PodcastIntel
Sign in Get Started Free
Latent Space: The AI Engineer Podcast
Latent Space: The AI Engineer Podcast

The First Mechanistic Interpretability Frontier Lab — Myra Deng & Mark Bissell of Goodfire AI

Feb 6, 2026 · 1h 8m
AI Summary
  • Goodfire converts mechanistic interpretability from 'peeking inside models' into production workflow via APIs and enterprise deployments, raising $150M Series B at $1.25B valuation
  • Core bet: AI lifecycle broken because only reliable control is data; post-training through RLHF/fine-tune hopes models absorb right behaviors while absorbing wrong ones silently
  • Solution is bi-directional human-model interface to read internals, edit surgically, and eventually use interpretability during training so customization isn't trapped in post-hoc supervision

More from Latent Space: The AI Engineer Podcast

View all episodes →

Get AI Summaries for Every New Episode

Subscribe to Latent Space: The AI Engineer Podcast and get AI summaries, guest tracking, and email digests delivered automatically.

Sign Up Free →