A thought-provoking piece from the Hoover Institution’s Andy Hall (who also teaches at the Stanford Graduate School of Business) is shared here: this is the Introduction, though the entire work is highly worth reading:
No period in modern American economics has matched the shock that AI industry leaders say is on the horizon. Dario Amodei warned of an unusually painful impact on workers, suggesting it could be “bigger than any before,” and predicting that AI might erase half of all entry-level white-collar jobs while driving unemployment to the 10–20 percent range within five years. He isn’t alone in this view. Both OpenAI and Anthropic have begun presenting, in expansive policy memos, the kind of social compact they foresee the post-AGI economy requiring—shorter workweeks, public wealth funds, and a thoroughly modernized taxation system. They insist that abundance is coming and they want to help figure out how to share it.
Can the tech sector successfully pre-empt American populism by drafting the post-AGI social contract before the public has even asked for one, and before we know whether speculative growth and job displacement are truly ahead? My answer, after months of collaborating with my coding agents to sift through polling data, policy proposals, and historical parallels, is that it cannot.
In the scenario the labs are sketching, the politics of AGI will be the politics of jobless prosperity. And that makes precise forecasting difficult. The economy could grow rapidly even as jobs vanish, more like the Industrial Revolution or the China Shock than a typical recession, with mass disruption occurring alongside the rapid enrichment of a small elite at the very top.
Voters in this world will not be anxious about a shrinking economy but furious at being excluded from a booming one, and they may very well prevent the boom from arriving at all. Jasmine Sun has shown how this anxiety is already hardening into nascent political anger, noting that “the anti-elite and nihilistic attitudes that have dominated US political culture in the last few years are transmuting into anger at AI billionaires.” Alex Imas, in “What will be scarce?”, presents the most careful economic argument for taking the underlying disruption seriously, even as he explains why both the short-run doomers and the long-run doomers may be mistaken about mass unemployment.
The labs foresee all of this, which explains why their policy memos have grown so ambitious. It would be easy to read this as good news, since the groups that would have to fund redistribution are already volunteering to do so in advance.
But this approach cannot work. First, social contracts tend to be extracted from the powerful by those they affect, rather than dictated from above to a public that has yet to decide what it wants. And second, we still do not know what the economic contours of AGI will look like—we aren’t even sure that it will produce jobs losses, let alone mass unemployment.
As we oscillate between prophecies of catastrophe and visions of plenty, I have arrived at three conclusions:
- The pushback against AI isn’t here yet. There is unease among American voters, but no populist backlash yet, because the structural conditions for one have not materialized. This gives a potentially narrow window to plan our response to job loss before it becomes a populist issue.
- A genuine backlash will arrive if and when job losses gain momentum. The backlash will properly materialize if unemployment climbs by a couple of percentage points—as I hypothesize—together with a clear public narrative blaming AI. If we don’t have a robust inventory of smart policy ideas by that moment, we risk being overwhelmed by poorly crafted populist alternatives.
- The labs should prioritize measurement over redistribution. Their best contribution during the window before any backlash is the infrastructure that makes this transition visible—usage data, displacement indicators, self-activating triggers—not premature social contracts lacking credibility and a coalition to enforce them. The eventual bargain should be negotiated by those directly affected, and the data and tools that enable them to bargain from a position of clear information are what the labs can build now.
