I'm a principal product manager. I build software with a team of AI agents, and I write about what I learn.
My day job is product management. Outside of it, I build small products for real problems, usually my own first. A meeting companion so nothing said in a meeting gets lost. A personal operating system for tasks, habits and the week ahead. A gym log I can use between sets. A security reviewer for small businesses that can't afford a security team.
I build them with AI agents doing most of the typing. I write down who the product is for, what the problem really is and how I'll know it worked; agents turn that into plans, code and tests; I make the taste calls. It's the same job as product management, just with a much shorter distance between deciding and shipping.
I'm exploring where that goes: what a product team looks like when most of the team is agents, which human jobs remain (I think taste, direction and saying no), and how to build AI products people can actually rely on.
How I work
- Start with the customer and the problem. Every project here begins with a person and a moment, not a feature.
- Pick a measure before building. "Log a set in one tap" is a target; "a better gym app" isn't.
- Ship small, use it daily, then decide what's next. Most of these went from idea to daily use in under three weeks.
Elsewhere
The best way to reach me is a reply or a DM on X.