What I'm doing now

I'm building AI-native software end to end: architecture, backend, apps, and the operations that keep them running after launch.

AI agents are part of my daily engineering process, next to the production code I still write myself. I write the brief, set the constraints, review what comes back and decide what ships. Agents draft; the design, the review and the accountability stay with me.

Two questions are holding my attention this month. How do you split one build across several agents without the pieces drifting apart? And what does a good review gate look like when the author is fast, confident and never tired?

I'm also writing more: engineering notes from production systems I have built since 2012, and from what agents change about how software gets built.