All200520062007200820092010201120122013201420152016201720182019202020212022202320242025
AllJanFebMarAprMayJunJulAugSepOctNovDec
All12345678910111213141516171819202122232425262728293031

[Linkpost] Jan Leike on three kinds of al­ign­ment taxes

Orpheus16Jan 6, 2023, 11:57 PM
27 points
2 comments3 min readLW link
(aligned.substack.com)

The Limit of Lan­guage Models

DragonGodJan 6, 2023, 11:53 PM
44 points
26 comments4 min readLW link

Why didn’t we get the four-hour work­day?

jasoncrawfordJan 6, 2023, 9:29 PM
141 points
34 comments6 min readLW link
(rootsofprogress.org)

AI se­cu­rity might be helpful for AI alignment

Igor IvanovJan 6, 2023, 8:16 PM
36 points
1 comment2 min readLW link

Cat­e­go­riz­ing failures as “outer” or “in­ner” mis­al­ign­ment is of­ten confused

Rohin ShahJan 6, 2023, 3:48 PM
93 points
21 comments8 min readLW link

Defi­ni­tions of “ob­jec­tive” should be Prob­a­ble and Predictive

Rohin ShahJan 6, 2023, 3:40 PM
43 points
27 comments12 min readLW link

200 COP in MI: Tech­niques, Tool­ing and Automation

Neel NandaJan 6, 2023, 3:08 PM
13 points
0 comments15 min readLW link

Ball Square Sta­tion and Rider­ship Maximization

jefftkJan 6, 2023, 1:20 PM
13 points
0 comments1 min readLW link
(www.jefftk.com)

Child­hood Roundup #1

ZviJan 6, 2023, 1:00 PM
84 points
27 comments8 min readLW link
(thezvi.wordpress.com)

AI im­prov­ing AI [MLAISU W01!]

Esben KranJan 6, 2023, 11:13 AM
5 points
0 comments4 min readLW link
(newsletter.apartresearch.com)

AI Safety Camp, Vir­tual Edi­tion 2023

Linda LinseforsJan 6, 2023, 11:09 AM
40 points
10 comments3 min readLW link
(aisafety.camp)

Kakistocuriosity

LVSNJan 6, 2023, 7:38 AM
7 points
3 comments1 min readLW link

AI Safety Camp: Ma­chine Learn­ing for Scien­tific Dis­cov­ery

Eleni AngelouJan 6, 2023, 3:21 AM
3 points
0 comments1 min readLW link

Me­tac­u­lus Year in Re­view: 2022

ChristianWilliamsJan 6, 2023, 1:23 AM
6 points
0 commentsLW link

UDASSA

Jacob FalkovichJan 6, 2023, 1:07 AM
27 points
8 comments10 min readLW link

The In­vol­un­tary Pacifists

CapybasiliskJan 6, 2023, 12:28 AM
11 points
3 comments2 min readLW link

Get an Elec­tric Tooth­brush.

CerveraJan 5, 2023, 9:08 PM
21 points
4 comments1 min readLW link

Dis­cur­sive Com­pe­tence in ChatGPT, Part 1: Talk­ing with Dragons

Bill BenzonJan 5, 2023, 9:01 PM
2 points
0 comments6 min readLW link

Trans­for­ma­tive AI is­sues (not just mis­al­ign­ment): an overview

HoldenKarnofskyJan 5, 2023, 8:20 PM
34 points
6 comments18 min readLW link
(www.cold-takes.com)

How to slow down sci­en­tific progress, ac­cord­ing to Leo Szilard

jasoncrawfordJan 5, 2023, 6:26 PM
134 points
18 comments2 min readLW link
(rootsofprogress.org)

Paper: Su­per­po­si­tion, Me­moriza­tion, and Dou­ble Des­cent (An­thropic)

LawrenceCJan 5, 2023, 5:54 PM
53 points
11 comments1 min readLW link
(transformer-circuits.pub)

Col­lapse Might Not Be Desirable

DzoldzayaJan 5, 2023, 5:29 PM
−2 points
9 comments2 min readLW link

Sin­ga­pore—Small ca­sual din­ner in Chi­na­town #6

Joe RoccaJan 5, 2023, 5:00 PM
2 points
1 comment1 min readLW link

[Question] Image gen­er­a­tion and al­ign­ment

rpglover64Jan 5, 2023, 4:05 PM
3 points
3 comments1 min readLW link

[Question] Ma­chine Learn­ing vs Differ­en­tial Privacy

IlioJan 5, 2023, 3:14 PM
10 points
10 comments1 min readLW link

Covid 1/​5/​23: Var­i­ous XBB Takes

ZviJan 5, 2023, 2:20 PM
21 points
18 comments15 min readLW link
(thezvi.wordpress.com)

Run­ning by Default

jefftkJan 5, 2023, 1:50 PM
112 points
40 comments1 min readLW link
(www.jefftk.com)

PSA: re­ward is part of the habit loop too

Alok SinghJan 5, 2023, 11:00 AM
22 points
2 comments1 min readLW link
(alok.github.io)

In­fo­haz­ards vs Fork Hazards

jimrandomhJan 5, 2023, 9:45 AM
68 points
16 comments1 min readLW link

Monthly Shorts 12/​22

CelerJan 5, 2023, 7:20 AM
5 points
2 comments1 min readLW link
(keller.substack.com)

The 2021 Re­view Phase

RaemonJan 5, 2023, 7:12 AM
34 points
7 comments3 min readLW link

Illu­sion of truth effect and Am­bi­guity effect: Bias in Eval­u­at­ing AGI X-Risks

RemmeltJan 5, 2023, 4:05 AM
−13 points
2 commentsLW link

When you plan ac­cord­ing to your AI timelines, should you put more weight on the me­dian fu­ture, or the me­dian fu­ture | even­tual AI al­ign­ment suc­cess? ⚖️

Jeffrey LadishJan 5, 2023, 1:21 AM
25 points
10 comments2 min readLW link

Why I’m join­ing Anthropic

evhubJan 5, 2023, 1:12 AM
118 points
4 comments2 min readLW link

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?

JoshuaFox and Zvi Schreiber
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

beren and Eric Winsor
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)