PERSON
Richard Bellman
The applied mathematician who gave sequential decision-making its grammar—inventor of dynamic programming and the Bellman equation, namer of the curse of dimensionality, and the unwitting architect of every AI agent that chooses a next action.
Richard Bellman is the most consequential AI theorist most people in AI have never read. Working at the RAND Corporation in the early 1950s, he distilled the logic of optimal sequential decision into a single recursive relation—now universally called the
Bellman equation—which says the value of where you stand equals the best you can do right now plus the value of wherever your best move lands you. That sentence, made precise, is the literal foundation of
reinforcement learning: every Q-table, every value function, every temporal-difference update is a descendant of one idea he wrote down to survive a hostile Secretary of Defense. He also named the obstacle that haunted his own method for decades: the
curse of dimensionality, the brute fact that state spaces grow exponentially with the number of variables, rendering his exact equation intractable on any problem that mattered. The astonishing thing about modern AI is that
deep learning found a way to slip the curse—not