Every­body Wants to Rule the Fu­ture—Is Longter­mism’s Man­date of Heaven by Arith­metic Jus­tified?

E.G. Blee-Goldman19 Jan 2026 23:31 UTC
6 points
10 comments9 min readLW link

What can Kick­starter teach us about goal com­ple­tion?

Elijah19 Jan 2026 22:03 UTC
13 points
0 comments4 min readLW link

All (Non-Triv­ial) De­ci­sions Are Undecidable

(M)ason19 Jan 2026 21:51 UTC
−9 points
1 comment1 min readLW link

Pre­train­ing on Aligned AI Data Dra­mat­i­cally Re­duces Misal­ign­ment—Even After Post-Training

RogerDearnaley19 Jan 2026 21:24 UTC
106 points
12 comments11 min readLW link
(arxiv.org)

Med­i­cal Roundup #6

Zvi19 Jan 2026 21:20 UTC
31 points
2 comments11 min readLW link
(thezvi.wordpress.com)

Could LLM al­ign­ment re­search re­duce x-risk if the first takeover-ca­pa­ble AI is not an LLM?

Tim Hua19 Jan 2026 18:09 UTC
25 points
2 comments6 min readLW link

AGI both does and doesn’t have an in­finite time horizon

Sean Herrington19 Jan 2026 16:57 UTC
15 points
0 comments4 min readLW link

Desider­ata of good prob­lems to hand off to AIs

Jozdien19 Jan 2026 16:55 UTC
29 points
1 comment4 min readLW link

Test­ing few-shot coup probes

Joey Marcellino19 Jan 2026 16:31 UTC
7 points
0 comments4 min readLW link

The Example

Valerii K.19 Jan 2026 15:27 UTC
10 points
0 comments10 min readLW link

How to think about en­e­mies: the ex­am­ple of Greenpeace

19 Jan 2026 11:02 UTC
18 points
22 comments10 min readLW link
(cognition.cafe)

Silent Agree­ment Evaluation

Graeme Ford19 Jan 2026 9:11 UTC
5 points
0 comments9 min readLW link

Grad­ual Paths to Col­lec­tive Flourishing

Nora_Ammann19 Jan 2026 7:52 UTC
50 points
10 comments13 min readLW link

“Le­murian Time War” by Ccru

Nathan Delisle19 Jan 2026 5:37 UTC
7 points
0 comments1 min readLW link

Five Th­e­ses on AI Art

jenn19 Jan 2026 4:24 UTC
61 points
16 comments8 min readLW link

@Last­bas­tionof­so­bri­ety & The Sin­gu­lar­ity

AdamLacerdo19 Jan 2026 0:45 UTC
4 points
0 comments16 min readLW link

VLAs as Model Or­ganisms for AI Safety

TheSigillite18 Jan 2026 22:40 UTC
16 points
0 comments6 min readLW link

“The first two weeks are the hard­est”: my first digi­tal declutter

mingyuan18 Jan 2026 22:04 UTC
219 points
11 comments2 min readLW link
(mingyuan.substack.com)

When the LLM isn’t the one who’s wrong

Julian Bradshaw18 Jan 2026 21:37 UTC
81 points
9 comments2 min readLW link

Lifelink™: Free­dom for your Child

TsviBT18 Jan 2026 20:35 UTC
9 points
1 comment3 min readLW link

How to Love Them Equally

Shoshannah Tekofsky18 Jan 2026 17:09 UTC
38 points
5 comments2 min readLW link
(shoshanigans.substack.com)

Mas­sive Ac­ti­va­tions in DroPE: Ev­i­dence for At­ten­tion Reorganization

David Africa18 Jan 2026 15:05 UTC
19 points
0 comments8 min readLW link

Ir­ra­tional­ity as a Defense Mechanism for Re­ward-hacking

Ashe Vazquez Nuñez18 Jan 2026 3:57 UTC
49 points
8 comments4 min readLW link

Blog­ging, Writ­ing, Mus­ing, And Thinking

sonicrocketman18 Jan 2026 3:28 UTC
11 points
4 comments3 min readLW link
(brianschrader.com)

Is METR Un­der­es­ti­mat­ing LLM Time Hori­zons?

andreasrobinson18 Jan 2026 1:19 UTC
40 points
6 comments17 min readLW link

Un­der­stand­ing Trust: Pro­ject Update

abramdemski17 Jan 2026 21:19 UTC
66 points
0 comments2 min readLW link

Fo­cus­ing on Flour­ish­ing Even When Sur­vival is Un­likely (Part I)

