Daron Acemoglu has been the target of an unusually harsh piece in The Economist, which prompted fellow economists to come to his aid. The magazine cast doubt on the data and methods underlying his research on institutions; it suggested that Acemoglu frames complex historical episodes through overly simple narratives, and that “institutions” in his hands is a category elastic enough to explain anything; and it dismissed his position on AI as “pessimistic to the point of losing their credibility.” Acemoglu called the article a hit piece and has replied, objecting among other things that the verdict rests on the testimony of economists the magazine did not name.
Most of those who came to his defense questioned the very idea that a magazine could challenge the scientific output of one of the most prominent social scientists of his generation—and a Nobel laureate. This is a curious line of defense, and an old one. Guilds have always held that a craft can be judged only from inside it.
Among those who joined the debate is Luigi Zingales of the University of Chicago. In a post on X, Zingales—who has himself “criticized important parts of Daron’s work”—framed the controversy as a by-product of a disagreement over AI, or more precisely as a kind of exemplary punishment meted out by The Economist to a very prominent AI critic:
“Everybody has heard of the Luddites, but few know Sir Thomas Maitland, the man who defeated them. Rather than using brute force to take them all out, Maitland simply identified the men he took to be ringleaders, had them arrested on evidence that was often thin, and executed them. Hit one, educate a hundred. This is the strategy that the British magazine The Economist seems to have adopted toward the new version of the Luddites: AI critics.”
The sentence bears two readings: that AI critics are the new Luddites, or merely that they are being treated as such. The weaker reading turns the post into a complaint about tone, which would be a waste of a good analogy. So I will take Zingales to mean the stronger thing—which is also to concede him more than he asks. But taking a label seriously means asking what it once denoted.
So: who were the Luddites?
The movement took its name from a Ned Ludd who never existed—yet protesters signed letters and proclamations against mechanization as Ned Ludd, King Ludd, or General Ludd. It began in Nottinghamshire in March 1811, peaked in 1812, and had flickered out by 1816. It was the most prominent, but neither the only nor the first, episode of machine-breaking.
Resistance to the introduction of machines in production was not new. It was strongest in old trades, such as wool and hosiery, which were long established in England and capable of expressing political demands. Mechanization in late eighteenth- and early nineteenth-century England was not merely the introduction of new machinery; it was also a breakthrough in that particular technology known as organization. People reacted not only to instruments that could abridge work and economize labor for capitalists, but also to the birth of the factory system. That system resulted in the centralization of many independent workshops and the abandonment of the cottage industry. Technology made obsolete a domestic organization of production that many felt embodied the wisdom of ages, stabilized family relationships, and established strong bonds between masters and apprentices.
Cotton was relatively new and had little pre-industrial history in England: Arkwright patented the water frame in 1769 and opened Cromford in 1771. Wool and silk did have such a history, and work accustomed to bargaining collectively. The Webbs noted that Luddism “apparently began among the Framework-knitters,” long organized in local ubs with some rudimentary federal. What these men lacked was a lawful channel to protest: the Combination Laws of 1799 and 1800, passed to limit the risk of sedition during the war with France, did not invent the offense—combination was already actionable at common law as conspiracy—but made prosecution summary and cheap. These were old trades, with a habit of collective organization and no lawful way to press a claim.
Eric Hobsbawm influentially described Luddism as “collective bargaining by riot”: a form of negotiation against employers whose costs were being cut by machinery. Another Marxist historian, E.P. Thompson, saw the Luddites as part of a wider working-class movement with a radical political consciousness. On Thompson’s account, they opposed the ‘factorization’ of their livelihoods as much as the machines themselves. Joel Mokyr, in The Enlightened Economy, cautions that “the Luddite riots were more complex than just a response to mechanization”: grievances differed from district to district, the movements were uncoordinated, and in Lancashire the attacks fell less on new techniques as such than on whatever lay nearest to hand in an unorthodox bargaining game between employers and men.
