All200520062007200820092010201120122013201420152016201720182019202020212022202320242025
AllJanFebMarAprMayJunJulAugSepOctNovDec
All12345678910111213141516171819202122232425262728293031

In­tro­duc­ing Shrubgrazer

jefftkDec 16, 2022, 2:50 PM
22 points
0 comments2 min readLW link
(www.jefftk.com)

Paper: Trans­form­ers learn in-con­text by gra­di­ent descent

LawrenceCDec 16, 2022, 11:10 AM
28 points
11 comments2 min readLW link
(arxiv.org)

Will Machines Ever Rule the World? MLAISU W50

Esben KranDec 16, 2022, 11:03 AM
12 points
7 comments4 min readLW link
(newsletter.apartresearch.com)

AI over­hangs de­pend on whether al­gorithms, com­pute and data are sub­sti­tutes or complements

NathanBarnardDec 16, 2022, 2:23 AM
4 points
0 comments3 min readLW link

AI Safety Move­ment Builders should help the com­mu­nity to op­ti­mise three fac­tors: con­trib­u­tors, con­tri­bu­tions and coordination

peterslatteryDec 15, 2022, 10:50 PM
4 points
0 comments6 min readLW link

Mask­ing to Avoid Miss­ing Things

jefftkDec 15, 2022, 9:00 PM
17 points
2 comments1 min readLW link
(www.jefftk.com)

Con­sider work­ing more hours and tak­ing more stimulants

Arjun PanicksseryDec 15, 2022, 8:38 PM
33 points
11 commentsLW link

We’ve stepped over the thresh­old into the Fourth Arena, but don’t rec­og­nize it

Bill BenzonDec 15, 2022, 8:22 PM
2 points
0 comments7 min readLW link

[Question] How is ARC plan­ning to use ELK?

jacquesthibsDec 15, 2022, 8:11 PM
24 points
5 comments1 min readLW link

How “Dis­cov­er­ing La­tent Knowl­edge in Lan­guage Models Without Su­per­vi­sion” Fits Into a Broader Align­ment Scheme

CollinDec 15, 2022, 6:22 PM
244 points
39 comments16 min readLW link1 review

High-level hopes for AI alignment

HoldenKarnofskyDec 15, 2022, 6:00 PM
58 points
3 comments19 min readLW link
(www.cold-takes.com)

Two Dog­mas of LessWrong

omnizoidDec 15, 2022, 5:56 PM
−7 points
155 comments69 min readLW link

Covid 12/​15/​22: China’s Wave Begins

ZviDec 15, 2022, 4:20 PM
32 points
7 comments10 min readLW link
(thezvi.wordpress.com)

The next decades might be wild

Marius HobbhahnDec 15, 2022, 4:10 PM
175 points
42 comments41 min readLW link1 review

Ba­sic build­ing blocks of de­pen­dent type theory

Thomas KehrenbergDec 15, 2022, 2:54 PM
49 points
9 comments13 min readLW link

AI Ne­o­re­al­ism: a threat model & suc­cess crite­rion for ex­is­ten­tial safety

davidadDec 15, 2022, 1:42 PM
67 points
1 comment3 min readLW link

Who should write the defini­tive post on Ziz?

Nicholas / Heather KrossDec 15, 2022, 6:37 AM
4 points
45 comments3 min readLW link

[Question] Is Paul Chris­ti­ano still as op­ti­mistic about Ap­proval-Directed Agents as he was in 2018?

Chris_LeongDec 14, 2022, 11:28 PM
8 points
0 comments1 min readLW link

«Boundaries», Part 3b: Align­ment prob­lems in terms of bound­aries

Andrew_CritchDec 14, 2022, 10:34 PM
72 points
7 comments13 min readLW link

Align­ing al­ign­ment with performance

Marv KDec 14, 2022, 10:19 PM
2 points
0 comments2 min readLW link

Con­trary to List of Lethal­ity’s point 22, al­ign­ment’s door num­ber 2

False NameDec 14, 2022, 10:01 PM
−2 points
5 comments22 min readLW link

Kol­mogorov Com­plex­ity and Si­mu­la­tion Hypothesis

False NameDec 14, 2022, 10:01 PM
−3 points
0 comments7 min readLW link

[Question] Stan­ley Meyer’s wa­ter fuel cell

mikbpDec 14, 2022, 9:19 PM
2 points
6 comments1 min readLW link

[Question] Is the AI timeline too short to have chil­dren?

YorethDec 14, 2022, 6:32 PM
38 points
20 comments1 min readLW link

Pre­dict­ing GPU performance

Marius Hobbhahn and Tamay
Dec 14, 2022, 4:27 PM
60 points
26 comments1 min readLW link
(epochai.org)

[In­com­plete] What is Com­pu­ta­tion Any­way?

