In September 2025 I spoke at the Developer Productivity Engineering Summit with my Netflix colleague Eric Wendelin, who leads our Application Development Platform. Our session was “Programming” GenAI into Developer Productivity.
We covered the decisions that actually determine whether a GenAI developer-productivity effort goes anywhere:
- Build versus buy. Where it is worth building something specific to your stack, and where you are just rebuilding a worse version of a product you could have bought.
- What we learned piloting AI coding agents. The real results, including the parts that did not go the way we expected.
- Fostering a culture of AI innovation through a champions community. How a dedicated community of practice does the adoption work that no amount of top-down rollout will do for you.
The part I care most about
The champions community is the thread that connects this talk to almost everything else I work on. You can buy every tool on the market and still see nothing change, because tools do not teach anyone anything.
What moves adoption is a group of people inside the org who are genuinely good at this, visible to their peers, and given the time to help. That is a program-design problem, not a procurement one, and it is the same argument I ended up writing a whole book about.
Session: “Programming” GenAI into Developer Productivity
With Eric Wendelin, Application Development Platform Lead at Netflix
Developer Productivity Engineering Summit, September 2025