Movement was suppressed by a combination of legal and military force. It is true that Maitland preferred intelligence to confrontation. In Yorkshire his troops never fired a shot in anger. But nothing about the operation was surgical. By May 1812 some fourteen thousand men had been sent north, a thousand of them stationed in Huddersfield alone—a force larger than Wellington’s army in Portugal in 1808 and several times greater than anything previously used against domestic unrest—supported by paid informers and by the systematic purchase of King’s Evidence. In March 1812 Parliament made frame-breaking a capital offense; Byron opposed the bill in the Lords that February and in 1816 composed his “Song for the Luddites” (“Down with all Kings but King Ludd!”). Of the sixty-four men brought before the York commission, seventeen were hanged—three on January 8, 1813, for the murder of the manufacturer William Horsfall, fourteen more on the 16th—and a few more were transported to Australia. The indictments were for murder, riot, robbery, and burglary as much as for breaking machines.
At any rate, not everyone in the workers’ movement opposed innovation. In A History of British Socialism, Max Beer reminds us that the radical William Cobbett “soon appealed to the Luddites to desist from destroying machinery and to join, instead, the movement for Parliamentary reform.” A few years later Thomas Hodgskin, defending labor against the claims of capital, would not target machines but stressed that human labor embodied the knowledge of their workings and was necessary to activate them. Hodgskin wrote for the Mechanics’ Magazine and edited The Chemist: both were part of an attempt to elevate labor through instruction, and to make human capital—a concept he came very close to articulating—more precious and valuable.
Zingales and Acemoglu are extremely accomplished scholars, not manual workers. Their positions are certainly useful to an interesting debate. Zingales, for example, supported Elon Musk’s suit against OpenAI on the grounds of corporate governance and the opacity of public benefit corporations, and in his podcast he has frequently questioned the AI hype, seeing it as fueled by unrealistic expectations. But the Luddites acted where every other channel was shut: no vote, no press, no lawful combination, and after March 1812 a capital statute them. Being handled roughly in a magazine is not being by Sir Thomas Maitland. If the analogy is meant to show that AI critics are being silenced, the visibility of the complaint refutes it. AI critics are not short of an audience, and the most recent case came from inside the trade. Jacob Coxon, who had done pretraining research at OpenAI and at Anthropic, resigned on September 9 with a post saying that the labs are gambling with our lives. It passed a hundred million views within a day. Perhaps the comparison holds—but in a very different sense.
Consider the narrative that the Luddites were up against something bigger and different than machines. They saw the collapse of a domestic mode of production into the factory, the of an arrangement that seemed to embody the wisdom of ages and to hold masters and apprentices together. Now ask which occupation today is most exposed to precisely that kind of reorganization. I would tend to agree with Zingales that it is still hard to see how transformative AI will be for our economies. What is quite apparent is that it will be transformative of our university system—or perhaps simply destructive of it, since education is hardly the most adaptable of industries. Are the people who live by teaching and research bound to be the new Luddites?
So far, most of them seem not to have smelled a coming storm. A Boston Fed survey found that the more educated the worker, the more he expects AI to raise productivity, and that doctorate holders were the least worried of all about their own prospects—though the survey was fielded at the end of 2024, and one wonders how it would read after two more years of Claude and Gemini. Outsiders are less sanguine: a 2026 poll of working-class voters found pluralities expecting AI to hurt teachers and education workers. The trade with the most to lose may indeed be education.
And it is a trade in the old sense. Accreditation, tenure, peer review, control of the credentials that license entry: the university is one of the last great guilds in the West, and the last great cottage industry, a confederation of small workshops each run by a master with his apprentices, held together by custom and by the wisdom of ages. Guilds have characteristic reflexes, one of which we saw at the start of this post: the conviction that a practitioner may be judged only by fellow practitioners. That reflex was not absurd in 1812 and it is not absurd now. It is just that it never once saved anybody. The trades that insisted only a framework-knitter could judge framework-knitting were not wrong about the knitting. They were arguing about the wrong thing, because what was being settled was not who understood the craft but who would organize it. Guilds rarely lose the argument about competence. They lose the question of relevance, and they lose it to whoever is reorganizing the work.
Hodgskin, if he were writing today, would presumably tell the profession what he told the machine-breakers of his own day: the knowledge of how the thing works is in you, and your business is to make that knowledge more valuable, not to smash the thing—or, in the modern version, not to demand that someone smash it on your behalf. Whether the profession will hear him is another matter. Will AI produce something that makes contemporary universities—Ivy League included—look like cottage industries? And will we see university professors marching on data centers?
(0 COMMENTS)
