PERSON
Abraham Wald
The refugee mathematician who spent his career attending to the data that was not there—inventing survivorship-bias analysis, sequential decision theory, and the minimax principle, and thereby laying the statistical foundations on which machine learning now stands without knowing his name.
Abraham Wald died in 1950, before any general-purpose computer had done anything memorable, yet the hardest problems of contemporary AI are, in the most precise sense available, Wald's problems in new clothes. His 1943 memorandum A Method of Estimating Plane Vulnerability Based on Damage of Survivors has been distilled into the legend of the bullet-holed bombers—look where the holes aren't—but the legend keeps the punchline and throws away the proof, which is itself a form of the error it describes. What Wald actually did was build an estimation method: given the distribution of damage on planes that came back, and a model of the selection process that determined which planes those were, he derived a way to recover the vulnerability of the planes that did not. The lesson is not counterintuitive wit. It is a structural discipline: your sample has been filtered by a process you did not control, and no amount of careful analysis of
Keep reading with YOU ON AI
Unlock the full book, 10,000+ field-guide entries, and a 1000+ thinker library. If you have a book code, register now — it takes a minute.