Writing
Every build ships with a write-up of what broke and design notes on why it works the way it does. They live next to the code on GitHub.
- Write-upSept 2026 A research agent that cites every sentence ↗ What broke, starting with zero evidence on the first run, and why the fix I planned was aimed at the wrong cause.
- Design notesSept 2026 How the research agent works, and why ↗ The six-step pipeline, four layers against made-up answers, and a cost ledger that found where the money went.
- Design notesSept 2026 How the coach works, and how a fake user tests it ↗ The agent loop and its guards, what the model sees, the ten-day simulator, and all 16 runs with their scores.
- LogOngoing The shipping log ↗ One entry per ship: what it does and the honest version of what broke.