This paper tells the story of adaptive learning. With origins in the rational expectations revolution, models of adaptive learning are as much about moving people’s expectations closer to rational expectations as further away. But relaxing the rational expectations requirement that subjective expectations are model consistent permits addressing a broader class of economic questions as well as resolving certain empirical puzzles. Importantly, models of optimal intertemporal decision making with adaptive learning are supported by a growing body of empirical evidence using surveys of professional, household, firm and policymaker forecasts.