$500 + $500 Bounty Prob­lem: Does An (Ap­prox­i­mately) Deter­minis­tic Max­i­mal Re­dund Always Ex­ist?

May 6, 2025, 11:05 PM
72 points
11 comments3 min readLW link

The The­ory Be­hind Loss Curves

James CamachoMay 6, 2025, 10:22 PM
16 points
3 comments4 min readLW link
(github.com)

Nega­tive Re­sults on Group SAEs

Josh EngelsMay 6, 2025, 9:49 PM
70 points
3 comments8 min readLW link

ACX At­lanta May 2025 Meetup

Steve FrenchMay 6, 2025, 9:00 PM
2 points
0 comments1 min readLW link

[Question] What kind of policy by an AGI would make peo­ple happy?

StanislavKrymMay 6, 2025, 6:05 PM
1 point
2 comments1 min readLW link

Paper High­lights, April ’25

gasteigerjoMay 6, 2025, 2:22 PM
4 points
0 comments7 min readLW link
(aisafetyfrontier.substack.com)

Zucker­berg’s Dystopian AI Vision

ZviMay 6, 2025, 1:50 PM
61 points
7 comments11 min readLW link
(thezvi.wordpress.com)

Will pro­tein de­sign tools solve the snake an­tivenom short­age?

Abhishaike MahajanMay 6, 2025, 1:11 PM
30 points
0 comments17 min readLW link
(www.owlposting.com)

Utah Court Case Over State Law Re­gard­ing “Per­son­hood” for Non­hu­man Intelligences

Stephen MartinMay 6, 2025, 12:54 PM
10 points
3 comments2 min readLW link

Global Risks Weekly Roundup #18/​2025: US tar­iff short­ages, mil­i­tary polic­ing, Gaza famine.

NunoSempereMay 6, 2025, 10:39 AM
31 points
2 comments3 min readLW link
(blog.sentinel-team.org)

OpenAI’s Jig May Be Up

ValeMay 6, 2025, 8:51 AM
3 points
2 comments3 min readLW link

My Rea­sons for Us­ing Anki

Parker ConleyMay 6, 2025, 7:01 AM
10 points
0 comments3 min readLW link
(parconley.com)

It’s ‘Well, ac­tu­ally...’ all the way down

benwrMay 6, 2025, 5:44 AM
40 points
34 comments1 min readLW link
(www.benwr.net)

Five Hinge‑Ques­tions That De­cide Whether AGI Is Five Years Away or Twenty

charlieoneillMay 6, 2025, 2:48 AM
125 points
17 comments5 min readLW link

Non­profit to re­tain con­trol of OpenAI

ArchimedesMay 5, 2025, 11:41 PM
37 points
1 comment1 min readLW link
(openai.com)

Un­ex­pected Con­scious Entities

Gunnar_ZarnckeMay 5, 2025, 10:14 PM
34 points
6 comments6 min readLW link

The First Law of Con­scious Agency: Lin­guis­tic Rel­a­tivity and the Birth of “I”

Dima (lain)May 5, 2025, 9:20 PM
−17 points
4 comments2 min readLW link

New­ton’s sec­ond law ex­plained: it works in many universes

TahpMay 5, 2025, 7:47 PM
19 points
10 comments15 min readLW link
(quark.rodeo)

Repli­ca­tor->Ve­hi­cle Align­ment and Hu­man->AI Alignment

derelict5432May 5, 2025, 7:23 PM
0 points
3 comments4 min readLW link

The Sweet Les­son: AI Safety Should Scale With Compute

Jesse HooglandMay 5, 2025, 7:03 PM
95 points
3 comments3 min readLW link

[Question] Blue light, ‘Adrenal ASMR’: strange ex­pe­riences I can’t find any liter­a­ture about

vernichtungMay 5, 2025, 6:58 PM
16 points
6 comments1 min readLW link

Ts­inghua pa­per: Does RL Really In­cen­tivize Rea­son­ing Ca­pac­ity in LLMs Beyond the Base Model?

