Boaz Barak shares full AI safety course videos, from Schulman to Altman
eigenron · x · 2026-11-09
Harvard professor Boaz Barak posted an update on his AI safety course on LessWrong: scribe notes for every lecture, videos where speakers permitted, and weekly student experiments — some of which may grow into papers.
The speaker lineup is remarkable: Nicholas Carlini on cyber capabilities, Dwarkesh Patel & Daniel Kokotajlo on recursive self-improvement, Chad Jones & Erik Brynjolfsson on economic impact, John Schulman on grading and reward hacking, a Q&A with Sam Altman, Nathan Lambert on open-source models, Helen Toner on frontier AI regulation, Jack Lindsey on interpretability, Jeff Sebo on model welfare, Jakub Pachocki on emerging risks, plus upcoming sessions on biosecurity and threat modeling.
All notes, videos, and readings are available on the course webpage, which Barak admits he maintains with ChatGPT.
More from Companies & People
- Jensen Huang accused of lying to Bloomberg reporters, 'everything is securities fraud' — ShakeelHashim · 2026-11-09
- Polymarket opens book on Jensen Huang exiting Nvidia CEO role by year-end at 10% odds — Polymarket · 2026-11-09
- Early Anthropic employee memoir: equity pledges, FBI briefings, traumatized bureaucrats — LessWrong 精选 · 2026-11-09
- Stanford rewrote CS146S for agents: open-source patches replace one-shot homework — Latent Space · 2026-11-09
- Axios publishes Jensen Huang's letter to Nvidia staff about his affair — KateClarkTweets · 2026-11-09
- Economist cover cast Jensen Huang as 'The Sorcerer of Silicon' just eight weeks ago — GaryMarcus · 2026-11-09