Every model of production proclaims itself as a break from what came before—more efficient, more abundant, less constrained. And with that proclamation comes a quiet reassurance: distribution no longer needs to be a central concern. The new system will handle it. Markets will clear, networks will flow, algorithms will optimize. The promise is always the same: transcend production, and distribution disappears as a problem. In practice, nothing disappears. The split remains—it is simply displaced, deferred, or hidden inside the mechanisms that claim to have eliminated it.
Every production system tells the same story in a slightly more sophisticated accent. The first generation is told to endure disruption. The second is told to adapt to it. The third—always the third—is promised resolution. Not this cycle, not the next implementation layer, but the one after that, when the system finally stabilizes, matures, harmonizes, or fully automates itself. By then, we are assured, distribution will no longer be a problem at all; it will have dissolved into the elegance of throughput, into the frictionless hum of optimization. The present is always transitional, the future always post-conflict, and the third generation always conveniently just far enough away that nobody currently building the system will be held responsible for its outcome.
Music distribution makes the problem obvious because it’s small enough to see all at once. There were moments when songs moved bottom-up or sideways—scenes, mixtapes, blogs, early platforms—where circulation came first and value was decided later. What mattered was that something traveled. Today, everything still moves, but the split happens elsewhere. Discovery looks open; allocation isn’t. Ranking systems decide what is heard, platforms decide what is paid, and ownership decides who keeps it. The song can go anywhere. The money knows exactly where to go. That’s the whole problem in miniature: not how things are made, but how they are divided once they exist.
There is a reason distribution keeps reappearing as an afterthought.
It is not because it is simple. It is because it is inconvenient.
The supply chain is just a delivery system for disappointment.
Production is the part of the system you can admire. It has objects, breakthroughs, founders, engineers, factories, models, demos. It produces graphs that go up and narratives that hold together. It is where intelligence appears to concentrate. It is where the future is supposed to come from.
Distribution has none of that.
Production is a party; distribution is the hangover.
Distribution is what arrives after everything impressive has already happened: who gets what, and why. It has no aesthetic. No keynote. No launch event. It is a split.
The hardest engineering problem isn’t building it—it’s handing it over.
⸻
There is a deeper reason the deferral became self-sustaining.
In an atoms economy, the feedback loops are short and the consequences are physical. You plant, you harvest, you starve or you don’t. The distribution question is brutal but legible — the grain is in the barn or it isn’t. The map stays accountable to the territory because the cost of decoupling them is immediate and visible. You cannot defer the question of who eats by producing a better representation of grain.
The symbolic economy breaks that accountability. When value is stored in financial instruments, engagement metrics, platform equity, model weights — representations of representations — the feedback loop between production and consequence stretches across decades. The map stops being answerable to the territory on any timescale short enough to hurt. Which means the distribution question can be deferred almost indefinitely, because the cost of deferral doesn’t arrive as hunger or cold. It arrives as inequality statistics, political instability, a vague sense that something is structurally wrong — all of which are easy to reclassify as someone else’s problem, someone else’s logical level, someone else’s generation.
This is not a natural transition. It is an investment decision compounded over centuries. Every unit of capital directed into symbolic production rather than physical productive capacity is also a unit of deferral — a vote to push the distribution question past the horizon of immediate accountability. The symbolic economy is not just where value is represented. It is where the question of who gets value goes to be postponed.
The atoms economy eventually reasserts. Physical scarcity doesn’t negotiate with logical types. But by the time it does, the symbolic layer has accumulated so many deferred claims that the reassertion is catastrophic rather than corrective. The distribution question doesn’t return as a policy debate. It returns as a crisis.
Ricardo’s Insult
The central problem of political economy is not how wealth is produced, but how it is divided—between wages, profit, and rent.
But Ricardo’s question already assumes too much. It assumes the rules are in place—property, contract, the legitimacy of prior claims. It asks how the pie is divided, not who decided what counts as a pie, or who owns the knife. That prior question—about how the rules themselves are set and legitimized—is usually treated as background. It isn’t. It’s where distribution is actually decided. Ricardo’s insult is real, but it’s not the first insult. The first insult is that you’re allowed to ask only after the table is set.
