Language models seem to be much better than humans at next-token prediction

Buck, Fabien and LawrenceC

11 Aug 2022 17:45 UTC

128 points

52 comments13 min readLW link

«Boundaries», Part 1: a key missing concept from utility theory

Andrew_Critch26 Jul 2022 23:03 UTC

113 points

16 comments7 min readLW link

Changing the world through slack & hobbies

Steven Byrnes21 Jul 2022 18:11 UTC

228 points

14 comments10 min readLW link

What should you change in response to an “emergency”? And AI risk

AnnaSalamon18 Jul 2022 1:11 UTC

290 points

59 comments6 min readLW link

Humans provide an untapped wealth of evidence about alignment

TurnTrout and Quintin Pope

14 Jul 2022 2:31 UTC

168 points

92 comments10 min readLW link

On how various plans miss the hard bits of the alignment challenge

So8res12 Jul 2022 2:49 UTC

248 points

81 comments29 min readLW link

ITT-passing and civility are good; “charity” is bad; steelmanning is niche

Rob Bensinger5 Jul 2022 0:15 UTC

137 points

32 comments6 min readLW link

Looking back on my alignment PhD

TurnTrout1 Jul 2022 3:19 UTC

283 points

58 comments11 min readLW link

It’s Probably Not Lithium

Natália Coelho Mendonça28 Jun 2022 21:24 UTC

412 points

179 comments28 min readLW link

What Are You Tracking In Your Head?

johnswentworth28 Jun 2022 19:30 UTC

215 points

73 comments4 min readLW link

Nonprofit Boards are Weird

HoldenKarnofsky23 Jun 2022 14:40 UTC

149 points

25 comments20 min readLW link

(www.cold-takes.com)

Security Mindset: Lessons from 20+ years of Software Security Failures Relevant to AGI Alignment

elspood21 Jun 2022 23:55 UTC

312 points

40 comments7 min readLW link

Where I agree and disagree with Eliezer

paulfchristiano19 Jun 2022 19:15 UTC

738 points

202 comments20 min readLW link

Humans are very reliable agents

alyssavance16 Jun 2022 22:02 UTC

249 points

35 comments3 min readLW link

AGI Ruin: A List of Lethalities

Eliezer Yudkowsky5 Jun 2022 22:05 UTC

681 points

638 comments30 min readLW link

Public beliefs vs. Private beliefs

Eli Tyre1 Jun 2022 21:33 UTC

132 points

25 comments5 min readLW link

Six Dimensions of Operational Adequacy in AGI Projects

Eliezer Yudkowsky30 May 2022 17:00 UTC

263 points

65 comments13 min readLW link

Benign Boundary Violations

Duncan_Sabien26 May 2022 6:48 UTC

197 points

85 comments18 min readLW link

Visible Homelessness in SF: A Quick Breakdown of Causes

alyssavance25 May 2022 1:40 UTC

194 points

40 comments2 min readLW link

[Intro to brain-like-AGI safety] 15. Conclusion: Open problems, how to help, AMA

Steven Byrnes17 May 2022 15:11 UTC

81 points

11 comments14 min readLW link