Con­tra Com­mon Knowledge

abramdemskiJan 4, 2023, 10:50 PM
52 points
31 comments16 min readLW link

Ad­di­tional space com­plex­ity isn’t always a use­ful met­ric

Brendan LongJan 4, 2023, 9:53 PM
4 points
3 comments3 min readLW link
(www.brendanlong.com)

List of links for get­ting into AI safety

zefJan 4, 2023, 7:45 PM
6 points
0 comments1 min readLW link

Open­ing Face­book Links Externally

jefftkJan 4, 2023, 7:00 PM
12 points
3 comments1 min readLW link
(www.jefftk.com)

Con­ver­sa­tional canyons

Henrik KarlssonJan 4, 2023, 6:55 PM
59 points
4 comments7 min readLW link
(escapingflatland.substack.com)

Progress links and tweets, 2023-01-04

jasoncrawfordJan 4, 2023, 6:23 PM
15 points
0 comments1 min readLW link
(rootsofprogress.org)

200 COP in MI: Analysing Train­ing Dynamics

Neel NandaJan 4, 2023, 4:08 PM
16 points
0 comments14 min readLW link

What’s up with ChatGPT and the Tur­ing Test?

Jan 4, 2023, 3:37 PM
13 points
19 comments3 min readLW link

2022 was the year AGI ar­rived (Just don’t call it that)

Logan ZoellnerJan 4, 2023, 3:19 PM
101 points
60 comments3 min readLW link

From Si­mon’s ant to ma­chine learn­ing, a parable

Bill BenzonJan 4, 2023, 2:37 PM
6 points
5 comments2 min readLW link

Ba­sic Facts about Lan­guage Model Internals

Jan 4, 2023, 1:01 PM
130 points
19 comments9 min readLW link

Ri­tual as the only tool for over­writ­ing val­ues and goals

mrcbarbierJan 4, 2023, 11:11 AM
41 points
24 comments32 min readLW link

Nor­malcy bias and Base rate ne­glect: Bias in Eval­u­at­ing AGI X-Risks

RemmeltJan 4, 2023, 3:16 AM
−16 points
0 commentsLW link

Causal rep­re­sen­ta­tion learn­ing as a tech­nique to pre­vent goal misgeneralization

PabloAMCJan 4, 2023, 12:07 AM
21 points
0 comments8 min readLW link

What makes a prob­a­bil­ity ques­tion “well-defined”? (Part II: Ber­trand’s Para­dox)

Noah TopperJan 3, 2023, 10:39 PM
7 points
3 comments9 min readLW link
(naivebayes.substack.com)

“AI” is an indexical

TW123Jan 3, 2023, 10:00 PM
10 points
0 comments6 min readLW link
(aiwatchtower.substack.com)

An ML in­ter­pre­ta­tion of Shard Theory

berenJan 3, 2023, 8:30 PM
39 points
5 comments4 min readLW link

Talk­ing to God

abramdemskiJan 3, 2023, 8:14 PM
30 points
7 comments2 min readLW link

My Ad­vice for In­com­ing SERI MATS Scholars

Johannes C. MayerJan 3, 2023, 7:25 PM
58 points
6 comments4 min readLW link

Touch re­al­ity as soon as pos­si­ble (when do­ing ma­chine learn­ing re­search)

LawrenceCJan 3, 2023, 7:11 PM
117 points
9 comments8 min readLW link1 review

Kolb’s: an ap­proach to con­sciously get bet­ter at anything

jacquesthibsJan 3, 2023, 6:16 PM
12 points
1 comment6 min readLW link

[Question] {M|Im|Am}oral Mazes—any large-scale coun­terex­am­ples?

DagonJan 3, 2023, 4:43 PM
24 points
4 comments1 min readLW link

Effec­tively self-study­ing over the Internet

libaiJan 3, 2023, 4:23 PM
11 points
1 comment4 min readLW link

Set-like math­e­mat­ics in type theory

Thomas KehrenbergJan 3, 2023, 2:33 PM
5 points
1 comment13 min readLW link

Monthly Roundup #2

ZviJan 3, 2023, 12:50 PM
23 points
3 comments23 min readLW link
(thezvi.wordpress.com)

Whisper’s Wild Implications

Ollie JJan 3, 2023, 12:17 PM
19 points
6 comments5 min readLW link

How to eat potato chips while typing

KatjaGraceJan 3, 2023, 11:50 AM
45 points
12 comments1 min readLW link
(worldspiritsockpuppet.com)

[Question] I have thou­sands of copies of HPMOR in Rus­sian. How to use them with the most im­pact?