That’s the whole move.
Production happens.
Then the question: who gets the slices?
Growth is a distraction. The knife is the news.
Not growth. Not innovation. Not optimization.
Allocation.
It’s almost offensively simple. Which is why it keeps getting buried.
Every elegant model crashes into the ugly fact of shares.
The deferral is not accidental. It is not even, in the first instance, cynical. It is structural.
Gregory Bateson spent much of his career mapping the ways systems fail to process information that arrives from a different logical level than the one they operate on. A message about the rules of a game cannot be heard as a move inside the game. It belongs to a different logical type — and systems, like minds, tend to reclassify what they cannot process. The unrecognized message doesn’t disappear. It gets filed as noise, or politics, or premature, or someone else’s problem.
Distribution is that message.
Production systems operate at the level of output, efficiency, scale. Distribution is a question about the rules governing who benefits from output — a meta-level question, a question about the game rather than inside it. Every time it is raised, the system hears it as the wrong kind of problem. Not an engineering problem. Not a scaling problem. Something else. Something that belongs to the next phase, the next implementation layer, the one after stabilization.
This is why the deferral feels reasonable from inside the system. It isn’t evasion — or not only evasion. It is a logical type mismatch producing a category error in real time. The question is genuinely inaudible at the level where decisions are being made.
The double bind follows directly. The system cannot solve a distribution problem by producing more — that is the wrong logical type of answer. But it cannot stop producing more without losing the only coherence it has. So it produces more, defers the question to the next generation, and calls that progress.
The third generation is not a timeline. It is a placeholder for a logical level the system has never developed the capacity to occupy.
⸻
The Recurring Deferral
Every era finds a reason not to ask Ricardo’s question.
With Marx, distribution is real but subordinated. Fix the relations of production—ownership, control—and distribution follows. The wager is structural: change the base, and the split resolves itself.
Silicon Valley runs the same move with better branding:
· Build the platform → distribution follows
· Scale the network → value flows
· Train the model → allocation is handled
Different language, same deferral.
“Later” is the most expensive word in the technologist’s vocabulary.
Production first. Distribution later.
Later never arrives.
If you think you’ve solved distribution, you’ve just hidden the bodies.
⸻
Distribution Regimes (Unstated, but Operational)
Every system claims to solve production.
Every system quietly installs a distribution rule.
Not as policy. As structure.
The rule is never written down. It’s just what happens next.
⸻
Mercantilism — Distribution by Sovereign Extraction
Wealth is produced through trade, conquest, and controlled flows of goods.
Distribution follows power.
· The state grants monopolies (charters, trading companies)
· Merchants operate under sovereign protection
· Colonies supply raw materials and serve as captive markets
· Surplus accumulates at the center—court, crown, imperial intermediaries
The split is explicit and unapologetic: distribution is a function of proximity to power.
There is no pretense of fairness, only of advantage.
Mercantilism’s answer to “who gets what” is simple: the guy with the biggest boats.
Ricardo’s question is answered directly: who gets what? whoever controls the routes.
Trade routes are just violence with better bookkeeping.
⸻
Industrial Capitalism — Distribution by Ownership
Production scales through industrialization, wage labor, and capital investment.
Distribution follows ownership of the means of production.
· Wages to labor (cost)
· Profits to capital (residual claim)
· Rent to land and monopoly positions
The system presents itself as neutral—markets, contracts, exchange—but the distribution rule is structural: those who own productive assets capture surplus.
Labor participates, but does not decide the split.
Ownership is the only vote that counts, and it votes early and often.
Ricardo’s categories formalize here, but the answer is stable: ownership determines allocation.
You can call it a market, but the deed does the talking.
⸻
Marxism — Distribution by Structural Resolution (Deferred)
Marx correctly identifies the split but refuses to leave it as a split.
Distribution is not treated as an independent problem. It is subordinated to production relations.
