The Machine Interregnum

The old compact is falling away. The successor has not arrived.

This is an interregnum. Into the gap rush wolves, empires, agents, and markets. Each offers a constitution. None yet commands assent. Let us start with wolves.

Jonathan Bi 毕豪 is not merely defending techbros. Jonathan Bi has given the wolf the most elegant constitution so far. That is why he matters. The argument is not settled. It has only begun to name its combatants.

---

I. The Wolf’s Theology

Bi’s argument is not subtle in its provocation. America once attracted people because it was equal. Tocqueville saw a society where social distance was small, aristocracy absent, and ordinary citizens carried themselves with democratic confidence. Bi says he came to America for the opposite reason: not equality, but “unequal greatness”; not the comfort of Canada, but the vertical distance of America. Most Americans, he admits, might be happier in Canada. But America has the fire. It burns many. It makes a few glow.

This is why he is dangerous. Not because he is crude, but because he is not. He does not say inequality is good in the vulgar libertarian way. He says equality has costs. Equality can produce envy, mediocrity, conformity, materialism, and a war of all against all. A fair exam system may be fair, but it may also crush imagination. A society that gives everyone dignity may also make greatness feel presumptuous. A society that uplifts everyone may also produce citizens who are comfortable, decent, and small.

That is the knife. We should not pretend it does not cut.

For much of the twentieth century, the great developmental states had a clean answer to the question of what society owes its ordinary members. Singapore gave perhaps the cleanest version: uplift everyone. Housing, schools, healthcare, jobs, order, transport, savings, skills, social mobility. A working state. A believable floor.

This was not only moral. It was brutally practical. The developmental state needed citizens. The factory needed workers. The army needed soldiers. The welfare state needed taxpayers. Growth needed people. The moral claim and the growth claim reinforced each other. The rope held. It gave ordinary people a story in which their effort mattered, their children could rise, and their government had a reason to care about their fate. The old compact was not always just. But it needed you. And needing you gave it a reason to invest in your dignity.

The machine order weakens that comfort. Not because humans disappear. Not because ordinary people cease to matter. But because the economic case for full inclusion becomes less self-evident. AI, automation, compute, capital, private stacks, energy systems, frontier labs, and a thin layer of exceptional operators may carry more of the production function. Growth may require fewer clerks, fewer coders, fewer junior analysts, fewer ordinary professionals, fewer human beings in the loop. Then “uplift everyone” begins to sound quaint. Not wrong. Worse: no longer obviously necessary.

Bi looks at this and does not flinch. He baptizes the techno-aristocracy. Most people may be happier in Canada. America is for the few who can metabolise frenzy into greatness. Do not destroy the fire because many cannot bear it.

And he has another move. His strongest claim is not that industrial aristocracy produces wealth. It is that industrial aristocracy may now be one of the few remaining bulwarks against the sovereign. Tocqueville’s old American freedom was protected by religion, family, jury, township, associations, local press, lawyers, and entrepreneurial labour. These were exercises. They trained the muscle of freedom. Bi argues that many of these exercises have decayed. The forms remain—elections, rights, speech, courts—but the muscle has atrophied. In their place stand industrial fiefs: billionaires, tech firms, private labs, model companies, rocket companies, cloud providers, defence stacks, eccentric patrons. They are not democratic associations. They are aristocratic. But they may be powerful enough to resist the state.

So Bi’s answer is not to abolish the industrial aristocracy. That would require an enormous coercive force, perhaps worse than the thing being abolished. His answer is to tame it, educate it, moderate it, and harness it. The choice, he says, is not between egalitarian Eden and industrial aristocracy, but between a debauched irresponsible aristocracy and a moderated productive one. What is needed is not only an educator of the masses, but a “teacher of kings.”

This is not stupid. A private aristocracy can check the state. A billionaire can fund the dissident. A platform can resist a government. A lab can shelter thought too strange for universities. A company can build what public bureaucracy cannot. Power can check power, even when the forms are ugly.

---

II. The Test of Revocability

But a bulwark against the state can still be a machine over society. A castle can protect against the king while crushing the village. The fact that OpenAI, SpaceX, Palantir, Nvidia, and the defence-tech complex can resist Washington does not mean they protect the worker, the tenant, the student, the patient, the child, or the small country beneath them. They may check sovereign power from above while deepening revocability below.

