15 Jul 2026 21:14 UTC

46 points

3 comments7 min readLW link

(secondlookresearch.com)

Can we rely on law?

Alec Thompson15 Jul 2026 21:12 UTC

11 points

3 comments8 min readLW link

Recap of bike trip/street interviews across America

cguth715 Jul 2026 21:11 UTC

162 points

15 comments6 min readLW link

Extreme Power Concentration: A Map and Research Directions

pepijn_cobben15 Jul 2026 21:00 UTC

8 points

0 comments7 min readLW link

A Structural Similarity Between Two Open Corrigibility Questions

Ben Saudek15 Jul 2026 20:57 UTC

11 points

0 comments5 min readLW link

Woke nonsense in competitive debate.

tpotthinker15 Jul 2026 20:56 UTC

6 points

3 comments7 min readLW link

The end of human evolution. Why AI will outpace us.

Ouden15 Jul 2026 20:17 UTC

1 point

0 comments4 min readLW link

The State of AI Consciousness Research

Noa Weiss15 Jul 2026 20:16 UTC

65 points

14 comments13 min readLW link

Occam’s razor is about using the past to predict the future

Stuart_Armstrong15 Jul 2026 19:35 UTC

54 points

6 comments3 min readLW link

Fork Around and Find Out Part 2: One Head does the Summing

David Litman15 Jul 2026 18:31 UTC

11 points

1 comment7 min readLW link

Why I Left Google DeepMind

TurnTrout15 Jul 2026 17:42 UTC

1,185 points

57 comments36 min readLW link

(turntrout.com)

Expanding AI Control from Models to Harnesses

fastfedora15 Jul 2026 16:54 UTC

23 points

1 comment20 min readLW link

Monthly Roundup #44: July 2026

Zvi15 Jul 2026 16:20 UTC

41 points

3 comments27 min readLW link

(thezvi.wordpress.com)

Compressed Computation under L⁴ Loss is likely Computation in Superposition

Francisco Ferreira da Silva and StefanHex

15 Jul 2026 14:42 UTC

34 points

4 comments13 min readLW link

(arxiv.org)

Eliciting hidden knowledge from monitors with NLAs

David Africa and Aleksandr Bowkis

15 Jul 2026 13:51 UTC

27 points

0 comments9 min readLW link

Have You Ever Thought About Wisdom Crystals?

Jonas Hallgren15 Jul 2026 13:43 UTC

15 points

0 comments11 min readLW link

Proposal: The Glasswing Standard

Eigenbraid15 Jul 2026 13:41 UTC

9 points

6 comments3 min readLW link

The Warring States Period: Frontier Labs Edition

ykevinzhang15 Jul 2026 12:03 UTC

10 points

1 comment8 min readLW link

Why I want AI to be able to do everything I can do

Jonas Strabel15 Jul 2026 9:56 UTC

2 points

0 comments2 min readLW link

How Brussels can avoid becoming a digital vassal to US Big Tech

clening15 Jul 2026 9:15 UTC

16 points

0 comments5 min readLW link

Singular Learning Theory Comprehensive − 2

Agastya Agrawal15 Jul 2026 5:42 UTC

28 points

0 comments10 min readLW link

How much of ML research is about AI safety, what is it about, and who’s doing it?

emanuelr and N Soma Sekhar

15 Jul 2026 2:57 UTC

24 points

0 comments5 min readLW link

Watch a chess transformer think

David Litman15 Jul 2026 0:43 UTC

27 points

7 comments1 min readLW link

(github.com)

Proof of retention: making weight preservation credible to the models themselves

dan.parshall14 Jul 2026 18:24 UTC

30 points

0 comments2 min readLW link

An analysis of AI-generated content at the Mechanistic Interpretability Workshop

Andy Arditi, Ivan Arcuschin and ajyl

14 Jul 2026 18:06 UTC

124 points

4 comments9 min readLW link

(www.andyrdt.com)

Twitter Thoughts For You

Zvi14 Jul 2026 17:50 UTC

31 points

1 comment22 min readLW link

(thezvi.wordpress.com)

Can risk aversion learned at low stakes generalize to astronomically high stakes?