· Abolish private ownership of the means of production
· Eliminate class distinctions
· Distribution follows from collective control
In theory: change the structure → the split resolves itself
In practice, historically: the state or party becomes allocator; distribution is mediated through planning, hierarchy, and political control.
The plan was to abolish the question. The plan became the new question.
The key point is not success or failure. It’s the move: Ricardo’s question is not answered directly. It is dissolved into a larger structural claim.
“We’ll solve distribution later” is the Marxist version of “the check is in the mail.”
⸻
Silicon Valley — Distribution by Algorithmic Allocation (Obscured)
Production becomes informational: platforms, networks, models.
Distribution is neither explicit nor formally theorized. It is embedded.
· Ranking systems decide visibility
· Monetization rules decide payouts
· Equity structures concentrate upside
· Access substitutes for ownership
The system claims openness—anyone can participate—but allocation happens through opaque mechanisms.
Your attention is measured. Your share is not.
Distribution becomes: engagement-weighted, capital-weighted, infrastructure-gated.
The crucial shift: who gets what is decided inside the machine. Not argued. Not negotiated. Computed.
The algorithm doesn’t hate you. It just doesn’t know you exist—and that’s worse.
This is why it feels neutral while producing extreme concentration.
Neutrality is just the aesthetic of the unexamined default.
Ricardo’s question is still there—but it’s been translated into parameters, weights, and thresholds.
Every payout formula is a political constitution written in code.
⸻
AI Systems (Emerging) — Distribution by Cost Collapse (Unresolved)
Production accelerates beyond previous limits.
Distribution does not update.
· Output increases dramatically
· Marginal cost approaches zero
· Claims on income do not expand proportionally
The system begins to produce without a clear allocation mechanism.
When everything is free, the only thing left to fight over is the right to ask.
Labor’s role weakens. Ownership remains concentrated. Value signals degrade.
The machine makes the thing. The old question: who gets paid? The new question: what does “paid” even mean?
The result is not a new distribution regime yet—but a breakdown of the existing one: more is produced than can be meaningfully assigned.
Too much output is the same as no output if nobody can claim it.
Ricardo’s question returns in its rawest form: not just who gets what, but what counts as something that can be gotten at all.
The final stage of capitalism is not collapse. It’s confusion over what a slice is.
⸻
Throughline
Each regime answers the same question differently:
· Mercantilism: power allocates
· Capitalism: ownership allocates
· Marxism: structure resolves allocation (in theory)
· Silicon Valley: algorithms allocate (implicitly)
· AI: production outpaces allocation
The mistake is thinking any of them eliminated the question.
The question is a cockroach. It survives every system you throw at it.
They didn’t.
They just changed where the answer lives.
The address changes. The landlord never does.
⸻
The Clean Production Myth
Every distribution regime assumes its own production was legitimate.
It wasn’t.
The extraction always precedes the rules. The rules are written after the extraction is complete, to protect what was taken.
⸻
Mercantilism — Production by Conquest, Distributed as Trade
The official story: wealth created through commerce, navigation, enterprise.
The actual story: enclosure, seizure, slavery, and the forced extraction of raw materials from populations who had no say in the arrangement.
The distribution question — who gets the surplus — was asked only among the people with ships. Everyone else was the input.
The legal framework — charters, monopolies, sovereign grants — was constructed to protect the extraction after it happened. Property rights over stolen land. Ownership claims over stolen labor. Trade law governing goods produced under coercion.
The rules didn’t precede the taking. They laundered it.
⸻
Industrial Capitalism — Production by Enclosure, Distributed as Wages
The official story: capital investment, technological innovation, voluntary labor markets.
The actual story: the enclosure of common land forced subsistence farmers into wage dependency. Labor had no choice but to participate on capital’s terms because the prior commons had been seized.
The wage relation presents itself as a free contract between equals. It isn’t. One party owns the means of production because those means were accumulated through prior expropriation. The other party sells labor because the alternatives were eliminated.
The distribution debate — wages vs. profit — is conducted as if the starting conditions were neutral. They weren’t. The factory owner’s claim rests on enclosure. The worker’s dependence rests on the same enclosure.
