Docs/Guide

On the Luddites

This essay is the extended reading for the video "Opposing AI = Luddite?". A three-minute video holds conclusions; it does not hold the source work. This is the full version: what actually happened in 1811, an earlier round of machine-breaking almost nobody cites correctly, every crack in the "wheel of history" argument, and the question of who the word "inevitable" is actually working for.

Like Philosophy, this essay holds exactly one position: the objection is to how machines are used, not to machines. Every factual claim is sourced at the end.

I. 1811: The People Called Luddites

Start with the mistake almost everyone makes: there was no new technology in the Nottingham conflict.

The stocking frame the Luddites smashed was invented by William Lee in 1589 — by 1811, framework knitters had been working on these machines for over two centuries. The machine was their livelihood; they were among the most skilled machine operators in England. The popular image of "Luddites hating new machines" is wrong from its very foundation.

The real chain of motive: in those years, a group of hosiers hired cheap labor to churn out shoddy cut-up stockings on wide frames, crashing prices — and the wages of people doing honest work fell below subsistence.

They tried the legal route first. In 1778 and 1779, the framework knitters petitioned Parliament twice for wage regulation — both petitions were rejected. And under the Combination Acts (1799–1800), unions and strikes were criminal offenses: petitioning was the only legal channel the system left them, and that channel was closed. Only after the legal road ran out did frames start breaking in the night.

The breaking was selective: hosiers who cut wages and produced trash lost their machines by the row; those who paid fair rates lost not a single frame. The historian Eric Hobsbawm gave this behavior a precise name — collective bargaining by riot. It was the last negotiating move of people with no other table left, not hatred of machinery.

Parliament's answer was the Frame Breaking Act (1812): death for machine-breaking. Byron's maiden speech in the House of Lords (February 27, 1812) was a defense of the weavers; it changed nothing. The government moved some twelve thousand troops into the Midlands — more, Hobsbawm notes, than Wellington's field army in the Peninsula at the time. In early 1813, the York Special Commission hanged seventeen men.

In the same period, the framework knitters' United Committee under Gravener Henson was still pushing a trade-regulation bill through Parliament — it passed the Commons and was killed in the Lords. The violence and the legal path ran in parallel, and both roads ended at a wall.

Mainstream historiography (E. P. Thompson's The Making of the English Working Class, Kevin Binfield's Writings of the Luddites) reaches one consistent verdict: they objected to how the machines were used, not to the machines.

II. The Earlier Round: The Spinning Jenny and the Spinners

The "spinning jenny" comes up constantly in today's comment sections, so it deserves its own section — because the Luddites never smashed a single jenny. The people who smashed jennies did so more than forty years earlier.

The Spinning Jenny, invented by James Hargreaves around 1764–65, was a multi-spindle hand-powered spinning frame: one person at one wheel drove eight spindles, and later models ran to dozens and beyond. It answered the yarn famine created when the flying shuttle (1733) sped up weaving — spinning was the bottleneck of the whole industry.

The people it displaced were hand spinners, a workforce largely made up of women working at home. Spinning was one of the very few respectable paid occupations open to women — the English word spinster originally meant a spinning woman, so common was the unmarried woman supporting herself at the wheel that the occupation became the word for the marital status.

In 1768, spinners in Blackburn broke into Hargreaves' house and destroyed his machines (accounts differ on the details); he moved to Nottingham. In 1779, much larger riots swept Lancashire — and note this detail: the rioters destroyed only the large jennies of roughly twenty spindles and up, and left the domestic-scale machines untouched. A small jenny could still live inside the family economy; a large one could only stand in a capitalist's workshop. Like the Luddites thirty years later, what they opposed was usage and ownership.

Parliament did not side with them either, for structural reasons: before the Reform Act of 1832, workers had no vote; laissez-faire was becoming the new orthodoxy, and the Tudor-era wage and apprenticeship protections were being repealed one by one (1809–1814); the beneficiaries were concentrated and articulate, the losers diffuse and voiceless. Economic history has a term for the transition — the Engels' pause (Robert C. Allen, 2009): output grew for roughly two generations while workers' real wages stood still. The aggregate story came true in the end; the bill was charged to people who did not live to see the dividend.

