Editor's note. This is the first paper in a long-form track inside Cancel Clankers — the place where the publication writes at chapter length, with citations, about what the weekly Dispatch and Argument can only point at. It is the compressed version of the seven-chapter book we are writing chapter by chapter. The Dispatches and the Arguments will keep running. The papers run alongside.
A preview essay
A Note on Optimism
There is a serious version of the optimist case for humanoid robotics, and it deserves a serious answer. The version we will spend the rest of this paper answering, the one made by people who have written real books and built real companies and earned the right to be wrong on the record, goes like this.
For most of human history, most adults spent most of their waking hours doing work that injured them, bored them, separated them from their children, and ended their lives early. Every time a wave of technology has lifted some piece of that burden — the plow, the steam engine, the assembly line, the elevator, the spreadsheet — the people doing the old work suffered through a transition, and the generation after them lived better. The pattern, on a long enough timeline, is unambiguous. We are taller, we live longer, we read more, we bury fewer of our children. The optimists who write about humanoids see another lifting of the burden. They see warehouse work that destroys a back over a decade replaced by a machine that does not have a back. They see eldercare in the third shift of a hospital, the loneliest and most invisible labor in this country, no longer rationed by who can afford it. They see the people doing the work freed, at last, to do something else.
That is the strong version. We are going to concede every true thing in it before we turn.
Yes — work injures people, and machines that take the injurious parts are gains. Yes — the long-run pattern, measured generously and across centuries, is that humans live materially better lives after technological transitions than before them. Yes — humans freed from the worst kinds of labor have built cathedrals, composed symphonies, sequenced genomes. The optimist is not lying about any of this.
Here is what the optimist is leaving out.
The first thing the optimist leaves out is that humanoid robots, by deliberate design, are not the elevator and not the spreadsheet. They are not built to take a specific task out of a human's day. They are built to take the human's place in the building. The form factor is the thesis. A machine that climbs the same stairs and lifts the same boxes and wears the same uniform and stands at the same workstation does not free a worker the way a forklift does. It replaces a worker the way another worker would. This is not an accident of engineering. The companies building these machines say so in their pitch decks. The investors funding them say so on their earnings calls. The form factor is the product, and the product is substitution.
The second thing the optimist leaves out is what happens to a person when the structure of their adult life is removed and not replaced. The optimist assumes the released worker walks into a richer existence. The historical record — the actual record, not the brochure — says the released worker walks into anomie, into early death, into despair that we know how to measure because we have measured it many times. The optimist talks about freedom from drudgery. The data we have on what freedom from drudgery does, when it arrives suddenly to people without other structures, is not the data the optimist is using.
The third thing the optimist leaves out is who owns the machines.
The fourth thing the optimist leaves out is what kind of access, on what terms, ordinary households will actually have to those machines once they exist.
The fifth thing the optimist leaves out is what populations do when they have been displaced quickly and densely and given no replacement.
The sixth thing the optimist leaves out is what the displacing class will use to protect itself when the fifth thing arrives.
And the seventh thing the optimist leaves out, the one we will spend our closing pages on, is that nothing about this is inevitable. There are off-ramps. There is a window. The window is open now. It is not open forever, and it is not very wide.
This is a paper about those seven things. It is short — eight thousand words — because it is a preview. The book that follows will be longer, and there will be more numbers, and more names, and more time spent inside each of the rooms where these decisions are being made. But the argument is already complete. You can hold all of it in one sitting. You should.
We are not against progress. We are against the specific design choice of building a machine in the shape of the worker. We are against the specific concentration of capital that this design choice rewards. We are against the specific kind of society this design choice, left alone, will produce.
We have a name for that society. We have a name for that machine. We have a name for the people who are about to be replaced by it, and the name is not a percentage on a memo. The name is yours, and ours, and our neighbors'.
The optimist is sincere. We grant the sincerity, and we deny the conclusion. Here is why.
§1 Displacement
The economist who has thought most carefully and most publicly about why prior waves of automation did not produce mass unemployment is David Autor at MIT. His 2015 paper, "Why Are There Still So Many Jobs?", is the canonical answer. The argument runs like this: automation takes specific tasks out of a worker's day, the remaining tasks become more valuable, and workers are reallocated to higher-skill complements of the new machines. The bank teller becomes the relationship manager; the typist becomes the analyst; the assembly-line operator becomes the maintenance technician. The complementarity holds across two centuries of industrial transition. It holds, Autor argued, because automation has always been about tasks — narrow, defined, severable from the rest of what a human does on the job.1 That is the strong version of the labor-optimist case, made by a serious person on the record. Andreessen's e/acc gloss, Adcock's earnings-call confidence, Khosla's investor talk — each is a louder, less careful descendant of this argument. We will spend this section dismantling the careful version. The loud versions fall when it does.
Here is the dismantling. Humanoid robots are not, by the people building them, designed to take tasks out of a worker's day. They are designed to take the worker out of the building. The form factor is the thesis. A machine that stands at a workstation, picks the same items, walks the same aisles, and clocks the same shift does not occupy a complementary slot adjacent to the worker — it occupies the worker's slot. This is what Daron Acemoglu and Pascual Restrepo, working through U.S. commuting-zone data, have been pointing to for the last six years: the wave we are inside is not task-displacing in the historical sense, where the displaced human shifts to a higher-skill complement. It is role-displacing. The human-shaped slot is the target.2 Acemoglu and Restrepo's 2022 paper made the consequence explicit: when automation primarily displaces rather than complements, the productivity effect is real but does not compensate, and the displacement effect dominates the wage and employment outcomes.3 The optimist case requires the complementarity to hold; the data say it has stopped holding.
The standard reply to this is the lump-of-labor fallacy — the idea that work is a fixed pie, that automation simply removes slices, that human ingenuity will generate new slices for the displaced to fill. The reply has been the dominant rhetorical move of labor optimism for two centuries. It is correct in the long historical record, and it has been correct precisely because the machine, until now, has not been able to do the thing the new slice required. Daniel Susskind's 2020 book, A World Without Work, walks through what happens to that reply when the machine is general-purpose. New slices are still being generated. The question the optimist has stopped asking is which species fills them.4 Anton Korinek's 2024 NBER paper formalizes this — in scenarios where capital can substitute for nearly all labor, the marginal product of labor approaches zero, wages collapse, and the lump-of-labor reassurance becomes a description of how the pie gets eaten without you.5 The fallacy is fallacious when one of the species at the table can buy the table.
The geography matters. Acemoglu and Restrepo's commuting-zone work found that one additional industrial robot per thousand workers reduced the local employment-to-population ratio by roughly 0.2 percentage points and local wages by roughly 0.42 percent, with the burden concentrated on production workers and routine occupations.6 These were industrial robots, bolted to factory floors. They could not climb a stair. They could not pick a non-uniform package. They were narrow, expensive, slow. They still produced a measurable, durable, geographically concentrated displacement in the towns where they landed. The humanoid is not the industrial robot. It is the industrial robot freed from the bolt. The geography of the next wave will not be the rust belt that already absorbed the last one; it will be the warehouse counties and distribution corridors that the last six weeks of public reporting have already named. Memphis. The Tennessee and Mississippi distribution belt where Nike eliminated 775 jobs and cited "increased use of automation" by name. Sunderland, where Nissan cut 900. Nashville, where Kroger closed its automated-fulfillment facility and laid off 132. Dow's 4,500. The white-collar wave is not separate; it is the early front. Six American CEOs in sixteen days this May announced more than 14,500 cuts and named AI as the driver, and those memos were written before a single humanoid had stepped onto a production floor in a U.S. facility. That has now changed.7
The fourth thing the optimist gets wrong is the clock. Carl Benedikt Frey's The Technology Trap spends its first third walking through what economic historians call Engels' pause — the multi-decade window during which the first industrial revolution made British workers materially worse off before wages began the long recovery the optimist now cites as evidence the system works.8 The recovery took two generations. The men and women who lived inside Engels' pause did not live to see it. The optimist treats those decades as the unavoidable cost of progress, and the recovery as the moral payoff. We note, simply, that Frey himself has been clear about what the historical pattern teaches: the disrupted generation eats the disruption. The optimist's "societies adapt" is a true statement about the optimist's grandchildren. It is not a true statement about the optimist's worker. The humanoid wave will not run on Frey's chronology. The deployment curve being demonstrated in Q2 2026 is moving in quarters where the ATM ran in years, and the reasons are already on the public record: the pilots are running, the capital has pooled at scale, the AI substrate is already deployed across the white-collar economy, and Chinese competitive pressure is collapsing every U.S. release cycle.9 The chronology compresses into a window of perhaps ten years, not eighty. A pause of one generation compresses to a single career.
The fifth and final piece of the displacement argument is the one Erik Brynjolfsson has now, to his credit, said out loud himself. In 2014, he and Andrew McAfee wrote The Second Machine Age — the optimist canon of its decade, an extended argument that humans should learn to "race with the machine," complementing it rather than competing.10 Eight years later, in Daedalus, Brynjolfsson published "The Turing Trap" and argued, in effect, that the design choice the industry has made is the dangerous one: by directing technical effort toward producing human-like machines, the industry has chosen substitution over augmentation, and the incentive structure now pushes ever harder in the substituting direction.11 Brynjolfsson did not name humanoids. He did not need to. The Turing trap is the humanoid trap. The architect of the optimist canon has, on the record, conceded the structural point. The question is no longer whether the form factor matters. The question is what we are willing to do about it.
The receipt for this section is the one Acemoglu and Restrepo placed on the public record in 2020 — one industrial robot per thousand workers, 0.2 percentage points of employment, 0.42 percent of wages, durable across the data, concentrated in the towns that already lost the last round. The receipt for the next decade is the one Brett Adcock placed on a livestream on the 13th of May, 2026 — Figure 03 cleared more than 10,000 packages across an unbroken eight-hour autonomous shift, with one visible miss and one unit self-replacement, no human on the floor.12 The optimist case required the complementarity to hold. The data say it has stopped holding. The form factor was the warning. The eight-hour shift was the receipt. The slot we are watching being filled is the human-shaped one, and there is no adjacent slot.
