A strange twist on the road to AGI

cveresOct 12, 2022, 11:27 PM
−8 points
0 comments1 min readLW link

Help out Red­wood Re­search’s in­ter­pretabil­ity team by find­ing heuris­tics im­ple­mented by GPT-2 small

Oct 12, 2022, 9:25 PM
50 points
11 comments4 min readLW link

Towards a com­pre­hen­sive study of po­ten­tial psy­cholog­i­cal causes of the or­di­nary range of vari­a­tion of af­fec­tive gen­der iden­tity in males

tailcalledOct 12, 2022, 9:10 PM
52 points
4 comments37 min readLW link

Six (and a half) in­tu­itions for KL divergence

CallumMcDougallOct 12, 2022, 9:07 PM
165 points
27 comments10 min readLW link1 review
(www.perfectlynormal.co.uk)

[MLSN #6]: Trans­parency sur­vey, prov­able ro­bust­ness, ML mod­els that pre­dict the future

Dan HOct 12, 2022, 8:56 PM
27 points
0 comments6 min readLW link

[Question] Pre­vi­ous Work on Re­cre­at­ing Neu­ral Net­work In­put from In­ter­me­di­ate Layer Activations

bglassOct 12, 2022, 7:28 PM
1 point
3 comments1 min readLW link

Be more effec­tive by learn­ing im­por­tant prac­ti­cal knowl­edge us­ing flashcards

StenemoOct 12, 2022, 6:05 PM
5 points
2 comments1 min readLW link

Ar­ti­cle Re­view: Google’s AlphaTensor

Robert_AIZIOct 12, 2022, 6:04 PM
8 points
4 comments10 min readLW link

Align­ment 201 curriculum

Richard_NgoOct 12, 2022, 6:03 PM
102 points
3 comments1 min readLW link
(www.agisafetyfundamentals.com)

Progress links and tweets, 2022-10-12

jasoncrawfordOct 12, 2022, 4:59 PM
8 points
0 comments1 min readLW link
(rootsofprogress.org)

Build­ing a trans­former from scratch—AI safety up-skil­ling challenge

Marius HobbhahnOct 12, 2022, 3:40 PM
42 points
1 comment5 min readLW link

In­stru­men­tal con­ver­gence in sin­gle-agent systems

Oct 12, 2022, 12:24 PM
33 points
4 comments8 min readLW link
(www.gladstone.ai)

Sin­ga­pore—Small ca­sual din­ner in Chi­na­town #5

Joe RoccaOct 12, 2022, 8:59 AM
3 points
1 comment1 min readLW link

A game of mattering

KatjaGraceOct 12, 2022, 8:50 AM
28 points
2 comments5 min readLW link
(worldspiritsockpuppet.com)

Cal­ibra­tion of a thou­sand predictions

KatjaGraceOct 12, 2022, 8:50 AM
57 points
7 comments5 min readLW link
(worldspiritsockpuppet.com)

My ar­gu­ment against AGI

cveresOct 12, 2022, 6:33 AM
7 points
5 comments1 min readLW link

Ac­tu­ally, All Nu­clear Famine Papers are Bunk

Lao MeinOct 12, 2022, 5:58 AM
113 points
37 comments2 min readLW link1 review

Contin­gency is not arbitrary

Gordon Seidoh WorleyOct 12, 2022, 4:35 AM
13 points
0 comments3 min readLW link

That one apoc­a­lyp­tic nu­clear famine pa­per is bunk

Lao MeinOct 12, 2022, 3:33 AM
110 points
10 comments1 min readLW link

As­tralCodexTen and Ra­tion­al­ity Meetup Or­ganisers’ Re­treat Asia Pa­cific region

Oct 12, 2022, 3:20 AM
14 points
4 comments2 min readLW link

Ab­bots Brom­ley Horn Dance History

jefftkOct 12, 2022, 2:10 AM
11 points
0 comments2 min readLW link
(www.jefftk.com)

Power-Seek­ing AI and Ex­is­ten­tial Risk

Antonio FrancaOct 11, 2022, 10:50 PM
6 points
0 comments9 min readLW link

From tech­noc­racy to the counterculture

jasoncrawfordOct 11, 2022, 7:37 PM
28 points
1 comment26 min readLW link
(rootsofprogress.org)

Pret­tified AI Safety Game Cards

abramdemskiOct 11, 2022, 7:35 PM
47 points
6 comments1 min readLW link

On the proper pi­lot­ing of flesh shoots

Mordecai WeynbergOct 11, 2022, 6:52 PM
−4 points
6 comments1 min readLW link

Why I think nu­clear war trig­gered by Rus­sian tac­ti­cal nukes in Ukraine is unlikely

Dave OrrOct 11, 2022, 6:30 PM
50 points
7 comments3 min readLW link

Anony­mous ad­vice: If you want to re­duce AI risk, should you take roles that ad­vance AI ca­pa­bil­ities?

