Call for con­trib­u­tors to the Align­ment Newsletter

Rohin ShahAug 21, 2019, 6:21 PM
39 points
0 comments4 min readLW link

Two senses of “op­ti­mizer”

Joar SkalseAug 21, 2019, 4:02 PM
35 points
41 comments3 min readLW link

Para­dox­i­cal Ad­vice Thread

HazardAug 21, 2019, 2:50 PM
13 points
10 comments1 min readLW link

Three Levels of Motivation

DragonGodAug 21, 2019, 9:24 AM
15 points
1 comment7 min readLW link

Odds are not easier

MrMindAug 21, 2019, 8:34 AM
9 points
6 comments1 min readLW link

GPT-2: 6-Month Fol­low-Up

lifelonglearnerAug 21, 2019, 5:06 AM
28 points
1 comment1 min readLW link

Lana Wa­chowski is do­ing a new Ma­trix movie

mako yassAug 21, 2019, 12:47 AM
5 points
3 comments1 min readLW link

[Question] What au­thors con­sis­tently give ac­cu­rate pic­tures of com­plex top­ics they dis­cuss?

seezAug 21, 2019, 12:09 AM
34 points
3 comments1 min readLW link

[Question] What are the rea­sons to *not* con­sider re­duc­ing AI-Xrisk the high­est pri­or­ity cause?

David Scott Krueger (formerly: capybaralet)Aug 20, 2019, 9:45 PM
29 points
27 comments1 min readLW link

[Question] Has Moore’s Law ac­tu­ally slowed down?

Matthew BarnettAug 20, 2019, 7:18 PM
14 points
7 comments1 min readLW link

Cere­bras Sys­tems un­veils a record 1.2 trillion tran­sis­tor chip for AI

avturchinAug 20, 2019, 2:36 PM
7 points
4 comments1 min readLW link
(venturebeat.com)

Lis­bon SSC Meetup #1

tamkin&popkinAug 20, 2019, 12:20 PM
1 point
0 comments1 min readLW link

Self-su­per­vised learn­ing & ma­nipu­la­tive predictions

Steven ByrnesAug 20, 2019, 10:55 AM
18 points
14 comments9 min readLW link

Nega­tive “eeny meeny miny moe”

jefftkAug 20, 2019, 2:48 AM
25 points
6 comments1 min readLW link

Why I Am Not a Technocrat

SpugpowAug 20, 2019, 2:06 AM
−3 points
4 commentsLW link
(radicalxchange.org)

A mis­con­cep­tion about immigration

limerottAug 19, 2019, 10:37 PM
1 point
9 comments4 min readLW link
(limerott.com)

[Question] Do We Change Our Minds Less Often Than We Think?

RaemonAug 19, 2019, 9:37 PM
20 points
5 comments1 min readLW link

Clas­sify­ing speci­fi­ca­tion prob­lems as var­i­ants of Good­hart’s Law

VikaAug 19, 2019, 8:40 PM
72 points
5 comments5 min readLW link1 review

Unstriving

Jacob FalkovichAug 19, 2019, 2:31 PM
38 points
7 comments6 min readLW link

Good­hart’s Curse and Limi­ta­tions on AI Alignment

Gordon Seidoh WorleyAug 19, 2019, 7:57 AM
25 points
18 comments10 min readLW link

Raph Koster on Vir­tual Wor­lds vs Games (notes)

RaemonAug 18, 2019, 7:01 PM
26 points
8 comments2 min readLW link

“Can We Sur­vive Tech­nol­ogy” by von Neumann

Ben PaceAug 18, 2019, 6:58 PM
33 points
2 comments1 min readLW link
(geosci.uchicago.edu)

Prokary­ote Mul­ti­verse. An ar­gu­ment that po­ten­tial simu­la­tors do not have sig­nifi­cantly more com­plex physics than ours

mako yassAug 18, 2019, 4:22 AM
0 points
5 comments2 min readLW link

Neu­ral Nets in Python 1

lifelonglearnerAug 18, 2019, 2:48 AM
10 points
3 comments8 min readLW link

In­spec­tion Para­dox as a Driver of Group Separation

ShmiAug 17, 2019, 9:47 PM
29 points
0 comments1 min readLW link

South Bay Meetup

David FriedmanAug 17, 2019, 7:56 PM
1 point
0 commentsLW link

Prob­lems in AI Align­ment that philoso­phers could po­ten­tially con­tribute to

Wei DaiAug 17, 2019, 5:38 PM
79 points
14 comments2 min readLW link

[Question] How can you use mu­sic to boost learn­ing?

