Why distance, density, and data requirements behave strangely as features grow—hurting generalization and making models harder to learn.
Measure average pairwise distances as you add random features—watch “nearness” vanish.
