N Di­men­sional In­ter­ac­tive Scat­ter Plot (ndisp)

TristanTrim15 Aug 2025 23:08 UTC
10 points
3 comments12 min readLW link

SE Gyges’ re­sponse to AI-2027

StanislavKrym15 Aug 2025 21:54 UTC
29 points
13 comments46 min readLW link
(www.verysane.ai)

Towards data-cen­tric in­ter­pretabil­ity with sparse autoencoders

15 Aug 2025 20:10 UTC
53 points
2 comments18 min readLW link

Mu­sic taste is (also) a next to­ken prediction

eamag15 Aug 2025 17:49 UTC
5 points
0 comments2 min readLW link
(eamag.me)

The­ory of cul­ture as waste.

Laureana Bonaparte15 Aug 2025 17:34 UTC
−3 points
15 comments2 min readLW link

Spend­ing Too Much Time At Airports

Zvi15 Aug 2025 16:10 UTC
57 points
24 comments7 min readLW link
(thezvi.wordpress.com)

How to make the fu­ture bet­ter (other than by re­duc­ing ex­tinc­tion risk)

wdmacaskill15 Aug 2025 15:40 UTC
19 points
1 comment3 min readLW link

Should you start a for-profit AI safety org?

KatWoods15 Aug 2025 13:52 UTC
8 points
4 comments1 min readLW link

How to get ChatGPT to re­ally thor­oughly re­search something

KatWoods15 Aug 2025 12:54 UTC
18 points
1 comment1 min readLW link

Thoughts on Grad­ual Disempowerment

Tom Davidson15 Aug 2025 11:56 UTC
62 points
32 comments19 min readLW link

Misal­ign­ment clas­sifiers: Why they’re hard to eval­u­ate ad­ver­sar­i­ally, and why we’re study­ing them anyway

15 Aug 2025 11:48 UTC
59 points
3 comments17 min readLW link

A Phy­logeny of Agents

15 Aug 2025 10:47 UTC
40 points
12 comments6 min readLW link
(substack.com)

My kids won’t be workers

Gauraventh15 Aug 2025 7:06 UTC
3 points
0 comments6 min readLW link
(y1d2.com)

Euro­pean Links (15.08.25)

Martin Sustrik15 Aug 2025 4:20 UTC
21 points
8 comments2 min readLW link
(www.250bpm.com)

Le­gal Per­son­hood—Three Prong Bun­dle Theory

Stephen Martin15 Aug 2025 4:13 UTC
13 points
6 comments4 min readLW link

Men­tal Gym­nas­tics.

Laureana Bonaparte15 Aug 2025 4:08 UTC
3 points
0 comments13 min readLW link

Rare AI and the Fermi Paradox

dawnstrata15 Aug 2025 4:05 UTC
11 points
6 comments9 min readLW link

Tris­tan’s Projects

TristanTrim15 Aug 2025 3:46 UTC
6 points
4 comments2 min readLW link

Tri­al­ing Far UVC and Gly­col Va­pors at BIDA

jefftk15 Aug 2025 2:20 UTC
19 points
1 comment2 min readLW link
(www.jefftk.com)

A philo­soph­i­cal ker­nel: bit­ing an­a­lytic bullets

jessicata15 Aug 2025 1:35 UTC
64 points
21 comments13 min readLW link
(unstableontology.com)

A let­ter to Kyle Fish on the Re­tire­ment of Claude 3 Sonnet

bridgebot15 Aug 2025 1:08 UTC
−4 points
3 comments5 min readLW link

Con­cep­tual Rhyme and Metaphor

Jordan Rubin15 Aug 2025 0:05 UTC
2 points
0 comments9 min readLW link
(jordanmrubin.substack.com)

Train­ing a Re­ward Hacker De­spite Perfect Labels

14 Aug 2025 23:57 UTC
132 points
45 comments4 min readLW link

AGI: Prob­a­bly Not 2027

Tomás B.14 Aug 2025 22:24 UTC
16 points
8 comments1 min readLW link
(www.verysane.ai)

