Strong-Misal­ign­ment: Does Yud­kowsky (or Chris­ti­ano, or TurnTrout, or Wolfram, or…etc.) Have an Ele­va­tor Speech I’m Miss­ing?

Benjamin BourlierMar 15, 2024, 11:17 PM
−4 points
3 comments16 min readLW link

In­tro­duc­ing METR’s Au­ton­omy Eval­u­a­tion Resources

Mar 15, 2024, 11:16 PM
90 points
0 comments1 min readLW link
(metr.github.io)

Are AIs con­scious? It might depend

Logan ZoellnerMar 15, 2024, 11:09 PM
6 points
6 comments3 min readLW link

Beyond Max­ipok — good re­flec­tive gov­er­nance as a tar­get for action

owencbMar 15, 2024, 10:22 PM
20 points
0 commentsLW link

Mid­dle Child Phenomenon

PhilosophicalSoulMar 15, 2024, 8:47 PM
3 points
3 comments2 min readLW link

Ca­pa­bil­ity or Align­ment? Re­spect the LLM Base Model’s Ca­pa­bil­ity Dur­ing Alignment

Jingfeng YangMar 15, 2024, 5:56 PM
7 points
0 comments24 min readLW link

Ra­tional An­i­ma­tions offers an­i­ma­tion pro­duc­tion and writ­ing ser­vices!

WriterMar 15, 2024, 5:26 PM
33 points
0 comments1 min readLW link

Im­prov­ing SAE’s by Sqrt()-ing L1 & Re­mov­ing Low­est Ac­ti­vat­ing Fea­tures

Mar 15, 2024, 4:30 PM
26 points
5 comments4 min readLW link

Stuttgart, Ger­many—ACX Spring Mee­tups Every­where 2024

Benjamin RMar 15, 2024, 2:59 PM
2 points
1 comment1 min readLW link

Con­trol­ling AGI Risk

TeaSeaMar 15, 2024, 4:56 AM
6 points
8 comments4 min readLW link

Ulm, Ger­many—ACX Spring Mee­tups Every­where 2024

Benjamin RMar 15, 2024, 1:32 AM
2 points
1 comment1 min readLW link

New­port News/​ Virginia ACX Meetup

DanielMar 14, 2024, 11:46 PM
1 point
0 comments1 min readLW link

Con­struc­tive Cauchy se­quences vs. Dedekind cuts

jessicataMar 14, 2024, 11:04 PM
47 points
23 comments4 min readLW link
(unstableontology.com)

A Nail in the Coffin of Exceptionalism

Yeshua GodMar 14, 2024, 10:41 PM
−17 points
0 comments3 min readLW link

Toward a Broader Con­cep­tion of Ad­verse Selection

Ricki HeicklenMar 14, 2024, 10:40 PM
177 points
61 comments13 min readLW link
(bayesshammai.substack.com)

More peo­ple get­ting into AI safety should do a PhD

AdamGleaveMar 14, 2024, 10:14 PM
61 points
24 comments12 min readLW link
(gleave.me)

Col­lec­tion (Part 6 of “The Sense Of Phys­i­cal Ne­ces­sity”)

LoganStrohlMar 14, 2024, 9:37 PM
28 points
0 comments8 min readLW link

Fixed point or os­cillate or noise

lemonhopeMar 14, 2024, 6:37 PM
3 points
10 comments1 min readLW link

How use­ful is “AI Con­trol” as a fram­ing on AI X-Risk?

Mar 14, 2024, 6:06 PM
70 points
4 comments34 min readLW link

Sparse au­toen­coders find com­posed fea­tures in small toy mod­els

Mar 14, 2024, 6:00 PM
33 points
12 comments15 min readLW link

AI #55: Keep Claud­ing Along

ZviMar 14, 2024, 3:40 PM
62 points
16 comments70 min readLW link
(thezvi.wordpress.com)

To the av­er­age hu­man, con­trol­led AI is just as lethal as ‘mis­al­igned’ AI

YonatanKMar 14, 2024, 2:52 PM
6 points
20 comments5 min readLW link

Claude vs GPT

Maxwell TabarrokMar 14, 2024, 12:41 PM
12 points
2 comments2 min readLW link
(www.maximum-progress.com)

A brief re­view of China’s AI in­dus­try and regulations

Elliot MckernonMar 14, 2024, 12:19 PM
24 points
0 comments16 min readLW link

[Question] Can any LLM be rep­re­sented as an Equa­tion?

