The notebook, volume № 01

What the lab is learning while it builds.

Positions we hold, methods we use, records we keep. A new note when the work earns one, not on a schedule, and any figure drawn from an argument rather than a measurement says so in its caption.

07
notes filed
05
registers
27
sources verified
Lead note§ 07 · 27 Jul 2026

Nobody owned the design

What two years of evidence says about the engineer's job in AI-native development, and about the part of it we still cannot hand over.

Read the note18 min · evidence · 27 sources
The reading that anchors it
737×

slower than gcc -O0 on SQLite, from a compiler that passes 99% of the GCC torture suite.

Evidence grade of the corpus4 peer-reviewed or official20 preprint or independent3 vendor-reported, flagged
7 notes
  1. § 0602 Jul 2026

    Hiring rarely breaks at the résumé screen. It breaks the moment the standard depends on who happens to be in the room.

    standardscalibration
    position6 min
  2. § 0518 Jun 2026

    A rubric drafted in an afternoon meets its first surprising answer within the hour. Notes on criteria that bend without breaking.

    rubricsauthoring
    method8 min
  3. § 0429 May 2026

    An AI can hold a standard steady all night. It still should not have the last word. Where the line sits, and why it stays there.

    judgmentlimits
    position5 min
  4. § 0307 May 2026

    Two candidates, one score, different work entirely. The number says they tied; the reasoning says which one you want.

    reasoningreview
    method7 min
  5. § 0210 Apr 2026

    Before Evals judges anyone, it sits its own exam: transcripts with known answers, counter-examples, and drift checks on every release.

    evaluationdrift
    engineering9 min
  6. § 0113 Mar 2026

    watchglass studies where human judgment belongs in a world that works with AI. Why the first instrument out of the lab is a hiring tool.

    serieslab
    series6 min