Mechanis­tic In­ter­pretabil­ity as Re­v­erse Eng­ineer­ing (fol­low-up to “cars and elephants”)

David Scott Krueger (formerly: capybaralet)Nov 3, 2022, 11:19 PM
28 points
3 comments1 min readLW link

[Question] Could a Supreme Court suit work to solve NEPA prob­lems?

ChristianKlNov 3, 2022, 9:10 PM
15 points
0 comments1 min readLW link

[Video] How hav­ing Fast Fourier Trans­forms sooner could have helped with Nu­clear Disar­ma­ment—Veritaserum

mako yassNov 3, 2022, 9:04 PM
17 points
1 commentLW link

Fur­ther con­sid­er­a­tions on the Ev­i­den­tial­ist’s Wager

Martín SotoNov 3, 2022, 8:06 PM
3 points
9 comments8 min readLW link

AI as a Civ­i­liza­tional Risk Part 6/​6: What can be done

PashaKamyshevNov 3, 2022, 7:48 PM
2 points
4 comments4 min readLW link

A Mys­tery About High Di­men­sional Con­cept Encoding

Fabien RogerNov 3, 2022, 5:05 PM
46 points
13 comments7 min readLW link

Why do we post our AI safety plans on the In­ter­net?

Peter S. ParkNov 3, 2022, 4:02 PM
4 points
4 comments11 min readLW link

Mul­ti­ple De­ploy-Key Repos

jefftkNov 3, 2022, 3:10 PM
15 points
0 comments1 min readLW link
(www.jefftk.com)

Covid 11/​3/​22: Ask­ing Forgiveness

ZviNov 3, 2022, 1:50 PM
23 points
3 comments6 min readLW link
(thezvi.wordpress.com)

Ad­ver­sar­ial Poli­cies Beat Pro­fes­sional-Level Go AIs

sanxiynNov 3, 2022, 1:27 PM
31 points
35 comments1 min readLW link
(goattack.alignmentfund.org)

K-types vs T-types — what pri­ors do you have?

Cleo NardoNov 3, 2022, 11:29 AM
74 points
25 comments7 min readLW link

In­for­ma­tion Mar­kets 2: Op­ti­mally Shaped Re­ward Bets

eva_Nov 3, 2022, 11:08 AM
9 points
0 comments3 min readLW link

The Ra­tional Utili­tar­ian Love Move­ment (A His­tor­i­cal Ret­ro­spec­tive)

Caleb BiddulphNov 3, 2022, 7:11 AM
3 points
0 commentsLW link

The Mir­ror Cham­ber: A short story ex­plor­ing the an­thropic mea­sure func­tion and why it can matter

mako yassNov 3, 2022, 6:47 AM
30 points
13 comments10 min readLW link

Open Let­ter Against Reck­less Nu­clear Es­ca­la­tion and Use

Max TegmarkNov 3, 2022, 5:34 AM
27 points
25 comments1 min readLW link

Lazy Python Ar­gu­ment Parsing

jefftkNov 3, 2022, 2:20 AM
20 points
3 comments1 min readLW link
(www.jefftk.com)

AI as a Civ­i­liza­tional Risk Part 5/​6: Re­la­tion­ship be­tween C-risk and X-risk

PashaKamyshevNov 3, 2022, 2:19 AM
2 points
0 comments7 min readLW link

[Question] Is there a good way to award a fixed prize in a pre­dic­tion con­test?

jchanNov 2, 2022, 9:37 PM
18 points
5 comments1 min readLW link

“Are Ex­per­i­ments Pos­si­ble?” Seeds of Science call for reviewers

rogersbaconNov 2, 2022, 8:05 PM
8 points
0 comments1 min readLW link

Hu­mans do acausal co­or­di­na­tion all the time

Adam JermynNov 2, 2022, 2:40 PM
57 points
35 comments3 min readLW link

Far-UVC Light Up­date: No, LEDs are not around the cor­ner (tweet­storm)

DavidmanheimNov 2, 2022, 12:57 PM
73 points
27 comments4 min readLW link
(twitter.com)

Hous­ing and Tran­sit Thoughts #1

ZviNov 2, 2022, 12:10 PM
35 points
5 comments16 min readLW link
(thezvi.wordpress.com)

Mind is uncountable

Filip SondejNov 2, 2022, 11:51 AM
18 points
22 commentsLW link

AI Safety Needs Great Product Builders

goodgravyNov 2, 2022, 11:33 AM
14 points
2 commentsLW link

Why is fiber good for you?

bracesNov 2, 2022, 2:04 AM
18 points
2 comments2 min readLW link

In­for­ma­tion Markets

eva_Nov 2, 2022, 1:24 AM
46 points
6 comments12 min readLW link

Se­quence Reread: Fake Beliefs [plus se­quence spotlight meta]

RaemonNov 2, 2022, 12:09 AM
27 points
3 comments1 min readLW link

Real-Time Re­search Record­ing: Can a Trans­former Re-Derive Po­si­tional Info?

