Resources reads the build in front of you. This page reads the career you are running while you build, across nine dimensions, on my own work first. Every post is one of the nine questions held against one piece of evidence, mine or someone else's. Research names what Data measures.
Which of the nine questions the evidence actually speaks to.
How well: supported, inferred, or still open — with what it cannot say stated beside what it can.
Whether the thing is being measured — on Data, pending, or not yet readable.
Each family holds the post I would hand you first.
My share of the words in my own transcripts fell from about 55 percent to a 7 percent week in June, the same summer I was spread across twice the repositories. Delegation or decline: the line cannot say.
From my archive · September 2026Payroll data through June 2026: workers aged twenty-two to twenty-five in AI-exposed occupations sit 19 percent below trend, through hiring, not firing. Experienced workers show no gap yet.
Worth reading · August 2026One model a year ago; five to seven a month now. The provider that carried everything carries a fifth. No fallback has ever been run, so the concentration number has no tested exit behind it.
From my archive · September 2026Nothing published on this dimension yet.
From my archive posts read the lines on Data: a year of my own agent transcripts, published there with a version-zero definition beside every number. Worth reading posts take a recent piece by a researcher or group of standing, state the finding the way its authors did, with their caveat, and read it against the nine dimensions.
Both kinds carry the same three rows, drawn from the inquiry's own evidence standard: what is under threat, what is observable, what is measured now. Both end with what I am doing about it, and both keep a section for what stays: a bad reading is a reason to change one thing, not a reason to stop building this way. Posts publish when they are ready, not on a schedule.
Definitions are dated and published before a reading goes public, so a metric can't be tuned to fit a result afterward. Every change to a definition is versioned and visible. The same commitment appears in full on Data.
The kit ships with the same instrument I am running on myself. Mentorship installs it with me in the room.
New Resources, Research, Data readings, and kit editions. This list is the only announcement I send.