A Bayesian Ex­pla­na­tion of Causal Models

Menotim27 Oct 2025 23:16 UTC
2 points
0 comments25 min readLW link

Brain­storm­ing Food on the Cheap + Healthy + Con­ve­nient + Edible Frontier

Morpheus27 Oct 2025 23:04 UTC
19 points
3 comments4 min readLW link

Trans­ac­tional method for non-trans­ac­tional re­la­tion­ship: Re­la­tion­ship as a Com­mon-pool Re­source problem

David H.27 Oct 2025 22:29 UTC
2 points
0 comments7 min readLW link

[Question] How Im­por­tant is In­vert­ing LLMs?

Maloew27 Oct 2025 20:59 UTC
8 points
1 comment1 min readLW link

Ask­ing (Some Of) The Right Questions

Zvi27 Oct 2025 19:00 UTC
31 points
3 comments14 min readLW link
(thezvi.wordpress.com)

life les­sons from trading

thiccythot27 Oct 2025 16:56 UTC
43 points
3 comments4 min readLW link

Agen­tic Mon­i­tor­ing for AI Control

LAThomson27 Oct 2025 16:38 UTC
9 points
0 comments9 min readLW link

Model Pa­ram­e­ters as a Stegano­graphic Pri­vate Channel

Lennart Finke27 Oct 2025 16:08 UTC
9 points
0 comments5 min readLW link

Ma­jor sur­vey on the HS/​TS spec­trum and gAyGP

tailcalled27 Oct 2025 14:31 UTC
22 points
3 comments8 min readLW link

Death of the Author

J Bostock27 Oct 2025 12:35 UTC
5 points
0 comments3 min readLW link

Ex­plor­ing the multi-di­men­sional re­fusal sub­space in rea­son­ing models

Le magicien quantique27 Oct 2025 9:03 UTC
5 points
2 comments4 min readLW link

AIs should also re­fuse to work on ca­pa­bil­ities research

Davidmanheim27 Oct 2025 8:42 UTC
150 points
20 comments3 min readLW link

Un­com­mon Utili­tar­i­anism #3: Bounded Utility Functions

Alice Blair27 Oct 2025 5:06 UTC
16 points
10 comments6 min readLW link

List of lists of pro­ject ideas in AI Safety

Veronica Gordi27 Oct 2025 1:28 UTC
6 points
0 comments14 min readLW link
(www.notion.so)

[Question] How valuable is money-in-mar­ket?

Hruss27 Oct 2025 0:47 UTC
6 points
1 comment1 min readLW link

Credit goes to the pre­sen­ter, not the inventor

Algon26 Oct 2025 23:55 UTC
42 points
5 comments3 min readLW link

On Flesh­ling Safety: A De­bate by Klurl and Tra­pau­cius.

Eliezer Yudkowsky26 Oct 2025 23:44 UTC
253 points
52 comments79 min readLW link

Re­sults of “Ex­per­i­ment on Bernoulli pro­cesses”

joseph_c26 Oct 2025 21:47 UTC
9 points
2 comments4 min readLW link

cer­tain ex­otic neu­ro­trans­mit­ters as SMART PILLS: or com­pounds that in­crease the ca­pac­ity for men­tal work in humans

azergante26 Oct 2025 20:51 UTC
4 points
0 comments22 min readLW link
(erowid.org)

Cancer has a sur­pris­ing amount of detail

Abhishaike Mahajan26 Oct 2025 20:33 UTC
127 points
18 comments11 min readLW link
(www.owlposting.com)

Sta­bil­ity of nat­u­ral la­tents in in­for­ma­tion the­o­retic terms

Aram Ebtekar26 Oct 2025 20:33 UTC
35 points
0 comments2 min readLW link

Les­sons from Teach­ing Ra­tion­al­ity to EAs in the Netherlands

Shoshannah Tekofsky26 Oct 2025 20:03 UTC
20 points
0 comments7 min readLW link
(forum.effectivealtruism.org)

Are We Their Chimps?

soycarts26 Oct 2025 16:04 UTC
−7 points
49 comments1 min readLW link

FWIW: What I no­ticed at a (Goenka) Vi­pas­sana retreat

David Gross26 Oct 2025 15:10 UTC
38 points
4 comments9 min readLW link

Brightline is Ac­tu­ally Pretty Dangerous

jefftk26 Oct 2025 12:51 UTC
53 points
12 comments3 min readLW link
(www.jefftk.com)

