Pathological Narcissism: The Pendulum Swing between Echoism and Sovereignism

Dawn Drescher22 Jun 2026 21:56 UTC

5 points

0 comments8 min readLW link

(impartial-priorities.org)

Speedup from AI Ghostwriting

Dawn Drescher22 Jun 2026 21:47 UTC

10 points

0 comments5 min readLW link

(impartial-priorities.org)

Functional Emotions and The Pope’s Encyclical on AI — Digital Minds Newsletter #3

22 Jun 2026 19:44 UTC

14 points

0 comments39 min readLW link

(www.digitalminds.news)

Planning for Preservation in the Age of AI

Raelifin22 Jun 2026 17:51 UTC

19 points

0 comments13 min readLW link

Advocates Can Influence LLM Values By Editing Wikipedia

jonahmattwoodward and Jasmine Brazilek

22 Jun 2026 15:47 UTC

5 points

4 comments8 min readLW link

Learning to Understand Evil

Yulia22 Jun 2026 14:43 UTC

5 points

0 comments7 min readLW link

Defeatism as Disempowerment

Ramya22 Jun 2026 14:24 UTC

4 points

0 comments3 min readLW link

A Mechanistic Explanation of Prompt Injection (and why you should study roles)

Charles Ye and lesscertain

22 Jun 2026 14:09 UTC

57 points

6 comments16 min readLW link

The AI Industrial Explosion — Part 4: Cheap power

djbinder22 Jun 2026 13:30 UTC

32 points

0 comments16 min readLW link

(defensesindepth.bio)

Not all features are created equal

enricobottazzi22 Jun 2026 13:20 UTC

17 points

0 comments9 min readLW link

Brittle model organisms obstructs deception elicitation work

Advik, Daniel Tan and Chloe Li

22 Jun 2026 10:48 UTC

19 points

3 comments7 min readLW link

We made a map of the doom debate. Here’s how the breakdown works.

Sean Herrington, Khai Tran, David Bravo, keivnc, Paul Hindoian, Josh Tuffy, mikaelacankosyan, Christopher A. Davis and Maryam Hampaei

22 Jun 2026 10:48 UTC

13 points

0 comments7 min readLW link

Naturally learned behaviors in deep MLPs resist detection by both human and learned algorithms

emanuelr22 Jun 2026 9:10 UTC

10 points

0 comments12 min readLW link

A US-China intervention because we can’t expect another Arkhipov/Petrov

Troy Tian22 Jun 2026 7:31 UTC

10 points

0 comments1 min readLW link

On revolutionary love in AI safety

Troy Tian22 Jun 2026 3:48 UTC

8 points

0 comments4 min readLW link

Do AI Biorisk Thresholds Need Intermediate Warning Levels?

Lukas Frei22 Jun 2026 1:09 UTC

9 points

0 comments3 min readLW link

NLA explanations can be shortened without harming reconstruction

loops22 Jun 2026 0:57 UTC

48 points

4 comments3 min readLW link

Introducing MonitoringBench

monika_j21 Jun 2026 18:43 UTC

44 points

0 comments6 min readLW link

How persona training could fail

Simon Lermen21 Jun 2026 16:38 UTC

13 points

0 comments4 min readLW link

A high-level model of AI bargaining

Anthony DiGiovanni21 Jun 2026 15:37 UTC

18 points

1 comment5 min readLW link