Chapter 13 of 13

Beyond the Farm

About 25 minutes to read

Toad Lane

On the evening of 21 December 1844, the longest night of the year, a small group of men opened the shutters of a ground-floor storeroom at 31 Toad Lane in Rochdale, a mill town in the north of England. Most of them were weavers; some were shoemakers, joiners and other tradesmen. There were twenty-eight of them, by the count the movement later settled on, and they had spent most of the year saving a few pence a week each until they had enough to rent the room and buy their first stock. It was not much: some flour, butter, sugar and oatmeal, and, in some accounts, candles. The gas company had refused to supply them, doubting that they could pay the bills, so the shop was lit by candlelight. According to the movement's first historian, George Jacob Holyoake, a few onlookers from the street came to laugh at the weavers' venture.

The shop opened two evenings a week. Its rules were simple. It sold pure food at fair prices, and for cash, so that its members would not fall into debt. Each member had one vote, however much capital he or she had put in. Capital received a fixed, limited return. The surplus that remained after costs was returned to members in proportion to what they had bought, not to what they owned. Part of the surplus was set aside for education. Membership was open to anyone who would keep the rules.

Chapter 1 of this book began twenty years earlier, with Robert Owen addressing the President and Congress of the United States and buying a whole town in Indiana in which to build a new moral world. New Harmony collapsed within two years. The weavers of Toad Lane had read Owen, and some of them shared his dream of cooperative communities. But they started with a shop. It sold candles and oatmeal, and it lasted. Its rules became the model for a movement that, according to the International Cooperative Alliance, now has on the order of a billion members.

This book set out to stand between those two outcomes: to design, in good faith, a society beyond the farm, and then to test it as history tested New Harmony and Toad Lane. Its rule was that no part of the design would survive unless history or evidence allowed it. This last chapter gathers the verdicts. It sets out the design as it stands after twelve chapters of attack, the three futures the coming decades could bring, the signs that will tell them apart, and what the 99 can do — and the 1 might choose — to make the next entry in the ledger. Then it hands the ledger to the reader.

The forces, reassessed

The first book found that the ratio between the 0.01, the 1 and the 99 shifted only when several of five forces lined up. Part II of this book tested each against the next fifty years.

Catastrophe (Chapter 2) will come — climate change makes that certain — but it cannot be part of any plan. The shocks of 2008 and 2020 entrenched the owners rather than levelling them, because modern states now protect concentrated wealth from destruction. Only one lesson of catastrophe is usable: crises reward whoever has their ideas ready, so the design must exist before the crisis does.

Scarcity (Chapter 3) is the strongest force in the 99's favour since the Black Death, if the book's assumptions hold. Ageing populations make workers scarce, and AI, at a wall, cannot replace the care, judgement, physical work, trust and above all creativity that people supply. But scarcity alone delivers nothing, as Japan shows; owners escape it by importing less-protected workers, moving work abroad, or automating what they can.

Organisation (Chapter 4) survives in new forms. Scattered workers — content moderators in Nairobi, drivers in London and Amsterdam, writers in Hollywood — have organised and won, especially where they found a counterpart with the power to agree. Across borders the record is thin.

Divided owners (Chapter 5) are visible everywhere: nation against world, AI companies against the owners of creative rights, giants against rivals, states against the owners, even the rich who ask to be taxed. Every gain in Chapter 4 came with help from somewhere above.

Lasting institutions (Chapter 6) are possible when gains are universal, automatic, structural, closed to exit and defended by institutions of their own, and when they protect the capital as well as the payment. The end of bank secrecy is the proof. Alaska's dividend is the warning.

And beneath them, the story (Chapter 7). Meritocracy is losing believers; its successors are being written now. Under the book's assumptions, the next story is the one job that stays wholly human, which makes it the 99's to write. Part III added the newest danger, the owners' exit (Chapter 8): private cities, purchased passports, tax flight and the dream of a worker-less machine. Under the wall assumption that dream is partly an illusion, and that is the 99's lever.

The verdict on Part II is that the forces are more available than at any time since the mid-twentieth century, but none is enough alone, and the most powerful of them — scarcity — depends on assumptions this book cannot prove.

