Writing
Field Notes
Notes from building AI agents in production environments. Not thought leadership. Not a newsletter. Things we noticed while working.
What actually makes an AI agent good at its job
Most agent failures are not model failures. They are scope failures. Here is what we mean by that, and how we define boundaries before writing any code.
Agent vs automation: the practical difference in a production environment
When rule-based is enough, and when it stops being enough.
How we structure agent workflows before writing any code
The planning step that prevents most production failures.
What we mean when we say autonomous — and what we do not
Clearing up a word that has been stretched in every direction.
What the EU AI Act actually means for businesses deploying agents
The obligations that are already active and the ones coming in 2026.