Lesson 0004 · Exercise
No new material here — just the whole Lecture 3 arc, played out as one decision at a time. Five steps, five chances to reason wrong before you reason right.
You have data on past patients: symptoms, plus whether they turned out to have the disease. A new patient walks in. Build the system that decides whether to flag them for follow-up testing — from choosing what a valid answer looks like, all the way to checking the deployed rule is fair.
Each step below gives you a few options. They're deliberately unlabeled — some are tempting mistakes pulled straight from the last three lessons, not random wrong answers. Pick one; if it's wrong you'll see exactly which reasoning error it was and get sent back to try again. Only one full sequence reaches the end.