[Question] Should an un­der­grad avoid a ca­pa­bil­ities pro­ject?

DoubleSep 12, 2023, 11:16 PM
4 points
2 comments1 min readLW link

[Linkpost] Con­tra four-wheeled suit­cases, sort of

Gunnar_ZarnckeSep 12, 2023, 8:36 PM
18 points
4 comments1 min readLW link
(dynomight.substack.com)

Seek­ing Feed­back on My Mechanis­tic In­ter­pretabil­ity Re­search Agenda

RGRGRGSep 12, 2023, 6:45 PM
3 points
1 comment3 min readLW link

Au­to­mat­i­cally find­ing fea­ture vec­tors in the OV cir­cuits of Trans­form­ers with­out us­ing probing

Jacob DunefskySep 12, 2023, 5:38 PM
16 points
2 comments29 min readLW link

Startup Roundup #1: Happy Demo Day

ZviSep 12, 2023, 1:20 PM
38 points
5 comments15 min readLW link
(thezvi.wordpress.com)

[Question] Is there some­thing fun­da­men­tally wrong with the Uni­verse?

Caerulea-LawrenceSep 12, 2023, 12:02 PM
6 points
80 comments2 min readLW link

Stu­pidity is also hard

walkthroughwallsSep 12, 2023, 2:45 AM
−8 points
4 comments2 min readLW link

Ap­ple Cider Baklava

jefftkSep 12, 2023, 2:10 AM
15 points
0 comments1 min readLW link
(www.jefftk.com)

How use­ful is Cor­rigi­bil­ity?

martinkunevSep 12, 2023, 12:05 AM
11 points
4 comments5 min readLW link

Con­tra Heighn Con­tra Me Con­tra Func­tional De­ci­sion The­ory

Bentham's BulldogSep 11, 2023, 7:49 PM
−10 points
14 comments6 min readLW link

Ma­chine Evolution

Sep 11, 2023, 7:29 PM
11 points
2 comments22 min readLW link

[Question] Is there a hard copy of the se­quences available any­where?

Cole WyethSep 11, 2023, 7:01 PM
3 points
1 comment1 min readLW link

Ama­zon KDP AI con­tent guidelines

ChristianKlSep 11, 2023, 6:36 PM
12 points
0 comments1 min readLW link

A Case for AI Safety via Law

JWJohnstonSep 11, 2023, 6:26 PM
20 points
12 comments4 min readLW link

Erdős Prob­lems in Al­gorith­mic Probability

Aidan RockeSep 11, 2023, 4:44 PM
13 points
4 comments2 min readLW link

PSA: The com­mu­nity is in Berkeley/​Oak­land, not “the Bay Area”

maiaSep 11, 2023, 3:59 PM
105 points
7 comments1 min readLW link

A Bat and Ball made me Sad

Darren McKeeSep 11, 2023, 1:48 PM
14 points
26 comments1 min readLW link

Fo­cus on the Hardest Part First

Johannes C. MayerSep 11, 2023, 7:53 AM
42 points
13 comments1 min readLW link

The Promises and Pit­falls of Long-Term Forecasting

GeoVaneSep 11, 2023, 5:04 AM
1 point
0 comments5 min readLW link

Log­i­cal Share Splitting

DaemonicSigilSep 11, 2023, 4:08 AM
93 points
16 comments9 min readLW link
(pbement.com)

[Question] High school advice

bohaskaSep 11, 2023, 1:26 AM
11 points
16 comments1 min readLW link

Seat­tle As­tral Codex Ten Monthly Social

a7xSep 10, 2023, 7:00 PM
1 point
0 comments1 min readLW link

[Question] What are some good lan­guage mod­els to ex­per­i­ment with?

tailcalledSep 10, 2023, 6:31 PM
16 points
3 comments1 min readLW link

Play­ing the game vs. find­ing a cheat code

MetacelsusSep 10, 2023, 6:11 PM
34 points
1 comment3 min readLW link
(open.substack.com)

