Host, Dwarkesh Podcast; author, The Scaling Era
Dwarkesh Patel
Profile
If you want to understand what AI researchers actually think — not the PR version, not the tweet version, the long-form unfiltered version — Dwarkesh Patel is where you go. His Dwarkesh Podcast has become the closest thing the field has to an oral history, conducted in real time while the scaling era unfolds. He interviews the people building frontier models, asks them things they didn’t expect to be asked, and makes them think on camera. The Economist called him “Silicon Valley’s favourite podcaster.” That’s about right.
Patel grew up between North Dakota, West Virginia, Maryland, and Texas, studied computer science at UT Austin, and started interviewing writers and technologists in 2020 under the name The Lunar Society. He was a kid with a microphone asking smart questions — and then the AI boom hit, and it turned out he was better prepared than almost any established journalist. He now spends a week or more researching each interview, reading every paper the guest has written, treating it more like a thesis defense than a press junket. This is why Mark Zuckerberg, Demis Hassabis, Andrej Karpathy, Ilya Sutskever, Dario Amodei, and Elon Musk all sit down with him for hours.
What makes him useful for people learning AI is that he’s learning in public. He isn’t pretending to be the smartest person in the room — he’s genuinely trying to figure out how transformers, scaling laws, RL, and agents actually work, and he drags his guests into explaining the mechanics instead of reciting talking points. His conversations with engineers like Sholto Douglas and Trenton Bricken on how LLMs are trained are the kind of technical deep-dives you rarely hear outside a lab. For a developer trying to catch up on modern AI, binge-listening to his back catalog is arguably more efficient than most online courses.
In 2025 he co-authored The Scaling Era with Gavin Leech — an oral history stitched from his interviews with Jared Kaplan, Leopold Aschenbrenner, Dario Amodei, and others, plus explainers and definitions for the non-specialist. Time named him to its 100 most influential people in AI list in 2024. He has no lab, no product, no company to sell. His leverage is attention and preparation, and in an industry drowning in both hype and silence, he’s carved out a weirdly essential role: the guy who gets the real answer on tape.
Books
Key Articles & Papers
Andrej Karpathy — AGI is still a decade away Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race Mark Zuckerberg — Llama 3, $10B models, Caesar Augustus Dario Amodei — We are near the end of the exponential Jensen Huang — TPU competition, selling chips to China, Nvidia's supply chain moat Sholto Douglas & Trenton Bricken — How to build and understand GPT-7's mind Elon Musk — In 36 months, the cheapest place to put AI will be space Thoughts on AI progress (Dec 2025) Dwarkesh Patel on Continual LearningYouTube
Spotify Podcasts
