Power, progress, and technology (with Daron Acemoglu)

Episode Summary

Executive Summary: The episode centers on Daron Acemoglu’s thesis that technology does not automatically create shared prosperity; its effects depend on who controls it and how power is distributed. Using historical and contemporary examples, he argues that automation, offshoring, and AI often enrich elites unless countervailing institutions—unions, regulation, worker voice, and democratic oversight—push innovation toward human-complementary uses.

Main Topics: Technology is not destiny (Priority: 5/5): Acemoglu argues that technological progress has only a loose, conditional relationship with broad welfare. Innovation can raise productivity without raising most workers’ living standards unless institutions force gains to be shared. Power and countervailing institutions (Priority: 5/5): The discussion emphasizes that economics often ignores power, but wage growth and prosperity depend on unions, regulation, democracy, and worker bargaining power. Historical examples of unequal tech gains (Priority: 5/5): The cotton gin and the Industrial Revolution show that major technologies can increase profits and output while worsening labor conditions, lowering wages, and prolonging exploitation for decades. Automation, offshoring, and labor decline (Priority: 4/5): The guests connect automation and offshoring as similar mechanisms that remove tasks from workers, depress wages, and reduce labor demand, especially for non-college workers. The postwar middle-class era and its reversal (Priority: 5/5): The 1945-1975 period is presented as an era of broadly shared prosperity produced by political and institutional choices, later reversed by shareholder-value ideology, deregulation, and weakened labor power. AI, data, and corporate concentration (Priority: 4/5): Generative AI is framed as a new battleground in which firms may use data, surveillance, and automation to intensify inequality unless policy redirects AI toward worker augmentation. Policy responses for pro-human innovation (Priority: 4/5): Proposals include changing tax incentives that favor machines over labor, regulating data use, strengthening worker voice, and potentially breaking up dominant tech firms.

Key Arguments: Technology’s benefits are not automatically shared; gains depend on institutional choices and power relations. Orthodox economics is incomplete because it assumes away power and treats distributional outcomes as largely deterministic. The cotton gin increased wealth for slaveholders while worsening conditions for enslaved laborers, showing that productivity gains can coexist with exploitation. Automation does not necessarily raise wages; if machines replace tasks rather than complement workers, labor demand and bargaining power fall. Offshoring and automation function similarly by removing tasks from domestic workers and shifting income toward capital. The postwar middle class grew because labor, regulation, and government acted as countervailing powers that forced sharing of productivity gains. Since the late 1970s, shareholder-value ideology, union decline, deregulation, and digital technologies combined to weaken labor and accelerate inequality. Countries like Germany and Scandinavia show that the same technologies can be used differently when unions, training systems, and cooperative institutions are stronger. AI is malleable and can be used for surveillance and automation or for worker productivity, autonomy, and new tasks. Policy should not slow innovation; it should redirect innovation toward human-complementary uses and broader prosperity.

Data Points: U.S. population doing very well: about 30% - Acemoglu says roughly this share of Americans is benefiting strongly in the current economy. High-income/specialized group: 10-15% - He estimates this share includes people with postgraduate degrees or scarce specialized skills who are doing especially well. Industrial Revolution wage impact: about 100 years - The guests note it took roughly a century for the Industrial Revolution’s gains to translate into broad worker benefits. Workweek increase during early Industrial Revolution: 20-25% more hours a week - Acemoglu describes deteriorating labor conditions in Britain, including longer hours. U.S. corporate stock buybacks: a trillion dollars a year - Hanauer cites buybacks as a sign of excess financialization and underinvestment in workers. Tax difference between labor and machines: about 25%+ vs about 5% - Acemoglu argues labor is taxed far more heavily than capital equipment, incentivizing automation. German manufacturing response: robots plus retraining into technical roles - He contrasts German firms’ practice of retraining displaced workers with U.S. job loss.

Pivotal Quotes: "There is only a loose correlation between technology and the welfare of the majority of citizens" — Nick Hanauer / transcript framing of the book's thesis: Introduces the episode’s central premise that innovation does not automatically produce shared prosperity. "This time is no different. This time is just like many other technological transitions" — Daron Acemoglu: He rejects the idea that the current tech wave is uniquely destined to benefit everyone. "It did not have to turn out this way. It was not deterministic." — Daron Acemoglu: He stresses that rising inequality and worker displacement were the result of choices, not inevitability.

Implications: The episode argues listeners should see AI and other technologies as political-economic choices, not neutral forces. Future prosperity depends on labor power, democratic oversight, and policies that make innovation complement workers rather than replace them.