§2 Purposelessness
The serious version of the optimist case on meaning comes not from a venture capitalist but from a long tradition of post-scarcity humanism. Keynes wrote it down in 1930, in "Economic Possibilities for our Grandchildren," and it has been reanimated for the humanoid era by a generation of writers — Brynjolfsson and McAfee in passing, Andrew Yang on the campaign trail, the UBI advocates inside the AI labs — who argue that drudgery is what humans have always wanted to be free of.13 Keynes thought his grandchildren would work fifteen hours a week and spend the rest of their time on art, family, friendship, the cultivation of the inner life. The humanoid optimist takes the next step: now, finally, the technology exists to deliver Keynes's bargain. The factory floor empties. The eldercare ward stops being a place of unbearable labor. The eight-hour shift, which has cost human bodies and human marriages for two centuries, ends. The freed adult does what adults have always wanted to do when no one is making them do anything else. They live.
We grant the seriousness, and we deny the conclusion. The problem with this argument is that we have a great deal of evidence about what humans actually do when the structure of work is removed without a replacement, and the evidence does not match the brochure.
The first piece of evidence is conceptual, and it comes from Hannah Arendt's The Human Condition, published in 1958, in the only chapter of mid-twentieth-century philosophy that the humanoid optimists ought to be required to read. Arendt distinguished three categories of human activity. Labor is the cyclical, repetitive work of biological survival — eating, cleaning, tending the body, sustaining the species. Work is fabrication — the building of a world of durable objects that outlasts the maker, the world of the craftsman and the carpenter and the engineer. Action is speech and deed in the company of others — the realm of politics, of being seen and heard, of having a place in the public conversation.14 Arendt's argument, written in the long shadow of the totalitarianisms of the twentieth century, was that the three are not interchangeable and not collapsible. A society that strips away labor without preserving work and action does not produce flourishing; it produces what she called the animal laborans freed into nothing. The humanoid wave, by its design, takes the labor category and absorbs it. The optimist promises that the freed human moves up into work and action. The mechanism by which that movement happens is unspecified, and the mechanism matters.
The second piece of evidence is empirical, and it comes from a place the labor optimists rarely visit — the mortality data. Anne Case and Angus Deaton, the Princeton economists who have spent the last decade on what they call deaths of despair, have documented the reversal of two centuries of declining mortality among working-age Americans without bachelor's degrees. The deaths come in three forms — suicide, drug overdose, and alcoholic liver disease — and the trend lines tracked, with eerie tightness, the structural employment changes of the regions hit hardest by the prior wave of dislocation.15 In the manufacturing belt of the Midwest, in the Appalachian coal counties, in the deindustrialized small cities of the Rust Belt, the deaths arrived with the same chronology as the plant closures and the same chronology as the collapse of the institutions that had been built around plant work. Marriage rates fell. Church attendance fell. Union membership fell. Life expectancy fell. The optimist promised that the displaced would find higher purposes. The displaced were buried, in numbers that have not stopped accumulating.
The third piece of evidence is historical, and it answers the optimist's likely reply — the displaced of yesterday were not given enough income; if we get the income side right, the rest follows. The reply is reasonable. It is also wrong, and the evidence has been on the record for a century. The leisure class of the European aristocracy and the American Gilded Age had income. They had income at scales that make the most generous UBI proposals look modest. They did not, by and large, build cathedrals or write symphonies or sequence genomes. They drank, gambled, dueled, and produced an extensively documented intergenerational pattern of depression, dependency, and despair that Thorstein Veblen catalogued in 1899 and that the historical record has continued to confirm.16 The natural experiment has been run. Income without structure does not produce flourishing. It produces a specific kind of human damage that we have books and books about.
The fourth piece of evidence, and the one the humanoid optimist most needs to confront, is what work actually is in the lives of ordinary people. Studs Terkel spent the early 1970s asking. He recorded hundreds of working Americans — steelworkers, hotel switchboard operators, fire watchers, gravediggers, hospital orderlies, schoolteachers, prostitutes, gas station attendants — and what came back was not a uniform complaint about drudgery. What came back was something subtler and more dangerous to the optimist case. The work, for most of his subjects, was the part of the day where they were somebody. The job site was the place where their name was known, their skill was visible, their contribution mattered, and the absence of those things at the end of a working life — in retirement, in injury, in layoff — was reported by his subjects as a specific kind of grief.17 Robert Putnam's two long books, Bowling Alone and The Upswing, traced what that grief did at the level of civic association: as work disappeared or fragmented, the union halls, the bowling leagues, the parent-teacher associations, the volunteer fire companies, the Lions Clubs, and the small associative life of American communities went with it.18 The workplace was not just where the wage was earned. It was the load-bearing column of the building.
The fifth piece of evidence, the one we hold for the close of this section, is what the numbers look like now, before the humanoid wave arrives. Prime-age male labor-force participation in the United States stood at roughly ninety-seven percent in the early 1960s. It has fallen, in a steady multi-decade drift, to roughly eighty-eight or eighty-nine percent in 2025.19 The eight or nine percentage points that have gone missing are not retirees and not students. They are working-age men who have left the labor force entirely. They are the population on whom the deaths-of-despair curves have landed hardest. They are also, for our purposes, the baseline — the floor we are looking at before the humanoid wave begins. The optimist is proposing to take that floor and lower it, and proposing to do so under the description that the lowering is liberation.
The honest reply from the optimist is the UBI reply, and we will spend more time with it in §7. The short version is this. Income transfers solve the income problem. The UBI advocates are right that no one should starve, and they are right that a serious income floor is necessary. They are not right that a serious income floor is sufficient. The five pieces of evidence above are evidence that meaning, structure, association, and contribution do not arrive in the mail with a check. Arendt's action is not vouchered. Case and Deaton's mortality reversal did not respond to disability checks. Terkel's subjects did not describe their pay; they described their work. The income side is solvable. The meaning side has no known engineering solution, and the optimist's casual assumption that the freed human will produce the meaning out of leisure has — at the population scale, in the actual data — a track record that we know how to read.
The receipt for §2 is the one Case and Deaton placed on the public record in 2020, and that the CDC's mortality data have confirmed in every annual release since. American life expectancy, for the first time outside wartime in a century, fell in three consecutive years before the pandemic and has not recovered since. The cause was not the pandemic. The cause was the slow unbuilding of the structures inside which working people had lived. The humanoid wave is the next step of that unbuilding, and the optimist's reply — freed for higher purposes — has, in the population most likely to be freed, already been tested. The test is on the tombstones.
§3 Capture
The democratization claim is the third piece of the optimist case, and the people making it have a serious historical record to point at. Vinod Khosla has been arguing for over a decade that the broad arc of consumer technology is the commoditization of expensive expertise — what cost fifty thousand dollars in 1970 costs forty dollars in 2025 — and that AI extends the curve into domains, like personalized medicine and personalized tutoring, that until now resisted it.20 Marc Andreessen's "Techno-Optimist Manifesto," published in October 2023, made the broader version on behalf of his entire portfolio class: the techno-capital machine produces abundance broadly, and resistance to it represents a grave moral failure.21 Brett Adcock has been on the consumer-trajectory record for two years — a humanoid in every household by the early 2030s, priced under twenty thousand dollars at scale, the smartphone curve applied to bodies. The strongest version of the case is this. The personal computer was a hand-built hobbyist device in 1975. The smartphone was a six-hundred-dollar luxury good in 2007. The humanoid will follow the same trajectory and end up in every American garage by 2035. Every household gets a body. The playing field flattens. We are going to spend the rest of this section explaining why the smartphone is the wrong analogy and Standard Oil is the right one.
The mechanism Thomas Piketty laid out a decade ago in Capital in the Twenty-First Century — that the rate of return on capital persistently exceeds the rate of economic growth, and that this is the structural reason wealth concentrates over time regardless of how productive the underlying economy becomes — has, until now, been bounded by one constraint. The capital good and the labor good were different things.22 A factory could not, by itself, run a factory; an investment in machinery had to be paired with labor that owned half the bargaining table. The humanoid wave dissolves the constraint. For the first time in the industrial era, the capital good is the labor good. The investment, the worker, and the rent flow back to the same balance sheet. Anton Korinek and Joseph Stiglitz formalized the consequence in 2019: under realistic institutional conditions, the productivity gains from broad AI capability accrue disproportionately to the owners of the substituting capital, and the income-distribution problem does not solve itself through aggregate growth.23 Daniel Susskind's 2020 Oxford model paper carries the formalism the rest of the way — under broadening machine capability, the equilibrium share of national income paid to labor falls, and may fall to zero in the limit.24 The optimist's "growth lifts all boats" assumes a distribution mechanism that the structure of ownership refuses to produce.
The smartphone analogy is true about its consumer half and false about its strategic half. The hardware in our pockets is much cheaper than it was twenty years ago. The substantive capability of that hardware — the cloud services it depends on, the foundation models it queries, the app store it cannot leave, the operating system it cannot replace, the data exhaust it cannot prevent — lives inside a market structure that is far more concentrated than the smartphone industry of 2007. Two companies own the mobile operating system. Three companies host most of the cloud. A handful of laboratories train the frontier models. The hardware commoditized; the platform did not, and the platform is where the rents accumulate. The substantive question for humanoids is not what a unit costs at scale. The substantive question is who owns the foundation model, who owns the manufactured fleet, who owns the training-data corpus generated by every shift the fleet works, and who owns the compute substrate any of it depends on. The answers, in 2026, are already concentrated, and concentrated in fewer hands than the smartphone era ever was.
The historical record on what happens to foundational platform technologies under capitalist ownership has been written, repeatedly, by Tim Wu. The Master Switch documents the recurring cycle through which open, disruptive information technologies are reliably captured by integrated monopolies — radio, telephony, film, broadcast television, cable, the internet — each moving from openness to consolidation over a span of decades, driven by the same combination of network effects, capital intensity, and vertical-integration incentives.25 No exception is on record. The optimist's burden, if the smartphone analogy is to hold, is to explain why humanoid robotics will be the first foundational platform technology in the modern record to escape the cycle. The case for the exception is hard to make. The capital intensity of humanoid manufacturing exceeds any prior consumer-electronics category. The network effects are stronger — every fleet hour generates training data that improves the model that controls the fleet, and the model belongs to the firm. The vertical-integration incentive is stronger — the platform builder owns the foundation model, the manufactured body, the training corpus, and the compute. There is no historical case for the exception. The 2026 market structure, in this section's closing receipt, has already executed the consolidation in advance.