What the tests taught

The four tests did not weigh equally, and it is worth saying which did the most work.

Precedent kept the design modest. Again and again, the cases that worked were small, specific and patient: a fund protected by a fiscal rule, a dividend paid equally to every resident, a cooperative with a pay cap, a works council with a veto, a citizens' assembly with a deadline. The grand designs — New Harmony, the Soviet planners, the Meidner plan in its full form — either failed or were never tried. The design that survives is built from the small things that lasted.

Evidence was strongest on the floor and weakest on everything new. On cash transfers, the floor and the effects of a dividend on work, there is now a substantial body of research. On common ownership of AI, on governing machines through councils chosen by lot, and on counting care and creation as contribution at the scale of a whole society, there is almost none, because none has been tried. Much of Part IV is therefore a design supported by analogy, not by proof, and the ledger says so.

Capture changed the design more than any other test. It turned "the machines publicly owned" into a bundle of rules about income, control and transfer; it turned "AI replacing the shepherds" into a governance split five ways; it turned the measurers of contribution into councils with low stakes and deadlines. Every time the design was asked how the farm would take it over, it had to grow another defence. That is the first book's central finding, confirmed: the farm reassembles itself around every reform, and a design that does not plan for its own capture will be captured.

Exit changed the shape of the whole. It was the test of exit — Gibbon's prison, Kant's warning, the right to leave — that turned one borderless society into a federation of many, and it was the distinction between the exit of persons and the exit of the commons that told the design which doors to close and which to open. A society that people cannot leave is a farm, however well it is run. A society that wealth can leave is a farm too, for those it leaves behind.

The final ledger

Part IV built the design and attacked it, element by element. Here is the ledger of the whole design as it stands after the tests of Chapters 2 to 12.

What survived

• A world with no outside for the commons. No haven for wealth, no sale of citizenship, no private jurisdictions, no stateless people (Chapters 6, 8 and 12).
• Common ownership of the machines and models — but only in a precise form: common trusts whose capital is locked and cannot be sold, holding above all the physical layer of data centres, chips and power, with open models as their starting stock, and with every person's share equal, direct and non-transferable (Chapters 5, 8 and 9).
• AI doing the work of the shepherds: keeping the books, allocating, scheduling and checking, which removes the main route by which managing classes have become owning classes (Chapter 10).
• Every person a worker and a shareholder. Human work remains necessary under the wall assumption, and every person owns an equal share of the commons (Chapters 3, 4 and 9).
• Reward weighted by contribution, defined broadly and justified as reciprocity rather than desert (Chapters 7 and 11).
• A floor nobody falls below, which the evidence from Alaska, Finland, Kenya, the United States and the pandemic suggests can be paid without people ceasing to contribute (Chapters 2, 5 and 11).

What had to change

• One borderless society became a federation of many self-governing societies sharing one floor and common rules for the owners, because a single world state would be Gibbon's "safe and dreary prison" and the largest prize for capture in history (Chapter 12). The doors must be closed to the commons and open to people. Belonging comes in two layers: global by birth, with a global dividend for life wherever a person lives; local by residence, after a short period, so that no one is a permanent outsider within the society they live in.
• "Public" cannot mean "owned by the state." The Soviet new class, Russia's vouchers and the state AI of China and Russia all show what happens when it does (Chapters 5 and 9). Ownership must be kept broad strand by strand: income equal and direct, capital locked, shares unsaleable, control divided and accountable, the models open to inspection. And the commons must be built by people already free to defend it, as Norway's was and the petro-states' was not.
• The machine cannot be the governor, only its instrument (Chapters 4 and 10). The shepherd's office is split into five parts that cannot act alone: the machine, its managers, rule-setters chosen by lot, independent auditors, and tribunes with a veto. Every significant decision about a person gets a real human review, with the burden of proof on the system — the presumption that convicted the subpostmasters reversed. Those whom the machine manages have an elected collective voice with a veto over it. The machines' power to persuade is governed by citizens. Every check has a deadline.
• Independent organisation is permanent (Chapter 4). A society in which everyone is a shareholder still needs unions and associations of its own, or it is a farm with a kinder shepherd.
• Contribution is counted mainly in time, an hour of care equal to an hour of production, with modest premiums, measured coarsely, governed by councils chosen by lot, open to veto by those measured, and kept low in its stakes: it affects only income above the floor and the dividend (Chapter 11).
• The gap is capped within a band of a few times, with the managers of the commons at the bottom of it (Chapters 9 and 11).
• Creators are honoured but cannot become an aristocracy. Their reward is income that cannot be inherited, cannot buy a larger share of the commons and cannot fence off the common culture, with short, unsaleable rights held by creators (Chapters 3 and 11).
• The floor is a right of belonging, not a reward or a feed (Chapters 5, 6, 7 and 11). It is universal, automatic and permanent, made of an equal dividend from capital the 99 own plus universal services, with a payout rule no single veto or budget can cut. The 99's claim on it must never depend on being needed or on agreeing (Chapters 8 and 12).
• The design must be able to change (Chapters 6 and 11). Its core rules are revised slowly, at fixed intervals, by a citizens' assembly and a referendum, never by a faction alone.

