Prediction Machines
Ajay Agrawal
"Prediction Machines" is a practical framework for understanding artificial intelligence through the lens of basic economics. Written by University of Toronto economists Ajay Agrawal, Joshua Gans, and Avi Goldfarb, the book strips away the sci-fi hype and treats AI as a commodity. It argues that the current revolution is actually just a massive drop in the cost of a single input: prediction.
The authors draw a parallel to the invention of electric light. When light became cheap, people did not just use it to see at night; they fundamentally redesigned their cities and workdays. Similarly, as AI drives the cost of prediction down, we will start using it for tasks we never considered prediction problems before. Autonomous driving is framed not as robot intelligence, but as a prediction problem where the car is simply predicting what a good human driver would do next.
The most critical insight is the shift in value. As machines master prediction, the economic value of the complementary skill rises. That skill is Judgment. While machines can predict the outcome of a decision, only humans can determine the payoff of that outcome. The book asserts that the future of work lies in determining what we should want, while leaving the calculation of how to get there to the machines.