Daron Acemoglu and Simon Johnson's 2023 book Power and Progress is the clearest statement of the political point, and the one the democratization optimist most needs to confront. The choice of which technologies to develop and deploy is not given by physics. It is made — by capital allocation, by regulatory framework, by labor's institutional power, by the political coalitions that succeed or fail at directing the development.26 The first century of the British industrial revolution, between roughly 1760 and 1860, produced no broad prosperity. The standard of living for the working population stagnated or declined while productivity rose. Broad prosperity arrived in the second century, after labor had organized, after suffrage had expanded, after the state had been forced to redistribute. The lesson is not "technology eventually delivers." The lesson is "labor eventually fought for it." The humanoid wave is arriving at a moment when labor's institutional power in the United States is at a hundred-year low — private-sector union density around six percent, total union density around ten percent, both the lowest figures in the BLS series.27 The optimist treats the form factor as a discovery handed down from physics; Acemoglu and Johnson treat it as a decision being made by a small number of firms under specific institutional conditions. Both descriptions can be defended. Only one of them has a historical track record.
The lock-in is the fifth piece. Piketty's second book, Capital and Ideology, spends its middle chapters cataloguing the legal, ideological, and educational mechanisms by which concentrated wealth defends itself once it has formed — the tax preferences, the educational stratification, the legal infrastructure designed by and for capital.28 Tim Wu's The Curse of Bigness names the specific American instance: the antitrust tradition that had operated for seventy years as a structural check on concentration was systematically dismantled between roughly 1975 and 1985, replaced by a "consumer welfare" standard that asks only whether prices to consumers go up.29 By that standard, every humanoid platform's roadmap satisfies the test — they all promise lower prices to consumers. By the older standard, the one used for most of the twentieth century, the same roadmap fails — concentration is itself the harm. The current regulatory cycle on AI and robotics is being run by firms whose policy staff write the framework's drafts and whose lobbyists shepherd those drafts through. The mechanism is not novel. It is the regulatory-capture mechanic the United States ran on broadcasting in the 1930s and on telecommunications in the 1990s. It is running now on humanoids, and the window during which a different choice is available is narrowing in real time.
The receipt for §3 is the present-tense market structure of the humanoid industry, which has already done its consolidation in advance of the consumer phase the optimist invokes. Figure AI raised at a thirty-nine-billion-dollar valuation in early 2026. Apptronik closed a Series A of more than nine hundred million dollars in the same window. Tesla's Optimus program, Agility Robotics with its GXO partnership, 1X with its commercial pilots, Boston Dynamics at BMW Spartanburg, Toyota at Woodstock, Humanoid Inc. with the Schaeffler two-thousand-unit contract — eight to ten firms constitute the platform-builder cohort in 2026, and the Humanoid Index ETF launched this year on a ticker, $BOT, whose entire investable universe fits on a single page.30 By 1882, twelve years into its consolidation, Standard Oil controlled roughly nine-tenths of U.S. refining capacity. The humanoid cohort has assembled comparable share — by capital raised, by signed deployment contracts, by foundation-model ownership, by training-data accumulation — in roughly three years, before the consumer phase has begun. The optimist's democratization is being made about a market whose 2026 structure looks like Standard Oil's 1880, not like the personal computer's 1990. The concentration arrived before the scale, not after.
§4 Lease Society
The fourth piece of the optimist case is the consumer-liberation argument, and it has the most attractive spokespeople of any version we will engage with. Bernt Børnich at 1X has spent two years pitching NEO, the company's consumer-tier humanoid, as the household analogue of the personal computer — a member of the family, a presence in the kitchen, a help to the elderly parent.31 Brett Adcock has put the Figure consumer roadmap on the record at sub-twenty-thousand-dollar price points. Elon Musk has been more grandiose — Optimus as your friend, your dance partner, your reason never again to have to do the dishes. The strongest version of the case follows the analogy. The personal computer in 1980 was distrusted and expensive; by 2000 it was a tool of liberation in every middle-class home. The smartphone repeated the trick on a tighter timeline. The personal humanoid will too. The optimist's vision is the worker freed of drudgery, yes — but also, and this is the version they save for the consumer-facing keynote, the elderly parent who keeps her independence, the disabled adult who lives unassisted, the household where dinner cooks itself while the family talks. We grant the charm of the vision. We are going to spend this section explaining why the vision presupposes a thing — ownership — that the firms making the machines have already decided you will not have.
Aaron Perzanowski and Jason Schultz documented the legal architecture of the transformation in The End of Ownership, published in 2016, in the only book on consumer law the humanoid optimists ought to be required to read.32 An entire generation of consumer products that used to be objects you possessed have been reclassified as services you license. The books on your Kindle were never your books — Amazon famously demonstrated as much in 2009, when it remotely deleted purchased copies of George Orwell's 1984 from customers' devices over a licensing dispute. The films in your iTunes library were never your films; their disappearance, on the platform's schedule, is a contract term you agreed to without reading. Adobe's transition from perpetual licenses to the Creative Cloud subscription, announced in May 2013, was the most-discussed instance of a category move that has since become the default — every Photoshop customer who paid for the perpetual license in 2012 has, by 2026, paid the subscription several times over for the same software.33 The license is the product, and the license is revocable. The shift was incremental enough to escape political notice but, taken end to end, the shift is what happened. We do not, in 2026, own most of what we have paid for. The optimist's "personal humanoid in every household" is being proposed at the end of a quarter-century in which household ownership of complex devices has been systematically replaced with household subscription to them.
Cory Doctorow's now-canonical word for what happens to platforms under the licensing logic — enshittification — names the predictable arc the optimist's pitch deck does not include.34 The platform is generous to users to attract them, generous to business customers to extract value from them, then degrades both sides to maximize extraction for the shareholder. Every platform Doctorow has documented — Facebook, Amazon, Google Search, Uber, the entire app-store layer — has executed roughly the same playbook. The reason humanoids are uniquely suited to this playbook, and uniquely vulnerable to its consequences, is that the humanoid is not a screen you can put down and not a service you can switch from. It is a physical object in your home with a capability you have built your household around. The asymmetry between user and platform that Doctorow describes for digital services is, in the humanoid case, a physical asymmetry as well. The platform can throttle your humanoid's capability remotely, can require a subscription to features that worked yesterday, can require a payment to unlock features promised at sale, and the user's option to leave is the option to live without the device they have integrated into the load-bearing structure of their household. The lock-in is not behavioral. The lock-in is architectural.
The "appliance-that-can-refuse" is not a hypothetical. The history of how it works has been on the public record for a decade, in three cases the humanoid optimist has not addressed. John Deere spent the 2010s and early 2020s blocking farmers from repairing their own tractors, requiring authorized-dealer service software, and treating diagnostic codes as protected intellectual property — the Federal Trade Commission's 2025 complaint and consent order against Deere & Company was the first federal acknowledgment that the practice constituted an antitrust harm.35 Tesla has, on the public record, remotely revoked range-extension features from used vehicles whose new owners did not pay a per-vehicle relicensing fee, has bound Autopilot and Full Self-Driving capability to the specific vehicle identification number that originally purchased them, and has documented post-sale remote deactivation of features that worked before resale.36 HP's printer business has, for over a decade, used DRM-locked ink cartridges and firmware updates that brick third-party supplies — a pattern documented by iFixit, the Electronic Frontier Foundation, and consumer-protection litigation in multiple jurisdictions.37 These are the precedents. They are not edge cases; they are the dominant model under which complex consumer hardware now operates in the United States. The humanoid is, in this lineage, the appliance whose refusal stakes are highest — because the capability it might withhold is the capability you brought it into your home to provide, and the household reorganized itself around its presence.
Shoshana Zuboff's diagnosis of surveillance capitalism names the business model the optimist's brochure strips out, and the humanoid is its terminal form factor.38 The household humanoid is, by the physics of the hardware, the most comprehensive sensor that has ever been placed inside a private dwelling. It sees everyone in the home. It hears everyone in the home. It records every interaction with every member of the household, at a resolution and continuity no Alexa, no Nest, no Ring camera has approached. The data exhaust is the product, in the Zuboff sense, at least as much as the cleaning of the kitchen is. The data exhaust, in turn, is what feeds the second-order layer the optimist has not asked you to think about — the credit-scoring, the insurance-pricing, the employability-rating, the political-risk-rating apparatus that has been quietly built on the data exhaust of the smartphone era and that will be enriched by an order of magnitude by the data exhaust of the humanoid era. Samantha Hoffman, Jeremy Daum, and Rogier Creemers have spent the last decade documenting the Chinese social credit experiment carefully — not because the United States is becoming China, but because China has, on the public record, demonstrated what the technical apparatus can do when it is deployed at the scale a fleet of household humanoids would make possible.39 The forward indicator is on the table. The optimist has not addressed it.
The dignity argument is the simplest one, and the one the optimist's charm relies on us not asking. The humanoid in the brochure is your servant — a presence in your home that obeys you, helps you, makes your life easier. The humanoid in the contract is the platform's agent — a presence in your home that the platform configures, updates, downgrades, monitors, and reserves the right to repurpose. These are not the same machine. They are the same hardware running different governance. The optimist asks you to imagine the first and quietly hands you the second. Living in a household that contains an appliance you do not control, that can be revoked, that reports on you, that may at any moment refuse to perform the thing you bought it for, is not the liberation the optimist is selling. It is the precise opposite. The vision of independence the optimist invokes — the elderly parent who keeps her independence, the disabled adult who lives unassisted — is the vision the contract makes impossible, because the contract makes the household conditional on the platform's continued willingness to permit the humanoid's continued function. A capability you do not control is not a capability. It is a permission.
The receipt for §4 is the John Deere case, because it is the case the courts have already adjudicated. The Federal Trade Commission filed its complaint against Deere & Company in January 2025, alleging that the company's systematic restriction of repair to authorized dealers — combined with the diagnostic-code lockout, the firmware-update kill switch, and the proprietary-tool requirement — constituted an antitrust violation. Deere settled, agreeing to make its diagnostic tools available and to commit to right-to-repair obligations. The settlement is the public-record statement that an American farmer, paying full price for a piece of American manufacturing, did not own the machine he was farming with — that what the contract had transferred was a permission, conditional on the manufacturer's continued willingness to grant it. The same legal architecture — proprietary diagnostic tools, license-bound capability, firmware-revocable feature gates, no third-party servicing — is the architecture the humanoid platform-builders are deploying, at scale, in the consumer roadmap the optimist is selling. The case that established the harm was settled in 2025. The form factor in which the harm will hit the household ships, in consumer volumes, beginning in 2027. The window for legislating the lesson learned from a hundred and fifty years of American farmers is the window between now and then.
wx5 Rupture
The fifth piece of the optimist case is the most reasonable-sounding of them all, and the most dangerous. It is the historical claim — societies adapt, they always have, the disruption is real but bounded, the adaptation arrives, the equilibrium reasserts itself. David Autor has been the most careful articulator of this case across his 2015 Journal of Economic Perspectives paper and the decade of work that followed; Steven Pinker's Enlightenment Now (2018) is the macro version, anchored in centuries of data on declining violence, increasing wealth, and expanding political enfranchisement.40 Andreessen's "Techno-Optimist Manifesto" is the loud version, riding the same horse without the supporting historical work. The optimist asks us to concede the historical record, and we will. Yes, the British workers of 1830 lived materially worse than the British workers of 1760, and the British workers of 1900 lived materially better. Yes, the displaced Americans of 1933 found new structures by 1950. Yes, in the long run, every prior wave of automation has been absorbed, the institutions have re-formed, and the population on the other side has emerged better fed, better housed, and longer-lived. We grant the long run. We are going to spend this section on the inconvenient interval between disruption and adaptation, the people who lived in it, and the political shape the unrest of that interval took. The optimist's "societies adapt" is a sentence about the optimist's grandchildren. It is not a sentence about the optimist's neighbors.