Domestic hand-spinning as paid work was essentially extinct within a generation. Hundreds of thousands of rural women lost their main income — by scale, one of the largest single occupational displacements of the industrial revolution, and it barely enters the mainstream narrative, because the victims had no guild, no surviving petitions, and no names.

Now put "artists should just go learn the jenny" back into history. It is wrong three times over: wrong machine (the Luddites broke stocking frames, shearing frames, and power looms — not one jenny); wrong trade (the jenny hit spinners; the Luddites were framework knitters — people who already lived on machines, for whom "go learn the machine" was noise); wrong era (the jenny-breakers were spinners in 1768, two generations before Luddism). Whoever compresses these two histories into one sneer has just demonstrated knowing only the nouns of either.

III. Five Cracks in the "Wheel of History"

With the facts laid out, the argument — "the wheel of history rolls forward, whoever gets crushed had it coming" — can be inspected crack by crack. There are five.

First: description is not judgment. "It got crushed" is a statement of fact; "it deserved crushing" is a value claim, and the second does not follow from the first. Judging a demand by its defeat means the losing demand was the wrong demand — yet it is mainstream historiography that vindicated the machine-breakers. The winners' own historians do not hold the winners' version.

Second: the object of the objection is misidentified. In every historical case cited, the workers objected to the terms of deployment — wages, quality, ownership, consent — and the proof is the selective breaking that appeared independently in two different generations. Today's artists' demands are likewise entirely procedural: obtain consent before training on a work (Consent), attribute what is used (Credit), share revenue where there is commercial gain (Compensation) — three demands filed in black and white in the class action against AI companies (Andersen v. Stability AI, filed 2023, ongoing). Not one of them prohibits a technology.

Third: the analogy breaks at the input. A car does not need the coachman's labor to run; the jenny did not eat the spinners' yarn. A generative model fed no human work generates nothing. In the entire history of the industrial revolution there is no precedent for this: a machine built from the expropriated work-product of the very people it displaces, without their consent. This is not the story-shape of "new technology retires an old trade"; the industrial revolution's library of metaphors has no shelf for it.

Fourth: "adapt" presumes the chairs still exist. The economic point of the machine is to reduce the number of paid humans. In the 1810s, "learn the jenny" failed arithmetic — one machine replaced dozens, and the factory spinning jobs that followed went mostly to men and children. "Learn AI" fails the same way today: the new seats are orders of magnitude fewer than the displaced ones. This is not a question of attitude; it is subtraction.

Fifth: only half the history is cited. The "wheel" narrative quotes the crushing half and omits the rule-making half: the chaos of the printing press produced the Statute of Anne (1710) — the world's first copyright law, the legal foundation on which every creator earns a living today; a dozen years after Luddites were hanged, the Combination Acts were repealed (1824) and unions became legal; the flood of portable cameras produced the right to privacy (Warren & Brandeis, 1890); the seatbelt (Bohlin, 1959) went from a cost automakers resented to a legal standard. Every generation's technology rules were forced into existence by people who were called obstacles to progress. And that half of history is repeating now: China's labeling regulation for AI-generated content took effect in September 2025 (AI-generated content must be labeled; the obligation falls on service providers and platforms); the class action has been accepted by the court; Adobe trains on licensed libraries and pays contributor bonuses, Shutterstock runs a contributor fund, and AI companies are signing training-data licenses with content owners one by one. Nobody pays for a futile cause. The rules curve is also "the trend" — whoever cites only the adoption curve is reading half the chart.

IV. Who "Inevitable" Works For

The five cracks point to one question: why has an argument this broken been used, and believed, for two hundred years? Because the word "inevitable" is working three jobs at once.

It is a rhetorical weapon. Declaring a thing "bound to happen" waives the obligation to argue that it should happen, and absolves everyone involved — an inevitability has no right or wrong, only the wise and the unwise. Dressing laissez-faire up as economic law rather than class choice then, and dressing displacement up as physical law rather than commercial terms now, is the same rhetorical move: disguising what someone decided as what nobody decided. The beneficiaries vanish from view, and the losers' protest becomes a tantrum against nature.

