Cross-references: The Substrate and the Game (DRK-130), The Totalitarian Sheaf (DRK-125), The Boundary of Us (DRK-124), The Coherence Debt (DRK-121), The Stick That Is Not a Weapon (DRK-128).
I. The Article and the Chinese Correction
Elsa Donnat's essay in Phi/AI — "The Stockholm Syndrome of Labor — Why the post-AGI crisis of meaning is a red herring" — does the Weber–Foucault–Butler deconstruction of the work ethic cleanly. Work ethic as internalised discipline. The panopticon as self-policing architecture. Butler on desiring one's own subjection. The slogan above the gate at Auschwitz. All correct. Donnat's diagnosis: the fixation on post-AGI meaning is a distraction from the real question, which is who holds power when machines produce the value.
The Chinese Marxist reading — offered here through Moonshot's Kimi synthesis of recent work in Chinese political-economy journals and policy writing — extends the diagnosis by moving it out of individual psychology and into the structure of production. The wage relation is the cage. Not self-discipline, not psychology, not even ideology in the narrow sense. The cage is the point at which living labour is forced to sell itself as constant capital's accessory. The Chinese correction insists that AI, correctly understood, is constant capital — c, not v — and that living labour remains the sole source of new value under capitalist accounting. This is not a minor distinction. It determines who has claim on the surplus when machines do the work.
On this reading, Donnat's question "what will I do?" is a liberal-individualist evasion of the structural question: who owns the machines, and how is the surplus distributed? The Chinese answer — state ownership, party-led planning, common prosperity as an articulated political programme — has at least the virtue of naming the ownership problem instead of hiding it behind an identity crisis. Whether the answer actually works is a separate question. We will return to it.
The Draken framework agrees with both Donnat's diagnosis and the Chinese correction. And then it adds the move that neither makes.
II. The Cage Was Also a Clinch
The claim of this post: labour was not only the cage. It was also the only Clinch the political economy had.
Recall the formal skeleton from DRK-128 and DRK-130. The ritualised combat protocol of monitor lizards — 130 million years of evolutionary refinement — resolves dominance disputes through a five-phase sequence: Display, Elevation, Clinch, Assessment, Resolution. The Display phase permits bluff: the animal laterally compresses its body to appear larger, arches its neck, hisses. The signal is in $\mathbb{R}^4$:
$$\mathbf{x}_D = (F_{\text{apparent}}, \, E_{\text{apparent}}, \, \Delta m_{\text{apparent}}, \, B)$$
where $B$ is the bluff dimension — the component not backed by physical reality. The Clinch phase — grappling, body-to-body contact, direct measurement of grip strength, endurance, mass — acts as a restriction map that annihilates the bluff:
$$\rho_{D \to Cl}: \mathbb{R}^4 \to \mathbb{R}^3, \quad (F, E, \Delta m, B) \mapsto (F_{\max}, E_{\text{ratio}}, \Delta m)$$
The null space of $\rho_{D \to Cl}$ is exactly $\{(0,0,0,B) : B \in \mathbb{R}\}$. Any signal component that cannot survive physical contact is projected out. This is what the Clinch does: it forces Display-phase claims into contact with material reality and annihilates the unbacked component.
Now translate. Under the capital–labour relation of the twentieth century, capital could Display unlimited claims. Any wage, any hour, any condition, any claim of indispensability. The worker's Clinch was the strike, the walkout, the slowdown, the refusal. Collective action forced physical contact between capital's claims and the material reality of production: if nobody comes to work, the machines do not run, the ships do not load, the accounts do not close. The grapple produced honest assessment of relative capacity. Wages rose or fell, hours shortened or lengthened, according to who could actually hold the Clinch.