Benjamin HiltonOct 11, 2022, 2:16 PM
54 points
9 comments1 min readLW link

Misal­ign­ment Harms Can Be Caused by Low In­tel­li­gence Systems

DialecticEelOct 11, 2022, 1:39 PM
11 points
3 comments1 min readLW link

[Sketch] Val­idity Cri­te­rion for Log­i­cal Counterfactuals

DragonGodOct 11, 2022, 1:31 PM
6 points
0 comments6 min readLW link

[Question] How much does the risk of dy­ing from nu­clear war differ within and be­tween coun­tries?

amaraiOct 11, 2022, 11:55 AM
4 points
7 comments1 min readLW link

Did you en­joy Ramez Naam’s “Nexus” tril­ogy? Check out this in­ter­view on neu­rotech and the law.

fowlertmOct 11, 2022, 11:10 AM
5 points
0 comments1 min readLW link

What “The Mes­sage” Was For Me

Alex BeymanOct 11, 2022, 8:08 AM
−3 points
14 comments4 min readLW link

Up­dates and Clarifications

SD MarlowOct 11, 2022, 5:34 AM
−5 points
1 comment1 min readLW link

What if hu­man rea­son­ing is anti-in­duc­tive?

Q HomeOct 11, 2022, 5:15 AM
4 points
2 comments13 min readLW link

Ful­l­ness to Indi­cate Cleanliness

jefftkOct 11, 2022, 12:40 AM
9 points
12 comments1 min readLW link
(www.jefftk.com)

[Question] What hap­pened to the an­nual LW de­mo­graphic sur­veys?

ROMOct 11, 2022, 12:19 AM
5 points
2 comments1 min readLW link

EA & LW Fo­rums Weekly Sum­mary (26 Sep − 9 Oct 22′)

Zoe WilliamsOct 10, 2022, 11:58 PM
13 points
2 comments1 min readLW link

Don’t ex­pect AGI any­time soon

cveresOct 10, 2022, 10:38 PM
−14 points
6 comments1 min readLW link

QAPR 4: In­duc­tive biases

Quintin PopeOct 10, 2022, 10:08 PM
67 points
2 comments18 min readLW link

Apollo

Jarred FilmerOct 10, 2022, 9:30 PM
46 points
0 comments3 min readLW link

[Question] Does biol­ogy re­li­ably find the global max­i­mum, or at least get close?

Noosphere89Oct 10, 2022, 8:55 PM
24 points
71 comments1 min readLW link

Disen­tan­gling in­ner al­ign­ment failures

Erik JennerOct 10, 2022, 6:50 PM
23 points
5 comments4 min readLW link

ACX meetup [Oc­to­ber]

sallatikOct 10, 2022, 5:23 PM
1 point
0 comments1 min readLW link

Nat­u­ral Cat­e­gories Update

Logan ZoellnerOct 10, 2022, 3:19 PM
33 points
6 comments2 min readLW link

When re­port­ing AI timelines, be clear who you’re defer­ring to

Sam ClarkeOct 10, 2022, 2:24 PM
38 points
6 comments1 min readLW link

Why Balsa Re­search is Worthwhile

ZviOct 10, 2022, 1:50 PM
87 points
12 comments8 min readLW link
(thezvi.wordpress.com)

Les­sons learned from talk­ing to >100 aca­demics about AI safety

Marius HobbhahnOct 10, 2022, 1:16 PM
216 points
18 comments12 min readLW link1 review

We can do bet­ter than argmax

Jan_KulveitOct 10, 2022, 10:32 AM
49 points
4 comments1 min readLW link

Vege­tar­i­anism and depression

MaggyOct 10, 2022, 9:11 AM
2 points
2 comments1 min readLW link

Re­sults from the lan­guage model hackathon

Esben KranOct 10, 2022, 8:29 AM
22 points
1 comment4 min readLW link