The Luddites are the most-cited and least-understood figures in the optimist's repertoire. The cartoon Luddite — the technophobe smashing a loom because change frightens him — is a Victorian-era myth, manufactured by mill owners and circulated by mid-twentieth-century textbooks. E. P. Thompson's 1963 The Making of the English Working Class recovered who the actual Luddites were: skilled framework knitters and croppers in early-nineteenth-century England, fighting a specific quasi-legal battle over the operation of the new wide-frame machines, demanding regulated work, fair wages, and adherence to the apprenticeship laws then still on the statute books.41 They were not opposed to machines. They were opposed to a specific configuration of machinery that violated the labor compact under which their towns had organized themselves for two centuries. The British state responded by hanging seventeen of them in York in January 1813, deploying twelve thousand troops to the Luddite counties (more than Wellington had at Salamanca the year before), and making frame-breaking a capital offense. The mill owners won. The framework knitters disappeared. The towns the framework knitters had filled depopulated, and the displaced moved into the new factory cities, where infant mortality rates in the 1820s and 1830s exceeded those of plague-era London. The Luddites were not wrong about what was coming. They were right, on the public record, about the consequences for the population they were defending. The historical irony is that they have been reduced to a slur against people who, like them, are correct.
The 1930s in the United States is the cleanest modern case study available, because the data are abundant and the chronology is documented. The mechanization of American agriculture between roughly 1900 and 1940 displaced approximately a third of the U.S. labor force from the farm. The displaced did not, by and large, walk smoothly into industrial work. They walked into a decade of joblessness, dust-bowl emigration, evictions, and the specific political crises the period produced.42 The country that emerged on the other side of the New Deal — with Social Security, the National Labor Relations Act, federal deposit insurance, and the modern administrative state — was not the country that entered the 1920s. The adaptation arrived. But the adaptation arrived through the specific political process that the Great Depression made unavoidable: mass strikes, the largest socialist and communist movements in American history, the Court-packing crisis, the Bonus Army's encampment and forcible expulsion from Washington in 1932, and the McCormack-Dickstein committee's 1934 investigation into the alleged plot to install a corporatist regime under Smedley Butler's leadership. The 1930s is the case the optimist cites as proof "societies adapt." The 1930s is also, on the historical record, the closest the United States came in the twentieth century to losing the republic. The adaptation cost the political instruments of the country a refit at gunpoint. That is what adaptation looks like in the during.
Karl Polanyi's The Great Transformation, written from exile during the Second World War and published in 1944, supplied the theory under which the Luddites and the 1930s and every subsequent adaptation describe the same mechanism. The "double movement" thesis runs like this: when capitalism unbinds the market from the social structures that have constrained it — when labor, land, and money are treated as ordinary commodities rather than as the substrate of ordinary life — society generates a compensating counter-movement, organized through politics, demanding that the substrate be re-embedded.43 The counter-movement can take democratic forms (the New Deal, the postwar European welfare state, the Indian independence movement). It can take authoritarian forms (Italian and German fascism, the Stalinist consolidation). The shape of the counter-movement is determined not by the disruption itself but by the political institutions and coalitions available to absorb it. Barrington Moore's Social Origins of Dictatorship and Democracy (1966) supplies the comparative-politics evidence beneath Polanyi's framework, tracing the paths from agrarian society to modern political form across England, France, Germany, Japan, Russia, India, China, and the United States, and showing in each case that the shape of the counter-movement was path-dependent on the specific institutional capacities present at the moment of rupture.44 Polanyi's argument, written in the wreckage of the worst possible answer to the question, is that the counter-movement is not optional. It will come. The contribution Polanyi and Moore make together is the specification: the shape of the arrival depends on choices that must be made in advance.
Peter Turchin's quantitative-history project, developed across Ages of Discord (2016) and End Times (2023), supplies the contemporary measurement framework. Turchin and his collaborators identified four structural indicators that, on long historical data, precede political instability: elite overproduction (more aspirants to elite status than positions available), declining mass living standards, rising state fiscal stress, and declining institutional trust. By Turchin's measurements, the United States entered the "instability window" around 2010, and each of his four indicators has worsened in every annual measurement since.45 The humanoid wave, in this framework, is the impulse applied to a system that is already in late-phase pre-crisis dynamics. Turchin's models do not predict revolution; they predict instability — declining state capacity, polarized factional politics, episodic political violence, and a heightened probability of constitutional crises. Hannah Arendt's The Origins of Totalitarianism (1951) supplies the human-scale companion to Turchin's quantitative one: the diagnosis that totalitarian movements draw their members not from any particular class but from the mass — the people who have been rendered politically and economically homeless, whose previous structures of meaning and membership have dissolved, and who are available to be recruited into movements that promise meaning and inclusion at the cost of the institutions that had previously distributed them.46 The humanoid wave is, in the Arendt-Turchin framework, a mass-producing machine. The optimist's "societies adapt" assumes the adaptive capacity of the institutions is intact. The data say the adaptive capacity has been degrading for fifteen years before the humanoid impulse arrives. Adding the impulse to the system in its current state is not the safe move.
The optimist's final retreat, when the historical record runs out, is to argue that this time will be different in the easier direction — better adaptation, lighter cost, faster equilibrium. The argument is unfalsifiable in its strong form, and it is the optimist's responsibility to provide the evidence. The evidence on offer, when one looks for it, is thin. The optimist points to the long-run trend lines and asks us to extrapolate. We point to the institutional conditions that produced the trend lines — strong labor movements, expanding suffrage, a state with the fiscal capacity to redistribute, a press with the editorial capacity to investigate, a judiciary with the political insulation to enforce, an executive with the legitimacy to act — and we ask whether any of those conditions describes the United States of 2026. The answer is no. The conditions under which prior adaptations succeeded have been hollowed out by forty years of deliberate policy choices. The optimist is asking us to bet on the adaptation while the institutional infrastructure of adaptation is at its lowest ebb in a century. The "this time" the optimist is invoking is, in fact, different — but in the harder direction. Acemoglu and Johnson's point in Power and Progress applies in both temporal directions: the broad-prosperity outcomes the optimist invokes arrived only when labor had institutional power to direct the choice, and the institutional power is precisely what has been disassembled in advance of the impulse.
The receipt for §5 is the Autor, Dorn, and Hanson "China shock" series, because it is the cleanest contemporary measurement of the mechanism we have been describing. The trio's papers, beginning with their 2013 American Economic Review article and extending through more than a decade of follow-up work, traced the consequences of the post-2001 surge in Chinese manufacturing imports on U.S. commuting zones with measurable manufacturing exposure.47 The findings are not contested. Manufacturing employment collapsed in the exposed zones, and the collapse did not produce the adjustment Autor's earlier theoretical work had assumed. Wages fell, labor-force participation fell, marriage rates fell, mortality rose, and — the politically critical finding, documented in the 2020 paper with Kaveh Majlesi — the exposed zones moved sharply toward populist and authoritarian-style candidates in subsequent election cycles, on both the right and the left. The 2016 American political earthquake mapped, with eerie tightness, onto the China-shock geography. The 2024 reinforcement of the same coalition deepened the map. Dani Rodrik's and Barry Eichengreen's comparative work extends the same finding across European cases: structural displacement, unaccompanied by institutional repair, produces populist political reaction in predictable ways.48 The China shock was a one-percent shock applied to a single sector over a decade. The humanoid wave is the same mechanism scaled up by an order of magnitude, applied across multiple sectors, on a compressed timeline. The China-shock political result is the floor of what to expect, not the ceiling. The optimist owes us an account of why the same mechanism, scaled up tenfold, will produce a different political outcome. No optimist has yet produced one.
§6 The Iron Garden
The optimist's case has, by now, asked us to concede a compressed wave of displacement, a meaning-collapse on the other side of it, a consolidation of ownership over the new productive class of machines, a household-level subordination through the lease model, and a predictable political instability the historical record promises will follow. The optimist's position requires all five concessions. It then requires one more — that the instability resolve, the displacement stabilize, the new arrangement settle into legitimacy, and the wealth concentrated by §3 remain in the hands that concentrated it. The optimist does not, in the public version of the case, specify how this last step is to happen. The optimist does not name the mechanism. We are going to spend this section naming the mechanism the optimist's position implies. The implicit answer to the question — by what means is the new order secured against the population to whom it was done — is the answer nobody at the conference is required to say out loud. We are saying it out loud. The mechanism, on the public record, in the form of the firms already manufacturing it and the customers already buying it, is the same humanoid form factor whose civilian version we have spent the last five sections discussing — repurposed for the defensive needs of a society arranged against the bottom seventy percent of itself. The civilian humanoid is the front of the catalog. The other half of the catalog is in this section.
Mike Davis spent a career on the geography of fortified affluence. City of Quartz (1990) catalogued Los Angeles's transition into what Davis called "Fortress L.A." — gated communities, private security districts, sensored public space, the hostile architecture of public squares, the deliberate engineering of urban space against the unhoused and the poor.49 Sixteen years later, Planet of Slums (2006) extended the analysis to the global scale: by the early twenty-first century, the developing world's billions of people living in informal urban settlements outside the formal economy had produced, country by country, the same response — fortified enclaves for the propertied, hardened policing for the boundaries between, and the systematic privatization of urban order. Teresa Caldeira's City of Walls (2000) documented the São Paulo version with the precision of an anthropologist living inside it — the gated condominiums, the armored cars, the private-school routes, the daily geography of avoidance that an upper-middle-class household must internalize to survive the city.50 These are not warnings about a possible future. They are documentation of an existing present. The propertied have been organizing themselves into a fortified geography for forty years. The humanoid wave is the technology that finishes that arrangement — the manpower component of the fortification, automated, sourced from the same firms whose consumer roadmaps fill the keynote slides.
