About

Somewhere between running a consultancy and building AI-native products, I discovered that the most interesting problems aren't in AI capability itself but in the intersection of human and AI cognition. How do you maintain strategic coherence across multiple AI agent sessions? How do you carry decisions, constraints, and open questions between conversations without losing the thread? What happens when you treat context not as something that resets with each session, but as a persistent, structured resource that grows richer over time?

This site is where I write about those questions. Not the marketing version, but the working notes. Architecture decisions that survived contact with reality, the ones that didn't, and the patterns that emerge when you're building systems that sit between human strategic thinking and AI execution.

I'm interested in reasoning interoperability, orchestration without control, and what happens when you give AI agents enough structure to be useful without so much that they become brittle. Most of what I write comes from building things and watching what breaks.