DragonGodDec 14, 2022, 4:17 PM
16 points
1 comment13 min readLW link
(arxiv.org)

Chair Hang­ing Peg

jefftkDec 14, 2022, 3:30 PM
11 points
0 comments1 min readLW link
(www.jefftk.com)

My AGI safety re­search—2022 re­view, ’23 plans

Steven ByrnesDec 14, 2022, 3:15 PM
51 points
10 comments7 min readLW link

Ex­tract­ing and Eval­u­at­ing Causal Direc­tion in LLMs’ Activations

Fabien Roger and simeon_c
Dec 14, 2022, 2:33 PM
29 points
5 comments11 min readLW link

Key Mostly Out­ward-Fac­ing Facts From the Story of VaccinateCA

ZviDec 14, 2022, 1:30 PM
61 points
2 comments23 min readLW link
(thezvi.wordpress.com)

Dis­cov­er­ing La­tent Knowl­edge in Lan­guage Models Without Supervision

XodarapDec 14, 2022, 12:32 PM
45 points
1 comment1 min readLW link
(arxiv.org)

[Question] COVID China Per­sonal Ad­vice (No mRNA vax, pos­si­ble hos­pi­tal over­load, bug-chas­ing edi­tion)

Lao MeinDec 14, 2022, 10:31 AM
20 points
11 comments1 min readLW link

Beyond a bet­ter world

DavidmanheimDec 14, 2022, 10:18 AM
14 points
7 comments4 min readLW link
(progressforum.org)

Proof as mere strong evidence

adamShimiDec 14, 2022, 8:56 AM
28 points
16 comments2 min readLW link
(epistemologicalvigilance.substack.com)

Try­ing to dis­am­biguate differ­ent ques­tions about whether RLHF is “good”

BuckDec 14, 2022, 4:03 AM
108 points
47 comments7 min readLW link1 review

[Question] How can one liter­ally buy time (from x-risk) with money?

Alex_AltairDec 13, 2022, 7:24 PM
24 points
3 comments1 min readLW link

[Question] Best in­tro­duc­tory overviews of AGI safety?

JakubKDec 13, 2022, 7:01 PM
21 points
9 comments2 min readLW link
(forum.effectivealtruism.org)

Ap­pli­ca­tions open for AGI Safety Fun­da­men­tals: Align­ment Course

Richard_NgoDec 13, 2022, 6:31 PM
49 points
0 comments2 min readLW link

What Does It Mean to Align AI With Hu­man Values?

AlgonDec 13, 2022, 4:56 PM
8 points
3 comments1 min readLW link
(www.quantamagazine.org)

It Takes Two Parac­eta­mol?

Eli_Dec 13, 2022, 4:29 PM
33 points
10 comments2 min readLW link

[In­terim re­search re­port] Tak­ing fea­tures out of su­per­po­si­tion with sparse autoencoders

Lee Sharkey, Dan Braun and beren
Dec 13, 2022, 3:41 PM
150 points
23 comments22 min readLW link2 reviews

[Question] Is the ChatGPT-simu­lated Linux vir­tual ma­chine real?

KenoubiDec 13, 2022, 3:41 PM
18 points
7 comments1 min readLW link

Ex­is­ten­tial AI Safety is NOT sep­a­rate from near-term applications

scasperDec 13, 2022, 2:47 PM
37 points
17 comments3 min readLW link

What is the cor­re­la­tion be­tween up­vot­ing and benefit to read­ers of LW?

banevDec 13, 2022, 2:26 PM
7 points
15 comments1 min readLW link

Limits of Superintelligence

Aleksei PetrenkoDec 13, 2022, 12:19 PM
1 point
5 comments1 min readLW link

Bay 2022 Solstice

RaemonDec 13, 2022, 8:58 AM
17 points
0 comments1 min readLW link

Last day to nom­i­nate things for the Re­view. Also, 2019 books still ex­ist.

RaemonDec 13, 2022, 8:53 AM
15 points
0 comments1 min readLW link

AI al­ign­ment is dis­tinct from its near-term applications

paulfchristianoDec 13, 2022, 7:10 AM
255 points
21 comments2 min readLW link
(ai-alignment.com)

Take 10: Fine-tun­ing with RLHF is aes­thet­i­cally un­satis­fy­ing.

Charlie SteinerDec 13, 2022, 7:04 AM
37 points
3 comments2 min readLW link

[Question] Are law­suits against AGI com­pa­nies ex­tend­ing AGI timelines?

SlowingAGIDec 13, 2022, 6:00 AM
1 point
1 comment1 min readLW link