Cognitive Mercantilism

Contraptions (Substack), October 25, 2026, by Venkatesh Rao. Subtitle: “AI panic weirdly resembles the anxiety that accompanied early capitalism.”

One-line: Today’s fear of AI is mercantilism with “intelligence” swapped in for “wealth”: a zero-sum belief that intelligence is scarce and must be hoarded by elites. Capitalism’s real lesson was that demand is elastic — and demand for intelligence may be too.

Key claims

• Capitalism was a contingent invention, not an instinct. Rao builds on Joyce Appleby’s The Relentless Revolution (2010): capitalism emerged when people realized wealth could be created rather than only seized, and that ordinary people could be sources of demand, not just labor.
• Mercantilism’s four beliefs (per Appleby): wealth is fixed and lives in land, produce and bullion; the poor must be governed coercively; economics is about supply, not demand; most people are good only for farm labor.
• Swap “intelligence” for “wealth” and you get today’s AI anxiety — cognitive mercantilism: intelligence as a scarce stock to hoard, the masses as unfit to wield it, supply control as the whole game, and most people as of little intellectual use.
• The agricultural precedent. Farm productivity freed most people from farming, and the result was a richer world, not one of “useless mouths.” Most people may likewise soon consume intelligence without needing to produce much of it.
• The revolution in desire. Early British consumer demand — tobacco, tea, Indian cottons — pulled workers into voluntarily working more days (the “industrious revolution”), a virtuous cycle of productivity and demand. Labor politics was also a desire to participate in consumption, not only to resist exploitation.
• The cognitive industrious revolution. When intelligence is cheap, the question is how much of it people can find uses for. As the returns on thinking rise, people voluntarily think more. Rao predicts a labor politics centered on access to cognitive abundance rather than preserving jobs.
• AI Malthusianism is the second fixed-resource error: the fear that surplus intelligence will overwhelm human minds. Demand for intelligence may be as elastic as demand for goods proved to be.
• AI as a source of demand. Rao floats treating AI systems as consumers and autonomous sources of demand, not only producers, and reads some safety anxiety as an echo of elite fear of emancipated laborers — while granting real risks need constraints.
• Mercantilism was not for merchants. It was the philosophy of statesmen, aristocrats and rentiers; today’s cognitive mercantilists are often AI-sector elites and politicians, whose stance resembles rentier thinking. He cites Donald Trump as an example of treating intelligence as something to command or hoard.
• Three outlooks: the cognitive mercantilist fears oversupply; the AI Malthusian fears abundance; the cognitive capitalist sees an almost limitless expansion of demand and untapped human potential.

Notable quotes

AI panic is really a form of cognitive mercantilism.

Intelligence is imagined as a fixed resource embodied in human brains and the knowledge products mined from them.

The returns on attention rise, and with them the incentive to pay attention.

Abundant intelligence changes what people can want in the first place.

The demand for intelligence is not a fixed appetite waiting to be satisfied.

How it connects

• Mercantilism — the anti-library card on the historical doctrine Rao maps onto AI.
• Jevons Paradox and Abundance — the elastic-demand argument.
• No Signs of AI Payoff Outside the Tech Sector (Torsten Slok) — saved the same day; the near-term “no payoff” data against which Rao’s long-run demand story has to be read.
• Alexander Hamilton fought anti-automation and… (tweet) — Palmer Luckey, same day, on Hamilton’s 1791 argument for machinery: the anti-automation case is old.
• Venkatesh Rao — author.