Seven-ish Words from My Thought-Language

Lorxus26 Oct 2025 4:30 UTC
68 points
13 comments4 min readLW link
(tiled-with-pentagons.blogspot.com)

Remembrancy

Algon25 Oct 2025 22:47 UTC
11 points
0 comments3 min readLW link

Pyg­mal­ion’s Wafer

Charlie Sanders25 Oct 2025 20:17 UTC
8 points
2 comments4 min readLW link
(www.dailymicrofiction.com)

De­bat­ing theism

Ivan25 Oct 2025 18:35 UTC
−21 points
0 comments25 min readLW link

[Question] Why is OpenAI re­leas­ing prod­ucts like Sora and At­las?

J Thomas Moros25 Oct 2025 17:59 UTC
16 points
10 comments1 min readLW link

Ori­gins and dan­gers of fu­ture AI ca­pa­bil­ity denial

Patrick Spencer25 Oct 2025 16:13 UTC
68 points
18 comments10 min readLW link

Do you com­pletely trust that you are com­pletely in the shit? - de­spair and in­for­ma­tion -

P. João25 Oct 2025 14:42 UTC
−2 points
17 comments3 min readLW link

Assess­ing Far UVC Positioning

jefftk25 Oct 2025 14:00 UTC
20 points
3 comments2 min readLW link
(www.jefftk.com)

Mus­ings on Re­ported Cost of Com­pute (Oct 2025)

Vladimir_Nesov24 Oct 2025 20:42 UTC
103 points
11 comments2 min readLW link

Re­gard­less of X, you can still just sign su­per­in­tel­li­gence-state­ment.org if you agree

Ishual24 Oct 2025 20:30 UTC
58 points
0 comments3 min readLW link

The Fu­ture of In­ter­pretabil­ity is Geometric

sbaumohl24 Oct 2025 18:32 UTC
23 points
0 comments5 min readLW link

New State­ment Calls For Not Build­ing Su­per­in­tel­li­gence For Now

Zvi24 Oct 2025 17:40 UTC
80 points
3 comments7 min readLW link
(thezvi.wordpress.com)

Notes on “Ex­plain­ing AI Ex­plain­abil­ity”

Eleni Angelou24 Oct 2025 17:22 UTC
20 points
0 comments6 min readLW link

Can Rea­son­ing Models Obfus­cate Rea­son­ing? Stress-Test­ing Chain-of-Thought Monitorability

24 Oct 2025 17:21 UTC
17 points
1 comment5 min readLW link

I will not sign up for cryonics

Syd Lonreiro_24 Oct 2025 16:56 UTC
−18 points
5 comments1 min readLW link

Dol­lars in poli­ti­cal giv­ing are less fun­gible than you might think

lincolnquirk24 Oct 2025 15:54 UTC
6 points
1 comment5 min readLW link
(lincolnquirk.substack.com)

Can AI Agents with Diver­gent In­ter­ests Learn To Prevent Civ­i­liza­tional Failures?

joao_abrantes24 Oct 2025 15:08 UTC
1 point
0 comments1 min readLW link

LW Re­acts pack for Dis­cord/​Slack/​etc

plex24 Oct 2025 13:20 UTC
65 points
13 comments1 min readLW link
(drive.google.com)

AI Timelines and Points of no return

Gabriel Alfour24 Oct 2025 11:15 UTC
36 points
8 comments1 min readLW link
(cognition.cafe)

In­tro­duc­ing Con­trolArena: A library for run­ning AI con­trol experiments

Mojmir24 Oct 2025 9:51 UTC
13 points
0 comments3 min readLW link
(www.aisi.gov.uk)

Can we steer AI mod­els to­ward safer ac­tions by mak­ing these in­stru­men­tally use­ful?

Francesca Gomez24 Oct 2025 9:18 UTC
5 points
0 comments2 min readLW link
(www.wiserhuman.ai)

Plan 1 and Plan 2

Towards_Keeperhood24 Oct 2025 8:18 UTC
50 points
22 comments3 min readLW link

Guys I might be an e/​acc

Taylor G. Lunt24 Oct 2025 3:25 UTC
14 points
29 comments4 min readLW link

How an AI com­pany CEO could quietly take over the world

Alex Kastner23 Oct 2025 23:33 UTC
52 points
13 comments11 min readLW link

Wor­lds Where Iter­a­tive De­sign Suc­ceeds?

Max Harms23 Oct 2025 22:14 UTC
23 points
5 comments8 min readLW link