The Nat­u­ral Ab­strac­tion Hy­poth­e­sis: Im­pli­ca­tions and Evidence

CallumMcDougallDec 14, 2021, 11:14 PM
39 points
9 comments19 min readLW link

Robin Han­son’s “Hu­mans are Early”

RaemonDec 14, 2021, 10:07 PM
11 points
0 comments2 min readLW link
(www.overcomingbias.com)

Ngo’s view on al­ign­ment difficulty

Dec 14, 2021, 9:34 PM
63 points
7 comments17 min readLW link

A pro­posed sys­tem for ideas jumpstart

Valentin2026Dec 14, 2021, 9:01 PM
4 points
2 comments3 min readLW link

Should we rely on the speed prior for safety?

Marc CarauleanuDec 14, 2021, 8:45 PM
14 points
5 comments5 min readLW link

ARC’s first tech­ni­cal re­port: Elic­it­ing La­tent Knowledge

Dec 14, 2021, 8:09 PM
228 points
90 comments1 min readLW link3 reviews
(docs.google.com)

ARC is hiring!

Dec 14, 2021, 8:09 PM
64 points
2 comments1 min readLW link

In­ter­lude: Agents as Automobiles

Daniel KokotajloDec 14, 2021, 6:49 PM
26 points
6 comments5 min readLW link

Zvi’s Thoughts on the Sur­vival and Flour­ish­ing Fund (SFF)

ZviDec 14, 2021, 2:30 PM
193 points
65 comments64 min readLW link1 review
(thezvi.wordpress.com)

Con­se­quen­tial­ism & corrigibility

Steven ByrnesDec 14, 2021, 1:23 PM
70 points
29 comments7 min readLW link

Mys­tery Hunt 2022

Scott GarrabrantDec 13, 2021, 9:57 PM
30 points
5 comments1 min readLW link

En­abling More Feed­back for AI Safety Researchers

frances_lorenzDec 13, 2021, 8:10 PM
17 points
0 comments3 min readLW link

Lan­guage Model Align­ment Re­search Internships

Ethan PerezDec 13, 2021, 7:53 PM
74 points
1 comment1 min readLW link

Omicron Post #6

ZviDec 13, 2021, 6:00 PM
89 points
30 comments8 min readLW link
(thezvi.wordpress.com)

Anal­y­sis of Bird Box (2018)

TekhneMakreDec 13, 2021, 5:30 PM
11 points
3 comments5 min readLW link

Solv­ing In­ter­pretabil­ity Week

Logan RiggsDec 13, 2021, 5:09 PM
11 points
5 comments1 min readLW link

Un­der­stand­ing and con­trol­ling auto-in­duced dis­tri­bu­tional shift

L Rudolf LDec 13, 2021, 2:59 PM
33 points
4 comments16 min readLW link

A fate worse than death?

RomanSDec 13, 2021, 11:05 AM
−25 points
26 comments2 min readLW link

What’s the back­ward-for­ward FLOP ra­tio for Neu­ral Net­works?

Dec 13, 2021, 8:54 AM
20 points
12 comments10 min readLW link

Sum­mary of the Acausal At­tack Is­sue for AIXI

DiffractorDec 13, 2021, 8:16 AM
12 points
6 comments4 min readLW link

Hard-Cod­ing Neu­ral Computation

MadHatterDec 13, 2021, 4:35 AM
34 points
8 comments27 min readLW link

[Question] Is “gears-level” just a syn­onym for “mechanis­tic”?

David Scott Krueger (formerly: capybaralet)Dec 13, 2021, 4:11 AM
48 points
29 comments1 min readLW link

Baby Nicknames

jefftkDec 13, 2021, 2:20 AM
11 points
0 comments1 min readLW link
(www.jefftk.com)

[Question] Why do gov­ern­ments re­fer to ex­is­ten­tial risks pri­mar­ily in terms of na­tional se­cu­rity?

Evan_GaensbauerDec 13, 2021, 1:05 AM
3 points
3 comments1 min readLW link

[Question] [Re­solved] Who else prefers “AI al­ign­ment” to “AI safety?”

