When they can perform a task, AIs are much cheaper than humans

djbinder27 Sep 2026 23:44 UTC

20 points

0 comments10 min readLW link

(defensesindepth.bio)

Securing AI Research Needs an Owner

Jan Michelfeit27 Sep 2026 23:10 UTC

10 points

0 comments10 min readLW link

Game-Theoretic Disempowerment: Loss of Control to Agencyless AI

RedMan27 Sep 2026 22:25 UTC

11 points

0 comments4 min readLW link

Sector: de-slopping AI-driven research

Taras Kutsyk27 Sep 2026 22:10 UTC

8 points

0 comments7 min readLW link

Why research personas despite RL scaling?

Cleo Nardo27 Sep 2026 20:29 UTC

38 points

4 comments1 min readLW link

Skeuomorphic AI Safety

Roko27 Sep 2026 2:32 UTC

11 points

2 comments4 min readLW link

My best anti-doom argument

Hazard_real27 Sep 2026 0:49 UTC

10 points

4 comments12 min readLW link

(hazard3.substack.com)

Why I expect AI replication incidents by 2027

Reworr R26 Sep 2026 23:58 UTC

9 points

1 comment4 min readLW link

Extinction does not feel as bad as it should

mxbutterfree26 Sep 2026 22:55 UTC

12 points

7 comments5 min readLW link

[Linkpost] Looking into the Swarm’s Eye

nwyin26 Sep 2026 22:03 UTC

15 points

2 comments1 min readLW link

(florianbrand.com)

Plan R+, Diversity, Escrow and Political Rights for ASICs

Roko26 Sep 2026 18:59 UTC

12 points

1 comment4 min readLW link

Existential Risk Is An Extraordinary Claim

James Brobin26 Sep 2026 14:44 UTC

−3 points

14 comments3 min readLW link

Writing a Theorem Prover from scratch

astle dsa26 Sep 2026 10:38 UTC

14 points

2 comments11 min readLW link

Poverty in the midst of abundance: AI will make goods cheaper, but your labor will get cheaper faster

cousin_it26 Sep 2026 7:54 UTC

92 points

14 comments1 min readLW link

Addictions are anesthesia

Chris Lakin26 Sep 2026 6:55 UTC

61 points

13 comments2 min readLW link

(chrislakin.blog)

Evidence about risk should be transparent

Ajeya Cotra25 Sep 2026 23:39 UTC

170 points

4 comments4 min readLW link

Plan R: AI Safety by ASICs

Roko25 Sep 2026 23:26 UTC

122 points

29 comments5 min readLW link

Alignment Forecasting: Predicting Misalignment from Training Data

Yueh Han "John" Chen, Bruce W. Lee, Ilia Sucholutsky and Tomek Korbak

25 Sep 2026 18:30 UTC

23 points

0 comments7 min readLW link

Spurious probes as a black-box alternative to activation probing

Ziqian Zhong25 Sep 2026 18:26 UTC

58 points

2 comments9 min readLW link

We need a better theory of polarization, because it’s failing to predict the AI debate

less_raichu25 Sep 2026 17:57 UTC

91 points

55 comments6 min readLW link