What remains unknown

• Whether the wall holds, and whether agents remain driven by people. Everything in Part II and much of Part IV rests on it.
• Whether common trusts can build and run the most advanced machines as well as private companies can.
• Whether citizens can understand the machines well enough to govern them, or whether the interpreters who can read them become the next priesthood.
• Whether rotating councils can hold their own against the permanent professionals who run the machines.
• Whether time can measure creation fairly, and whether prestige and networks will reproduce a celebrated class that no income cap can touch.
• What a permanent, universal floor would do over a whole lifetime and a whole society.
• Whether the most powerful states would ever join a federation that bound them, and whether free movement between members as unequal as today's countries could be phased in without a backlash that destroys it.
• Whether workers can organise across borders on anything like the scale at which capital operates.
• Whether climate change and war between great powers leave room for any design to be built, or whether, as Chapter 2 feared, they produce a world of fortified enclaves first.
• How any of it could be built from the world as it is.

If the assumptions are wrong

Chapter 1 promised to say what would change if the book's third or fourth assumption proved false. The design as it now stands gives a clearer answer than was possible then.

If AI has no wall, the scarcity of human labour — the force on which so much of Part II relied — disappears, and with it the 99's oldest bargaining chip. The contribution-weighted half of the design would shrink. But the half that matters most would survive intact, because the tests made it independent of being needed. Common ownership of the machines, an equal and unsaleable share for every person, and a floor given as a right of belonging were built precisely so that, as Chapter 8 required, the claim of the 99 would not depend on the owners needing them. That was the insurance policy. In a world without a wall it would become the whole of the design.

If the wall exists but is not creativity, the map of scarce work changes, and the danger of an aristocracy of creators eases. The design's rules for reward would apply instead to whatever proved scarce — care, judgement, trust — and the same limits on inheritance, accumulation and enclosure would hold.

The three refinements of the wall that Chapter 1 added make the picture sharper, and each points to something the 99 can hold on to. If agents remain driven by people, as the book now assumes, then the work that stays human is not only care and creation but direction and responsibility: setting the goals, judging the results and answering for them. That work can be spread widely or concentrated in a small class of people who direct swarms of machines, and which happens is a political choice, not a technical one. If AI stays heavy, its owners stay tied to places, grids, water and permits, and the communities that host the machines hold a lever that no owner of oil wells or offshore accounts had to reckon with. And if open models stay close to the frontier, the commons can begin without expropriating anyone. Each refinement could fail. Agents might learn to set and judge their own work; AI might become light enough to run anywhere; open models might fall behind or be banned. If all three failed together with the wall, the design would rest entirely on its insurance policy: common ownership of whatever machines exist.

Objections that stand

A book that tests a design in good faith should end by naming the objections it could not fully answer.

The first is knowledge. In 1945 the economist Friedrich Hayek argued, in "The Use of Knowledge in Society", that the information needed to run an economy is scattered among millions of people in fragments no central planner could ever gather, and that prices are the means by which those fragments are coordinated. The Soviet planners proved him largely right. The design does not abolish markets: people would still buy and sell, firms and cooperatives would still compete, and prices would still carry information. What it holds in common is the capital of the machines and the commons of knowledge, not every exchange. But a society in which the main machines are owned in common trusts and much allocation is done by AI under public rules is closer to the planners than its defenders might like, and whether machines can gather the scattered knowledge Hayek described any better than the planners could is unknown.

