AI Takeover Sce­nario with Scaled LLMs

simeon_cApr 16, 2023, 11:28 PM
42 points
15 comments8 min readLW link

My ex­pe­rience get­ting fund­ing for my biolog­i­cal research

MetacelsusApr 16, 2023, 10:53 PM
78 points
10 comments5 min readLW link
(denovo.substack.com)

Top les­son from GPT: we will prob­a­bly de­stroy hu­man­ity “for the lulz” as soon as we are able.

ShmiApr 16, 2023, 8:27 PM
63 points
28 comments1 min readLW link

On ur­gency, pri­or­ity and col­lec­tive re­ac­tion to AI-Risks: Part I

DenreikApr 16, 2023, 7:14 PM
−10 points
15 comments5 min readLW link

Effi­cient Learn­ing: Memorization

Alvin ÅnestrandApr 16, 2023, 5:58 PM
4 points
2 comments5 min readLW link
(forum.effectivealtruism.org)

Mechanis­ti­cally in­ter­pret­ing time in GPT-2 small

Apr 16, 2023, 5:57 PM
68 points
6 comments21 min readLW link

La Crosse, WI Ra­tion­al­ity Meetup

Daniel UebeleApr 16, 2023, 5:33 PM
1 point
0 comments1 min readLW link

The Soul of the Writer (on LLMs, the psy­chol­ogy of writ­ers, and the na­ture of in­tel­li­gence)

rogersbaconApr 16, 2023, 4:02 PM
11 points
1 comment3 min readLW link
(www.secretorum.life)

Pos­si­bi­liz­ing vs. actualizing

TsviBTApr 16, 2023, 3:55 PM
31 points
2 comments5 min readLW link

Hu­man Ex­tinc­tion by AI through eco­nomic power

ChristianKlApr 16, 2023, 12:15 PM
8 points
1 comment8 min readLW link

Bit Flip

Charlie SandersApr 16, 2023, 7:30 AM
−2 points
11 comments11 min readLW link

Dou­ble-nega­tion as framing

Stuart JohnsonApr 16, 2023, 6:59 AM
25 points
9 comments6 min readLW link

[Link/​cross­post] [US] NTIA: AI Ac­countabil­ity Policy Re­quest for Comment

Kyle J. LuccheseApr 16, 2023, 6:57 AM
8 points
0 comments1 min readLW link
(forum.effectivealtruism.org)

[Question] Who is test­ing AI Safety pub­lic out­reach mes­sag­ing?

yanni kyriacosApr 16, 2023, 6:57 AM
13 points
2 comments1 min readLW link

Fea­tures of Emacs that I only re­cently discovered

EmacsScrubApr 16, 2023, 6:57 AM
12 points
5 comments3 min readLW link

ACX meetup in Prague (16th of May)

Jiří NádvorníkApr 16, 2023, 6:25 AM
4 points
0 comments1 min readLW link

Smar­tyHead­erCode: anoma­lous to­kens for GPT3.5 and GPT-4

AdamYedidiaApr 15, 2023, 10:35 PM
71 points
18 comments6 min readLW link

Open-source LLMs may prove Bostrom’s vuln­er­a­ble world hypothesis

Roope AhvenharjuApr 15, 2023, 7:16 PM
1 point
1 comment1 min readLW link

[linkpost] Elon Musk plans AI start-up to ri­val OpenAI

HatfieldApr 15, 2023, 7:06 PM
11 points
11 comments1 min readLW link
(www.ft.com)

FLI re­port: Poli­cy­mak­ing in the Pause

Zach Stein-PerlmanApr 15, 2023, 5:01 PM
15 points
3 comments1 min readLW link
(futureoflife.org)

Reflec­tive jour­nal en­tries us­ing GPT-4 and Ob­sidian that de­mand less willpower.

Solenoid_EntityApr 15, 2023, 12:45 PM
56 points
24 comments7 min readLW link

An ex­am­ple ele­va­tor pitch for AI doom

laserficheApr 15, 2023, 12:29 PM
2 points
5 comments1 min readLW link

AI as Con­tact with our Col­lec­tive Unconscious

Scott BroockApr 15, 2023, 2:11 AM
−4 points
6 comments4 min readLW link

The Truth About False

Thoth HermesApr 15, 2023, 1:01 AM
−21 points
4 comments17 min readLW link
(thothhermes.substack.com)

The ‘ pe­ter­todd’ phenomenon

mwatkinsApr 15, 2023, 12:59 AM
192 points
50 comments38 min readLW link1 review

[Question] Con­cave Utility Question

Scott GarrabrantApr 15, 2023, 12:14 AM
55 points
36 comments2 min readLW link

List of re­quests for an AI slow­down/​halt.

