Process notes, including the failures — conclusions are cheap, but the path matters.
| Date | Note | Tags |
|---|---|---|
| 2026-11-08 | Root-causing the kettle-impersonates-AC misclassification | modeling |
| 2026-10-26 | Rewriting the DNS sampler: threads to asyncio, and what broke | tooling |
| 2026-09-30 | PWM duty vs RPM: where the datasheet lies | hardware |
| 2026-06-14 | Why probe ordering bias nearly ruined the DNS numbers | methodology |
| 2026-03-02 | Feature sketch v2 for switch-on transients | modeling |
| 2025-11-19 | First month of plug-level sampling: gaps, fixes, lessons | data |