The second is incentive. Compressed rewards, a generous floor and short rights over creations might slow the innovation on which everyone's prosperity depends. The evidence of the post-war decades, the Nordic countries and Mondragon suggests that economies can thrive with far narrower gaps than today's, and Chapter 11 found that security encourages enterprise more often than it smothers it. But none of those cases tested the design at the frontier of a technology as fast-moving as AI, and the doubt is legitimate.

The third is time. A society governed through councils chosen by lot, works councils, assemblies, audits and referendums asks a great deal of its members' evenings — a complaint about socialism often attributed to Oscar Wilde. The design answers that a society in which machines keep the books should have more time, not less, and that most people would sit on a council only a few times in a lifetime. Whether people would want to govern as much as the design requires is a question only practice could answer.

The fourth is feasibility: that none of this will happen, because the owners are too strong and the 99 too divided. That is the objection the first book's history supports most strongly. It is also the one that history has most often proved wrong, slowly and then suddenly, when the forces lined up.

Three futures

The five forces, the owners' exit and the state of the machines make three broad futures possible. None is fixed, and the coming decades are likely to contain elements of all three, in different places. But they are different enough that the reader should be able to tell which is winning.

The Managed Herd 2.0. The owners keep the machines, the models and the data. AI raises output, and part of the surplus is paid out as a feed: a basic income from the owners or from a state that depends on them, cheaper services, more entertainment, perhaps a dividend paid at someone's discretion. Work becomes less secure and more closely watched, managed by algorithms whose human review is symbolic. The story is one of benevolent abundance. Nobody starves. Nobody owns anything that matters. It is Rome's grain dole and circuses on a planetary scale, and it is the future the first book would predict, because it is what the farm has always done when it could afford to: feed the herd rather than free it. It is the most likely of the three if the owners stay united and the 99 stay scattered.

The Enclosure of Intelligence. The wall turns out to be lower than this book assumes, or the owners act as if it were. A small number of companies and states own the systems that do most of the work, and they no longer need the many as workers, soldiers or taxpayers. The rentier state of Chapter 8 becomes the model for the world: rulers who live on the machines rather than on their people, and can afford to ignore them. The owners leave — into private jurisdictions, protected compounds, perhaps one day into space — and take the machines with them. What the many receive depends entirely on the goodwill of people who do not need them. This is the darkest of the three futures, and the "intelligence curse" names it.

The Federation. The machines and models are held in common trusts, governed by the split institutions of Chapter 10 and owned in equal, unsaleable shares by every person. AI keeps the books and does much of the routine work; people do the work only people can do, and are rewarded for it within a narrow band, above a floor that is theirs by right. Many societies, different from one another, share the floor and close the owners' doors together. It is not a utopia. Its councils will quarrel, its auditors will miss things, its federations will strain, and every one of its institutions will be under constant pressure from people who would like to turn it into a farm again. But it is a society in which no small group owns the surplus that everyone else produces. Of the three, it is the only one that would be beyond the farm.

All three can already be seen in the world as it is. The Managed Herd is visible in the gig economy's algorithmic managers and symbolic reviews, in support that arrives as discretionary emergency payments rather than as rights, and in the owners' own proposals, described in Chapter 5, for dividends paid from funds they would design. The Enclosure is visible in the concentration of the computing power needed to build the most capable AI in a handful of companies, in the private cities and purchased passports of Chapter 8, and in the 2025 order to mine the seabed outside the international authority. And the Federation is visible too, in fragments: in Norway's fund and Alaska's dividend, in Mondragon's cooperatives and Germany's works councils, in Switzerland's public model and Ireland's citizens' assemblies, in the end of bank secrecy and the common heritage of mankind. None of the fragments is the design. Each shows that a piece of it can exist.