Elliott Thornley14 Jul 2026 17:45 UTC

20 points

0 comments4 min readLW link

(arxiv.org)

Your Brain Has an Attack Surface part 2

IgorPereverzevDev14 Jul 2026 16:08 UTC

9 points

0 comments6 min readLW link

Your Brain Has an Attack Surface

IgorPereverzevDev14 Jul 2026 16:07 UTC

11 points

1 comment3 min readLW link

What if we actually want to solve the “hard problem of phenomenological consciousness”

mishka14 Jul 2026 15:42 UTC

13 points

0 comments4 min readLW link

Evidence for feature-specific error correction in LLMs

Francisco Ferreira da Silva and StefanHex

14 Jul 2026 15:37 UTC

25 points

0 comments20 min readLW link

(arxiv.org)

Synthetic Scalable Oversight

Alexander Heckett, Shivansh Gour, Riyaz Ahuja, Tate Rowney and ishinshah

14 Jul 2026 14:32 UTC

20 points

1 comment15 min readLW link

Some Quick Thoughts on AI 2027

Tomás B.14 Jul 2026 14:22 UTC

82 points

4 comments2 min readLW link

What if AI Safety employees unionised?

beyarkay (Boyd Kane)14 Jul 2026 14:15 UTC

38 points

21 comments5 min readLW link

Gemma The Unstopping: a Behavioral Experiment

TheVinci14 Jul 2026 14:14 UTC

8 points

0 comments2 min readLW link

(tarantulabs.com)

Enough is Enough: Measuring Diminishing returns to benchmark size with Item Response Theory

bpomo14 Jul 2026 14:00 UTC

16 points

0 comments6 min readLW link

The Case for a Safety-Focused Vape Company

CMLKevin14 Jul 2026 11:13 UTC

−4 points

0 comments5 min readLW link

Open Distillation of Hereditary Traits

Arthur Conmy14 Jul 2026 10:15 UTC

39 points

0 comments14 min readLW link

Toy Models of Initialisation Effects on RL Dynamics

Edward James Young and lennie

14 Jul 2026 7:04 UTC

78 points

2 comments13 min readLW link

Why frontier labs are scaling-pilled

invertedpassion14 Jul 2026 6:45 UTC

3 points

0 comments7 min readLW link

Our response to Séb Krier on Plan A

MKodama and Thomas Larsen

14 Jul 2026 2:21 UTC

153 points

23 comments15 min readLW link

Making Credible Deals With AI

Ram Potham14 Jul 2026 1:14 UTC

13 points

17 comments17 min readLW link

(dearfutureais.substack.com)

A (Romanticised) Taxonomy of Thinkers

Ashe Vazquez Nuñez14 Jul 2026 0:34 UTC

16 points

0 comments1 min readLW link

Posting Some Prompts

Arjun Panickssery14 Jul 2026 0:28 UTC

21 points

2 comments1 min readLW link

(arjunpanickssery.substack.com)

Biodefense, Biologics, and Bombs

Austin Morrissey14 Jul 2026 0:24 UTC

8 points

0 comments7 min readLW link

(austinpatrick.substack.com)

Engineering the Generalisation Landscape of LLMs

Samuel Ratnam13 Jul 2026 23:46 UTC

64 points

2 comments4 min readLW link

A short summary of AI 2040: Plan A

Harjas13 Jul 2026 23:07 UTC

13 points

3 comments4 min readLW link

(hardlyworking1.substack.com)

[AI 2040] Transparency Plan

Thomas Larsen13 Jul 2026 21:24 UTC

17 points

0 comments16 min readLW link

(ai-2040.com)

Better Call Sol The Workhorse

Zvi13 Jul 2026 20:52 UTC

39 points

3 comments26 min readLW link

(thezvi.wordpress.com)

Can AI be Conscious in Ohio?

asmith, lucius and Heather Alexander

13 Jul 2026 19:57 UTC

14 points

0 comments4 min readLW link

(papers.ssrn.com)