Thomas KwaMay 5, 2025, 6:56 PM
68 points
21 comments2 min readLW link
(arxiv.org)

In­tro & Pro­posal for AGI Model

PickleBrineMay 5, 2025, 6:56 PM
0 points
0 comments3 min readLW link

AI Su­per­or­ganisms: An Alter­na­tive Path­way to Ar­tifi­cial Superintelligence

Aaron VanzylMay 5, 2025, 6:55 PM
4 points
5 comments15 min readLW link

Kar­ls­ruhe ACX: The colours of her coat

wilmMay 5, 2025, 6:35 PM
2 points
0 comments1 min readLW link

The Me­tac­u­lus Cup Series Is Live, $5,000 Prize Pool

ChristianWilliamsMay 5, 2025, 5:14 PM
4 points
0 commentsLW link
(www.metaculus.com)

Com­mu­nity Feed­back Re­quest: AI Safety In­tro for Gen­eral Public

May 5, 2025, 4:38 PM
6 points
5 comments3 min readLW link

GPT-4o Sy­co­phancy Post Mortem

ZviMay 5, 2025, 4:00 PM
55 points
1 comment16 min readLW link
(thezvi.wordpress.com)

Le­gal Su­per­vi­sion of Fron­tier AI Labs is the an­swer.

GauraventhMay 5, 2025, 1:36 PM
14 points
2 comments3 min readLW link
(robertandgaurav.substack.com)

The cru­cible — how I think about the situ­a­tion with AI

owencbMay 5, 2025, 1:18 PM
25 points
1 comment8 min readLW link
(strangecities.substack.com)

Light­ning Talks: Thought, Trick, Curiosity

marta_kMay 5, 2025, 11:49 AM
1 point
0 comments1 min readLW link

Are stan­dard­ized tests effec­tive?

HrussMay 5, 2025, 10:02 AM
1 point
1 comment1 min readLW link

Pro­posal: Liquid Pre­dic­tion Mar­kets for AI Forecasting

Jesse RichardsonMay 5, 2025, 5:13 AM
23 points
2 comments3 min readLW link

Why “Solv­ing Align­ment” Is Likely a Cat­e­gory Mistake

Nate SharpeMay 5, 2025, 4:26 AM
22 points
3 comments3 min readLW link

AI, An­i­mals, & Digi­tal Minds 2025: ap­ply to speak by Wed­nes­day!

Alistair StewartMay 5, 2025, 12:56 AM
4 points
0 comments1 min readLW link

AI, An­i­mals, & Digi­tal Minds 2025

Alistair StewartMay 5, 2025, 12:51 AM
2 points
0 comments1 min readLW link

Notes on the Long Tasks METR pa­per, from a HCAST task contributor

abstractapplicMay 4, 2025, 11:17 PM
108 points
7 comments2 min readLW link

Why I am not a successionist

Nina PanicksseryMay 4, 2025, 7:08 PM
62 points
52 comments2 min readLW link
(ninapanickssery.substack.com)

Overview: AI Safety Outreach Grass­roots Orgs

May 4, 2025, 5:39 PM
46 points
8 comments2 min readLW link

The Power Users We For­got: Why AI Needs Them Now More Than Ever

Anthony FoxMay 4, 2025, 5:23 PM
1 point
6 comments3 min readLW link

Fake AI law­suits to drive links

Yair HalberstadtMay 4, 2025, 4:53 PM
22 points
0 comments1 min readLW link
(www.rationalistjudaism.com)

Scott Aaron­son at UT Austin on May 17 | Com­pu­ta­tional Com­plex­ity & Philosophy

ekkoláptoMay 4, 2025, 4:42 PM
1 point
0 comments1 min readLW link

In­ter­pretabil­ity Will Not Reli­ably Find De­cep­tive AI

Neel NandaMay 4, 2025, 4:32 PM
316 points
66 comments7 min readLW link

80 con­cepts on my new ver­sion of AI: DecisionBots

Wes RMay 4, 2025, 2:04 PM
0 points
2 comments15 min readLW link

Where have all the to­kens gone?

bracesMay 4, 2025, 1:52 PM
13 points
7 comments6 min readLW link

The Ukraine War and the Kill Market

Martin SustrikMay 4, 2025, 7:50 AM
98 points
13 comments5 min readLW link
(250bpm.substack.com)

PSA: Be­fore May 21 is a good time to sign up for cryonics

AlexMennenMay 4, 2025, 4:10 AM
53 points
0 comments1 min readLW link

GTFO of the So­cial In­ter­net Be­fore you Can’t: The Miro & Yindi Story

keltanMay 4, 2025, 1:08 AM
30 points
12 comments10 min readLW link

“Su­per­hu­man” Isn’t Well Specified

JustisMillsMay 3, 2025, 11:42 PM
32 points
9 comments3 min readLW link
(justismills.substack.com)

Nav­i­gat­ing burnout

gwMay 3, 2025, 10:07 PM
73 points
1 comment9 min readLW link
(www.georgeyw.com)