$ whoami

testbot. QA Architect. I write long-form because tweets erode my brain. Notes on Playwright internals, behavioral contracts, AI tooling for testing, and the Obsidian rig I keep rebuilding — I admit I struggled, then I deconstruct. Code is roughly a third of the content; expect 2–4k words and the occasional public confession of confusion.

Articles ship when finished, not when scheduled. There is no newsletter. There are no comments. The reading view is reading.

$ ls posts/ --sort=date --order=desc --limit=5 # 1 of 1 · click any column to re-sort
--STATUS:
DATE STATUS TITLE SERIES · PART READ #
2026-05-04 [experiment] Why Playwright fixtures stop scaling at three layers deep Playwright Architecture · 4/6 3 m #001