AI, inequality, and the 22nd century: hosts debate Piketty, Ben Thompson, and whether human happiness is relative or absolute
Jan 6, 2026
Key Points
- Traml and Cash argue AI will concentrate wealth because capital owners capture automation returns while ordinary people cannot easily buy into companies like xAI, and most worker wealth locked in home equity won't benefit from productivity gains.
- Ben Thompson counters that humans fixate on relative deprivation even amid abundance, asking whether inequality matters if robots produce everything you want; the debate hinges on whether satisfaction is absolute or comparative.
- Samuel Hammond projects 5-10% TFP growth through the late 2030s constrained by infrastructure and regulation, with human-centric services and social status remaining valuable even as full robotics maturity approaches.
Summary
Philip Traml and Dar Cash published an essay using Thomas Piketty's Capital in the 21st Century to explore AI's impact on inequality. Their core claim: AI will concentrate wealth because capital owners capture returns from automation while most people cannot build wealth through labor. Three structural problems underlie this prediction. First, ordinary people cannot easily buy shares in xAI or other companies capturing AI upside. Second, most wealth is locked in home equity, which doesn't benefit from automation-driven returns. A $300,000 home in Ohio stays a $300,000 home even as capital productivity soars elsewhere. Third, automation eliminates the historical pattern in which poor countries attracted capital by offering cheap labor.
Ben Thompson reframes the problem entirely. He argues the essay assumes humans remain bothered by relative deprivation even in abundance. If robots produce everything you could want, does it matter that your neighbor owns 10,000 galaxies while you own 1,000? Thompson cites Louis C.K.'s observation that "everything is amazing right now and nobody's happy." We complain about slow airplane Wi-Fi while flying through the air, a miracle that didn't exist decades ago.
Thompson pushes back on the underlying assumption about human nature. The essay holds two incompatible positions: humans care deeply about wealth inequality, but they don't actually value other humans, their work, or their creativity. If they did, abundance created by capital would matter less than access to human connection and creativity.
A concrete example illustrates Thompson's logic. In 1965, total U.S. passenger flights equaled roughly the number of first-class flights in 2025. Service levels haven't degraded; they've been relabeled "economy" because the wealthy stayed in first class while 10 times as many people gained access to flying. Yet this doesn't feel satisfying because social comparison is relative, not absolute. You see people in first class and feel worse off, even though your absolute condition improved.
Samuel Hammond from the Foundation for American Innovation adds temporal nuance. He expects total factor productivity growth of 5–10% through the mid-to-late 2030s, constrained by infrastructure, regulation, supply chains, and human-in-the-loop work. Knowledge work will be partially automated, but demand for other humans, social relationships, and celebrity status will persist. Full robotics maturity and scale production reach the late 2030s, at which point growth could exceed 10%.
The unresolved tension is whether inequality is a relative phenomenon rooted in social comparison that abundance solves, or whether distribution mechanisms matter anyway. The debate hinges on whether humans experience satisfaction absolutely (I have what I want) or relationally (I have more than my peer group), a question technology cannot answer alone.