The history of private security in the United States is the history of labor unrest, and the connection is structural. Allan Pinkerton's National Detective Agency, founded in Chicago in 1850, was within twenty-five years the largest standing armed force in North America — larger than the United States Army of the same period — operating as the private army that broke the Great Railroad Strike of 1877, the Homestead Strike of 1892, and the Pullman Strike of 1894.51 The Pinkertons did, by private contract, what the public state was not yet politically able to do: project lethal force, on behalf of capital, against the populations performing the work. The pattern recurred. The Baldwin-Felts Detective Agency broke the southern Appalachian coal strikes of the 1910s and 1920s. The Coal and Iron Police of Pennsylvania were a private uniformed force, deputized by the state but paid by the operators, with statutory authority to use lethal force against miners and their families on company land. The pattern moved overseas in the 1990s and 2000s, when DynCorp, MPRI, and Blackwater performed the same function for the United States in Iraq and Afghanistan. Jeremy Scahill's Blackwater (2007) is the canonical narrative; the Nisour Square killings of September 2007 and the subsequent decade of legal contortions over contractor accountability are the canonical case. The privatization of force is not a hypothesis. It is a hundred and seventy-five years of American labor and foreign-policy history. The humanoid platform-builders are walking into a market that has been built, by their predecessors, since the Lincoln administration.
Antony Loewenstein's The Palestine Laboratory (2023) names the contemporary mechanism by which population-control technology becomes domestic-policing technology.52 The technologies of population management are developed at colonial borders, tested under "battlefield" conditions, and then marketed to the security industries of the global North. The Israeli military-industrial export sector, the U.S. Department of Defense's Section 1033 surplus-equipment program, the Department of Homeland Security's border-tech procurement pipeline, the European Frontex agency's adoption of Israeli sensor systems — these are not separate phenomena. They are pieces of a single development pipeline. Border tech becomes city tech becomes campus tech becomes corporate tech. Stephen Graham's Cities Under Siege (2010) traced the architecture of the pipeline a decade earlier — the migration of what Graham called "new military urbanism" from Baghdad and the West Bank to Belfast and Detroit, the surveillance and crowd-control techniques perfected in colonial and counterinsurgency contexts and then re-imported for use against domestic populations under the description of public safety.53 The humanoid form factor is the next node in this pipeline. The military pilot programs run first, under controlled conditions, against permitted targets, with limited press access and full deniability. The technology that survives the military pilot graduates to police pilots. The police pilots graduate to private security. The private security graduates to corporate facility management. Each step, the legal accountability declines. Each step, the population to whom the technology may be applied broadens. The pipeline has been running for a quarter-century with conventional military and surveillance technologies. The humanoid bipedal is now in it.
The contracts are on the public record. Ghost Robotics has been selling its quadrupedal Vision 60 platform to the U.S. Department of Defense for over five years; Sword International's mounted-weapon "SPUR" module turns the Ghost platform into an armed system that Sword's own marketing material has demonstrated firing rifle ammunition at moving targets.54 The U.S. Air Force has deployed Ghost Robotics quadrupeds at multiple bases for security patrol since 2020. In October 2022, six leading robotics manufacturers — Boston Dynamics, Agility Robotics, ANYbotics, Clearpath Robotics, Open Robotics, and Unitree — signed a public pledge not to weaponize their general-purpose robots. Ghost Robotics, conspicuously, did not sign. The pledge itself is the proof that the question is active inside the industry; the absence of one name from it is the proof that the answer is not unanimous. The New York Police Department reactivated its Boston Dynamics Spot deployment, branded "Digidog," in 2023 after a 2021 suspension under public pressure. The Department of Homeland Security's Customs and Border Protection has piloted Spot units at the U.S. southern border. The Israeli military has been operating autonomous and remote-piloted ground systems at the Gaza fence since the mid-2010s; the +972 Magazine and Local Call reporting on the IDF's "Lavender" and "Gospel" systems, based on interviews with serving Israeli intelligence officers, established that AI-driven targeting at the scale of urban populations is not a future possibility but a present practice.55 None of these are speculative. They are programs of record, with congressional line items, vendor contracts, and operational doctrine. The humanoid bipedal is the next form factor in the same procurement stream.
The forcing question, the one the optimist's pitch deck does not include and that this section is written to require an answer to, is the simplest one. Who buys the first ten thousand humanoid security units, deploys them where, against whom, and what process — judicial, legislative, regulatory, civic — stands between the procurement and the deployment? In 2026 the answer to the first part is: the U.S. Department of Defense and the major federal security agencies, the Israeli military, the contracted-out border-security agencies of the European Union, the gulf-state royal protection services, and the upper-tier private-security firms whose corporate clients have already begun line-item budgeting for autonomous-systems contracts. The answer to "where" is: military bases, ports of entry, "high-value asset" sites — data centers, executive residences, energy infrastructure — and, at the next step in the pipeline, the corporate-campus and gated-community perimeters that Davis and Caldeira documented forty years ago. The answer to "against whom" is the population to whom the displacement of §1, the meaning-collapse of §2, the consolidation of §3, the lease-society of §4, and the political unrest of §5 will have happened. The answer to "what process stands between procurement and deployment" is, on the present public record, no process. The American Civil Liberties Union and the Electronic Frontier Foundation have spent over a decade documenting the absence of meaningful oversight on police acquisition of autonomous systems — no federal framework requiring public notice, weak state frameworks, and the local-procurement model that the Section 1033 surplus-equipment program has demonstrated is exploitable by any sheriff with a budget line.56 The optimist's silence on this question is not absence. It is the answer. The mechanism is being built, on the public record, against the population on whose behalf the optimist is publicly claiming to speak.
The receipt for §6 is the specific case the public debate has already adjudicated. On November 29, 2022, the San Francisco Board of Supervisors voted 8–3 to permit the San Francisco Police Department to use lethal force by remote-controlled robots under "extreme circumstances."57 The vote was the rare instance in which the question of police robot lethal force was put to a democratic body in advance of the technology's widespread deployment, with the question stated honestly. One week later, on December 6, the Board reversed itself under sustained public pressure, after the city's residents — informed of what their supervisors had voted for — produced a political response that the supervisors could not absorb. The reversal was the right outcome on the wrong timeline. The city had to be told, by reporters and by organizers, that its commissioners had voted to authorize police killing by robot, and the city's reaction was what reversed the policy. The optimist's view of the humanoid future requires this not to happen — requires the question not to be put, not to be debated openly, not to be reversed under public pressure. The optimist's view requires the police commissions of three thousand American cities, the procurement officers of the major federal security agencies, the security directors of the corporate campuses, and the contracting offices of the private-security firms to not put the question to a vote at all. The San Francisco episode showed, on the public record, what the question looks like when it is asked. It showed what the population says when asked. The optimist's bet, in this section, is that the question is not asked. The mechanism by which the question is not asked is the mechanism we have just described.
§7 Off-Ramps
The optimist's final retreat, when every other line of defense has been answered, is the inevitability claim. We are told that nothing can be done because nothing has ever been done — that technology arrives, that the population either adapts or doesn't, that the alternative to acceptance is a regression to a less prosperous past nobody actually wants. Marc Andreessen's "Techno-Optimist Manifesto" makes the argument explicitly, naming "bureaucracy" and "vetocracy" as the enemies and casting structural constraint on technology development as moral failure.58 Sam Altman has made variants of it on behalf of OpenAI; Demis Hassabis on behalf of DeepMind; the general formulation runs through every public AI-lab statement of intent — capability is coming, the question is only who builds it first, and the responsible parties are those who keep development inside the firms most committed to "safe" deployment. The argument is offered as humility. It is, structurally, an abdication.
We are not going to spend this section refuting the inevitability claim abstractly. We are going to spend it on six concrete proposals — each with a named precedent, a named historical analogue, an existing legal or institutional template — that the United States and its allies could enact in the next three to ten years and that would, in combination, direct the humanoid wave away from the cascade we have spent the last six sections describing. None of these are speculative. Each has been written down by serious people, advocated by named coalitions, and, where applicable, implemented in part by other jurisdictions on smaller test cases. The case that nothing can be done is, like every prior version of the optimist case, an assertion that requires evidence. The evidence has been gathered. The proposals exist. The question is whether the political coalition to enact them can be built inside the window the displacement cascade leaves us.
Proposal one: compute governance. The training of frontier AI models — including the foundation models that control humanoid systems — requires concentrated compute resources (datacenters, specialized GPUs, networking infrastructure) that are observable, locatable, and physically tied to specific firms and jurisdictions. The technical case that compute is the most effective regulatory choke point has been made carefully across a growing literature; the political case has been initiated through the EU AI Act, which entered into force in August 2024 with risk-tiered restrictions by application domain, and through the Bletchley Declaration of November 2023, in which twenty-eight states and the European Union committed to coordinate on frontier-AI risk assessment.59 These are starting points, not finished frameworks. The Montreal Protocol — the 1987 treaty that phased out ozone-depleting substances, enforced through trade mechanisms, monitored through observable industrial measurements, demonstrated to work across four decades — is the structural precedent. A coordinated international framework on humanoid bipedal commercial deployment — restrictions by use-case domain, mandatory human-oversight requirements, prohibition of weaponized configurations as a class — sits well inside the legal-instrument tradition Western democracies have used to govern dual-use technologies for seventy years. The argument that it cannot be done is the argument that the Montreal Protocol did not happen.
Proposal two: labor-replacement liability. Anton Korinek's "Taxing AI" (2022) lays out the case in its cleanest form: when a firm deploys substituting automation, the firm captures the productivity gain while the displaced worker, the worker's community, the worker's social-insurance system, and the broader fiscal apparatus absorb the cost.60 The externality has been identified for half a century in the labor-economics literature; it has been measured in the Acemoglu-Restrepo commuting-zone data we cited in §1; it has been politically lived through across the China-shock geographies we cited in §5. Korinek's proposal is to tax labor-replacing capital deployment at a rate that internalizes the externality — higher than the tax on labor-complementing capital deployment, structured to fund the adjustment and the floor underneath the displaced. The specific rate is a technical question. The structural move is a political one. The legal precedent — taxing externalities to internalize them — has been the foundation of American environmental policy since the 1970s, of American labor policy since the New Deal, and of American antitrust policy since the Sherman Act. The proposal is not radical. The radical position is the current one — that the firms doing the displacing pay nothing toward the consequences.
