AI

Building with models, agents, and the limits of automation.

2 pieces

AI

We Stopped Building Apps and Started Building an AI Organization

A shift in how we think about AI: not isolated applications, but a governed engineering organization where every role has a purpose and every decision is traceable.

· 2 min read
AI In progress

Letting Agents Write, Keeping Humans Responsible

I'm letting AI agents author real work — content, code, analysis — but committing nothing without a human gate. Here's the pattern so far, and what I'm still unsure about.

· 1 min read