01
Workflow
Break ambitious builds into human-sized checkpoints.
AI can move faster than our ability to review it. Small, testable milestones preserve context, control, and good judgment.
Field notes / working principles
What the work is teaching me.
Short observations from building with AI in public, at work, and at home. Weekly exploration posts will live here next.
02
Prototyping
The useful prototype is the one grounded in real work.
The best experiments begin with a workflow I already understand—then remove friction without hiding how the system behaves.
03
Practice
Documentation is part of the product.
Build logs, decisions, and failed approaches make an experiment repeatable. They also make the next version much smarter.