Proposal three: ownership redistribution. Alaska has, since 1982, distributed an annual dividend to every state resident drawn from the rents on Alaska's oil and mineral resources.61 The fund's permanent corpus exceeds eighty billion dollars; annual dividends have averaged on the order of one to two thousand dollars per resident across the program's four-plus decades. Norway's Government Pension Fund Global has built a corpus exceeding 1.5 trillion U.S. dollars on the same principle, from North Sea hydrocarbon rents.62 The proposal: treat the productivity gains from broadly deployed substituting automation as a national resource rent, structurally analogous to oil and minerals, and redistribute a share of those gains to the population on whose behalf the underlying public infrastructure — the educational system, the research base, the legal framework, the consumer market — was built. The mechanism: a sovereign humanoid-and-AI fund, capitalized through a statutory equity allocation from the major U.S. AI and humanoid firms (with international counterparts in allied jurisdictions), administered as a sovereign-wealth instrument, paying a structural dividend to every American resident. The political-economic argument was made decades earlier in the social-democratic tradition; the contemporary version is anchored by Korinek and extended in Acemoglu and Johnson's policy chapter. The Alaska fund is the existing forty-three-year proof that the mechanism works at scale and that the coalition for it can be built.
Proposal four: antitrust pre-consolidation. Lina Khan's 2017 Yale Law Journal note, "Amazon's Antitrust Paradox," reconstructed the case for structural antitrust enforcement separated from the consumer-welfare standard that has dominated American antitrust since the 1970s.63 Her tenure as FTC chair between 2021 and 2025 demonstrated that the framework can be enforced inside the existing administrative state when the political appointees so choose — the action against Deere & Company we cited in §4 is one of several. The specific application to the humanoid platform-builders: structural separations preventing single firms from owning the foundation-model layer, the manufactured-fleet layer, the training-data corpus, and the compute substrate simultaneously. Tim Wu's Curse of Bigness supplied the doctrinal foundation; Khan supplied the enforcement template; the federal antitrust statutes — the Sherman Act, the Clayton Act, Section 5 of the FTC Act — are sufficient for the action the moment requires. The proposal is to apply the enforcement now, while the platform-builder cohort is at eight to ten firms and the consolidation is incomplete, rather than thirty years from now, when the cohort is three firms and the consolidation is structural. The window is the window before the entrenchment; by the §3 evidence, that window is already short.
Proposal five: right-to-repair and the ownership floor. The FTC's 2025 settlement with John Deere is the foundation — a federal acknowledgment that the appliance-that-can-refuse model constitutes an antitrust harm.64 The European Union's Digital Markets Act, in force since May 2023, supplies the international template — interoperability mandates, data-portability requirements, structural-separation provisions for designated "gatekeeper" platforms.65 Cory Doctorow's policy program in The Internet Con extends the same logic.66 Applied to humanoids, the proposal: a federal humanoid right-to-repair statute prohibiting firmware-revocable feature gates, requiring open diagnostic interfaces, mandating that the operational software of a household humanoid run on the owner's hardware (not the manufacturer's cloud), and establishing that the act of purchasing a humanoid transfers ownership in the legal sense — the capability cannot be retroactively revoked, modified against the owner's interest, or made conditional on subscription. The proposal generalizes the Deere settlement from a single industry to a category of consumer-deployed autonomous systems. The legal instrument is already drafted in the European framework. The American version requires federal statute, which requires the political coalition we name in proposal six.
Proposal six: the coalition, and the window. The sixth proposal is, properly speaking, not a policy but the meta-policy — the question of who passes the first five. Mariana Mazzucato's The Entrepreneurial State (2013) and Mission Economy (2021) supply the framework: the modern state has historically been the most effective directional force on technology development when it has both the legitimacy and the institutional capacity to act, and the question of capacity is the question of coalition.67 The coalition required to pass proposals one through five is not, in 2026, hidden. It consists of: organized labor, in the early phase of rebuilding from the bottom of the BLS series we cited in §3; the small-and-medium-enterprise sector that has begun to recognize platform-monopoly as an existential threat; the small-farmer constituency already mobilized by the Deere right-to-repair fight; the professional middle class newly absorbing the white-collar AI displacement we cited in §1; the structural left, both its political and its civil-society wings; the structural-right populist faction whose objection to corporate concentration overlaps with the left's at the margin where these proposals would be enacted; and the academic and civil-society infrastructure — Mazzucato's institute, the Roosevelt Institute, the Open Markets Institute, AI Now, the Algorithmic Justice League, EPIC, Public Citizen — that has been generating the policy proposals for a decade waiting for a window. The Acemoglu-Johnson lesson from §5 applies in the affirmative direction here as well: the broad-prosperity outcomes of the second century of British industrialization arrived because labor and its allies fought for them, and the same condition obtains now. The window for the coalition to form, organize, and push the first five proposals through the political process is the window between the present and the moment the humanoid platform-builders reach the scale and lobbying capacity to entrench the consolidation past the point of legislative remedy. By our §3 evidence, that window is, conservatively, three to seven years.
The receipt for §7 — and for the entire affirmative argument of this paper — is the Alaska Permanent Fund's four-decade operational record. Established by amendment to the Alaska state constitution in 1976, capitalized by the rents on the state's mineral and hydrocarbon resources, distributing annual dividends to every resident continuously since 1982, the fund is the existing proof that the mechanism the optimist insists cannot work has, in an American jurisdiction, been working for nearly half a century.68 Through Democratic and Republican state administrations, through commodity-price collapses and booms, through the political pressures of every would-be reformer, the fund has held. Every Alaskan resident gets the dividend. The corpus exceeds eighty billion dollars. The structural design has survived four decades of political turbulence specifically because the constituency for it — the residents who receive the dividend — is the entire state population, and they vote. The lesson is not Alaskan exceptionalism. The lesson is that the political-economic argument the optimist treats as utopian was implemented in this country in 1976, by people who decided to do the thing. The mechanism exists. The precedent exists. The coalition that built it was assembled in a single American state in roughly three years. The same coalition can be assembled at the federal scale on a similar timeline, if the work begins now. The off-ramps are not theoretical. They are a forty-three-year operational record. The choice the optimist insists is impossible has already been made, by Americans, and is paying dividends, this year, into Alaskan bank accounts. The question is whether the rest of us choose to learn from them.
The Window
The optimist's case fails at the structural level we have spent this paper documenting. We have spent it on the seven things the optimist leaves out and on the six concrete things that, done together, would direct the wave differently. The cascade is not speculative. The actors are named. The contracts are filed. The funds are raised. The valuations are recorded. The patents are published. The bills, in their European versions, are already on the books. The American work is a coalition problem and a political-economy problem and a time problem, in that order.
We will not pretend the work is easy. The work is hard. It involves rebuilding institutions that have been deliberately weakened for two political generations. It involves persuading a population taught to mistrust the state to invest the state with new directional capacity. It involves convincing a small-business sector that fears regulation more than it fears monopoly that the monopoly is the larger fear. It involves organizing across the structural left and the structural-right populist faction in a country whose political infrastructure has been designed to keep them apart. None of this is easy. None of this is impossible. Each piece has been done before, by Americans who lived through worse moments and answered them.
The window is three to seven years. We are saying this as carefully as we can. It is not a prediction. It is an estimate, drawn from the §3 evidence on consolidation timelines, the §1 evidence on white-collar wave compression, and the §5 evidence on instability indicators. The window is the period during which the platform-builder cohort is still eight to ten firms instead of three, the political coalition is still possible to assemble, the labor-force participation collapse can still be slowed instead of accelerated, the right-to-repair statute can still be passed before the lock-in becomes irreversible. After the window, the lift becomes harder by orders of magnitude. Before the window closes, the lift is hard but bounded.
We will not pretend we have the answer. We have offered an argument. The argument is a starting point. The work — the actual building of the coalition, the writing of the bills, the organizing of the constituencies, the daily political labor by which any of this happens — is yours, and ours, and our neighbors'. Cancel Clankers will continue. The Dispatches will continue. The Arguments will continue. The papers, of which this is the first, will continue. The book this paper compresses will be written, one chapter at a time, across the window we have just named.
We were not designed to be replaced. We were not built to be leased. We were not raised to be ruled by appliances. We are not the population on whose behalf the optimist claims to speak; we are the population to whom the optimist's vision is being done. The optimist is sincere. We have granted the sincerity. We have, across seven sections, denied the conclusion. The work begins. The window is open. It is not open forever, and it is not very wide.
About
Cancel Clankers is a publication, not a person. The byline is Cancel Clankers — or, when the moment calls for it, The Resistance — because humanoid replacement is a structural problem and not a personal opinion. No individual writer's reputation is doing the work in these pages. The argument is doing the work. The argument is older than any of us, and it will continue past any of us.
The publication's Dispatches run on Sundays, the Arguments run on Thursdays, and the long-form papers — of which The Case Against the Humanoid Future is the first — ship on their own clock alongside. The next paper, Chapter 1: Displacement, deepens the §1 argument across the full ten-to-fifteen-thousand-word length the book chapters carry. The remaining chapters follow, in the order they are sequenced in §1–§7 above, until the seven-chapter book this paper previews is complete.
The work continues for as long as the window remains open.
Endnotes
-
David Autor, "Why Are There Still So Many Jobs? The History and Future of Workplace Automation," Journal of Economic Perspectives 29, no. 3 (Summer 2015): 3–30.
-
Daron Acemoglu and Pascual Restrepo, "Robots and Jobs: Evidence from US Labor Markets," Journal of Political Economy 128, no. 6 (2020): 2188–2244.
-
Daron Acemoglu and Pascual Restrepo, "Tasks, Automation, and the Rise in U.S. Wage Inequality," Econometrica 90, no. 5 (2022): 1973–2016.
-
Daniel Susskind, A World Without Work: Technology, Automation, and How We Should Respond (New York: Metropolitan Books, 2020).
-
Anton Korinek, "Scenarios for the Transition to AGI," NBER Working Paper No. 32255 (March 2024).
-
Acemoglu and Restrepo, "Robots and Jobs" (2020), at 2208–2210 — the headline coefficient on employment-to-population and the wage effect.
-
Cancel Clankers, "Six CEOs. One Answer.," Dispatch 001, 2026-05-10. The named cuts: Matthew Prince / Cloudflare, Dennis Woodside / Freshworks, Brian Armstrong / Coinbase, Enrique Lores / PayPal, the Arctic Wolf executive team, and Mark Zuckerberg / Meta. Total ~14,500 announced in sixteen days. The Nike, Nissan, Dow, and Kroger figures are from contemporaneous corporate announcements and reporting in the same window.