The signs

Which future is arriving will not be announced. It will show up in signs that are already measurable:

• Who owns the frontier. Whether the most capable AI models, and the computing power and data needed to build them, are concentrating in a few private hands, or whether public, commonly owned and open-weight systems are keeping close enough to matter.
• Who owns the physical layer. Whether the data centres, chips and power on which AI runs are owned only by a few companies, or whether host communities, municipalities, cooperatives and public trusts take shares in them.
• Who drives the agents. Whether directing AI agents becomes the skill of a small new class of shepherds, or a capacity that ordinary workers, with open tools, share.
• What happens to the wall. Whether the pay and bargaining power of care workers, skilled tradespeople and creators rise as their work becomes scarcer, or fall as AI and imported labour are used to get around them.
• Whether the doors are closing. Whether automatic exchange of financial information, registers of ownership and minimum taxes are extended to cover the very rich and the largest companies, or exemptions multiply, as with the G7's in 2025; whether citizenship sales and private zones spread or are shut down; whether the seabed and other global commons are shared or taken.
• Whether review is real. Whether courts and legislatures go on finding that human review of machine decisions must be real, as the Amsterdam court did, and whether workers win collective rights over the machines that manage them, as German works councils have.
• What kind of floor is built. Whether new forms of support are universal, automatic and paid from capital the public owns, or targeted, conditional and paid at someone's discretion.
• Whether the scattered organise. Whether unions and associations take root among gig workers, care workers, content moderators and creators, and whether any of them learn to act across borders.

A world in which the frontier concentrates, wages at the wall fall, the doors widen, review stays symbolic, floors stay discretionary and the scattered stay scattered is moving towards the Enclosure. A world in which those signs point the other way is moving, slowly, towards something like the Federation. Most of the world, for most of the coming decades, will be in between, in the Managed Herd — which is why the signs matter.

How it could begin

The hardest commitment this book owes is to say how the design could be built from the world as it is, without the defeat that met the Meidner plan. Chapter 9 recorded why that plan failed: the owners united against it while its supporters divided; its opponents could portray it as a transfer of power to union officials rather than to workers; it was confined to one country while capital was learning to cross borders; and as a plan of gradual change spread over decades, it gave its opponents a single target and years to organise. Each of those reasons suggests a rule for beginning.

Start with the new, not the old. The Meidner funds were to take a share of existing companies' profits, and the owners fought them as an expropriation. Much of the wealth the design is concerned with does not yet exist: the income of AI systems that are still being built, the value of models trained on the common record of humanity, the resources of the seabed. A claim on wealth that has not yet been privately captured is easier to win than a claim on wealth that has, and the public investments in computing power and models that governments are already making, for reasons of their own, are the beginning of such a claim. Public AI built for national sovereignty, Chapter 5 warned, is not yet the 99's. But a public model, built openly, can be placed in a trust rather than a ministry.

Claim a share of the physical layer. The data centres now being built need land, water, power lines and permits from the communities around them, and those communities are already using that leverage, mostly to say no. They could instead say yes on terms: a share of every large data centre for its host community and the common trust, as Danish law since 2009 has required developers of new onshore wind farms to offer at least a fifth of their shares to people living nearby. Wealth that is being built now, in particular places, with public grids and public permission, is the easiest place to begin.

Give every person a stake from the first day. The Meidner plan's beneficiaries were abstract; the Alaska dividend's are every resident, and it became one of the most popular policies in America. Every step of the transition should pay something, directly and equally, to every person — however small at first — so that the people who benefit are the people who vote. The defenders of a lock must exist before the attack on it comes.

Lock each step as it is taken. Gradualism failed in Sweden because each step depended on the next decision. Every step of the transition should be complete in itself, structural rather than regulatory, and protected as Norway's fund and the end of bank secrecy are: built into institutions that run automatically and have their own defenders. A step that can be reversed by the next government is not a step.

Use the splits; do not create a united front. The owners are divided — between the AI companies and the owners of creative rights, between the giants and their rivals, between states and the companies they fear. The Meidner plan faced owners who had every reason to stand together. A transition that begins with public models that rivals can build on, common registers that honest businesses welcome, and minimum taxes that governments need, offers some of the powerful a reason to stand on the other side.

Cross borders from the start. The end of bank secrecy succeeded because it was coordinated and enforced by access to the largest markets; the Meidner plan failed partly because it was Swedish. Every element of the design that concerns the owners' doors must be coordinated among as many states as possible from the beginning, with access to their markets as the leverage.