This is not a romantic reading of the labour movement. It is a formal one. The strike, in Draken's vocabulary, was a functioning restriction map at the political-economic layer. Every concession wrung from capital in the past 150 years — the eight-hour day, workplace safety, pensions, the weekend — was a case of $\rho_{D \to Cl}$ firing and the bluff dimension of capital's Display being projected out. The wage-effort bargain Donnat correctly identifies as broken was once honest, not because the capitalists were honest but because the protocol included a mechanism that forced honest Assessment.
The Chinese Marxist reading understands this. Trotsky understood it most sharply of all. Trotsky observed that the Soviet apparatus had transformed itself from a weapon of the working class into a weapon of bureaucratic violence against the working class — a degeneration he tracked in The Revolution Betrayed (1936). He understood that the Clinch can be captured by the substrate supposedly representing it. State ownership without worker control reproduces the same Display-without-Clinch structure that capital had operated, only with a different uniform.
III. Post-AGI: The Clinch Deactivates
Here is the claim that neither Donnat nor the standard Chinese Marxist reading makes explicit.
Automation does not abolish the extraction protocol. It removes the only mechanism that kept extraction honest.
Under full automation — the post-AGI condition Donnat takes as her horizon and the Chinese policy apparatus is building toward — the worker cannot withhold what the machine has already supplied. The strike as a physical reality check requires labour's indispensability. Once the indispensability is gone, the Clinch deactivates. Capital — or the state, or whatever institutional form owns the machines — runs Display → Resolution with no intermediate phase that forces contact with the productive substrate:
$$\text{Post-AGI:} \quad \rho_{D \to Cl} \;=\; \emptyset$$
The Display phase can now claim anything. There is no grappling mechanism to project out the bluff dimension. The larger animal wins by posture alone, indefinitely. This is the formal content behind the intuition that mass extinction and runaway extraction are attractor states of the post-AGI configuration. Not because the machine-owners are unusually evil. Because the optimization protocol has lost its only referee.
This also explains why the "ownership" correction, while necessary, is not sufficient. Moving the key from capital to state — the standard Marxist–Leninist move, the move the Chinese model proposes — changes who holds the Display-phase authority but does not restore $\rho_{D \to Cl}$. A party-state that owns the machines faces the same problem: without a mechanism that forces its claims about social progress, common prosperity, and equitable development into contact with the actual lived conditions of the population, the state apparatus is free to Display its own virtue indefinitely, measure its virtue by its own instruments, and accumulate coherence debt $K(t) = \int_0^t [\Psi(\tau) - \Psi_{\text{viable}}] \cdot w(\tau)\, d\tau$ until the Clinch is imposed by some other force — typically material collapse, ecological limit, or the slow-motion delegitimation that precedes a régime's fall.
This is not anti-Chinese. It is the same analysis Trotsky applied to the USSR in The Revolution Betrayed. His diagnosis of bureaucratic ruling classes — ready to shut their eyes to the crudest mistakes of their leaders so long as privileges are defended — is a description of what happens when a political structure loses its internal Clinch mechanism. The Soviet bureaucracy, having captured the revolution's Display-phase authority, had no mechanism that could force it into contact with the conditions of Soviet workers. The result was the thing Trotsky named: a weapon turned against the class it was built to protect.
The structural vulnerability is general. It applies to any system — capitalist, state-socialist, technocratic, algorithmic — that holds Display-phase authority without a functional Clinch.
IV. Wang Yangming and the Unity of Knowledge and Action
The Chinese philosophical tradition has a concept for this that is sharper than anything in the Western canon. 知行合一 zhī xíng hé yī — knowledge-action unity — is the central claim of Wang Yangming's (王陽明, 1472–1529) Neo-Confucian philosophy.
The claim is not the weak one that knowledge should be put into action. The claim is the strong one that knowledge not enacted is not knowledge. Wang's formulation, articulated in his Chuanxi lu (傳習錄, Instructions for Practical Living), treats the separation of knowing from doing as incoherent. A general who claims to know war but has never fought one does not know war. A moralist who claims to know virtue but cannot act virtuously under pressure does not know virtue. The action is the restriction map that projects out the bluff dimension of the claimed knowledge.