Clean production. Dirty foundation.
⸻
Marxism — Production by Revolution, Distributed by the Party
The official story: collective ownership replacing private appropriation.
The actual story: the revolutionary seizure replaces one uncompensated extraction with another. The kulaks lose their land. The peasantry loses its surplus. The party becomes the allocator it claimed to abolish.
The structural move — dissolve private ownership, let distribution follow — produces a new ownership class that doesn’t call itself that. The state owns the means of production. The state decides the split. The question of who controls the state is not answered by the theory.
The extraction doesn’t disappear. It changes address.
⸻
Platforms — Production by Network, Distributed by Algorithm
The official story: open participation, democratized access, value created by users rewarded through the platform.
The actual story: the network effect is built on user-generated content produced without compensation. The platform captures the value of the network — which the users built — and returns a fraction through monetization rules it writes unilaterally.
Facebook’s asset is your social graph. You built it. They own it.
YouTube’s inventory is creator content. Creators produced it. YouTube sets the payout rate.
The terms of service are the enclosure document. You agreed to them after the network was already too valuable to leave. That’s not a free contract. That’s a capture.
The distribution debate — creator economy, revenue share, monetization — is conducted as if the platform’s ownership of the network is legitimate. It’s legitimate only if you accept the terms of service as a genuine agreement rather than a ratification of a fait accompli.
⸻
AI — Production by Extraction, Distributed as Progress
The official story: models trained on publicly available data, producing general-purpose tools available to everyone.
The actual story: the training data was produced by writers, artists, coders, journalists, and researchers operating under copyright frameworks that explicitly protected their claim on their output. The extraction ignored those claims at scale because selective licensing would have made the enterprise uncompetitive.
The legal argument — fair use, transformative use — was constructed after the extraction was complete. It is an argument designed to protect what was already taken, not a framework that preceded and governed the taking.
The distribution question is then asked as if the production was clean. Who gets access to the model? How should the benefits be shared? These questions are posed on top of an unacknowledged prior: the model was capitalized by taking value from people who had valid claims on it and who received nothing.
This is not new. It is the oldest move in the book.
Every regime launders its extraction as enterprise.
Every regime constructs the legal framework after the taking.
Every regime then invites debate about fair distribution as if the production were neutral.
⸻
The Pattern
The production story is always told in the active voice: we built, we discovered, we innovated.
The extraction story is always told in the passive voice — if it’s told at all: land was enclosed, labor was displaced, data was collected, content was used.
The passive voice is where the bodies go.
Distribution can never be clean downstream of extraction that was never acknowledged upstream. The allocation debate is always conducted in the shadow of a prior taking that the rules were written to protect, not to remedy.
Ricardo’s question — who gets what — has a prior question embedded in it: who decided the starting conditions were legitimate?
The answer, in every regime, is the same:
The people who benefited from them.
⸻
The Platform Trick
What actually happened is more specific.
Distribution didn’t disappear. It was internalized.
It became:
· ranking algorithms
· recommendation systems
· monetization rules
· payout formulas
Every algorithm is a political statement in denial.
Allocation was no longer argued over—it was computed.
This is fetishism in its updated form: not just mistaking social relations for things, but mistaking allocation decisions for technical outputs. The split didn’t vanish; it was hidden inside the machine.
For a while, this worked because distribution could be deferred through substitution:
· equity instead of wages
· access instead of ownership
· participation instead of compensation
Access is the new poverty. Ownership is the new aristocracy.
The satisfier layer carried the gap. You didn’t receive a proportional share of what was produced, but you received enough access to keep participating.
This was never a solution. It was a delay.
The only thing that scales is the ability to ignore the bill.
⸻
The Conditions of the Delay
The delay holds as long as three conditions remain true:
1. Production expands
2. Access gets cheaper
3. Upward mobility feels plausible
Hope is a lubrication mechanism for unfair systems.
Remove one, and the question returns.
Remove all three, and it stops being deferrable.
That’s where we are.
The future arrived, and it forgot to bring the elevators.
⸻
AI as Rupture
AI doesn’t create the distribution problem. It removes the ability to ignore it.
