Next up
A coding-agent preflight, a small eval loop, and a comparison of where AI tools help in real app work.
Lab
Experiments I ran, not claims I made.
Small Swift projects, local model tests, prompts, scripts, and tool notes. The point is to leave enough detail that I can repeat the work later.
How notes should read
Question, setup, what happened, what broke, and whether I would use it again. No theatre.