The Economics of AI

There is no separate “economics of AI,” Peter Klein argues, any more than there was an economics of the internet—there’s just economics, applied to specific goods and services. He demystifies what today’s models actually do (predict the next word), places them in a long history of “new economy” hype that never required new economic theory, and works through the industry with standard tools: subjective value, the theory of the firm, network effects, and the incumbents’ lobbying for regulation to raise rivals’ costs. He closes on the deeper question—whether an AI can act in Mises’s sense—and argues it cannot: an AI exercises what he calls derived judgment, executing tasks on an entrepreneur’s behalf, but it can never decide whether to be an entrepreneur at all.

Recorded at the Mises Institute in Auburn, Alabama, on July 24, 2026.

Mises University is the world’s leading instructional program in the Austrian School of economics, and is the essential training ground for economists who are looking beyond the mainstream.