AI DevOps Podcast Paul Duvall: AI-Driven Development - Episode 390
24 snips
Feb 23, 2026 Paul Duvall, award-winning author and longtime DevOps leader who helped pioneer continuous integration. He talks about how AI reshapes engineering practices and the need for solid CI habits. He outlines spec-driven workflows, evolving code review with AI, and tools for generating implementations from specs. He also covers observability, cost expectations for AI tooling, and how teams and careers adapt.
AI Snips
Chapters
Books
Transcript
Episode notes
Two Year Medical Project Sparked CI Passion
- Paul Duvall's first big project was a global medical logistics system that took two years to reach production and involved about 100 people across six dev teams.
- Seeing users immediately suggest five to six improvements convinced him to deliver faster and drove his interest in CI and behind-the-scenes engineering practices.
Make CI Real With Trunk Based Development
- Do practice trunk-based development, commit early and often, run automated tests, and stop the line when the build breaks.
- Paul emphasizes local builds before push, short-lived branches, and making a clean build the top team priority.
AI Raises Stakes For Core Engineering Practices
- AI accelerates coding but makes engineering practices like frequent commits and automated tests even more critical.
- Paul built an open-source AI development patterns repo (about 22 patterns) to codify these evolved practices.
