LessWrong’s UX may not be living up to its ideas

neo18 Mar 2026 20:16 UTC

5 points

0 comments2 min readLW link

Two Skillsets You Need to Launch an Impactful AI Safety Project

Luc Brinkman and plex

18 Mar 2026 20:09 UTC

17 points

0 comments6 min readLW link

“Act-based approval-directed agents”, for IDA skeptics

Steven Byrnes18 Mar 2026 18:47 UTC

39 points

2 comments5 min readLW link

“Lost in the Middle” Replicates

Stephen Martin18 Mar 2026 16:09 UTC

18 points

0 comments1 min readLW link

Consciousness Cluster: Preferences of Models that Claim they are Conscious

James Chua, Owain_Evans, Sam Marks and Jan Betley

18 Mar 2026 16:06 UTC

55 points

2 comments5 min readLW link

Solving Slop

Noam Makavy18 Mar 2026 5:18 UTC

−1 points

0 comments1 min readLW link

Sycophancy Towards Researchers Drives Performative Misalignment

Taywon Min, rustem17, David Vella Zarb and Shi

18 Mar 2026 4:59 UTC

70 points

0 comments21 min readLW link

Newcomb’s Paradox Simulation

shakelush18 Mar 2026 4:30 UTC

6 points

2 comments1 min readLW link

(www.onlineide.pro)

The Psychopathy Spectrum

Dawn Drescher17 Mar 2026 21:36 UTC

36 points

0 comments1 min readLW link

(impartial-priorities.org)

LLMs as Giant Lookup-Tables of Shallow Circuits

niplav and Claude+

17 Mar 2026 21:35 UTC

85 points

25 comments7 min readLW link

Requiem for a Transhuman Timeline

Ihor Kendiukhov17 Mar 2026 21:27 UTC

145 points

11 comments5 min readLW link

There is No One There: A simple experiment to convince yourself that LLMs probably are not conscious

Peter Kuhn17 Mar 2026 17:26 UTC

12 points

17 comments5 min readLW link

Research note on window shifting training

Kei Nishimura-Gasparian and np_x

17 Mar 2026 15:58 UTC

26 points

1 comment15 min readLW link

Mahalo, Kailua-Kona

jenn17 Mar 2026 15:54 UTC

28 points

0 comments9 min readLW link

[Linkpost] “Exploiting” the In-Kind Donation Bias

jknowak17 Mar 2026 8:27 UTC

1 point

0 comments1 min readLW link

(epsilonaligned.substack.com)

[Linkpost] Let’s Start Eventmaxxing

jknowak17 Mar 2026 8:22 UTC

0 points

0 comments1 min readLW link

(epsilonaligned.substack.com)

[Linkpost] Many Faces Of Problem Decomposition

jknowak17 Mar 2026 8:17 UTC

1 point

0 comments1 min readLW link

(epsilonaligned.substack.com)

How to not do decision theory backwards

Anthony DiGiovanni17 Mar 2026 7:22 UTC

14 points

0 comments16 min readLW link

I made a job-level AI capability estimator by asking “Where is AI doing similar work today?”

Alex A17 Mar 2026 4:45 UTC

6 points

1 comment1 min readLW link

The tech bro ethos, and the end of my job

PatrickDFarley17 Mar 2026 4:13 UTC

11 points

5 comments3 min readLW link