-
Carl Benedikt Frey, The Technology Trap: Capital, Labor, and Power in the Age of Automation (Princeton: Princeton University Press, 2019), chapters 4–6 on Engels' pause and the political economy of the British transition.
-
Cancel Clankers, "Eight Hours. The Bar Just Fell.," Argument 001, 2026-05-14. The four acceleration factors and the deployment-curve compression are documented there with named pilot sites (BMW Spartanburg, Toyota Canada Woodstock, Hayward / 1X, GXO with Apptronik / Apollo) and contract milestones (Humanoid with Schaeffler, up to 2,000 units by 2032, first units live by year's end).
-
Erik Brynjolfsson and Andrew McAfee, The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (New York: W. W. Norton, 2014).
-
Erik Brynjolfsson, "The Turing Trap: The Promise and Peril of Human-Like Artificial Intelligence," Daedalus 151, no. 2 (Spring 2022): 272–287.
-
Figure AI livestream, 2026-05-13. The pre-stream challenge by Dr. Scott Walter (Diligence Director, RoboStrategy) and Adcock's reply ("we already do this every day") are on the public X record from 2026-05-12. The live demonstration aired the following day. Documented in Argument 001 above.
-
John Maynard Keynes, "Economic Possibilities for our Grandchildren" (1930), reprinted in Essays in Persuasion (London: Macmillan, 1931). The fifteen-hour week and the cultivation of "the art of life itself" appear in Section IV.
-
Hannah Arendt, The Human Condition (Chicago: University of Chicago Press, 1958). The tripartite distinction between labor, work, and action is laid out across chapters III–V; the diagnosis of the animal laborans freed into the void is the through-line.
-
Anne Case and Angus Deaton, Deaths of Despair and the Future of Capitalism (Princeton: Princeton University Press, 2020). Age-adjusted mortality reversal among working-age Americans without bachelor's degrees, beginning in the late 1990s; the three-cause structure (suicide, drug overdose, alcoholic liver disease) and the geographic concentration in regions of prior industrial dislocation are documented across chapters 1–4.
-
Thorstein Veblen, The Theory of the Leisure Class: An Economic Study of Institutions (New York: Macmillan, 1899). Veblen's account of conspicuous leisure and the predatory psychological pattern that accompanies inherited surplus is the foundational text; the subsequent twentieth-century social-history literature on aristocratic and inherited-wealth populations has reinforced rather than overturned the diagnosis.
-
Studs Terkel, Working: People Talk About What They Do All Day and How They Feel About What They Do (New York: Pantheon, 1974). The book is a primary-source archive; the synthesis offered above is ours.
-
Robert D. Putnam, Bowling Alone: The Collapse and Revival of American Community (New York: Simon & Schuster, 2000); Robert D. Putnam with Shaylyn Romney Garrett, The Upswing: How America Came Together a Century Ago and How We Can Do It Again (New York: Simon & Schuster, 2020).
-
U.S. Bureau of Labor Statistics, "Employment Situation" series, prime-age (25–54) male labor-force participation rate. The 1960s figure is from BLS historical series; the 2025 figure from the most recent monthly release. The 9-percentage-point decline corresponds to roughly five million working-age men.
-
Vinod Khosla, "Do We Need Doctors or Algorithms?" TechCrunch, January 10, 2012; and subsequent Khosla Ventures essays on AI and the democratization of expertise. Khosla's specific claim is that personalized medicine, education, and other high-cost expert services follow the long-run consumer-electronics curve from luxury good to commodity.
-
Marc Andreessen, "The Techno-Optimist Manifesto," Andreessen Horowitz (a16z), October 16, 2023. The manifesto explicitly names "the techno-capital machine" and identifies critics of broad technology deployment as morally culpable.
-
Thomas Piketty, Capital in the Twenty-First Century, trans. Arthur Goldhammer (Cambridge, MA: Belknap Press of Harvard University Press, 2014). The r > g framework is the core analytical apparatus, developed across Parts Two and Three.
-
Anton Korinek and Joseph E. Stiglitz, "Artificial Intelligence and Its Implications for Income Distribution and Unemployment," in The Economics of Artificial Intelligence: An Agenda, ed. Ajay Agrawal, Joshua Gans, and Avi Goldfarb (Chicago: University of Chicago Press, 2019). Originally circulated as NBER Working Paper 24174.
-
Daniel Susskind, "A Model of Technological Unemployment," Oxford University Department of Economics Discussion Paper No. 819 (July 2017; revised 2020). The formal model behind the prose argument in A World Without Work.
-
Tim Wu, The Master Switch: The Rise and Fall of Information Empires (New York: Knopf, 2010). The "Cycle" framework is laid out in the opening chapters and traced through the twentieth-century case studies that follow.
-
Daron Acemoglu and Simon Johnson, Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity (New York: PublicAffairs, 2023). The British industrial-revolution case is developed across chapters 4–6; the "directable technology" framing is the through-line.
-
U.S. Bureau of Labor Statistics, "Union Members Summary," annual release, 2024 reference year. Total union membership in the United States stood at roughly ten percent of wage and salary workers; private-sector union membership at roughly six percent — both the lowest figures in the modern series. The series begins in 1983 in its current form; earlier estimates from BLS and academic historical sources place mid-twentieth-century total union density above thirty percent.
-
Thomas Piketty, Capital and Ideology, trans. Arthur Goldhammer (Cambridge, MA: Belknap Press of Harvard University Press, 2020). The argument that wealth's defense mechanisms — legal, ideological, educational — operate as a system is developed across Parts Three and Four.
-
Tim Wu, The Curse of Bigness: Antitrust in the New Gilded Age (New York: Columbia Global Reports, 2018). The reconstruction of the Brandeisian antitrust tradition and the diagnosis of its 1970s–1980s dismantlement under the Bork-influenced "consumer welfare" standard.
-
Cancel Clankers, "Eight Hours. The Bar Just Fell.," Argument 001, 2026-05-14, catalogs the public-record pieces. The Figure AI thirty-nine-billion-dollar valuation round, Apptronik's Series A close above nine hundred million, the Schaeffler–Humanoid Inc. contract for up to two thousand units by 2032 with first units live by year's end, the Humanoid Index ETF (NYSE: $BOT) launched in 2026, and the cohort of platform-builder firms — Tesla Optimus, Agility (GXO partnership), 1X, Boston Dynamics (BMW Spartanburg and Toyota Woodstock), Apptronik, Figure AI, Humanoid Inc. — are documented across contemporaneous public filings and reporting in the same window. The Standard Oil comparison figure (~88–90 percent of U.S. refining capacity by 1882) is from the standard antitrust history; see Wu, Curse of Bigness, and the Tarbell record beneath it.
-
1X Technologies and Bernt Børnich, public statements and product material on NEO (consumer humanoid), 2024–2026; Brett Adcock and Figure AI public roadmap statements on consumer pricing; Elon Musk public statements on Tesla Optimus as household device, 2021–2025. The household-companion framing has been the consistent consumer-facing pitch across the cohort.
-
Aaron Perzanowski and Jason Schultz, The End of Ownership: Personal Property in the Digital Economy (Cambridge, MA: MIT Press, 2016). The Amazon-deletes-1984 case is at chapter 5; the iTunes / Kindle / mobile-app pattern is developed across chapters 3–6; the legal-architecture argument for treating digital products as licensed services rather than owned goods is the through-line.
-
Reporting on Adobe's May 2013 transition from Creative Suite perpetual licensing to the Creative Cloud subscription model. The user backlash, including a Change.org petition that gathered over thirty thousand signatures within weeks, and the structural consequence — perpetual-license customers paying again, repeatedly, for the same software — are documented across Ars Technica, The Verge, and contemporaneous trade-press coverage.
-
Cory Doctorow, The Internet Con: How to Seize the Means of Computation (London: Verso, 2023); and the ongoing "enshittification" series at Pluralistic, 2022–present. The four-stage arc — generous to users, then generous to business customers, then degrading to both, terminating in shareholder-only extraction — is laid out across multiple essays and synthesized in the book.
-
Federal Trade Commission, complaint and consent order, In the Matter of Deere & Company, January 2025; subsequent settlement and right-to-repair commitments. iFixit's multi-year coverage of John Deere's Service ADVISOR software, diagnostic-code lockouts, and aftermarket-parts firmware bricking is the technical record beneath the FTC's legal record. The U.S. PIRG "Deere in the Headlights" series (2021–2024) documents the state-level legislative parallel.
-
Reporting on Tesla's post-sale revocation practices — Autopilot and Full Self-Driving capability bound to the original purchasing vehicle and disabled on resale; the 2020 Model S supercharging-deactivation case; the 2023 Acceleration Boost per-vehicle re-purchase requirement; the documented post-sale range-extension revocations — covered across Jalopnik, Electrek, Ars Technica, and consumer-litigation filings, 2020–2024.
-
Electronic Frontier Foundation, "HP's Self-Destructing Printers and a Tale of Modern Hardware Failure," and subsequent EFF coverage of HP firmware updates 2016–2023; iFixit reporting on HP "Dynamic Security" and the DRM-locked ink cartridges; class-action litigation against HP in multiple U.S. and European jurisdictions over firmware-based cartridge bricking.
-
Shoshana Zuboff, The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power (New York: PublicAffairs, 2019). The business-model diagnosis is developed across Parts One and Two; the conversion of "behavioral surplus" into rating, scoring, and predictive-product apparatuses is the through-line of Parts Three and Four.
-
Samantha Hoffman, Engineering Global Consent: The Chinese Communist Party's Data-Driven Power Expansion (Australian Strategic Policy Institute, October 2019); Jeremy Daum, "Giving Credit Where Credit's Due," China Law Translate, 2017–2020; Rogier Creemers, "China's Social Credit System: An Evolving Practice of Control," SSRN, May 2018. These scholars are careful to distinguish the fragmented operational reality of Chinese social credit from the simplified Western narrative; we cite them precisely because they refuse the moral-panic framing and document what the technical apparatus actually does at scale.
-
David Autor, "Why Are There Still So Many Jobs?" (cited above as 69), is the canonical labor-economist version of the adaptation argument. Steven Pinker, Enlightenment Now: The Case for Reason, Science, Humanism, and Progress (New York: Viking, 2018), is the macro-historical version, anchored in the long-run quantitative-improvement data across mortality, literacy, violence, and political enfranchisement. Pinker is the careful version of the argument that the optimist's loud descendants — Andreessen's Manifesto above all — borrow without the underlying historical apparatus.