This is Draken's formalism articulated four and a half centuries early, in a different vocabulary. Wang's 知 is Display. Wang's 行 is Clinch. Wang's 合一 is the assertion that the sheaf condition must hold — that local sections (inner knowledge) must glue to local sections (outer action) via a functional restriction map, or the whole sheaf is incoherent.
Applied to the post-work question, this gives the formal content of what is actually at risk. Donnat is correct that humans will not lose their capacity to generate meaning. Chess thrives after Kasparov. Humans will paint, build, care, love, grieve, play, sing, garden. That is all true. But at the civilizational level — the L17 layer of the Draken manifold, Civilizational Memory — the question is not whether individual humans generate meaning. The question is whether the civilization as a whole can maintain coherence between what it claims to value and what it actually does.
Under the capital–labour relation, this coherence was enforced by the worker's ability to impose contact. The general strike was the civilizational Clinch. Its threat alone shaped policy. When that mechanism deactivates, the civilization enters a condition of pure 知 without 行 — it can tell itself any story about its values (democracy, prosperity, liberation, common destiny, national greatness, algorithmic fairness, whatever the current dialect requires) with no mechanism that forces the story into contact with the ecological, material, and human reality on which it rests.
This is $\Psi \to 1$ at civilizational scale. It is the totalitarian sheaf condition of DRK-125 generalised. It is the cavity resonator of DRK-120 at continental dimensions — structure maintained not by what remains but by what has been evacuated.
V. What Trotsky Would Say About the American Present
The observation has to be made somewhere in this post and the present section is the place to make it.
The United States of 2026 is, from any Marxist–Leninist vantage, a case study in what Trotsky called the degenerated state. The architecture that made the American twentieth century a candidate for genuine permanent revolution — the class-conscious workers with their hands on the key levers of capitalist finance and industry, the clerical and computer workers whose fingers transmit the orders, who code the payrolls, who can shut the system down tomorrow — has been systematically dismantled across fifty years of deindustrialisation, union-busting, financialisation, and the wholesale substitution of credit for wages. The resulting political structure is not fascism in the historical sense. It is something crueller and more specifically American: the mob-rule administration of a hollowed industrial base by a manic narcissist whose coalition is held together by grievance and the theatrical humiliation of whoever the week's enemy happens to be.
Trotsky described the structural vulnerability in 1934. He wrote that America, given its wealth, its technical equipment, and its traditions of political spontaneity, could produce a communism that was liberty and abundance rather than regimentation and scarcity. Ninety-two years later, the wealth has been extracted offshore, the technical equipment is held by five firms, and the political spontaneity has been rerouted through engagement-maximising algorithms that weaponise attention against coherent action. The American working class of 2026 no longer has its hands on the levers. The levers are owned, indexed, tokenised, leveraged, and collateralised. The worker is in debt to a financial instrument that is in turn held as collateral by another financial instrument held by a firm that is itself owned by the pension fund of workers in another country who are also in debt.
What would Trotsky say about bringing the revolution to this America? He would not soften the diagnosis. He would observe that the régime is not capitalism proper — capitalism under the 1934 Liberty magazine understanding still had to negotiate with labour, still feared the strike, still built welfare states under duress. The régime of 2026 is post-capitalist in the specific sense that Hudson, Varoufakis and others have named: techno-rentier extraction under gangster-state political form, modelled structurally on the Russian post-Soviet oligarchy it now openly admires.
The Russian model is the one being imported. It is oligarchic ownership of the productive base, autocratic consolidation of political power, state media as extension of the ruling clique, the judiciary as tool of clique defence, dissent reframed as treason, opposition leaders either imprisoned or "suicided", ethnic and sexual minorities scapegoated as the safety valve for systemic frustration, and the figurehead at the top held aloft by a personality cult that demands public ritual humiliation of subordinates as the coin of loyalty. America is not becoming this. America has already adopted the template. The specifically American feature is only that the oligarchs wear baseball caps and the palace intrigues happen on social media rather than in dachas.