Matthew BarnettAug 17, 2019, 6:59 AM
11 points
1 comment1 min readLW link

A Primer on Ma­trix Calcu­lus, Part 3: The Chain Rule

Matthew BarnettAug 17, 2019, 1:50 AM
12 points
4 comments6 min readLW link

Nashville SSC Septem­ber Meetup

friedelcraftinessAug 16, 2019, 3:16 PM
1 point
0 comments1 min readLW link

Beliefs Are For True Things

Davis_KingsleyAug 15, 2019, 11:23 PM
8 points
5 comments3 min readLW link

[Question] What ex­per­i­ments would demon­strate “up­per limits of aug­mented work­ing mem­ory?”

RaemonAug 15, 2019, 10:09 PM
33 points
6 comments2 min readLW link

Clar­ify­ing some key hy­pothe­ses in AI alignment

Aug 15, 2019, 9:29 PM
79 points
12 comments9 min readLW link

Tesser­cube — OpenPGP Made Mobile

Suji YanAug 15, 2019, 9:34 AM
4 points
0 comments1 min readLW link

A Primer on Ma­trix Calcu­lus, Part 2: Ja­co­bi­ans and other fun

Matthew BarnettAug 15, 2019, 1:13 AM
22 points
7 comments7 min readLW link

Par­tial sum­mary of de­bate with Ben­quo and Jes­si­cata [pt 1]

RaemonAug 14, 2019, 8:02 PM
89 points
63 comments22 min readLW link3 reviews

“De­sign­ing agent in­cen­tives to avoid re­ward tam­per­ing”, DeepMind

gwernAug 14, 2019, 4:57 PM
28 points
15 commentsLW link
(medium.com)

Subagents, trauma and rationality

Kaj_SotalaAug 14, 2019, 1:14 PM
113 points
4 comments19 min readLW link

Pre­dicted AI al­ign­ment event/​meet­ing calendar

rmoehnAug 14, 2019, 7:14 AM
29 points
14 comments1 min readLW link

Nat­u­ral laws should be ex­plicit con­straints on strat­egy space

ryan_bAug 13, 2019, 8:22 PM
8 points
6 comments1 min readLW link

Dis­tance Func­tions are Hard

Grue_SlinkyAug 13, 2019, 5:33 PM
31 points
19 comments6 min readLW link

Book Re­view: Sec­u­lar Cycles

Scott AlexanderAug 13, 2019, 4:10 AM
62 points
10 comments16 min readLW link1 review
(slatestarcodex.com)

A Primer on Ma­trix Calcu­lus, Part 1: Ba­sic review

Matthew BarnettAug 12, 2019, 11:44 PM
25 points
4 comments7 min readLW link

[Question] What ex­plana­tory power does Kah­ne­man’s Sys­tem 2 pos­sess?

Richard_NgoAug 12, 2019, 3:23 PM
31 points
2 comments1 min readLW link

Mesa-Op­ti­miz­ers and Over-op­ti­miza­tion Failure (Op­ti­miz­ing and Good­hart Effects, Clar­ify­ing Thoughts—Part 4)

DavidmanheimAug 12, 2019, 8:07 AM
15 points
3 comments4 min readLW link

Ad­jec­tives from the Fu­ture: The Dangers of Re­sult-based Descriptions

Pradeep_KumarAug 11, 2019, 7:19 PM
19 points
8 comments11 min readLW link

[Question] Could we solve this email mess if we all moved to paid emails?

Bird ConceptAug 11, 2019, 4:31 PM
29 points
50 comments4 min readLW link

AI Safety Read­ing Group

Søren ElverlinAug 11, 2019, 9:01 AM
16 points
8 comments1 min readLW link

[Question] Does hu­man choice have to be tran­si­tive in or­der to be ra­tio­nal/​con­sis­tent?

jmhAug 11, 2019, 1:49 AM
9 points
6 comments1 min readLW link

Di­ana Fleischman and Ge­offrey Miller—Au­di­ence Q&A

Jacob FalkovichAug 10, 2019, 10:37 PM
38 points
6 comments9 min readLW link