The robots are coming… what now? (with Heidi Shierholz and Daron Acemoglu)

Episode Summary

Executive Summary: The episode argues that automation is not a new threat and will not eliminate all jobs, but it can displace workers and worsen inequality when power is concentrated with employers and capital. Heidi Shierholz and Daron Acemoglu stress that outcomes depend on institutions, labor standards, union power, and whether technology is used to augment labor or replace it. The hosts conclude that the real issue is worker displacement, not automation itself.

Main Topics: Automation panic in historical context (Priority: 5/5): The hosts open by showing that fears of robots taking jobs recur across decades, from agriculture to modern AI, and are often recycled headlines rather than new phenomena. Automation does not eliminate all jobs (Priority: 5/5): Heidi Shierholz argues that technological change has historically shifted employment across sectors rather than producing sustained mass unemployment, citing agriculture as the clearest example. Worker displacement and job quality (Priority: 5/5): Both guests emphasize that automation can still hurt workers who lose good jobs, especially when replacement jobs pay less, lack benefits, or offer worse conditions. Power, unions, and labor standards (Priority: 5/5): A central theme is that the real driver of bad outcomes is declining worker power, weakened unions, low labor standards, and employer dominance, not technology alone. Technology as a platform: automation vs augmentation (Priority: 4/5): Acemoglu explains that technologies can be directed toward replacing labor or enhancing labor productivity, and that social norms and institutions shape which path dominates. Policy responses to disruption (Priority: 4/5): The discussion centers on solutions such as stronger unions, better safety nets, labor standards, low unemployment policy, and training—framed as the 'old school' fix of taking care of workers. Long-run social and political stakes (Priority: 4/5): The hosts argue that neoliberalism and shareholder primacy use technology to concentrate gains and socialize losses, making automation a political choice rather than an inevitability.

Key Arguments: Automation has repeatedly displaced tasks, but historically it has also created new jobs and tasks, so the total number of jobs has not suffered sustained decline. The key issue is not automation itself but displacement, especially when workers move from high-wage, benefit-rich jobs into low-wage precarious work. Good labor-market institutions—unions, labor standards, strong macro policy—help workers adapt and make automation benefits broadly shared. Technology can be designed and deployed either to automate labor away or to reinstate and augment labor; the direction depends on power and policy. A society with stronger worker bargaining power can absorb automation more fairly, because wage gains and productivity gains translate into more demand and more jobs. Corporations and tech ecosystems often favor automation because it maximizes control, cost cutting, and shareholder returns rather than broad social benefit. The decline in worker power since the 1970s helps explain wage stagnation, inequality, and weaker productivity gains from technological change. If displaced workers have access to good replacement jobs, the harm from automation is greatly reduced; if not, the transition becomes socially damaging.

Data Points: US agricultural employment share in early 1800s: over 80% - Heidi Shierholz uses agriculture as the classic example of automation-driven labor reallocation. US agricultural employment share today: less than 2% - Used to show that a massive sectoral decline did not eliminate jobs overall. US population in 1810: 7 million - Nick Hanauer contrasts early US labor structure with the modern economy. US population today: about 330 million - Used to underscore how employment scaled even as agriculture collapsed as a share of jobs. Commercial tractor trailers in the US: around 2 million - Acemoglu’s example of why autonomous vehicle adoption will likely be gradual. Value of non-autonomous tractor trailers: $300 billion - Calculated from 2 million trucks at about $150,000 each, illustrating slow replacement dynamics. Labor share trend: sharp drop starting around the 1980s - Acemoglu describes a break in the long-run pattern of labor share stability. Time period of wage/productivity divergence: last four and a half decades - Shierholz notes erosion of labor standards and stagnant wages over this period. Median family income loss from rising inequality: $59,000 vs. $105,000-$110,000 - Hanauer argues the median family would be far better off absent inequality-driven power shifts. Era of stronger US worker prosperity relative to Europe: 1965 - Acemoglu says the US median family was then more prosperous than any other median family globally.

Pivotal Quotes: "The answer to that, in my view, is a pretty clear no." — Heidi Shierholz: Her direct response to whether robots will take all the jobs. "We cannot end up in a circumstance where we fear automation." — Nick Hanauer: The hosts frame automation as beneficial if workers are protected. "The problem there was not the technological change. It is the erosion of worker power." — Nick Hanauer: A key exchange on why displaced workers end up worse off when unions and standards weaken.

Implications: The episode argues for a pro-worker response to automation: strengthen unions, standards, safety nets, and bargaining power so technology raises living standards instead of enriching owners while displacing workers.