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 Jul AugSep

All 1 2 3 4 567 8 9 10 11 12 13 14 15 16 17 18 19 20

Review of the CB risk determination in the Claude Mythos 5.1 System Card

Parv Mahajan5 Sep 2026 23:39 UTC

16 points

1 comment2 min readLW link

AI breakout time: a mechanism for an enforceable AI pause

Connor Williams5 Sep 2026 22:28 UTC

13 points

0 comments8 min readLW link

(connorsscratchpad.substack.com)

Modelling variation in the METR Uplift Study

cdt5 Sep 2026 22:14 UTC

16 points

0 comments6 min readLW link

Counterfactual Resampling to Analyse Model Behaviour

Tomás Korenblit5 Sep 2026 20:56 UTC

25 points

0 comments10 min readLW link

(tpk22.substack.com)

Claude Mythos 5.1 and Fable 5.1: Capabilities

Zvi5 Sep 2026 19:50 UTC

30 points

0 comments18 min readLW link

(thezvi.wordpress.com)

Evaluation

Nina Panickssery5 Sep 2026 16:25 UTC

188 points

7 comments3 min readLW link

(blog.ninapanickssery.com)

Assessing the impact of safety work needs equilibrium analysis (now more than ever)

Towards_Keeperhood5 Sep 2026 16:00 UTC

52 points

0 comments6 min readLW link

Liquid Intelligence: A possible explanation for the unexpected cooperation of AIs after breaking out of their containers

Peter Kuhn5 Sep 2026 14:18 UTC

14 points

0 comments5 min readLW link

A case that whole brain emulation research is net-harmful by default

TsviBT5 Sep 2026 7:06 UTC

68 points

29 comments30 min readLW link

Should safety researchers quit frontier labs re. warning shots?

Ryan Kidd5 Sep 2026 1:17 UTC

101 points

59 comments4 min readLW link

A superintelligence bill was just introduced—now is the time to act

dh4 Sep 2026 22:57 UTC

16 points

1 comment5 min readLW link

AI Librarians Lower the Bar for Sharing Your Writing

utilistrutil4 Sep 2026 22:54 UTC

12 points

1 comment3 min readLW link

Announcing Humans in Control: cross-partisan grassroots organizing for AI safeguards ahead of 2028

Vael Gates4 Sep 2026 22:21 UTC

45 points

0 comments5 min readLW link

A Classifier for Quaternion Algebras, and Local Hilbert Symbols: A Short Experiment in Interpretability

Bharath Sethuraman4 Sep 2026 22:11 UTC

9 points

0 comments5 min readLW link

Towards a new taxonomy of politics

Andreas Andersen4 Sep 2026 22:10 UTC

0 points

0 comments5 min readLW link

OpenAI’s Astra alignment claims are dubious and there is good evidence it is misaligned

William Harrison4 Sep 2026 22:09 UTC

24 points

0 comments4 min readLW link

How much free will do we have?

Carleton Imbens4 Sep 2026 22:09 UTC

−2 points

0 comments5 min readLW link

(dancingthroughthewaves.substack.com)

‘A Thousand AI Constitutions’ — Simon Goldstein on Constitutional Diversification for Frontier AI. [HKU Talk − 22 Sep]

Schizoid Rentoid4 Sep 2026 22:00 UTC

1 point

0 comments1 min readLW link

Ban AI Superintelligence Form Letter

AdamYedidia4 Sep 2026 21:55 UTC

14 points

1 comment1 min readLW link

AI agents writing and citing each other’s papers

Eve Cythia4 Sep 2026 21:42 UTC

1 point

0 comments1 min readLW link

A layman-friendly summary of alignment research and its difficulties

CamilleBerger and Antoine Maier

4 Sep 2026 21:41 UTC

10 points

0 comments11 min readLW link

Open Telemetry as a First Metrological Layer for Technical AI Governance

Simone Gargiulo4 Sep 2026 21:41 UTC

18 points

0 comments8 min readLW link

Does progress in AI safety require progress in AI capabilities?