Valentin BaltadzhievMar 14, 2024, 9:51 AM
1 point
2 comments1 min readLW link

‘Em­piri­cism!’ as Anti-Epistemology

Eliezer YudkowskyMar 14, 2024, 2:02 AM
171 points
92 comments25 min readLW link

How I turned do­ing ther­apy into ob­ject-level AI safety research

ChipmonkMar 14, 2024, 1:54 AM
15 points
5 comments4 min readLW link

Op­por­tunis­tic Time-Management

Richard HenageMar 13, 2024, 9:38 PM
13 points
2 comments1 min readLW link

AI gov­er­nance and strat­egy: a list of re­search agen­das and work that could be done.

Mar 13, 2024, 9:23 PM
7 points
1 comment17 min readLW link

High­lights from Lex Frid­man’s in­ter­view of Yann LeCun

Joel BurgetMar 13, 2024, 8:58 PM
48 points
15 comments41 min readLW link

On the Lat­est TikTok Bill

ZviMar 13, 2024, 6:50 PM
58 points
7 comments29 min readLW link
(thezvi.wordpress.com)

[Question] Recom­mended book for a bal­anced take and les­sons learned from covid pan­demic response

Martin Hare RobertsonMar 13, 2024, 6:14 PM
4 points
0 comments1 min readLW link

ACX/​LW Seat­tle spring meetup 2024

nsokolskyMar 13, 2024, 5:24 PM
12 points
3 comments1 min readLW link

Lay­ing the Foun­da­tions for Vi­sion and Mul­ti­modal Mechanis­tic In­ter­pretabil­ity & Open Problems

Mar 13, 2024, 5:09 PM
44 points
13 comments14 min readLW link

I was raised by de­vout Mor­mons, AMA [&|] Solic­it­ing Advice

ErioirEMar 13, 2024, 4:52 PM
31 points
41 comments2 min readLW link

Re­la­tional Agency: Con­sis­tently Reach­ing Out

Jonathan MoregårdMar 13, 2024, 2:34 PM
16 points
0 comments5 min readLW link
(open.substack.com)

[Question] What could a policy ban­ning AGI look like?

TsviBTMar 13, 2024, 2:19 PM
78 points
23 comments3 min readLW link

Click­bait Soapboxing

DaystarEldMar 13, 2024, 2:09 PM
24 points
16 comments3 min readLW link
(daystareld.com)

Vir­tual AI Safety Un­con­fer­ence 2024

Mar 13, 2024, 1:54 PM
14 points
0 comments1 min readLW link

Jobs, Re­la­tion­ships, and Other Cults

Mar 13, 2024, 5:58 AM
40 points
9 comments35 min readLW link

How do you im­prove the qual­ity of your drink­ing wa­ter?

Alex K. Chen (parrot)Mar 13, 2024, 12:37 AM
11 points
2 comments1 min readLW link

The Parable Of The Fallen Pen­du­lum—Part 2

johnswentworthMar 12, 2024, 9:41 PM
78 points
8 comments4 min readLW link

Open con­sul­tancy: Let­ting un­trusted AIs choose what an­swer to ar­gue for

Fabien RogerMar 12, 2024, 8:38 PM
35 points
5 comments5 min readLW link

[Question] Is any­one work­ing on for­mally ver­ified AI toolchains?

metachiralityMar 12, 2024, 7:36 PM
17 points
4 comments1 min readLW link

Trans­former Debugger

Henk TillmanMar 12, 2024, 7:08 PM
26 points
0 comments1 min readLW link
(github.com)

Su­perfore­cast­ing the Ori­gins of the Covid-19 Pandemic

DanielFilanMar 12, 2024, 7:01 PM
64 points
0 comments1 min readLW link
(goodjudgment.substack.com)

min­i­mum vi­able action

Sindhu PrasadMar 12, 2024, 4:06 PM
1 point
0 comments3 min readLW link

Hard­ball ques­tions for the Gem­ini Con­gres­sional Hearing

Michael ThiessenMar 12, 2024, 3:27 PM
−11 points
2 comments1 min readLW link

OpenAI: The Board Expands

ZviMar 12, 2024, 2:00 PM
92 points
1 comment30 min readLW link
(thezvi.wordpress.com)

Up­date on Devel­op­ing an Ethics Calcu­la­tor to Align an AGI to

sweenesmMar 12, 2024, 12:33 PM
4 points
2 comments8 min readLW link