Extinction Rate R_ext · Ch 1How fast generic work goes extinct. Above 5 you're landfill the moment tooling catches up.
Worked example — A four-person internal tools team
Last quarter they touched about a hundred files at a regional insurer. Roughly 85 were CRUD screens, form validation and REST wrappers — the kind a model writes correctly in one shot. About 15 encoded the underwriting rules nobody outside that building understands.
It came out at 11.3. Red — the book's line is 5.
They stopped hand-writing the CRUD and moved two of the four engineers onto turning the underwriting rules into a tested service. Next quarter: 40 generic, 45 authentic, same AI velocity — 1.8, green. Same team, same headcount. Different denominator.
Your turn — change one number
Result 11.33
Red. Landfill trajectory. Everything generic is already gone.
Book's benchmark: 🟢 < 2 · 🟡 2–5 · 🔴 ≥ 5 (you are landfill the moment tooling catches up).
⌁ ask what this number means for me
Authenticity Half-Life A_hl · Ch 1How long your work resists being replicated by a machine. Higher is safer.
Worked example — A freelance analyst's monthly report
Every month she sends fourteen slides of churn charts. The charts take three days. The last slide — call these three accounts this week, here's what to say — takes twenty minutes, and is the only reason the client renews.
It came out at 2.6. Yellow — weeks, not years.
The fourteen slides are the problem: they're the replicable part, and they're most of the invoice. She cut the deck to that last slide plus an appendix and raised the price. Replicability drops to 3 — 6.0, green. Less work, more moat.
Your turn — change one number
Result 2.57
Yellow. Half-life in weeks. Add taste, context, or craft.
Book's benchmark: 🟢 ≥ 4 (months→years) · 🟡 1.5–4 (weeks→months) · 🔴 < 1.5 (days).
⌁ ask what this number means for me