The cycle that began with [YOU] on AI asks what it means to see the machine clearly—without the narcotic of hype or the paralysis of fear. Galbraith is the cycle's anatomist of the hype: the thinker who showed, over a long career, how an economic consensus is manufactured by the people who benefit from it, and how to ask the rude question that cuts through it. Where does the power actually sit? Who decided what we are told to want? His answers, transposed from mid-century American capitalism onto the frontier AI industry, do not require adjustment so much as re-labeling.
His most portable contribution to the cycle is the concept of the dependence effect—the observation that in an affluent economy, producers increasingly create the desires they then satisfy. The AI industry's campaign to convince every business and every individual that they cannot afford to be without these tools is the dependence effect at its most sophisticated: want manufactured alongside product, urgency sold by the people who profit from the feeling. Galbraith had no word for the attention economy, but he described its logic in 1958, and the logic has only sharpened as the manufacture became personalized, continuous, and algorithmic.
He stands in the cycle's gallery as the thinker who supplies the political-economic lens that the others reach for in their own idioms. Where classical economics sees markets, Galbraith sees power. Where the tech optimist sees neutral tools, he sees the technostructure extending its reach. Where the industry narrative speaks of inevitability, he hears the conventional wisdom of a class that profits from that belief. His errors—especially his confidence that countervailing power arises automatically—are not embarrassments but warnings, the most useful part of the inheritance precisely because they are the questions still open.
Born in 1908 in rural Ontario, Galbraith came to economics through agricultural science, earning his doctorate at Berkeley and joining Harvard, where he would teach for nearly fifty years. During the Second World War he ran price control for the United States as deputy administrator of the Office of Price Administration, holding the prices of a wartime economy more or less stable—a feat he considered the major achievement of his life and a standing argument that the market did not regulate itself without deliberate management. He served as John F. Kennedy's ambassador to India, advised four presidents, and wrote more than forty books, all of them arguing that the textbook picture of sovereign consumers and competitive markets was a fairy tale maintained by its beneficiaries.
The shape of his suspicion was formed by what he saw from the inside of power. He observed that the dominant ideas in any era are the ones convenient to the dominant institutions, that economic theory had been built to describe markets that suited established interests, and that the most dangerous beliefs were the sincerely held ones—the conventional wisdom—because they required no defense and were therefore immune to argument. His recurring move was to ask of any economic arrangement who benefits from believing this, and the question never became less productive. By the time he wrote The New Industrial State in 1967, he had developed a full theory of the mature corporation as a planning organization run by specialists—the technostructure—that had escaped the discipline of the market and pursued growth and technical virtuosity as ends in themselves.
He was also wrong about important things, and he was wrong in ways consistent with his deepest temperament. He overestimated the permanence of the arrangements he described. He thought the planning corporation had permanently superseded the market; the corporate raiders of the 1980s and the shareholder-value revolution showed it had not. He thought countervailing power would emerge more or less automatically to balance concentrated corporate power; it largely did not, and the lax antitrust enforcement his thinking helped legitimate cleared the ground for the most dominant firms in history. These failures are not distractions from his legacy. They are the live, unsettled questions he handed forward—the questions most urgently open about artificial intelligence today.
The Conventional Wisdom. Galbraith's most portable concept: the body of beliefs a society accepts not because they have survived scrutiny but because they are familiar, reassuring, and convenient to those who prosper from them. The conventional wisdom is not a lie—it is sincerely held—which makes it far more resistant to argument than any lie could be. It is not undone by better reasoning but by events that make it untenable. The AI consensus—that the technology is inevitable, transformative, and beyond serious question—is conventional wisdom in his exact sense, manufactured by the people who sell it and accepted with a social cost to doubt it that he would have recognized immediately.
The Technostructure. In The New Industrial State (1967), Galbraith argued that the complexity of modern technology had transferred power inside the large firm from owners to specialists: engineers, scientists, planners, and managers whose pooled knowledge the enterprise could not function without. The technostructure pursued growth and technical virtuosity rather than profit maximization, because growth and virtuosity were what served the group. Applied to the frontier AI lab, the framework barely requires translation: the engineers who understand a frontier model hold power that no shareholder, board member, or regulator can effectively contest, and the objective—build the most powerful system, be first, expand the scope of the organization—matches Galbraith's description almost word for word.
The Dependence Effect. His most subversive claim: in an affluent economy, producers increasingly create the desires they then satisfy. The dependence effect inverts the orthodox picture of sovereign consumers choosing freely. Advertising and salesmanship manufacture the want alongside the product, so that the economy produces not to meet demand but to generate it. The AI industry is the dependence effect at industrial scale: vast capital floods into models and chips while a parallel campaign manufactures the conviction that no business or individual can afford to be left behind. The urgency is part of the product. The want is being created by the process that satisfies it.
Private Opulence, Public Squalor. Galbraith's most famous image—the gleaming automobile on the crumbling road, private abundance coexisting with neglected common goods—describes a structural tendency of markets to overprovide private goods (whose returns can be captured) and underprovide public ones (whose returns cannot). Applied to AI: the gains concentrate in firms and shareholders while the costs—displaced workers, strained energy grids, degraded information commons—are socialized. Private opulence, public squalor now arrives in silicon.
Countervailing Power. His most hopeful and most mistaken idea: that concentrated power on one side of a market tends to call forth organized power on the other. Galbraith thought the existence of Big Steel would reliably produce big unions and big buyers to balance it. Countervailing power often did not arise automatically, and where Galbraith's thinking encouraged complacency about concentration it helped produce the very monopolies it predicted would be balanced. The lesson is not that the theory is wrong but that countervailing power must be deliberately built—and that a technology specifically designed to automate away the human constituencies that have historically constituted such power makes the building harder than anything Galbraith faced.