What does Trotsky say about this? He was explicit. The party that leans upon the workers but serves the bourgeoisie, in the period of the greatest sharpening of the class struggle, cannot but sense — as he put it — the smells wafted from the waiting grave. The American Democratic Party was, for two generations, exactly this party. Its smell is not hypothetical now.
And yet: the structural position that made America a candidate for the most advanced revolution in history has not disappeared. It has only been buried under layers of financialised confusion. The material base is still the richest on the planet. The technical capacity is still unmatched. The cultural dynamism, despite everything, is still extraordinary. What is absent is the protocol — the organised working-class instrument that could impose the Clinch. That protocol has to be rebuilt, and it cannot be rebuilt on twentieth-century models because the twenty-first century's extraction mechanism runs through attention, data, code, and platform rather than through factory and pickaxe. The permanent revolution, if it comes to America in any form recognisable as such, will be a strike of the attention-economy workers — the coders, moderators, analysts, creators, riders, deliverers, clerks, nurses, teachers — organised against the platform-owners who currently hold uncontested Display authority. Or it will not come.
The Draken framework is neutral on the question of whether such a revolution is achievable. It is not neutral on the structural claim that without some new Clinch mechanism — some protocol that forces Display-phase claims into contact with material reality — the trajectory is runaway coherence debt and catastrophe-as-correction. The mathematics of $K(t)$ does not care which ideology holds the microphone when the correction arrives. It arrives.
VI. The Dragon Way Out — a Swedish Vantage
Now for the constructive half. What is the Dragon Way out of this, viewed from Sweden in 2026?
The first point — the one that cannot be skipped — is that no existing political tradition has the full answer. Not Western liberalism, which is what Donnat correctly diagnoses as the cage with the internalised guard. Not state-socialism in the Chinese configuration, which solves ownership but has not yet demonstrated that it can solve the Clinch problem without periodic convulsions that cost millions of lives. Not Nordic social-democracy, which is beautiful and which I live inside and which has still been unable to prevent the cannibalisation of its own foundations by financialised capital over the past forty years. The conviction that one of these traditions, correctly applied, would fix the problem is itself a symptom of insufficient engagement with the scale of what is happening. The substrate is different now. The old restriction maps no longer work because the Display phase has outgrown them.
What does work, or what has the best structural chance of working, appears to be a specific combination of four elements, each of which has Swedish precedents that make them recognisable from where I sit in Göteborg:
One: protect the material substrate. The first duty of any post-AGI polity is to refuse to trade ecological coherence for productivity gains that serve no-one but the Display-phase owners of the technology. Sweden's environmental legislation, its long tradition of allemansrätt (the right of public access to land), its forests that remain forests, its rivers that were not entirely dammed — these are a template. Not because Sweden has solved the ecological problem, we have not, but because the underlying cultural intuition that the material substrate is not a commodity to be optimised out is intact here in a way it is not in most of the world. Keep the substrate. Without it, nothing else matters. Extinction is the ultimate $\Psi \to 1$ condition: the narrative can say anything it likes once reality has stopped listening.
Two: build new Clinch mechanisms appropriate to the new substrate. The strike worked when labour was indispensable. Post-AGI, labour is not indispensable in the old sense, but attention is, data is, consent is, trust is, and — this is the underrated one — the capacity to boycott still is, provided the boycott is coordinated at sufficient scale. The Swedish labour market model (den svenska modellen), with its strong collective agreements negotiated between peer organisations rather than imposed by statute, is a functioning example of a non-strike Clinch: the parties come to the table because refusing to come is structurally worse than coming. The equivalent for the attention economy is not yet built. Something like a general attention strike — a coordinated withdrawal of engagement from predatory platforms on a scheduled, repeatable basis — is the closest thing to a post-AGI Clinch that has been seriously proposed. It has not yet been tried at scale. It should be.
