AI Safety Fact vs. Fiction | YC Paper Club

Why watch

Stanford professors from the Center for AI Safety discuss rigorous scientific approaches to AI safety, covering the OpenAI-Hugging Face incident, using Lean for proving code correctness, adaptive stress testing methods borrowed from aviation, and the need for an NTSB-style accident investigation body for AI. Ideal for engineers, researchers, and technical leaders seeking evidence-based understanding of AI safety beyond media narratives.