Neel NandaNov 1, 2022, 11:56 PM
69 points
16 comments1 min readLW link
(youtu.be)

All AGI Safety ques­tions wel­come (es­pe­cially ba­sic ones) [~monthly thread]

Robert MilesNov 1, 2022, 11:23 PM
68 points
105 comments2 min readLW link

[Question] Which Is­sues in Con­cep­tual Align­ment have been For­mal­ised or Ob­served (or not)?

ojorgensenNov 1, 2022, 10:32 PM
4 points
0 comments1 min readLW link

AI as a Civ­i­liza­tional Risk Part 4/​6: Bioweapons and Philos­o­phy of Modification

PashaKamyshevNov 1, 2022, 8:50 PM
7 points
1 comment8 min readLW link

Open & Wel­come Thread—Novem­ber 2022

MondSemmelNov 1, 2022, 6:47 PM
14 points
46 comments1 min readLW link

Mildly Against Donor Lotteries

jefftkNov 1, 2022, 6:10 PM
10 points
9 comments3 min readLW link
(www.jefftk.com)

Progress links and tweets, 2022-11-01

jasoncrawfordNov 1, 2022, 5:48 PM
16 points
4 comments3 min readLW link
(rootsofprogress.org)

On the cor­re­spon­dence be­tween AI-mis­al­ign­ment and cog­ni­tive dis­so­nance us­ing a be­hav­ioral eco­nomics model

Stijn BruersNov 1, 2022, 5:39 PM
4 points
0 comments6 min readLW link

Threat Model Liter­a­ture Review

Nov 1, 2022, 11:03 AM
78 points
4 comments25 min readLW link

Clar­ify­ing AI X-risk

Nov 1, 2022, 11:03 AM
127 points
24 comments4 min readLW link1 review

Au­dit­ing games for high-level interpretability

Paul CologneseNov 1, 2022, 10:44 AM
33 points
1 comment7 min readLW link

Re­mem­ber to trans­late your thoughts back again

brookNov 1, 2022, 8:49 AM
25 points
11 comments3 min readLW link
(forum.effectivealtruism.org)

Con­ver­sa­tions on Al­co­hol Consumption

AnnapurnaNov 1, 2022, 5:09 AM
20 points
6 comments9 min readLW link

ML Safety Schol­ars Sum­mer 2022 Retrospective

TW123Nov 1, 2022, 3:09 AM
29 points
0 commentsLW link

EA & LW Fo­rums Weekly Sum­mary (24 − 30th Oct 22′)

Zoe WilliamsNov 1, 2022, 2:58 AM
13 points
1 commentLW link

Cau­tion when in­ter­pret­ing Deep­mind’s In-con­text RL paper

Sam MarksNov 1, 2022, 2:42 AM
105 points
8 comments4 min readLW link

What sorts of sys­tems can be de­cep­tive?

Andrei AlexandruOct 31, 2022, 10:00 PM
16 points
0 comments7 min readLW link

“Cars and Elephants”: a hand­wavy ar­gu­ment/​anal­ogy against mechanis­tic interpretability

David Scott Krueger (formerly: capybaralet)Oct 31, 2022, 9:26 PM
51 points
25 comments2 min readLW link

Su­per­in­tel­li­gent AI is nec­es­sary for an amaz­ing fu­ture, but far from sufficient

So8resOct 31, 2022, 9:16 PM
132 points
48 comments34 min readLW link

San­ity-check­ing in an age of hyperbole

Ciprian Elliu IvanofOct 31, 2022, 8:04 PM
2 points
4 comments2 min readLW link

Why Aren’t There More Schel­ling Holi­days?

johnswentworthOct 31, 2022, 7:31 PM
63 points
21 comments1 min readLW link

The cir­cu­lar prob­lem of epistemic irresponsibility

Roman LeventovOct 31, 2022, 5:23 PM
5 points
2 comments8 min readLW link

AI as a Civ­i­liza­tional Risk Part 3/​6: Anti-econ­omy and Sig­nal Pollution

PashaKamyshevOct 31, 2022, 5:03 PM
7 points
4 comments14 min readLW link