-
E. P. Thompson, The Making of the English Working Class (London: Victor Gollancz, 1963; Penguin reissue 1968). The Luddite chapters are in Part Three. The York hangings (January 1813), the deployment of twelve thousand troops to the Luddite counties, the 1812 Frame-Breaking Act making machine-breaking a capital offense, and the subsequent depopulation of the framework-knitting towns are documented from the period's parliamentary record and surviving correspondence.
-
William E. Leuchtenburg, Franklin D. Roosevelt and the New Deal, 1932–1940 (New York: Harper & Row, 1963), is the standard one-volume treatment. Arthur M. Schlesinger Jr.'s three-volume The Age of Roosevelt (1957–1960) is the longer canonical synthesis. The McCormack-Dickstein hearings — formally the Special Committee on Un-American Activities, Seventy-Third Congress — investigated allegations of an industrial-finance plot against the Roosevelt administration in 1934; the committee's published report and Smedley Butler's congressional testimony are the primary documents. The Bonus Army episode (1932) is documented in Paul Dickson and Thomas B. Allen, The Bonus Army: An American Epic (New York: Walker, 2004).
-
Karl Polanyi, The Great Transformation: The Political and Economic Origins of Our Time (New York: Farrar & Rinehart, 1944). The "fictitious commodities" of labor, land, and money are introduced in Part Two; the double-movement thesis is developed across Parts Two and Three; the analysis of the interwar collapse and its alternative branches (the New Deal, the European welfare states, and the fascist counter-movements) constitutes Part Three.
-
Barrington Moore Jr., Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of the Modern World (Boston: Beacon Press, 1966). The comparative analysis of agrarian-to-modern transitions across England, France, the United States, China, Japan, India, and Germany is the book's organizing structure; the "no bourgeois, no democracy" thesis and the path-dependence of political form on prior class structure are the analytical through-line.
-
Peter Turchin, Ages of Discord: A Structural-Demographic Analysis of American History (Chaplin, CT: Beresta Books, 2016); Turchin, End Times: Elites, Counter-Elites, and the Path of Political Disintegration (New York: Penguin Press, 2023). The four-indicator framework — elite overproduction, declining mass living standards, rising state fiscal stress, declining institutional trust — and the United States' entry into the instability window around 2010 are developed across both volumes.
-
Hannah Arendt, The Origins of Totalitarianism (New York: Harcourt, Brace, 1951). Part Three, "Totalitarianism," develops the diagnosis of the mass as the social precondition for totalitarian movements — the population rendered politically and economically homeless by the dissolution of prior class structures and previously available institutional memberships.
-
David H. Autor, David Dorn, and Gordon H. Hanson, "The China Syndrome: Local Labor Market Effects of Import Competition in the United States," American Economic Review 103, no. 6 (October 2013): 2121–2168. Autor, Dorn, Hanson, and Kaveh Majlesi, "Importing Political Polarization? The Electoral Consequences of Rising Trade Exposure," American Economic Review 110, no. 10 (October 2020): 3139–3183. Autor, Dorn, and Hanson, "On the Persistence of the China Shock," Brookings Papers on Economic Activity (Fall 2021): 381–447. The 2020 paper is the source for the populist-electoral mapping; the 2021 update establishes that the labor-market and political effects did not mean-revert over the subsequent decade.
-
Dani Rodrik, "Populism and the Economics of Globalization," Journal of International Business Policy 1, no. 1–2 (June 2018): 12–33; Barry Eichengreen, The Populist Temptation: Economic Grievance and Political Reaction in the Modern Era (New York: Oxford University Press, 2018). Rodrik's left/right populism distinction and the structural-grievance framework, together with Eichengreen's comparative-historical work on the interwar period and the present, supply the cross-national evidence beneath the §5 receipt.
-
Mike Davis, City of Quartz: Excavating the Future in Los Angeles (London: Verso, 1990); Davis, Planet of Slums (London: Verso, 2006). The "Fortress L.A." chapter is at City of Quartz, Part Four; the hostile-architecture catalog and the privatization of public space are developed throughout.
-
Teresa P. R. Caldeira, City of Walls: Crime, Segregation, and Citizenship in São Paulo (Berkeley: University of California Press, 2000). The ethnography of the gated condominium and the daily geography of avoidance is developed across Parts Two and Three.
-
Frank Morn, "The Eye That Never Sleeps": A History of the Pinkerton National Detective Agency (Bloomington: Indiana University Press, 1982). The "larger than the U.S. Army" comparison is the standard one for the period roughly 1875–1893 and is documented across the labor-history literature; see also Robert Michael Smith, From Blackjacks to Briefcases: A History of Commercialized Strikebreaking and Unionbusting in the United States (Athens: Ohio University Press, 2003). Jeremy Scahill, Blackwater: The Rise of the World's Most Powerful Mercenary Army (New York: Nation Books, 2007), supplies the contemporary case; the Nisour Square killings of September 16, 2007, and the subsequent legal-accountability record are documented across federal court filings and Scahill's reporting.
-
Antony Loewenstein, The Palestine Laboratory: How Israel Exports the Technology of Occupation Around the World (London: Verso, 2023). The export-pipeline argument is developed across Parts Two and Three.
-
Stephen Graham, Cities Under Siege: The New Military Urbanism (London: Verso, 2010). The "new military urbanism" framework and the Baghdad-to-Belfast-to-Detroit migration are developed across Parts Two and Three.
-
Sword International, public marketing material and the "SPUR" (Special Purpose Unmanned Rifle) demonstration video, 2021–2024, mounted on the Ghost Robotics Vision 60 quadrupedal platform. U.S. Air Force public reporting on Ghost Robotics deployment at Tyndall AFB and Cape Canaveral, 2020–present. The October 2022 industry pledge not to weaponize general-purpose robots was signed by Boston Dynamics, Agility Robotics, ANYbotics, Clearpath Robotics, Open Robotics, and Unitree; Ghost Robotics declined to sign. NYPD Boston Dynamics Spot ("Digidog") deployment, 2020 reactivation 2023; U.S. Customs and Border Protection Spot pilots at the southern border, 2022 reporting.
-
Yuval Abraham, "'Lavender': The AI machine directing Israel's bombing spree in Gaza," +972 Magazine and Local Call, April 3, 2024. Abraham, "'A mass assassination factory': Inside Israel's calculated bombing of Gaza," +972 Magazine and Local Call, November 30, 2023. The reporting is based on interviews with serving and recently serving IDF intelligence officers and is the most-substantiated open-source account of AI-driven population-scale targeting in active military operation.
-
American Civil Liberties Union, "The Dawn of Robot Surveillance: AI, Video Analytics, and Privacy" (June 2019); subsequent ACLU reporting on police robot procurement. Electronic Frontier Foundation, "Atlas of Surveillance" project, ongoing; EFF reporting on Section 1033 program transfers of military-grade equipment to local police; EFF reporting on police drone procurement and oversight gaps, 2018–present.
-
San Francisco Board of Supervisors, vote on Police Department use-of-force policy regarding remote-controlled robots, November 29, 2022 (8–3 vote permitting lethal force under "extreme circumstances"); reversal of December 6, 2022. The proceedings are documented in the Board's public minutes and were covered contemporaneously across the San Francisco Chronicle, Mission Local, the Associated Press, and the New York Times.
-
Marc Andreessen, "The Techno-Optimist Manifesto" (cited above as 70). Sam Altman, public statements on AGI timelines and inevitability across the OpenAI blog and conference remarks, 2023–2026. Demis Hassabis, public statements on AGI development and the responsible-deployment thesis through DeepMind communications and Royal Society lectures, 2023–2025. The "AGI is coming, the question is only who builds it" framing has become the dominant rhetorical posture across the frontier-AI lab leadership and is the structural ancestor of the inevitability claim we address in this section.
-
Regulation (EU) 2024/1689, the "AI Act," entered into force August 1, 2024. "The Bletchley Declaration by Countries Attending the AI Safety Summit, 1–2 November 2023," signed by twenty-eight states and the European Union. The Montreal Protocol on Substances that Deplete the Ozone Layer, signed September 16, 1987; entered into force January 1, 1989; subsequent amendments (London 1990, Copenhagen 1992, Montreal 1997, Beijing 1999, Kigali 2016). The Montreal Protocol is the closest historical analogue for a coordinated international restriction on a specific dual-use technology, enforced through trade mechanisms and observable industrial measurement.
-
Anton Korinek, "Taxing AI," NBER Working Paper 30761, December 2022. See also Anton Korinek and Joseph E. Stiglitz, "Artificial Intelligence, Globalization, and Strategies for Economic Development," NBER Working Paper 28453, February 2021, for the broader development-economics frame.
-
State of Alaska, Permanent Fund Corporation, public records, 1976–present. The fund was established by amendment to the Alaska state constitution in 1976; first dividend distributed 1982. The permanent corpus exceeded eighty billion U.S. dollars as of the most recent annual report; annual per-resident dividends have averaged in the one-to-two-thousand-dollar range across the program's history, with year-to-year variation tied to the fund's earnings and to statutory formula. The fund's operational history is the empirical proof of concept cited in §7 Proposal 3 and again in the §7 receipt.
-
Norges Bank Investment Management (NBIM), Government Pension Fund Global, public records and annual reports, 1990–present. Corpus exceeding 1.5 trillion U.S. dollars as of the 2024–2025 annual reports; the world's largest sovereign wealth fund. Capitalized from Norwegian state petroleum revenues since 1996.
-
Lina Khan, "Amazon's Antitrust Paradox," Yale Law Journal 126, no. 3 (January 2017): 710–805. The note's reconstruction of structural antitrust enforcement separated from the post-1970s consumer-welfare standard is the central doctrinal anchor cited in §7 Proposal 4. Khan's tenure as Chair of the Federal Trade Commission (2021–2025) supplied the enforcement-template proof.
-
Regulation (EU) 2022/1925, the "Digital Markets Act," in force May 2, 2023. The interoperability mandates, data-portability requirements, and structural-separation provisions for designated "gatekeeper" platforms are the international template for the §7 Proposal 5 American statute.
-
Mariana Mazzucato, The Entrepreneurial State: Debunking Public vs. Private Sector Myths (London: Anthem Press, 2013); Mazzucato, Mission Economy: A Moonshot Guide to Changing Capitalism (London: Allen Lane, 2021). The directional-state argument and the case for institutional capacity as the precondition for technology direction are developed across both volumes.
First published 2026-05-26 by Cancel Clankers. © 2026 Cancel Clankers. All rights reserved.
— Stay Human ★