Scoping Assessment
Scope is a wall you build together — exclusions by name, acceptance criteria, change orders before work.
Class recording
The lecture recording posts here after the class is taught.
Class summary
Scope creep is measured, not folklore: over half of projects report it, with average budget overruns around a quarter. AI makes it worse, because "can’t the AI just also do X" sounds free and is not. The counter is a wall with four bricks — exclusions named explicitly, deliverables listed, acceptance criteria written down, and a change-order process that runs before any extra work starts.
The economics of narrow scope are shown with a receipt from Palantir’s S-1: the acquire-phase cohort at $0.6M revenue against a $65.4M contribution loss, and phase-gating as the answer. The class ends in a role-play — "just have the AI also do X" — with the wrong answer and the scripted right one, practised out loud.