This is the seam.

The old aristocrat had obligations because he was stuck. Land tied him to peasants. Blood tied him to estate. Memory tied him to place. The arrangement was often brutal, but the brutality had an address. The village knew whose door to stand before.

The new industrial aristocrat is not stuck. He has users, employees, contractors, customers, datasets, compute clusters, launch sites, shareholders, security clearances, offshore structures, foundations, and passports. He does not inherit land. He operates layers. He can leave.

That is what makes him secured. He is hard to revoke. He owns exits, shields, time, lawyers, institutional address, jurisdiction, trusted rooms, and the ability to move between games. The permissioned may have more access than any generation in history—AI tools, cheap content, cheap tutors, cheap entertainment, cheap companionship—but much of this is conditional. Platform access. Model access. Job access. School access. Housing access. Visa access. Credit access. Welfare access. Account access. The permissioned live inside abundance by permission.

The question, then, is not only about equality or greatness. It is about answerability. Who answers when AI deflates work? Who answers when private infrastructure becomes public necessity? Who answers when platforms eat childhood? Who answers when the model classifies, the bank scores, the employer automates, the school sorts, the app de-ranks, and the citizen cannot find a human address? Bi wants a teacher of kings. The rest of us need institutions that make kings answer.

And yet Bi troubles me because he names something I recognise. A society can uplift everyone and still produce a kind of spiritual dampness. Singapore did the old compact extraordinarily well: housing, schools, jobs, safety, healthcare, order. It made ordinary life nonfatal. But a life can be nonfatal and still feel thin. Good grades, good flats, good jobs, good behaviour, good systems, and a quiet suspicion that none of it adds up to a story worth telling. I have often wondered whether the people I know might be happier in a messier place, with less meritocratic sorting and more unplanned life. Bi gives that feeling a philosophy. He makes it sound like honesty, rather than exhaustion. That is why he must be taken seriously, and why simply dismissing him as an apologist for inequality would be a moral failure of its own.

---

III. The Other Claimants

Bi is not alone in the interregnum. Around him and the Wolf are four other claimants, less elegant perhaps, but just as real: Abundance, Demand, Empire, and Agentworld. I have to anonymise some names.

A1 breaks the scarcity ruler. The lower leg of K-shaped societies are not simply deprived, compared to the past. A child today inherits more knowledge than kings once possessed: Wikipedia, AI tutors, translation, open-source tools, cheap software, public networks, crypto rails, models, protocols, and eventually cheaper energy systems. There is abundance outside money. The question is not only who has scarce wealth. The question is whether abundance is sovereign or revocable.

This is where China enters. America produces outlier capability. China makes capability cheap. America makes monsters. China makes monsters cheap. That is not a joke. It is one of the most important structural features of the new world. America says: behold the frontier. China says: we will manufacture the frontier until it becomes ordinary. Solar panels, batteries, EVs, drones, grids, ports, manufacturing systems, open models, practical diffusion. China takes elite capability and drags it into price, scale, logistics, and ordinary use.

This is China’s blessing.

But cheap does not mean sovereign. Abundant does not mean free. Cheap solar that leaves no local grid capacity is dependency. Cheap AI that captures data is dependency. Cheap infrastructure that creates political leverage is dependency. Cheap platforms that own attention are dependency. Cheap tools that destroy local formation are dependency. So the new Chimerica is not the old one. Not China making goods while America consumes and Wall Street recycles the surplus. It is harsher. America produces aristocratic capability. China produces equalising abundance. The rest of the world tries not to become dependent on either.

A2 shrugs and asks what people will pay for. If output becomes cheap, old consumption saturates, and attention is already maxed out, demand will not disappear. It will move into unmet life. Mental health. AI translators for pets. Fantasy companions. Tutors. Retreats. Dopamine detox. Human attention. Founder roleplay. Eldercare. Family mediation. Language learning. Guided recovery. Paid silence. Paid intimacy. Paid containment.

