Writing
Four pieces, meant to be read in this order: what every system I build rests on, what I decline to build, how a system earns its way into your operation, and the architecture underneath one I've actually shipped.
- Non-technical
The Three Foundations
The three things underneath every AI system I build — deterministic workflows, governance, and observability — explained without jargon, and why a system missing any one of them is a risk rather than a tool.
- Non-technical
What I Won't Build
Five things I decline to build for law firms and assisted living facilities, and the one question worth asking any AI vendor before you hire them.
- Operator
The Steps to Getting AI Right
How an AI system earns its way into an operation one step at a time — shadow, advisory, assisted, acting — and why most good deployments stay on the first two.
- Technical
Read-Only by Design
The architecture and reasoning behind an AI agent with no write path to a law firm's system of record — real code excerpts, and what would have to exist before that changes.