Descriptions:
Dwarkesh Patel interviews two leading economists working at the intersection of AI and economic theory: Alex Imas, Director of AGI Economics at Google DeepMind and Professor of Economics at the University of Chicago, and Phil Trammell, Head of Economics at Epoch AI and a research scholar at Stanford. The conversation systematically examines what economic frameworks and current data tell us about a world of increasingly advanced automation — covering labor markets, scarcity, wages, and redistribution policy.
The guests invoke David Ricardo’s Industrial Revolution predictions as a case study in the difficulty of structural economic forecasting: Ricardo correctly predicted that automation would eliminate specific jobs, but failed to anticipate the creation of entirely new employment categories. Imas and Trammell use this to argue for epistemic humility and the value of prediction markets over individual expert forecasts. They then work through specific mechanisms: the O-ring model of task bundles, how elasticity of demand determines whether automation expands or contracts total employment, and the concept of a “relational sector” — services where human involvement is intrinsically part of the product value.
On current data, Imas pushes back on the “white-collar bloodbath” narrative, noting that prime-age employment is near historic highs and that junior software engineering job trends show slower growth rather than decline. The conversation also covers optimal taxation of AI-generated wealth, Jevons paradox, and how aggregate prediction markets might produce better forecasts than any individual economist. A dense, academically grounded discussion for anyone tracking the long-term economic implications of AI deployment.
📺 Source: Dwarkesh Patel · Published June 04, 2026
🏷️ Format: Interview
