar 2025 12:46 UTC

8 points

0 comments20 min readLW link

(www.thefloatingdroid.com)

Do reasoning models use their scratchpad like we do? Evidence from distilling paraphrases

Fabien Roger11 Mar 2025 11:52 UTC

127 points

23 comments11 min readLW link

(alignment.anthropic.com)