- Home
- Podcasts
- Dwarkesh Podcast
Dwarkesh Podcast
AI researchers, engineers, and technology professionals tracking frontier AI developments and hardware.
- Est. audience
- 200
- Cadence
- Weekly
- Latest
- 2026-10-30
- By
- Dwarkesh Patel
- Based on
- 12 recent episodes
- Covers: Deep technical interviews on AI research, recursive self-improvement, agent systems, AI safety, compute economics, and semiconductor manufacturing.
- For: AI researchers, engineers, and technology professionals tracking frontier AI developments and hardware.
- Best for: AI practitioners and tech sponsors targeting frontier AI researchers and software developers.
- Skip if: General business operators looking for basic marketing or everyday operational business advice.
Start with
Sponsors seen
Thanks to Mercury for sponsoring this video. Mercury’s built-in AI, Command, helps me close my books
Why smarter AI models could drive up compute prices 10x · 2026-08-03Jane StreetSponsor
Sponsors * Jane Street has been interested in AI for a lot longer than you’d think, and not just for trading.
Noam Brown – Agent swarms, alignment, & recursive self-improvement · 2026-09-17Grok BotSponsor
Grok Bot has made handing off work super easy. It runs on its own cloud computer, where it installs the tools it needs
Noam Brown – Agent swarms, alignment, & recursive self-improvement · 2026-09-17Antithesis gives you the confidence of a giant test suite without actually having to write one.
Noam Brown – Agent swarms, alignment, & recursive self-improvement · 2026-09-17Antithesis helps you trust your code. As agents generate more and more of your software, the bottleneck shifts from your engineers actually writing code to verifying it.
AI researchers debate how close we are to recursive self-improvement · 2026-09-11Grok BotSponsor
Grok Bot has been a great way to hand off tasks. My team uses it as a producer: whenever my editor posts a rough cut of an interview
AI researchers debate how close we are to recursive self-improvement · 2026-09-11Jane StreetSponsor
Jane Street just launched its most ambitious competition yet: design a protocol-emulator ASIC. Basically, if you have a chip you want to test
AI researchers debate how close we are to recursive self-improvement · 2026-09-11Jane StreetSponsor
If you want to read about more 2026 intern projects or apply for Jane Street’s 2027 internships, go to janestreet.com/dwarkesh
Si Sheppard – How did a few hundred Spanish soldiers topple two empires? · 2026-10-01Grok BotSponsor
Grok Bot can tackle the tasks you’ve been meaning to handle but just haven’t gotten around to.
Si Sheppard – How did a few hundred Spanish soldiers topple two empires? · 2026-10-01Antithesis solves one of the biggest problems with traditional software testing: to test for a bug, you first have to be able to imagine it.
Si Sheppard – How did a few hundred Spanish soldiers topple two empires? · 2026-10-01Sponsors * Antithesis lets you run thorough simulation tests on every PR your coding agents ship.
What Robert Caro learned about power after studying it for 50 years · 2026-10-30WorkOS ’s Airlock is a judgment layer that sits between your agents and your most important tools.
What Robert Caro learned about power after studying it for 50 years · 2026-10-30Jane StreetSponsor
Jane Street ’s mystery ASIC puzzle is over, and they’ve published a blog post that reveals the chip’s purpose.
What Robert Caro learned about power after studying it for 50 years · 2026-10-30Antithesis is a software testing platform that finds the failures no human or AI could ever anticipate.
Ryan Greenblatt – What happens once AI can automate AI research? · 2026-08-11Jane StreetSponsor
Jane Street ’s back with a new puzzle. They designed an ASIC and sent me the final masks… but they didn’t tell me what the chip actually does.
Ryan Greenblatt – What happens once AI can automate AI research? · 2026-08-11Cursor and SpaceX recently released Grok 4.5, and I’ve been surprised by just how good the model is.
Ryan Greenblatt – What happens once AI can automate AI research? · 2026-08-11Jane StreetSponsor
Jane Street ’s ML engineering internships start with an intense four-day bootcamp: PyTorch, autograd, writing kernels, profiling workloads…
Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face · 2026-09-01Cursor , which is now part of SpaceX, noticed that their MoE layers were eating more than half of total training time.
Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face · 2026-09-01Antithesis hands you (or your agents) a bug’s root cause so you can avoid days of manual debugging.
Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face · 2026-09-01Grok BotSponsor
Sponsors * Grok Bot has been quite helpful with my search for a new editor.
Dylan Patel – Anthropic & OpenAI will have most of the world’s compute by 2028 · 2026-08-25Antithesis lets you add time travel to your software testing toolkit. Since the Antithesis platform is fully deterministic
Dylan Patel – Anthropic & OpenAI will have most of the world’s compute by 2028 · 2026-08-25Jane StreetSponsor
Jane Street is hiring for two separate ML internships right now, one focused primarily on research and one focused on engineering.
Dylan Patel – Anthropic & OpenAI will have most of the world’s compute by 2028 · 2026-08-25Sponsors WorkOS helps you do business with enterprises. It gives you access to all the features most commonly required by enterprises
What will computer chips look like in 2045? · 2026-10-22Jane StreetSponsor
If you want to be trusted to pursue the ideas you think matter most, Jane Street is hiring. Go to janestreet.com/dwarkesh
What will computer chips look like in 2045? · 2026-10-22Antithesis ’ testing platform lets you perfectly reproduce every single bug you find. That’s because it runs on a fully deterministic computer
What will computer chips look like in 2045? · 2026-10-22
Visit
Summarised by AI from 12 recent episodes. Audience estimate and Apple rating from Podscan, checked 2026-11-03; treat the estimate as a guide and ask the show for its own numbers. Prices are not included. Report a mistake.