Lex Fridman Podcast
Long-form interviews with the people building AI. Technical but accessible.
3 episodes curated
Episodes
The State of AI Research and What Comes Next
A technical but accessible deep dive into where AI research is heading. Covers reasoning capabilities, multimodal models, and the gap between benchmarks and real-world performance. Includes discussion of AI safety research and alignment progress.
Hallucinations, Bias, and Trust: The AI Reliability Problem
Deep dive into why AI models hallucinate, what's being done about it, and how professionals should calibrate their trust in AI outputs. Includes practical frameworks for verification and cross-checking that any professional can use immediately.
The Constitutional AI Approach: Building Safer Systems
Interview with Anthropic researchers about Constitutional AI — the training methodology behind Claude. How it differs from RLHF, why it produces less sycophantic responses, and what it means for the future of AI safety. Technical depth with clear explanations.