Writing
How to think about AI agents.
Short pieces on frameworks, patterns, and the judgment calls that decide whether an AI agent actually earns its keep. The full implementation lives in the guide; these are the ideas that frame it.
-
Picking the model tier for a specific job
Most agent work runs cheap. A small named set of jobs justify the flagship. The rule for telling them apart.
-
Rungs of autonomy: how to decide what an agent can do on its own
A three-rung framework for deciding which actions your AI agent runs itself, which it drafts for approval, and which it asks before touching.
New posts land here. There’s no newsletter yet — bookmark the page, or follow updates on the guide’s launch list.