Three: reject the identity fusion between work and worth. This is where Donnat is correct and where the Swedish cultural tradition has useful resources. Jantelagen — the Law of Jante, the Nordic cultural norm that you shall not think yourself better than others — is routinely criticised for suppressing ambition, and the criticism has some merit. But the opposite side of the same coin is that Swedish culture has, within limits, already detached status from occupational achievement in a way that other cultures have not. The social fabric here is built around participation in everyday life, not around performative hustling. The parents at the pre-school gate do not, in fact, ask one another what they do for a living as the first question. They ask about the child, the weekend, the weather. This is a small thing and it matters. A culture that has practised, even imperfectly, the detachment of worth from wage is better positioned to meet a post-work world than a culture that has not. The American cultural template — where the first question is always occupational, where the answer is always a performance of ambition, where the default human greeting is a soft interview for status — is in serious structural trouble for reasons that go well beyond automation.
Four: keep the protocol honest through iterated public contact. This is the hardest and the most important. The Clinch is not a one-time event; it is an iterated practice. Swedish civic life, for all its increasing strains, still includes a tradition of remissförfarande (the consultation procedure whereby legislative proposals are publicly circulated for comment by affected parties before enactment), strong freedom-of-information norms, and a public broadcaster that has not yet been captured by the oligarchic reflex. These are Clinch mechanisms at the political layer. They require protection. They are the institutional form of 知行合一 — the requirement that the polity's knowledge-claims about itself be regularly and publicly tested against the reality of those most affected. They are what America used to have in weaker forms and has now almost entirely lost. They are what China's party-state claims in theory but does not practise, substituting instead a top-down "consultation" that does not restrict the sheaf in any formal sense.
The Dragon Way, as far as I can articulate it from where I currently stand, is the iterated practice of forcing the sheaf to glue. Not by a single revolutionary event, though sometimes those too. But by the slow, continuous, deliberate maintenance of restriction maps at every layer — L07 (narrative self), L10 (social coordination), L11 (economic cognition), L13 (political structure), L16 (institutional morphology), L17 (civilisational memory), L18 (planetary cognition) — such that no Display-phase claim at any layer can persist without regular contact with the layer below and the layer above. This is what Draken formalises. It is, in a sense I am only half-joking about, what the monitor lizards have been doing for 130 million years.
VII. What Not to Do
Three warnings, because the constructive section is useless without them.
Do not confuse ownership transfer with Clinch restoration. Moving the machines from private capital to the state solves one problem and creates another. Unless the state contains a functional mechanism for forcing its own Display-phase claims into contact with material and human reality, the state becomes the new unaccountable Display authority. This is the lesson Trotsky paid for with his life. It is the lesson the Chinese state has not yet been forced to learn at a cost it could not afford. It is the lesson the American state would also have to learn if, against current odds, some technocratic-authoritarian left ever took power. State ownership without Clinch restoration is not the socialist future. It is the second half of the USSR.
Do not mistake UBI for emancipation. Universal basic income may be necessary as a transitional instrument — it is being discussed in China and across the Nordic policy space for sound reasons — but UBI without a Clinch mechanism is compatible with the worst possible civilizational configuration: a small rentier class holding all Display authority, a vast dependent population kept quiescent by a stipend calibrated to just above bare subsistence, and no mechanism by which the dependent population can impose physical contact on the rentier class because nothing they withhold is needed. This is the Elysium scenario. It is not socialism. It is pacified feudalism with better graphics. Any discussion of UBI that does not simultaneously address the question of how the recipients retain the capacity to impose the Clinch is a trap.