Cleo Nardo17 Jan 2026 18:47 UTC
24 points
3 comments4 min readLW link

The truth be­hind the 2026 J.P. Mor­gan Health­care Conference

Abhishaike Mahajan17 Jan 2026 17:28 UTC
83 points
35 comments9 min readLW link
(www.owlposting.com)

Ja­pan is a bank

bhauth17 Jan 2026 16:33 UTC
22 points
2 comments1 min readLW link
(www.bhauth.com)

Turn­ing Down the Over­think­ing: How Catho­dal Brain Stim­u­la­tion Could Trans­form Stut­ter­ing Therapy

Rudaiba17 Jan 2026 14:54 UTC
9 points
0 comments8 min readLW link

What Wash­ing­ton Says About AGI

Zephaniah Roe17 Jan 2026 5:43 UTC
134 points
7 comments6 min readLW link

Light­cone is hiring a gen­er­al­ist, a de­signer, and a cam­pus op­er­a­tions co-lead

habryka17 Jan 2026 1:47 UTC
118 points
0 comments5 min readLW link

Ap­ply­ing to MATS: What the Pro­gram Is Like, and Who It’s For

17 Jan 2026 0:25 UTC
24 points
1 comment5 min readLW link

Forfeit­ing Ill-Got­ten Gains

jefftk17 Jan 2026 0:20 UTC
47 points
6 comments1 min readLW link
(www.jefftk.com)

Is It Rea­son­ing or Just a Fixed Bias?

Sriram Kiron16 Jan 2026 21:43 UTC
14 points
0 comments1 min readLW link
(ramaway.com)

Fu­ture-as-La­bel: Scal­able Su­per­vi­sion from Real-World Outcomes

Ben Turtel16 Jan 2026 21:21 UTC
−1 points
2 comments1 min readLW link

Com­par­ing your­self to other people

dominicq16 Jan 2026 20:31 UTC
10 points
3 comments2 min readLW link
(sundaystopwatch.eu)

Elic­it­ing Fron­tier Model Char­ac­ter Training

avikrishna16 Jan 2026 20:15 UTC
1 point
0 comments7 min readLW link

Prece­dents for the Un­prece­dented: His­tor­i­cal Analo­gies for Thir­teen Ar­tifi­cial Su­per­in­tel­li­gence Risks

James_Miller16 Jan 2026 18:43 UTC
165 points
15 comments63 min readLW link

Why fal­ling la­bor share ≠ fal­ling employment

Lydia Nottingham16 Jan 2026 17:27 UTC
−5 points
3 comments2 min readLW link
(lydianottingham.substack.com)

Digi­tal Minds: A Quick­start Guide

16 Jan 2026 17:15 UTC
10 points
1 comment22 min readLW link
(aviparrack.substack.com)

The cul­ture and de­sign of hu­man-AI interactions

zef16 Jan 2026 17:11 UTC
2 points
0 comments4 min readLW link
(bloodsteel.substack.com)

Con­fes­sion: I pranked Inkhaven to make sure no one fails

Mikhail Samin16 Jan 2026 16:03 UTC
52 points
1 comment10 min readLW link
(open.substack.com)

Monthly Roundup #38: Jan­uary 2026

Zvi16 Jan 2026 15:10 UTC
22 points
3 comments26 min readLW link
(thezvi.wordpress.com)

Scal­ing Laws for Eco­nomic Im­pacts: Ex­per­i­men­tal Ev­i­dence from 500 Pro­fes­sion­als and 13 LLMs

Ali Merali16 Jan 2026 13:40 UTC
21 points
6 comments4 min readLW link

[Pre-print] Build­ing safe AGI as an er­gonomics problem

16 Jan 2026 13:18 UTC
1 point
0 comments1 min readLW link
(doi.org)

Pow­er­ful mis­al­igned AIs may be ex­tremely per­sua­sive, es­pe­cially ab­sent mitigations

Cody Rushing16 Jan 2026 8:08 UTC
68 points
5 comments14 min readLW link

How to Use Foam Earplugs Correctly

Morpheus16 Jan 2026 7:47 UTC
8 points
2 comments1 min readLW link
(www.tassiloneubauer.com)

Should con­trol down-weight nega­tive net-sab­o­tage-value threats?

Fabien Roger16 Jan 2026 4:18 UTC
35 points
0 comments10 min readLW link

The De­fault Con­tra Dance Week­end Deal

jefftk16 Jan 2026 0:50 UTC
12 points
0 comments5 min readLW link
(www.jefftk.com)