It is a psychological shelter. "Inevitable" is not only the strong party's waiver; it is the weak party's painkiller. The shock is industry-wide and everyone is afraid; shouting "you deserved it, you failed to adapt" at the first visible casualties is a form of self-soothing — as long as "eliminated = failed to adapt" holds, the one shouting is still safe. Cognitive science has formal names for the mechanism: identity-protective cognition, and motivated reasoning. Once a position is welded to the self, rebutting the position becomes an attack on the person; evidence is no longer weighed, only repelled. This is not a defect of some other kind of person — it is everyone's factory setting, including the author of this essay. The only difference is whether one installs an error-correction routine.

It is narrative politics. The word "Luddite" is itself the losing side named by the winning side. The machine-breakers' own account — petition first, break selectively, the last negotiation — never entered the language; what entered was the winners' cartoon: fools who opposed progress. The spinning women never even qualified for a name; they simply vanish from the record. When the word is thrown at artists two hundred years later, the throwers no longer know where it came from — because the provenance was erased by the same exercise of naming power. History's verdicts are not discovered; they are written. All this essay does is take the pen back for one round.

And the way to resist "inevitable," across every case here, is exactly one move: unbundle "inevitable" back into "terms." That the technology will exist is indeed not up for a vote; but who benefits, who is credited, who gets fed into the machine, and who writes the rules — every one of those is negotiated, and every one of them has historically been moved.

V. Brakes

One last thing, more basic than any position.

This essay criticizes a posture of argument from start to finish, so it owes an answer: why believe its author isn't making the same mistake? By the "position welded to the self" standard, nobody is more welded than I am — my position is attached to my livelihood, my product, and my public identity.

The answer is not in the position; it is in the procedure. To tell whether someone is debating or defending a wall, don't measure their volume — count their concessions. A position that can grant "the AI shock is real, the industry may well contract, the fear is genuine," and holds only two lines — the extinction claim overreaches, and "they had it coming" does not follow — is load-bearing on evidence. A position that cannot concede a single step is load-bearing on faith, because conceding any step admits the whole wall is negotiable.

A person who rejects all external feedback is a car that has lost its brakes — the heading doesn't matter, because the road will turn and he has given up the ability to steer. The opposite of fatalism was never optimism. It is revisability.

Which is why this essay, that video, and this software are saying the same sentence: the ethics of a technology are not waited for — they are designed. And the first premise of design is admitting that the terms can be discussed.

— CreatorAris

Sources

E. J. Hobsbawm, "The Machine Breakers", Past & Present, no. 1 (1952) — origin of "collective bargaining by riot"
E. P. Thompson, The Making of the English Working Class (1963)
Kevin Binfield (ed.), Writings of the Luddites (2004)
Stocking frame: William Lee, 1589; petitions rejected 1778–79; Henson's bill: passed Commons, killed in Lords (1812)
Frame Breaking Act (1812); Byron's maiden speech (Feb 27, 1812); York Special Commission executions (1813)
Spinning Jenny: James Hargreaves, c. 1764–65, patented 1770; Blackburn attack of 1768 (accounts differ on details); Lancashire riots of 1779 (the spindle-count exemption line is reported variously as 20–24)
Robert C. Allen, "Engels' pause: Technical change, capital accumulation, and inequality in the British industrial revolution", Explorations in Economic History 46 (2009)
Statute of Anne (1710); repeal of the Combination Acts (1824)
Samuel Warren & Louis Brandeis, "The Right to Privacy", Harvard Law Review (1890)
Nils Bohlin's three-point seatbelt (Volvo, 1959, patent opened)
Andersen v. Stability AI (N.D. Cal., filed January 2023, ongoing — this essay asserts nothing about the outcome); Karla Ortiz, written testimony to the U.S. Senate Judiciary Committee (July 12, 2023)
Measures for Labeling AI-Generated Synthetic Content (Cyberspace Administration of China et al., effective September 1, 2025; original text at cac.gov.cn)