Noah Smith
· Noahpinion
· July 17, 2026 at 03:52
· ⏱ 44 min read
| Read on Substack ↗
Summary
Noah Smith's review of 'Power and Progress' argues that the book's historical examples are flawed, its definition of 'power' is tautological, and its central policy recommendation—steering innovation toward worker-complementing technologies—is unsupported and impractical. For markets, the article reinforces skepticism about narratives that blame automation for wage stagnation, but offers no actionable trade ideas.
•The book claims the Haber-Bosch process 'brought nothing like shared prosperity' despite sustaining half of Earth's population.
•Acemoglu and Johnson argue early Industrial Revolution textile automation immiserated workers, but Smith counters that similar technologies boosted wages in China and Bangladesh.
•The book's definition of power equates persuasion with compulsion, which Smith calls empirically useless.
•Smith notes that the authors cite no historical example where unions or governments successfully redirected innovation toward worker-complementing technologies.
•The book lacks footnotes, making it hard to verify claims like 'AI threatens to impoverish billions' which contradicts Acemoglu's own 2022 paper finding no detectable labor market effects.
•Several studies (Mann & Püttmann, Dixon et al., Koch et al.) find robot adoption correlates with higher employment, contradicting Acemoglu and Restrepo's 2020 findings.
•Real wages for production workers have risen strongly since the early 2010s, while inequality has flatlined—contra the book's pessimistic automation narrative.
•Smith argues that simple ex-post policies like wage subsidies are superior to trying to predict and redirect innovation ex-ante.