Mikhail SaminJan 3, 2023, 10:21 AM
26 points
3 comments1 min readLW link

Is re­cur­sive self-al­ign­ment pos­si­ble?

No77eJan 3, 2023, 9:15 AM
5 points
5 comments1 min readLW link

On the nat­u­ral­is­tic study of the lin­guis­tic be­hav­ior of ar­tifi­cial intelligence

Bill BenzonJan 3, 2023, 9:06 AM
1 point
0 comments4 min readLW link

SF Se­vere Weather Warning

stavrosJan 3, 2023, 6:04 AM
3 points
3 comments1 min readLW link
(news.ycombinator.com)

Sta­tus quo bias; Sys­tem jus­tifi­ca­tion: Bias in Eval­u­at­ing AGI X-Risks

Jan 3, 2023, 2:50 AM
−11 points
0 comments1 min readLW link

200 COP in MI: Ex­plor­ing Poly­se­man­tic­ity and Superposition

Neel NandaJan 3, 2023, 1:52 AM
34 points
6 comments16 min readLW link

The need for speed in web frame­works?

Adam ZernerJan 3, 2023, 12:06 AM
19 points
2 comments8 min readLW link

[Si­mu­la­tors sem­i­nar se­quence] #1 Back­ground & shared assumptions

Jan 2, 2023, 11:48 PM
50 points
4 comments3 min readLW link

Lin­ear Alge­bra Done Right, Axler

David UdellJan 2, 2023, 10:54 PM
57 points
6 comments9 min readLW link

MacArthur BART (Filk)

Gordon Seidoh WorleyJan 2, 2023, 10:50 PM
10 points
1 comment1 min readLW link

Knottiness

abramdemskiJan 2, 2023, 10:13 PM
43 points
4 comments2 min readLW link

[Question] De­fault Sort for Short­forms is Very Bad; How Do I Change It?

DragonGodJan 2, 2023, 9:50 PM
15 points
0 comments1 min readLW link

MAKE IT BETTER (a po­etic demon­stra­tion of the ba­nal­ity of GPT-3)

rogersbaconJan 2, 2023, 8:47 PM
7 points
2 comments5 min readLW link

Re­view of “Make Peo­ple Bet­ter”

MetacelsusJan 2, 2023, 8:30 PM
10 points
0 comments3 min readLW link
(denovo.substack.com)

Prepar­ing for Less Privacy

jefftkJan 2, 2023, 8:30 PM
23 points
1 comment2 min readLW link
(www.jefftk.com)

Large lan­guage mod­els can provide “nor­ma­tive as­sump­tions” for learn­ing hu­man preferences

Stuart_ArmstrongJan 2, 2023, 7:39 PM
29 points
12 comments3 min readLW link

On the Im­por­tance of Open Sourc­ing Re­ward Models

elandgreJan 2, 2023, 7:01 PM
18 points
5 comments6 min readLW link

Pre­dic­tion Mar­kets for Science

VaniverJan 2, 2023, 5:55 PM
27 points
7 comments5 min readLW link

Why don’t Ra­tion­al­ists use bidets?

LakinJan 2, 2023, 5:42 PM
31 points
33 comments2 min readLW link

Soft op­ti­miza­tion makes the value tar­get bigger

Jeremy GillenJan 2, 2023, 4:06 PM
119 points
20 comments12 min readLW link

Re­sults from the AI test­ing hackathon

Esben KranJan 2, 2023, 3:46 PM
13 points
0 commentsLW link

In­duc­tion heads—illustrated

CallumMcDougallJan 2, 2023, 3:35 PM
128 points
12 comments3 min readLW link

Op­por­tu­nity Cost Blackmail

adamShimiJan 2, 2023, 1:48 PM
70 points
11 comments2 min readLW link
(epistemologicalvigilance.substack.com)