A lab with a thesis about judgment could have started anywhere: code review, grading, editorial. We chose hiring because it is where the cost of inconsistent judgment is highest and most invisible. A bad merge gets reverted. A bad read on a person compounds for years, and nobody ever sees the control group.
Hiring has the honest shape for a first instrument
A standard that can be written down, a conversation that can be conducted the same way every time, and a decision that must stay human. If our thesis, that instruments hold the standard while people keep the last word, fails here, it fails everywhere, and we would rather find out early.
The discipline of the series matters as much as the instrument
We publish when a thing is dependable, not when it is demoable, and one at a time, because every instrument added before the last one is trusted subtracts from all of them.
Instrument № 02 exists in the workshop and will stay there until it stops surprising us. These notes are the record in the meantime.
watchglass (2026). Notes on instrument № 01. The notebook, volume № 01, § 01. watchglass.app.