Cruxes on US lead for some do­mes­tic AI regulation

Zach Stein-PerlmanSep 10, 2023, 6:00 PM
26 points
3 comments2 min readLW link

Us­ing Nega­tive Hal­lu­ci­na­tions to Man­age Sex­ual Desire

Johannes C. MayerSep 10, 2023, 11:56 AM
−2 points
24 comments1 min readLW link

Fea­ture pro­posal: Ex­port ACX meetups

ViliamSep 10, 2023, 10:50 AM
11 points
7 comments1 min readLW link

Bet­ting and forecasting

CarlJSep 9, 2023, 8:03 PM
2 points
0 comments1 min readLW link

AI pres­i­dents dis­cuss AI al­ign­ment agendas

Sep 9, 2023, 6:55 PM
217 points
23 comments1 min readLW link
(www.youtube.com)

Prob­a­bil­is­tic ar­gu­ment re­la­tion­ships and an in­vi­ta­tion to the ar­gu­ment map­ping community

lunatic_at_largeSep 9, 2023, 6:45 PM
13 points
4 comments10 min readLW link

How teams went about their re­search at AI Safety Camp edi­tion 8

Sep 9, 2023, 4:34 PM
28 points
0 comments13 min readLW link

Panel dis­cus­sion on AI con­scious­ness with Rob Long and Jeff Sebo

Aaron BergmanSep 9, 2023, 3:38 AM
10 points
0 commentsLW link
(www.youtube.com)

Pos­si­ble Diver­gence in AGI Risk Tol­er­ance be­tween Selfish and Altru­is­tic agents

Brad West Sep 9, 2023, 12:23 AM
1 point
1 comment2 min readLW link

Cap­ture the Flag Mechanis­tic In­ter­pretabil­ity Challenges

Sep 8, 2023, 11:00 PM
24 points
0 comments7 min readLW link

[Question] What is to be done? (About the profit mo­tive)

Connor BarberSep 8, 2023, 7:27 PM
1 point
21 comments1 min readLW link

What is the op­ti­mal fron­tier for due dili­gence?

Sep 8, 2023, 6:20 PM
41 points
1 comment1 min readLW link

Progress links di­gest, 2023-09-08: The Con­ser­va­tive Fu­tur­ist, cargo air­ships, and more

jasoncrawfordSep 8, 2023, 5:48 PM
14 points
7 comments5 min readLW link
(rootsofprogress.org)

The AI apoc­a­lypse myth.

Spiritus DeiSep 8, 2023, 5:43 PM
−22 points
12 comments2 min readLW link

Sum-thresh­old attacks

TsviBTSep 8, 2023, 5:13 PM
238 points
55 comments10 min readLW link
(tsvibt.blogspot.com)

De­bate se­ries: should we push for a pause on the de­vel­op­ment of AI?

XodarapSep 8, 2023, 4:29 PM
39 points
1 commentLW link

AI Prob­a­bil­ity Trees—Joe Car­l­smith (2022)

Nathan YoungSep 8, 2023, 3:40 PM
12 points
1 comment8 min readLW link

In­vad­ing Aus­tralia (End­less Former­lies Most Beau­tiful, or What I Learned On My Holi­day)

Oliver SourbutSep 8, 2023, 3:33 PM
12 points
1 comment8 min readLW link
(www.oliversourbut.net)

Ex­plain­ing grokking through cir­cuit efficiency

Sep 8, 2023, 2:39 PM
101 points
11 comments3 min readLW link
(arxiv.org)

Have At­ten­tion Spans Been De­clin­ing?

niplavSep 8, 2023, 2:11 PM
72 points
22 comments17 min readLW link1 review

Ex­plained Sim­ply: Quantilizers

brookSep 8, 2023, 12:54 PM
15 points
5 commentsLW link
(aisafetyexplained.substack.com)

Cross­ing the Ru­bi­con.

Spiritus DeiSep 8, 2023, 4:19 AM
−4 points
5 comments13 min readLW link

[Question] What EY and LessWrong meant when (fill in the blank) found them.

Bill BenzonSep 8, 2023, 1:42 AM
1 point
0 comments1 min readLW link

Bring back the Colosseums

lcSep 8, 2023, 12:09 AM
18 points
28 comments1 min readLW link

The Löbian Ob­sta­cle, And Why You Should Care

lukemarksSep 7, 2023, 11:59 PM
18 points
6 comments2 min readLW link

Science to Be Done In­ter­na­tion­ally Us­ing Blockchain

Victor PortonSep 7, 2023, 11:29 PM
−18 points
0 comments2 min readLW link
(science-dao.org)