← All summaries

David Shor and Byrne Hobart on the Politics of a White-Collar Wipeout

Odd Lots · Tracy Alloway, Joe Weisenthal — David Shor, Byrne Hobart · March 24, 2026 · Original

Most important take away

AI-driven white-collar job displacement is coming faster than most people expect, and the political consequences may be more significant than the economic ones. Roughly 70% of Americans already believe large-scale AI job loss is likely within five years, and polling shows the public is far more radical in its desired policy response (income guarantees, job guarantees, eviction protections) than politicians or commentators realize. The window for proactive political action is closing because once displacement is underway, the political system will be too slow to respond.

Chapter Summaries

The Speed and Scale of AI Progress

David Shor describes how AI capabilities are improving at an exponential rate, with autonomous operating time doubling roughly every 112 days. He notes a major disconnect between people who actively use tools like Claude Code and those who do not. Anthropic’s revenue last year was approximately 2X what even AI-optimistic experts predicted. AI adoption is outpacing every previous general-purpose technology rollout.

What Happens to White-Collar Jobs

Byrne Hobart argues that AI will follow the pattern of previous general-purpose technologies like electricity: mean compensation in affected fields may rise, but median compensation for current workers likely falls as many get washed out. The best-positioned workers will be those in regulated professions with constrained supply and excess demand (e.g., healthcare), where AI amplifies their output rather than replacing them. The economy may become more “guildified,” with human liability and accountability becoming an economically valuable function.

The Political Landscape Around AI

Shor’s polling reveals a clear demographic split: young people, men, educated people, and Black and Latino voters are more optimistic about AI, while older, working-class, and non-white populations are more pessimistic. AI has risen faster than any of the other 39 issues his firm tracks. Two-thirds of the public think the economy is rigged, only 35% feel financially secure, and voters are deeply skeptical of the claim that AI will create new jobs.

Republicans vs. Democrats on AI

Republicans have painted themselves into a corner with Trump on record saying AI will create tons of jobs and Vance pledging never to regulate AI. Democrats are in a better position to capitalize on public anxiety. The most effective political messaging combines AI concern with populism — “AI populism” tests better than either issue alone.

The COVID Analogy for AI Disruption

Both guests compare AI’s potential disruption to COVID: it will hit fast, scramble political coalitions, and overwhelm the political system’s ability to respond. Even if only 3% of people lose their jobs due to AI, it will become the biggest political issue in the world because of the pattern of diffuse benefits and concentrated losers.

Policy Responses the Public Actually Wants

The public is far more radical than the political class expects. Price controls now poll at 2-to-1 support. A policy package guaranteeing income up to $150,000, job guarantees, and eviction protections tested better than 98% of clips produced by Democratic professionals, polling at +30 overall and +15 among Trump voters. Shor argues that without a new social contract providing economic security, the alternative is a chaotic sector-by-sector regulatory battle that benefits no one.

Deepfakes, Media, and AI in Democracy

Hobart makes a contrarian case that deepfakes may be net positive because they erode the power of selectively edited real footage to manipulate opinion. AI-powered targeting could also allow coordination of interest groups around quality-of-life issues (like spam texts or click-to-cancel) that currently get ignored because only 5% of the population drives the majority of political content.

Summary

Actionable Insights and Investment Considerations:

  1. AI adoption is a sector-wide phenomenon, not a stock-picking category. Hobart compares AI to electricity circa 1925 — early on, investors fixate on pure-play “AI stocks,” but eventually every company becomes an AI company. This suggests that over time, the investment edge shifts from picking AI leaders to identifying which traditional businesses best leverage AI for productivity gains, similar to how electrification eventually favored companies that redesigned their operations around the new technology.

  2. Regulated professions with constrained supply are the biggest winners. Healthcare is called out specifically: doctors are rapid adopters of AI, supply is artificially constrained, and demand is effectively unlimited. Workers and businesses in regulated fields where a human must be “in the loop” for liability or compliance reasons will capture disproportionate value. Investing in healthcare services, professional services, and compliance-adjacent businesses could benefit from this dynamic.

  3. Data center and energy infrastructure remain a long-term play but are fully priced in near-term. Hobart notes all capacity in the data center supply chain is booked out years into the future. New supply is inelastic. AI companies may start guaranteeing delivery from energy companies (Siemens Energy, GE Vernova mentioned contextually) through 2032 and beyond. However, the near-term investment opportunity is less about these infrastructure plays and more about the organizational transformations they enable.

  4. The political risk to AI and tech companies is underpriced. Shor’s data shows AI concern is the fastest-rising political issue, and the public supports radical interventions. A data center ban, while unlikely in isolation, is politically plausible. More importantly, income guarantees, price controls, and heavy regulation of AI deployment poll extremely well across party lines. Investors should factor in meaningful regulatory risk to AI-exposed companies, particularly as the 2026-2028 election cycles approach.

  5. AI populism is the emerging political force to watch. The combination of AI anxiety and economic populism tests better than either alone. This is likely to produce bipartisan pressure for new regulations, taxes on AI-driven productivity, or mandated human-in-the-loop requirements. Companies that position themselves as pro-worker in their AI deployment may fare better politically and reputationally.

  6. Short-term labor market signal: Shor’s own firm has already shifted hiring away from copy editors and toward engineering and person-centric roles. This pattern is likely to accelerate across white-collar industries. Workers and investors should note that the jobs most at risk are those involving routine text production, translation, and copy editing, while roles requiring human judgment, client relationships, and regulatory accountability are more durable.

  7. Consumption inequality is decreasing even as income inequality rises. Hobart points to DoorDash usage being heaviest among lower-income people as evidence that AI-driven services disproportionately benefit consumers at lower income levels. Companies providing AI-enhanced services to mass-market consumers (delivery, financial tools, healthcare access) may see sustained demand growth even in a politically hostile environment.