Four Axes of Hunger

Brendan Long14 Aug 2025 19:03 UTC
25 points
3 comments2 min readLW link

Some­body in­vented a bet­ter bookmark

Alex_Altair14 Aug 2025 17:57 UTC
173 points
22 comments2 min readLW link

In defense of the amy­loid hypothesis

dsj14 Aug 2025 17:52 UTC
43 points
0 comments1 min readLW link
(www.astralcodexten.com)

A Prac­ti­cal Tool for Map­ping and Quan­tify­ing Belief Networks

Zack Friedman14 Aug 2025 17:22 UTC
7 points
0 comments1 min readLW link

AI #129: Com­i­cally Unconstitutional

Zvi14 Aug 2025 14:10 UTC
47 points
3 comments55 min readLW link
(thezvi.wordpress.com)

Health­care as education

Coafos14 Aug 2025 13:31 UTC
4 points
0 comments3 min readLW link

About Stress

Gabriel Alfour14 Aug 2025 10:33 UTC
25 points
0 comments1 min readLW link
(cognition.cafe)

Le­gal Per­son­hood—The “En­force­ment Gap”

Stephen Martin14 Aug 2025 6:07 UTC
8 points
0 comments3 min readLW link

Sleep­ing Machines: Why Our AI Agents Still Be­have Like Ta­lented Children

Michal Barodkin14 Aug 2025 2:31 UTC
23 points
4 comments8 min readLW link

Ex­plor­ing the “Anti-TESCREAL” Ide­ol­ogy and the Roots of (Anti-)Progress

Ottokar Hochman14 Aug 2025 2:30 UTC
23 points
2 comments2 min readLW link
(recapitulation.substack.com)

A YouTube Video Will Prob­a­bly Never Help You Quit YouTube

boundary_condition14 Aug 2025 0:59 UTC
26 points
11 comments10 min readLW link

Should you make stone tools?

Alex_Altair14 Aug 2025 0:15 UTC
190 points
48 comments3 min readLW link

METR Re­search Up­date: Al­gorith­mic vs. Holis­tic Evaluation

David Rein13 Aug 2025 22:47 UTC
101 points
7 comments1 min readLW link
(metr.org)

In­te­ri­ors can be more fun

Nina Panickssery13 Aug 2025 22:42 UTC
34 points
6 comments4 min readLW link
(blog.ninapanickssery.com)

Against Epistemic Democ­racy: A Epistemic Tier List of What Ac­tu­ally Works

Linch13 Aug 2025 21:28 UTC
9 points
3 comments1 min readLW link
(linch.substack.com)

Good Faith Arguments

Gordon Seidoh Worley13 Aug 2025 20:50 UTC
1 point
0 comments3 min readLW link
(uncertainupdates.substack.com)

Do­ing A Thing Puts You in The Top 10% (And That Sucks)

Brendan Long13 Aug 2025 19:50 UTC
74 points
23 comments2 min readLW link

In­trigu­ing Prop­er­ties of gpt-oss Jailbreaks

13 Aug 2025 19:42 UTC
14 points
0 comments10 min readLW link
(xlabaisecurity.com)

ChatGPT Caused Psy­chosis via Poisoning

Adele Lopez13 Aug 2025 19:15 UTC
18 points
2 comments1 min readLW link

Tech Tree for Se­cure Mul­tipo­lar AI

13 Aug 2025 17:18 UTC
11 points
3 comments2 min readLW link

Launch­ing new AIXI re­search com­mu­nity web­site + read­ing group(s)

Cole Wyeth13 Aug 2025 17:09 UTC
46 points
2 comments1 min readLW link

AI de­vel­op­ment as the first fully-au­to­mated job

tailcalled13 Aug 2025 16:45 UTC
17 points
4 comments1 min readLW link

Prob­ing Power-Seek­ing in LLMs

Moksh Nirvaan13 Aug 2025 16:04 UTC
6 points
0 comments12 min readLW link

GPT-5s Are Alive: Synthesis

Zvi13 Aug 2025 14:10 UTC
44 points
1 comment31 min readLW link
(thezvi.wordpress.com)

Books, maps, and teachings

Richard_Kennaway13 Aug 2025 11:44 UTC
14 points
1 comment3 min readLW link

En­light­en­ment AMA

lsusr13 Aug 2025 9:11 UTC
68 points
131 comments1 min readLW link