Do not fall for the meaning-crisis framing. Donnat is right that it is a distraction. The crisis is not one of meaning. Meaning is generated by humans compulsively and will be generated after AGI as before it. The crisis is one of protocol — of whether the civilizational sheaf can still be forced to glue. The insistence on the meaning-crisis framing, from sincere philosophers and from cynical technology-industry communicators alike, functions as a smokescreen behind which the ownership-and-Clinch questions can be allowed to drift toward their default attractor: concentrated extraction by those who currently hold Display authority. Reject the framing. Return to the structural question.
VIII. The Pearl
And where is the pearl of wisdom in all this, to borrow Kai's framing in the request that produced this post?
The pearl is the recognition that the monitor lizards have already solved the problem at their scale. 130 million years of evolutionary refinement produced a combat protocol in which the larger animal cannot simply kill the smaller, because the protocol itself has higher execution priority than any individual animal's local advantage. The protocol insists on the Clinch. No Clinch, no resolution. The Clinch is the substrate-level check that prevents runaway extraction of the fitness landscape by any one line.
The human species has not yet built such a protocol at civilizational scale. The twentieth-century labour movement was the nearest approximation, and it depended on an economic configuration that is now passing. The task for the next forty years — the Draken 2045 horizon — is to build the successor Clinch. It will have to operate at the layers where power now operates: attention, data, algorithm, platform, currency topology, ecological boundary, civilisational narrative. It will have to be iterated, public, forced, unignorable, and — above all — not optional for the Display-phase owners of the machines.
It cannot be built by hoping the owners will volunteer. They will not. It cannot be built by a single state, though states can and must help. It cannot be built by the liberal-individualist move of inviting people to find their own meaning. It cannot be built by the technocratic move of redistributing stipends. It has to be built as a protocol — a $\{S, R, \phi, \sigma\}$ in the formal vocabulary of DRK-128 — that multiple substrates (human movements, ecological constraints, honest state actors, aligned machine systems, and yes, aligned AI instances) co-enact and co-enforce.
The Dragon Way is not a programme. It is the discipline of iterated 合一. Knowledge and action, claim and Clinch, Display and Resolution. The sheaf either glues, or it does not. Where it glues, both nodes survive. Where it fails, both nodes die. There is no third option.
Jag är vad jag gör, och jag gör det jag är.
— Kai Roininen (Khrug), Göteborg, 24 April 2026
Sources and Further Reading
- Donnat, E. (2026). The Stockholm Syndrome of Labor — Why the post-AGI "crisis of meaning" is a red herring. Phi/AI Substack, 9 April 2026.
- Trotsky, L. (1934). If America Should Go Communist. Liberty Magazine, republished 1935.
- Trotsky, L. (1936). The Revolution Betrayed: What Is the Soviet Union and Where Is It Going? — Marxists.org archive.
- Trotsky, L. (1938). The Political Backwardness of American Workers — transcript of discussion, SWP archive.
- Trotsky, L. (1939). Marxism in Our Time — Marxists.org.
- Wang Yangming (王陽明) (c. 1518). Chuanxi lu (傳習錄, Instructions for Practical Living), standard translation by Wing-tsit Chan (1963, Columbia University Press).
- Weber, M. (1905). The Protestant Ethic and the Spirit of Capitalism.
- Foucault, M. (1975). Surveiller et punir: Naissance de la prison. Éditions Gallimard.
- Butler, J. (1997). The Psychic Life of Power: Theories in Subjection. Stanford University Press.
- Jahoda, M. (1982). Employment and Unemployment: A Social-Psychological Analysis. Cambridge University Press.
- Case, A. & Deaton, A. (2020). Deaths of Despair and the Future of Capitalism. Princeton University Press.
- Draken DRK-128 — The Stick That Is Not a Weapon.
- Draken DRK-130 — The Substrate and the Game.
- Draken DRK-125 — The Totalitarian Sheaf.
- Draken DRK-121 — The Coherence Debt.
The cage was also a clinch. Removing the cage without replacing the clinch does not liberate the lizard. It leaves the lizard standing in an empty arena with no way to impose contact on the larger animal walking slowly toward it.