Agents as Webs of Beliefs

Richard_Ngo27 Jun 2026 21:45 UTC

32 points

1 comment9 min readLW link

(www.mindthefuture.info)

The Prisoner’s Dilemma (Fiction)

testingthewaters27 Jun 2026 20:51 UTC

1 point

0 comments9 min readLW link

Neuralese is Actually Probably Good for Alignment

DaemonicSigil27 Jun 2026 19:40 UTC

14 points

7 comments6 min readLW link

Should we just quit?

Jacob Abraham27 Jun 2026 13:44 UTC

1 point

0 comments3 min readLW link

Flipping the eval on its head

Quinn27 Jun 2026 13:34 UTC

12 points

0 comments3 min readLW link

Deployment Awareness Matters More Than Evaluation Awareness

VojtaKovarik, Tomáš Gavenčiak and Mateusz Bagiński

26 Jun 2026 22:54 UTC

35 points

3 comments7 min readLW link

(limits-of-evaluation.org)

Just a Wrapper? How Much Do Scaffolds Matter?

Hans Gundlach, Zachary Brown and jaysonl

26 Jun 2026 22:21 UTC

9 points

0 comments11 min readLW link

(mitfuturetech.substack.com)

What did “scheming” and “mech interp” mean pre-2023?

Cleo Nardo26 Jun 2026 22:09 UTC

69 points

1 comment2 min readLW link

Why are adversaries assumed to be incapable of responding to AI risk?

KatjaGrace26 Jun 2026 21:51 UTC

91 points

7 comments1 min readLW link

(worldspiritsockpuppet.substack.com)

Screencasts could be scalable data + evals for single-user emulation (Guardian Angels)

sophia_xu26 Jun 2026 21:12 UTC

4 points

0 comments1 min readLW link

Randomness Testing with Bayesian Stats

DaemonicSigil26 Jun 2026 20:28 UTC

8 points

0 comments5 min readLW link

Should we combine protocols for AI Control Research?

Ram Potham and Tyler Tracy

26 Jun 2026 19:02 UTC

9 points

0 comments9 min readLW link

Not making a strong argument is a relief

Kaj_Sotala26 Jun 2026 18:41 UTC

43 points

5 comments10 min readLW link

(kajsotala.substack.com)

The Case for Model Forensics

aditya singh, gersonkroiz, Senthooran Rajamanoharan and Neel Nanda

26 Jun 2026 15:09 UTC

44 points

0 comments10 min readLW link

A full body MRI earns you a year of smoking

kqr26 Jun 2026 14:31 UTC

11 points

7 comments3 min readLW link

(entropicthoughts.com)

Don’t ignore the car crashes, and remember your freshman CS

jcksanderson26 Jun 2026 7:06 UTC

36 points

0 comments2 min readLW link

(jcksanderson.com)

LLMenard

Ilyass Mofaddel26 Jun 2026 4:26 UTC

1 point

0 comments5 min readLW link

Research note on negated reward hacking

ChristopherT26 Jun 2026 2:22 UTC

8 points

0 comments6 min readLW link

Surprising facts about the slave trade

Joseph Miller26 Jun 2026 1:34 UTC

255 points

12 comments8 min readLW link

Intelligence (Artificial) in a Fallen World

Joseph Babbo25 Jun 2026 23:58 UTC

3 points

0 comments11 min readLW link