Podcasts & talks
If you learn better by listening, these fifteen shows and talks will keep you current with how the people inside frontier labs actually think.
Subscribe to one podcast you'll actually listen to weekly rather than five you won't. The talks are evergreen and worth revisiting as your own understanding grows.
- 01
Dwarkesh Podcast
PodcastDwarkesh Patel · ongoingLong-form interviews with frontier-lab CEOs, researchers, and historians of technology.
Why read this. The single highest signal-density AI podcast.
- 02
Machine Learning Street Talk
PodcastTim Scarfe et al. · ongoingTechnical, sometimes contrarian conversations with leading ML researchers.
Why read this. Best for technical depth and disagreement.
- 03
80,000 Hours Podcast
PodcastRob Wiblin / Luisa Rodriguez · ongoingThree-hour interviews with researchers across AI safety, biosecurity, and policy.
Why read this. Patient, careful, and unusually thorough.
- 04
Cognitive Revolution
PodcastNathan Labenz · ongoingPractitioner-focused interviews on the state of AI capabilities and deployment.
Why read this. Useful for builders and product people, not just researchers.
- 05
Latent Space
Podcastswyx & Alessio · ongoingAI engineering podcast covering the production stack and emerging tools.
Why read this. Where you hear about new tools first.
- 06
No Priors
PodcastSarah Guo & Elad Gil · ongoingInvestor and founder interviews on AI markets and infrastructure.
Why read this. The business-side complement to research podcasts.
- 07
Ilya Sutskever — Welcome and Opening Talk (NeurIPS 2024)
TalkIlya Sutskever · 2024Sutskever's reflections on the end of pretraining as we knew it.
Why read this. A short, dense statement from one of the field's defining figures.
- 08
Andrej Karpathy — Intro to LLMs
TalkAndrej Karpathy · 2023A one-hour systems-level explanation of how LLMs work.
Why read this. The clearest single talk to share with non-technical colleagues.
- 09
Yoshua Bengio — Catalyzing AI Safety Research
TalkYoshua Bengio · 2024Bengio's case for rigorous AI safety research, from a Turing-award holder who pivoted into it.
Why read this. Honest and personal.
- 10
Demis Hassabis — The Future of AI
TalkDemis Hassabis · 2024DeepMind's CEO on scientific applications and the road to AGI.
Why read this. Inside view from the lab behind AlphaFold.
- 11
Geoffrey Hinton — Why I left Google to warn about AI
TalkGeoffrey Hinton · 2023Hinton explains his shift toward AI risk concern.
Why read this. First-hand context for one of the most consequential public statements in the field.
- 12
Anthropic — Interpretability research talks
TalkAnthropic · ongoingTalks and walkthroughs from Anthropic's interpretability and alignment teams.
Why read this. Best source for current interpretability progress.
- 13
AI Snake Oil — talks and book
TalkArvind Narayanan & Sayash Kapoor · 2024Princeton researchers on how to distinguish substantive AI progress from hype.
Why read this. Useful immunisation against breathless coverage.
- 14
Lex Fridman Podcast (AI episodes)
PodcastLex Fridman · ongoingLong interviews with AI researchers, founders, and adjacent thinkers.
Why read this. Wide-ranging; uneven; but archives most major figures.
- 15
Future of Life Institute Podcast
PodcastFuture of Life Institute · ongoingConversations on AI safety, policy, and existential risk.
Why read this. Steady policy and safety coverage.