Open in PodwiseOpen

Podcast cover

LessWrong · Philosophy

LessWrong (30+ Karma)

Audio narrations of LessWrong posts.

Episodes

[Linkpost] “The Psychopathy Spectrum” by Dawn Drescher

“Sycophancy Towards Researchers Drives Performative Misalignment” by Taywon Min, rustem17, David Vella Zarb

“Extracting Performant Algorithms Using Mechanistic Interpretability” by Ihor Kendiukhov

“Requiem for a Transhuman Timeline” by Ihor Kendiukhov

“Adding Typos Made Haiku’s Accuracy Go Up” by bira

“LLMs as Giant Lookup-Tables of Shallow Circuits” by niplav, Claude+

“Medical Roundup #7” by Zvi

“Types of Handoff to AIs” by Daniel Kokotajlo

“You can’t imitation-learn how to continual-learn” by Steven Byrnes

“PSA: Predictions markets often have very low liquidity; be careful citing them.” by Eye You

“AICRAFT: DARPA-Funded AI Alignment Researchers — Applications Open” by Mike Vaiana, Diogo de Lucena, Judd Rosenblatt

“Customer Satisfaction Opportunities” by Tomás B.

“LLM Misalignment Can be One Gradient Step Away, and Blackbox Evaluation Cannot Detect It.” by Yavuz Bakman

“Compradorization” by Benquo

“Models differ in identity propensities” by Jan_Kulveit, Raymond Douglas, vgel, owencb, David Duvenaud

“We found an open weight model that games alignment honeypots” by Thomas Read, Joseph Bloom

“Terrified Comments on Corrigibility in Claude’s Constitution” by Zack_M_Davis

“We Started Lens Academy: Scalable Education on Superintelligence Risk” by Luc Brinkman, meriton, pleiotroth, Chris-Lons

“Mini-Munich Succeeds Where KidZania Fails” by Novalis

“Inputs, outputs, and valued outcomes” by Kaj_Sotala

Page 68

Follow this podcast in Podwise

Sign in to get AI summaries, transcripts and mind maps for any episode, including new ones.

Open in Podwise