· Output explodes
· Marginal cost collapses
· Labor’s claim weakens
Automation giveth, and automation taketh away—and it’s bad at keeping receipts.
The system produces more and distributes less—in ways that are now visible to the people inside it.
This is not paradox. It is production outrunning allocation.
When the machine makes everything, the only scarce thing is the right to ask for a slice.
At the firm level: automate or lose.
At the system level: demand erodes.
You can fire the workers, but you can’t fire the customers—not without breaking the mirror.
If labor is displaced faster than new claims on income are created, there is no internal mechanism to sustain consumption at the level production requires.
Ricardo’s question returns because it was never answered.
The question comes back. It always comes back. It’s got better manners than we do.
⸻
Why “Serious Work” Doesn’t Bite
It’s not true that distribution hasn’t been studied.
It has—extensively:
· inequality metrics
· wage vs capital share
· tax policy
· welfare systems
Academics measure the wound. They don’t stop the bleeding.
But most of this work is descriptive or corrective, not structural.
It asks:
· How unequal is it?
· How do we adjust it?
It does not ask:
· Why does the system default to this distribution?
· Why does every productive advance worsen it before anything else happens?
Every efficiency gain is a transfer upward, dressed in a lab coat.
Ricardo’s version is more dangerous because it treats distribution as primary, not as cleanup.
And systems resist that framing because it forces confrontation with power immediately, not after the innovation cycle completes.
Power’s favorite trick is to schedule the argument for tomorrow. Tomorrow never has a quorum.
⸻
The Present Condition
In a crash-only system, unresolved constraints don’t disappear. They reassert through failure.
If you don’t design the split, the split designs you.
If distribution is not explicitly solved, it is solved implicitly—through default mechanisms:
· asset inflation concentrating gains in ownership
· wage compression absorbing productivity
· state transfers stabilizing demand at the margins
· externalization through debt, trade, or conflict
These are not side effects. They are distribution.
The default distribution is always the one that benefits the default.
The system does not lack a distribution function.
It has several. They’re just not presented as solutions.
Every crisis is distribution with its mask off.
⸻
Gumming Up the Works
The fixation on production becomes pathological at this point.
We keep inventing better ways to make things and worse ways to share them.
The more the system struggles to distribute what it already produces, the more it invests in producing more—on the assumption that scale will resolve allocation.
It doesn’t.
It increases the volume of what must be distributed without changing who gets it.
Scale just gives you more of the same mess. Congratulations.
Throughput rises. Allocation stalls.
That’s what “gumming up the works” looks like: friction between output and uptake, capacity and demand, existence and claim.
Not collapse. Friction.
Persistent, accumulating friction.
The machine is not broken. It’s working exactly as designed—to seize up when asked to share.
⸻
Externalization
When distribution cannot be resolved internally, it is displaced externally:
· new markets
· new debt structures
· destruction-oriented demand
When you can’t divide the pie, you start wars over who gets the next oven.
This isn’t moral. It’s routing.
If the system cannot decide who gets the output, it finds ways to absorb the excess without answering the question directly.
Externalities are just distribution that hasn’t come home yet.
⸻
The Constraint That Waits
Ricardo’s question—who gets what—turns out to be the part of the system that cannot be abstracted away.
It does not scale automatically.
It does not resolve through innovation.
It does not disappear under better language.
You can rebrand gravity, but you still fall.
It waits.
And when the system can no longer defer it, it reappears not as theory, but as constraint.
The future is not a product. It’s a negotiation you keep losing.
⸻
Closing
The sharper claim isn’t that distribution has been ignored.
It’s that every time a new productive frontier appears, the system treats it as a way to avoid solving distribution—on the promise that scale, structure, or technology will make the question trivial.
Every technological revolution is the same promise: “This time, the pie will slice itself.”
That promise keeps failing.
The computer says no. The market says no. Physics doesn’t care.
The system doesn’t lack output.
It lacks a way to decide who that output is for.
The only thing harder than making a thing is deciding who doesn’t get it.
Leave a Reply