Model Thinking: See the World Like a Many-Model Thinker

Why does segregation emerge even when no one wants it? Why does an innovation suddenly
tip at a critical threshold? Why does a crowd of biased ordinary people, averaged, often
beat the smartest expert? These unrelated-looking questions share one answer: thinking
with many small models. This is a flagship build-it-yourself course — no slides to skim,
no conclusions to memorize. Every model is one you assemble in pure numpy, tune by dragging
parameters, and actually run in your own sandbox: you'll run Schelling's grid until it
self-segregates and watch mild preferences amplify into macro segregation; nudge percolation
density and watch a giant connected cluster appear at the critical point; drive the Solow
model to its steady state and see why saving alone can't sustain growth; converge a Markov
chain to its stationary distribution and prove the "the start doesn't matter" convergence
theorem; run replicator dynamics and pull apart the diversity prediction theorem —
crowd error = average individual error − diversity. The course follows Scott Page's Model
Thinking module skeleton, but every session is original hands-on teaching rebuilt for the
online sandbox: typeset formulas, diagrammed models, draggable parameters, code that really
runs. You leave not with "24 memorized models" but with the feel of many-model thinking —
facing a new problem, you know which model to reach for.