Latent Space: The AI Engineer Podcast
Also known as Latent Space podcast
Interview podcast for AI engineers, hosted by swyx and Alessio Fanelli since 2023.
At a glance
Episodes to start with
Commoditizing the Petaflop — with George Hotz of the tiny corp, Code Interpreter == GPT 4.5 (w/ Simon Willison, Alex Volkov, Aravind Srinivas, Alex Graveley, et al.), The End of Finetuning — with Jeremy Howard of Fast.ai …
See all 9Connections
swyx, Alessio Fanelli and 4 others
See all 6
Also on this page: Key facts · Links · Sources · Related pages
- Latent Space: The AI Engineer Podcast is the podcast of Latent Space.
- Episodes are long technical interviews with founders, researchers and engineers at AI labs and tooling companies.
- The feed also carries news recaps and conference recordings.
- Episode titles in the AI for Science series carry a microscope emoji.
- Full show notes are published on latent.space.
Why it matters
- For engineers who want first-hand accounts of how models, agents and developer tools are built.
- Guests have included Jeff Dean, Bret Taylor, George Hotz and the creators of the Model Context Protocol.
Key facts
- Started
- 20234
- Co-founded by
- swyx and Alessio Fanelli7
- AI for Science series announced
- 28 Jan 20268
- Episodes in RSS feed
- 229 (30 Sep 2026)1
- Chart peak
- #10 in US Tech podcasts (as of February 2024)5
- Publisher on Apple Podcasts
- Latent.Space2
Episodes to start with
The line under a date fills as it gets older: empty when new, full at 2 years.
-
Commoditizing the Petaflop — with George Hotz of the tiny corp (opens latent.space)
George Hotz on how tiny corp is taking on Nvidia, Google and PyTorch, and on building with AMD.
-
A live group discussion of ChatGPT's Code Interpreter launch with Simon Willison and others.
-
The End of Finetuning — with Jeremy Howard of Fast.ai (opens latent.space)
Jeremy Howard on fast.ai's mission, inventing ULMFiT and why he now thinks it is wrong.
-
How NotebookLM Was Made (opens latent.space)
Raiza Martin and Usama Bin Shafqat of Google Labs on building NotebookLM and its Deep Dive audio feature.
-
The AI Architect — Bret Taylor (opens latent.space)
Sierra CEO and OpenAI chairman Bret Taylor on the future of software engineering and building products.
-
The Creators of Model Context Protocol (opens latent.space)
Anthropic's David Soria Parra and Justin Spahr-Summers on the origin and design of MCP.
-
Owning the AI Pareto Frontier — Jeff Dean (opens latent.space)
Jeff Dean on Google's search stack, sparse trillion-parameter models and co-designing TPUs with ML research.
-
Codex from 0 to 10M Users: Building ChatGPT Work — Akshay Nathan, OpenAI (opens latent.space)
OpenAI's core product engineering lead on building ChatGPT Work.
-
Claude Code’s Next Era — Thariq Shihipar, Anthropic (opens latent.space)
Anthropic's Thariq Shihipar on shipping Claude Code alongside the Opus and Sonnet 5.5 models.
Connections
Links
Sources
- Latent Space: The AI Engineer Podcast (RSS feed) (opens api.substack.com)
- Latent Space: The AI Engineer Podcast on Apple Podcasts (opens podcasts.apple.com)
- Latent Space: The AI Engineer Podcast on Spotify (opens open.spotify.com)
- About - Latent.Space (opens latent.space)
- One Year of Latent Space — Alessio Fanelli (opens alessiofanelli.com)
- Latent Space archive (opens latent.space)
- Alessio Fanelli - AI Engineer speaker page (opens ai.engineer)
- It's Time to Science - Latent.Space (opens latent.space)
Last checked against its sources 7 weeks ago
Is something on this page wrong or out of date? Tell us
The line under a date fills as time passes: empty when new, then one equal share each for a day, a week, a month and a year. For episodes and other items it is full at 2 years; for "last checked", at a year. A striped line means the date is known only to the month or year.