Cleo NardoApr 14, 2023, 11:55 PM
46 points
6 comments1 min readLW link

[linkpost] “What Are Rea­son­able AI Fears?” by Robin Han­son, 2023-04-23

Arjun PanicksseryApr 14, 2023, 11:26 PM
26 points
16 commentsLW link

“Do X be­cause de­ci­sion the­ory” ~= “Do X be­cause bayes the­o­rem”

lcApr 14, 2023, 8:57 PM
39 points
1 comment2 min readLW link

LLMs and hal­lu­ci­na­tion, like white on rice?

Bill BenzonApr 14, 2023, 7:53 PM
5 points
0 comments3 min readLW link

GPT-4 is eas­ily con­trol­led/​ex­ploited with tricky de­ci­sion the­o­retic dilem­mas.

scasperApr 14, 2023, 7:39 PM
6 points
4 comments2 min readLW link

On Car­ing about our AI Progeny

PeterMcCluskeyApr 14, 2023, 7:32 PM
22 points
5 comments1 min readLW link
(bayesianinvestor.com)

Moder­a­tion notes re: re­cent Said/​Dun­can threads

RaemonApr 14, 2023, 6:06 PM
50 points
560 comments2 min readLW link

What we’ve learned so far from our tech­nolog­i­cal temp­ta­tions project

Richard Korzekwa Apr 14, 2023, 5:46 PM
15 points
4 comments11 min readLW link
(aiimpacts.org)

[Question] How does con­scious­ness in­ter­act with ar­chi­tec­ture?

FinalFormal2Apr 14, 2023, 3:56 PM
5 points
3 comments1 min readLW link

Iqisa: A Library For Han­dling Fore­cast­ing Datasets

niplavApr 14, 2023, 3:16 PM
27 points
0 commentsLW link

What’s this prob­a­bil­ity you’re re­port­ing?

EOC and SCP
Apr 14, 2023, 3:07 PM
19 points
10 comments3 min readLW link

Nav­i­gat­ing AI Risks (NAIR) #1: Slow­ing Down AI

simeon_cApr 14, 2023, 2:35 PM
11 points
3 comments1 min readLW link
(navigatingairisks.substack.com)

[Question] What would the FLI mora­to­rium ac­tu­ally do?

ChristianKlApr 14, 2023, 1:14 PM
17 points
7 comments1 min readLW link

Re­search Re­port: In­cor­rect­ness Cascades

Robert_AIZIApr 14, 2023, 12:49 PM
19 points
0 comments10 min readLW link
(aizi.substack.com)

The self-un­al­ign­ment problem

Apr 14, 2023, 12:10 PM
155 points
24 comments10 min readLW link

AI Safety Europe Re­treat 2023 Retrospective

Magdalena WacheApr 14, 2023, 9:05 AM
43 points
0 comments2 min readLW link

[Question] What’s the differ­ence be­tween Wis­dom and Ra­tion­al­ity?

Yoav RavidApr 14, 2023, 6:22 AM
8 points
4 comments1 min readLW link

Shap­ley Value At­tri­bu­tion in Chain of Thought

leogaoApr 14, 2023, 5:56 AM
106 points
7 comments4 min readLW link

A fresh­man year dur­ing the AI midgame: my ap­proach to the next year

BuckApr 14, 2023, 12:38 AM
154 points
15 commentsLW link1 review

Against AI Un­der­stand­ing and Sen­tience: Large Lan­guage Models, Mean­ing, and the Pat­terns of Hu­man Lan­guage Use

Jonathan YanApr 13, 2023, 11:29 PM
−1 points
0 comments1 min readLW link
(philsci-archive.pitt.edu)

Fi­nan­cial Times: We must slow down the race to God-like AI

trevorApr 13, 2023, 7:55 PM
113 points
17 comments16 min readLW link
(www.ft.com)

R0 Is Not Counterfactual

jefftkApr 13, 2023, 7:50 PM
33 points
9 comments2 min readLW link
(www.jefftk.com)

Sub­scripts for Probabilities

niplavApr 13, 2023, 6:32 PM
67 points
9 comments5 min readLW link

The Virus—Short Story

Michael SoareverixApr 13, 2023, 6:18 PM
4 points
0 comments4 min readLW link