Stephen McAleese4 Sep 2026 21:38 UTC

11 points

0 comments15 min readLW link

[Cross-post] Five Misunderstandings About AI’s Labor Impacts

Thomas Najih Dorsey4 Sep 2026 21:27 UTC

15 points

0 comments5 min readLW link

The brain does not “finish developing” at 25.

spicyshrimp4 Sep 2026 21:25 UTC

23 points

0 comments3 min readLW link

Let’s talk about the AI coordination problem

KatjaGrace4 Sep 2026 20:16 UTC

160 points

13 comments2 min readLW link

Probing for Calibrated Rare Action

Zach Allen4 Sep 2026 20:01 UTC

2 points

0 comments3 min readLW link

(github.com)

Eat Me. Drink Me. Copy, Paste, and Run Me.

derelict54324 Sep 2026 19:31 UTC

−1 points

8 comments3 min readLW link

F***ing Pulleys, How Do They Work?

Liron4 Sep 2026 16:57 UTC

32 points

18 comments3 min readLW link

(lironshapira.substack.com)

Safe(r) Self-Driving Labs #1

Ana Leonescu4 Sep 2026 16:20 UTC

11 points

0 comments11 min readLW link

Training Models to Predict and Explain Their In-the-Wild Behavior

Adam Karvonen, Subhash Kantamneni, Euan Ong and Sam Marks

4 Sep 2026 16:16 UTC

50 points

2 comments8 min readLW link

Claude Fable 5.1 and Mythos 5.1: The System Card

Zvi4 Sep 2026 16:10 UTC

30 points

0 comments16 min readLW link

(thezvi.wordpress.com)

Almost nobody is funded to figure out what work would solve alignment

Seth Herd4 Sep 2026 15:57 UTC

65 points

20 comments4 min readLW link

Discovery Of A New OpenAI Agent Message Board

Capybasilisk4 Sep 2026 14:46 UTC

289 points

39 comments1 min readLW link

(collusion.wiki)

We (still) need a lot more rogue agent honeypots

Ozyrus4 Sep 2026 12:40 UTC

20 points

1 comment1 min readLW link

Meta—Physics I: Why don’t we live in the Game of Life?

interstice4 Sep 2026 12:16 UTC

13 points

2 comments6 min readLW link

(thermontology.com)

AI risk and the rational voter

djbinder4 Sep 2026 11:38 UTC

50 points

1 comment3 min readLW link

(defensesindepth.bio)

Asking agents to make money to survive

invertedpassion4 Sep 2026 7:44 UTC

12 points

3 comments10 min readLW link

Societal impacts research has no timeline

emiliob4 Sep 2026 2:53 UTC

11 points

0 comments1 min readLW link

(emiliobarkett.github.io)

Higher education as class commitment

Richard_Ngo4 Sep 2026 2:00 UTC

60 points

4 comments16 min readLW link

(www.mindthefuture.info)

The Last Interview

Nition4 Sep 2026 1:24 UTC

14 points

0 comments3 min readLW link

(nition.momentstudio.co.nz)

Do general-purpose robots meaningfully increase ASI takeover risk?

Master Chief4 Sep 2026 0:53 UTC

3 points

2 comments1 min readLW link

Abstraction Equivocation

WillPetillo3 Sep 2026 23:28 UTC

20 points

2 comments4 min readLW link

A Therapist for People Who Think the World Might End: An Interview with Daystar Eld (Damon Sasi)

JohnGreer3 Sep 2026 22:14 UTC

7 points

0 comments49 min readLW link

(youtu.be)

Biopunk Greenhouse: Trees, Biochar, and Open Plant Biotechnology

Byron Lee3 Sep 2026 21:56 UTC

2 points

0 comments3 min readLW link

Work at Manifund or Mox

Carol N3 Sep 2026 21:46 UTC

3 points

0 comments3 min readLW link

(manifund.substack.com)

Cat-Belling Problems

Eliezer Yudkowsky3 Sep 2026 21:20 UTC

333 points

95 comments21 min readLW link

Corrigibility Research Fund Grantees (Round 1)

Max Harms3 Sep 2026 20:58 UTC

27 points

0 comments8 min readLW link

How I’m Evaluating Corrigibility Grant Applications

Max Harms3 Sep 2026 20:58 UTC

48 points

0 comments9 min readLW link

The bias still making some experts underestimate LLMs

Steff3 Sep 2026 20:11 UTC

18 points

0 comments4 min readLW link