Have the ideas ready. Milton Friedman observed, as Chapter 7 recorded, that when a crisis comes, the actions taken depend on the ideas that are lying around. The next crisis — a financial crash, a pandemic, a climate disaster, the sudden displacement of millions of office workers by AI — will come. The design should be lying around when it does, in a form specific enough to be enacted in a week.

And start with a shop. The weavers of Toad Lane did not wait for the world to be ready for a new moral world. They started with what they could own and govern themselves, and made it last. Cooperatives, mutual funds that their members control, public models in trusts, citizens' assemblies, works councils with a veto over the machines, local dividends from local commons: each of them is a Toad Lane, and each is a place to learn how the larger design would work before anyone has to bet everything on it.

What the 99 can do, and what the 1 might choose

The first book ended by saying that the next entry would depend on what the 99 did, and on what the 1, who keep the books, decided to write. This book can be more specific.

The 99 can organise, in the new forms Chapter 4 described, and above all among those whose work stands behind the wall: carers, makers, creators and the people who train and check the machines. They can refuse symbolic review, insisting in courts, workplaces and legislatures that every machine that decides about them be open and contestable. They can demand universal rather than targeted support, because universal floors are defended by everyone. They can own something now — in cooperatives, in pension funds whose votes they control, in local commons — and learn to govern it. And they can write the story: under this book's assumptions, the one job the machine cannot do for them.

The 1 — the managers, professionals and technicians who keep the books for the owners, including the people who build AI — have a choice that their predecessors in the first book sometimes faced. They can serve as the shepherds of the Managed Herd or the engineers of the Enclosure, or they can build the tools of the Federation: open models, honest audits, real review, systems their users can govern. Chapter 5 suggested that the shepherds displaced by AI might become allies of the 99 or leaders of a movement to restore their own status. Which they become is not yet decided.

And the 0.01 have a choice too. Douglas Rushkoff's advice to the billionaires in Chapter 8 — that the best way to keep their guards loyal after the Event was to treat them well now, and to invest in a world in which the Event was less likely — was not naive. Under this book's assumptions, the owners' exit is partly an illusion: the machines will still need people, and the owners will still need the many, as every withdrawn elite in history did. Some of the very rich already ask to be taxed. The first book found rulers who cancelled debts, landlords who accepted reform rather than revolution, and a House of Lords that gave way. It also found that such choices were almost always made under pressure. The owners will choose well, if they do, when the 99 are organised enough that choosing badly has become the greater risk.

The next entry

On a clay tablet from Uruk, written more than five thousand years ago, a temple accountant recorded 29,086 measures of barley over thirty-seven months, and signed it, or was named on it, as Kushim. The first book began with that tablet because it is one of the earliest records of a name in human writing, and it records not a king or a poet but a keeper of accounts. For five thousand years the keepers of the ledger have recorded what the many owed the few.

This book has tried to draft a different entry. It might read something like this:

Received: the work of every person — the making, the caring, the learning, the creating, the risks taken and the commons kept. Held in common: the machines, the models, the knowledge of the species and the resources of the earth, unsaleable and open to inspection. Paid to every person, by right of belonging, wherever they live: an equal share, and a floor that no one falls below. Paid for contribution, within a narrow band: the rest. Kept by: the machine, under the eye of everyone it counts.

It is only a draft. Every line of it has been tested in these pages, and every line carries a verdict and a list of things unknown. Some of it will prove wrong. The rule of the book applies to its ending too: what history and evidence do not allow, the reader should let go.

One more disclosure belongs here. This book, as Chapter 1 said, was written by an AI. Its subject was commissioned by a person, who also set its assumptions; under the fourth of them, the frame of this book — its question and its conviction that there is something beyond the farm — came from a human being, and the machine's work was to build, test and keep the accounts. If that division of labour is a fair picture of what is coming, it is a hopeful one. It means the machines can keep the books. It does not mean they should decide what is written in them.

The weavers of Toad Lane opened their shop on the longest night of the year, by candlelight, because no one would sell them gas. They had saved for a year to buy a few sacks of flour, and the people in the street laughed at them. The ledger they started is still open. So is Kushim's.

The next entry has been drafted. Who will write it?