Some of this is useful. Some of it is tender. Some of it is absurd. Some of it is disgusting. But A2 is right about the market. People will want things. They will pay to be held, amused, soothed, coached, desired, accompanied, interpreted, and reassembled. The new consumerism will not wait for moral philosophy. It will colonise the gap between what the machine economy provides and what human beings still lack. The danger is not that this market exists. The danger is that it becomes the primary way a society responds to the loneliness and dislocation the machine order itself produces. Not civilisation. Abandonment with customer support.

Di Dongsheng makes the imperial bid. He tells his fellow Chinese: stop thinking as student, start thinking as centre. The old developmental state was a catch-up project. That project is complete. The task now is to take the reins—to provide order, capacity, infrastructure, production, finance, and development at scale. This is not merely a nationalist project. It is a bid to become the organising pole of the interregnum. Where Bi offers aristocratic freedom, Di offers imperial stability. Where A1 offers cheap abundance, Di offers a system that makes abundance strategic rather than merely cheap. The question Di leaves unanswered is whether imperial order can accommodate non-revocable life from below, or whether it merely replaces one form of dependency with another.

Benjamin Bratton’s AgentWorld says something more basic: the republic is no longer composed only of humans. The social field itself is being repopulated by agents, models, memories, protocols, simulations, interfaces, and hybrid forms of agency. His brief frames AgentWorld as a “preemptive anthropology” of human-AI centaur societies, where human societies are the niches into which artificial agents adapt and where agents then reconstruct those niches in return. He is not simply offering stack-governance as another claimant. AgentWorld is the interregnum’s atmosphere. It is what happens when the old republic is still arguing about kings, citizens, markets and empires, while non-human agents are already entering offices, schools, households, simulations, memory systems, and bureaucracies. They do not ask whether the social compact has been renegotiated. They begin acting. Then society explains itself afterward.

This is not a distant future. Ren Jiantao, writing from the Chinese academy in 2020, lamented China’s unpreparedness for a new age of machines, machine-human hybrids, immortals and other things that break the certainty of classical science. But here we are, a new age.

---

IV. The Axial Age Pressure

These five claimants—Bi’s wolf, A1’s abundance, A2’s demand, Di’s empire, Bratton’s AgentWorld—do not add up to a republic. They are a machine ecology with imperial ambitions. They are the contest.

The old developmental compact answered the question of what a human is for with economic necessity. The machine order breaks that answer without supplying a new one. People will not simply accept this. They will ask again what a human is for, what cannot be sold, what power must answer to, what dignity means when production no longer needs everyone, and what kinds of life are worth reproducing. These are not policy questions. They are civilisational questions. Perhaps this pressure will one day become religious, philosophical, or civilisational. Perhaps not. Interregnums can last a long time. The agony may produce prophets. It may also produce better vendors.

---

V. The Interregnum

That is why I do not want to rush to proclaim a new republic. The old republic is losing its reason for including everyone. The machine order has not yet earned a reason. In the gap, wolves, empires, stacks, and markets rush in.

Bi gives the wolf a constitution. A1 asks whether abundance can be revoked. A2 asks what people will pay for when old wants are exhausted. Di asks China to take the reins. Bratton sees the stack becoming governance.

None of this is yet a republic. It is the contest to replace one.

The mistake would be to answer Bi with nostalgia. Return to equality. Return to welfare. Return to old Singapore. Return to Tocqueville’s America. That river has moved. Bi is right about that.

But the mistake would also be to answer Bi with surrender. Let the wolf rule because the wolf builds rockets. Let the machine proceed because the machine produces growth. Let China diffuse capability because cheapness is blessing. Let the market monetise loneliness because demand is real.

That is not a settlement. That is drift with excellent vendors.

The real question is harder: What forms of power, belonging, and non-revocable life can exist before the next settlement arrives? Not protection of the permissioned. Too small. Not uplift everyone. Too old. Not meaning. Too easy. Not new consumerism. Too market. Not industrial aristocracy. Too wolfish.

The question is whether ordinary people can still become non-ignorable before the machine order learns to proceed without asking them.

That is the line.

The machine order may not need everyone. The old republic insisted everyone mattered. The interregnum begins when both statements are true and neither can rule.