Evan_GaensbauerDec 13, 2021, 12:35 AM
5 points
8 comments1 min readLW link

Work­ing through D&D.Sci, prob­lem 1

Pablo RepettoDec 12, 2021, 11:10 PM
8 points
2 comments1 min readLW link
(pabloernesto.github.io)

Teaser: Hard-cod­ing Trans­former Models

MadHatterDec 12, 2021, 10:04 PM
74 points
19 comments1 min readLW link

The Three Mu­ta­tions of Dark Rationality

DarkRationalistDec 12, 2021, 10:01 PM
−15 points
0 comments2 min readLW link

Red­wood’s Tech­nique-Fo­cused Epistemic Strategy

adamShimiDec 12, 2021, 4:36 PM
48 points
1 comment7 min readLW link

For and Against Lot­ter­ies in Elite Univer­sity Admissions

Sam EnrightDec 12, 2021, 1:41 PM
10 points
2 comments3 min readLW link

[Question] Nu­clear war anthropics

smountjoyDec 12, 2021, 4:54 AM
11 points
7 comments1 min readLW link

Some ab­stract, non-tech­ni­cal rea­sons to be non-max­i­mally-pes­simistic about AI alignment

Rob BensingerDec 12, 2021, 2:08 AM
70 points
35 comments7 min readLW link

Magna Alta Doctrina

jacob_cannellDec 11, 2021, 9:54 PM
60 points
7 comments28 min readLW link

EA Din­ner Covid Logistics

jefftkDec 11, 2021, 9:50 PM
17 points
7 comments2 min readLW link
(www.jefftk.com)

Trans­form­ing my­opic op­ti­miza­tion to or­di­nary op­ti­miza­tion—Do we want to seek con­ver­gence for my­opic op­ti­miza­tion prob­lems?

tailcalledDec 11, 2021, 8:38 PM
12 points
1 comment5 min readLW link

What on Earth is a Series I sav­ings bond?

rossryDec 11, 2021, 12:18 PM
11 points
7 comments7 min readLW link

D&D.Sci GURPS Dec 2021: Hun­ters of Monsters

J BostockDec 11, 2021, 12:13 PM
20 points
21 comments2 min readLW link

Anx­iety and com­puter architecture

Adam ZernerDec 11, 2021, 10:37 AM
13 points
8 comments3 min readLW link

[Question] Rea­sons to act ac­cord­ing to the free will paradigm?

Maciej JałochaDec 11, 2021, 8:44 AM
−3 points
5 comments1 min readLW link

Ex­trin­sic and In­trin­sic Mo­ral Frameworks

lsusrDec 11, 2021, 5:28 AM
14 points
5 comments2 min readLW link

Moore’s Law, AI, and the pace of progress

VeedracDec 11, 2021, 3:02 AM
128 points
38 comments24 min readLW link

What role should evolu­tion­ary analo­gies play in un­der­stand­ing AI take­off speeds?

anson.hoDec 11, 2021, 1:19 AM
14 points
0 comments42 min readLW link

[Question] Non­ver­bal thinkers: how do you ex­pe­rience your in­ner critic?

Phoenix EliotDec 11, 2021, 12:40 AM
9 points
2 comments1 min readLW link

The Plan

johnswentworthDec 10, 2021, 11:41 PM
260 points
78 comments14 min readLW link1 review

[Linkpost] Chi­nese gov­ern­ment’s guidelines on AI

RomanSDec 10, 2021, 9:10 PM
61 points
14 comments1 min readLW link

Un­der­stand­ing Gra­di­ent Hacking

peterbarnettDec 10, 2021, 3:58 PM
41 points
5 comments30 min readLW link

There is es­sen­tially one best-val­i­dated the­ory of cog­ni­tion.

abramdemskiDec 10, 2021, 3:51 PM
89 points
33 comments3 min readLW link

The Promise and Peril of Finite Sets

davidadDec 10, 2021, 12:29 PM
42 points
5 comments6 min readLW link

Are big brains for pro­cess­ing sen­sory in­put?

lsusrDec 10, 2021, 7:08 AM
42 points
20 comments3 min readLW link

Com­bin­ing Forecasts

jsteinhardtDec 10, 2021, 2:10 AM
10 points
1 comment6 min readLW link
(bounded-regret.ghost.io)