W. Edwards Deming was a physicist, statistician, and quality theorist who taught post-war Japanese manufacturers what eventually became the Toyota Production System - and, decades later, DevOps. John Willis, one of the founders of the DevOps movement and the author of a book on Deming, walks Whitney and Coté through that lineage: Deming's system of profound knowledge (theory of knowledge, variation, psychology, and systems thinking), how it landed at Toyota, and how it threads through Lean software development into modern delivery practice. From there, the conversation turns to what the industry is getting wrong about AI: bragging about K-LOC and token counts instead of value, treating probabilistic systems with old deterministic notions of risk, and forgetting the social-technical lessons we already paid for. Along the way: VC moats and the buy-versus-build conversation inside large organizations, David Foster Wallace's "This is Water" and the ladder of inference, Jevons Paradox and whether AI gets us a three-day work week or a six-day one, and what CS students should be learning besides how to code. Also a brief detour into why John would want fifteen minutes with Bill Clinton.

You can watch the video version of this episode as well, if you prefer that kind of thing.

Mentions: