The Counter-Thesis

TAM-CV.12 · The Capital View · The Approximate Mind

This arc has been written from the position of capital. The PE firm, the operating partner, the deal structure, the exit math. The view from Marcus’s office, where the trawler sits on the windowsill and the numbers describe a transition that is being financed, deliberately, by people who see its shape clearly and are writing checks based on the answer.

The view from capital is clear and structurally precise and honest about what it sees. It is also incomplete in ways that capital’s own instruments cannot measure. This essay attempts to hold the counter-thesis: the honest case against capital’s position, made from outside capital’s frame, using evidence that capital’s instruments are not designed to capture.

The counter-thesis is not that capital is wrong. It is that capital is correct within a frame that excludes the variables most likely to determine the outcome.

Capital’s Case, Stated Fairly

#

The strongest version of the capital argument, the one Marcus would make if pressed, runs like this.

The coordination problem is real. Mid-market firms carry management layers that are expensive, imperfect, and now addressable by AI. Someone is going to address them. If PE does not deploy the management strip, someone else will: a VC-backed platform, a Big Tech feature, a competitor who saw the math first. The management layer is going away regardless of who removes it. The question is whether the removal is organized thoughtfully, with severance and transition planning and operational continuity, or whether it happens chaotically, one company at a time, as individual firms adopt AI coordination tools without any institutional support for the people displaced.

PE’s argument is that organized displacement is better than chaotic displacement. Not better in the sense of painless. Better in the sense of structured, predictable, and manageable. Kevin gets eight weeks of severance and a clear timeline. Kevin in a company that adopts AI coordination on its own gets a gradually diminishing role and an eventual layoff with no institutional support at all.

This argument is not cynical. It is, within its frame, defensible. The transition is structural and demographic and technological. It is not caused by PE. PE is organizing a transition that would happen without it. The value that PE captures is the organizational premium: the difference between structured and unstructured displacement, priced through the fund’s return.

The argument has a second leg. The companies PE transforms become more competitive. Their cost structures are lower. Their operations are more consistent. Their ability to serve customers improves because the AI coordination layer handles scheduling, inventory, and routing better than the human management layer did. The economy, in aggregate, functions better when coordination overhead is reduced across thousands of mid-market firms. The gains are real. They accrue to customers, to the remaining employees whose jobs are more productive, and to the economy through lower prices and better service.

PE would also note, if pressed hard enough, that the cooperative alternative has a track record. It is mostly a track record of failure. Worker cooperatives have existed for two centuries. They remain a marginal economic form. Mondragon is the exception, not the rule. The typical cooperative fails within five years, dies of governance dysfunction, undercapitalization, or the simple exhaustion of trying to do collectively what a hierarchy does through authority. The cooperative is a beautiful idea that keeps breaking on the same rocks.

Capital’s case is not that cooperatives are wrong. It is that cooperatives do not scale, and the transition will not wait for them to learn how.

This is the strongest version. It deserves a serious response.

The Cooperative’s Case, Stated Fairly

#

The cooperative’s case does not begin with economics. It begins with a question that capital’s instruments cannot formulate.

Who should benefit from the removal of the management layer?

Capital’s answer is implicit in its structure: the investors who funded the removal. This is not an answer anyone deliberated on. It is a consequence of the ownership structure. The fund owns the company. The fund deploys the AI layer. The fund captures the margin. The question of who should benefit was answered by the capital structure before anyone thought to ask it.

The cooperative’s answer is explicit and deliberate: the people who do the work. The margin that management consumed is returned to labor because labor owns the enterprise. This is not an unintended consequence of the ownership structure. It is the reason for the ownership structure.

The distinction between implicit and explicit answers to distributional questions is the cooperative’s deepest argument. Capital distributes value according to its structure. The cooperative distributes value according to its design. Structure does not require justification. Design does. And the requirement to justify the distribution, to argue about it in meetings with plastic chairs and insufficient ventilation, is precisely what makes the cooperative a more honest institution than the fund.

The cooperative has a second argument that is economic rather than moral, and it is the one that capital’s analysts should find most unsettling.

Alignment.

In the PE model, the fund’s interests and the workers’ interests coincide when the company is growing and diverge when it is not. The fund optimizes for exit value. The workers optimize for stable income. When the fund deploys the management strip, the divergence is explicit: the value creation comes from eliminating positions. The remaining workers know this. They know that the same logic that removed Kevin could remove them if the AI layer’s capability improves. They work in an environment where the technology that makes their work better is also the technology that might make their work unnecessary. The awareness is corrosive. It does not appear in the EBITDA analysis. It appears in the quality of attention the worker brings to Tuesday afternoon, in the willingness to flag the problem the system did not catch, in the difference between doing the job and caring about the job.

In the cooperative, the AI that removes the management layer is owned by the workers. The savings flow to them. The technology that might extend to their roles is governed by them. The decision about how far the automation goes is theirs to make. They might make it badly. They might resist automation that would improve their operations because the automation threatens their jobs. But the resistance is a governance question, resolved through their own process, not an imposition from an owner whose interests diverge from theirs.

The cooperative’s alignment advantage is not sentimental. It is structural. Aligned organizations produce better outcomes because the people inside them have reason to care about the outcomes.

There is evidence for this. Mondragon’s productivity rates are comparable to conventional firms in the same industries. Worker cooperative survival rates beyond year five, for cooperatives that survive the initial governance learning curve, are higher than conventional firm survival rates. The evidence is not overwhelming. The sample sizes are small. The confounding variables are enormous. But the directional signal is consistent: aligned organizations, when they survive their governance challenges, perform at least as well as conventional firms and often better.

Capital’s response is that the governance challenges are the point. Most cooperatives do not survive them. The evidence that surviving cooperatives perform well is a survivorship bias that tells you nothing about the cooperatives that failed, which is most of them.

This is fair. It is also incomplete.

What Scale Means

#

The cooperative’s historical failure rate reflects the conditions under which cooperatives have historically formed, and those conditions are about to change.

Previous cooperatives faced two structural disadvantages that the AI coordination layer eliminates.

The first was the coordination cost of collective governance applied to operational management. Every decision that a manager would make unilaterally had to be made collectively, which meant meetings, debates, votes, and the grinding overhead of democratic process applied to questions that did not require it. Should we change the supplier? Should we adjust the shift schedule? Should we buy the new equipment? Each question consumed collective attention that could have been spent on the work itself.

The AI coordination layer handles operational management. The supplier decision is informed by price-quality analysis across the network. The shift schedule is optimized. The equipment decision is modeled against production data. The collective does not need to debate these questions because the AI handles them competently. What the collective governs is the strategic direction, the surplus distribution, the values that the AI’s optimization serves. The governance burden drops from everything to the things that matter.

This is a genuine structural change in the viability of the cooperative form. The historical failure rate of cooperatives reflects a world where collective governance was applied to operational management. In a world where AI handles operational management, collective governance is applied only to strategic direction. The burden is different. The failure mode may be different. The historical track record may be a poor predictor of what is now possible.

The second historical disadvantage was capital formation. Cooperatives had difficulty raising capital because institutional investors are structured to invest in entities that generate returns for external shareholders, and cooperatives by design do not have external shareholders. The capital available to cooperatives was limited to member contributions, retained earnings, and the handful of patient capital sources willing to accept cooperative governance structures.

This disadvantage has not changed. It has, if anything, deepened. As PE and VC move into the coordination economy, the capital available for proprietary platforms grows while the capital available for cooperative infrastructure does not. The asymmetry compounds.

But the capital requirement has changed. The AI coordination layer is not a factory. It is software. The cost of deploying a coordination platform across a cooperative drops every year as the underlying AI technology becomes cheaper and more capable. Sunita’s fourteen crore rupees, roughly $1.7 million, funds a pilot for an entire manufacturing cluster. The capital required to establish cooperative coordination infrastructure is a fraction of what it was five years ago, and it will be a fraction of the current fraction five years from now.

The cooperative’s historical failure at scale reflects conditions that are changing. The question is whether the conditions are changing fast enough.

The Three Futures

#

There are three plausible outcomes, and the honest assessment is that none of them is clearly more probable than the others.

The first future is capital dominance. PE, VC, and Big Tech establish proprietary ownership of the coordination infrastructure within the current investment cycle. The mid-market economy runs on private coordination platforms. Cooperatives exist at the margins, the way they exist now, as admirable but marginal alternatives to the dominant form. Kevin’s severance runs out. Kevin’s children enter a labor market where the management layer has been permanently removed and the savings flow to the owners of the platforms that replaced it. The toll booth economy, described in TAM-033 and TAM-051, persists in a new form: not the intermediary’s toll but the platform’s subscription.

This future is the most likely if nothing changes. Capital has money, speed, institutional infrastructure, and the gravitational advantage of existing market structure. The cooperative alternative requires active construction. The capital version happens by default.

The second future is bifurcation. Capital dominates in some markets and cooperatives establish themselves in others. The United States, with its private digital infrastructure and its weak cooperative tradition, develops a capital-dominated coordination economy. India, with its public digital rails and Sunita’s line item, develops a cooperative coordination economy. Europe, with its stronger cooperative tradition and its regulatory willingness to constrain platform monopolies, develops a hybrid. The coordination economy is not one thing. It is several things, shaped by the institutional context in which it develops.

This future is plausible because institutional contexts are genuinely different. The race between capital and cooperatives is not conducted on a single field. It is conducted on multiple fields with different rules, and the outcome on each field is determined by the rules as much as by the competitors.

The third future is the one that neither capital nor the cooperative movement is prepared for. Big Tech wins by default. The coordination function becomes a feature of general-purpose AI platforms rather than a specialized product or a cooperative infrastructure. Microsoft, Google, and Amazon embed operational coordination into their existing platforms. The mid-market company does not choose a coordination tool. It uses the one that is already in its stack. PE’s specialized platform is competed away. The cooperative’s independent infrastructure is outperformed. Both lose to the gravitational pull of platforms so large that the coordination economy is a rounding error in their revenue.

In this future, the question of who owns the coordination layer is answered by technology companies that were not trying to answer it. The enclosure is incidental. The toll is embedded. The alternatives are not defeated. They are rendered unnecessary by a platform that provides the function for free, or nearly free, as a feature designed to increase the stickiness of a larger subscription.

This is the future that should worry everyone, because it is the future in which the question of ownership never gets asked. The coordination economy is absorbed into the platform economy before anyone builds an alternative.

What Marcus Cannot See

#

Marcus is smart, careful, and honest within his frame. His frame is deal structure and return modeling and the specific mechanics of value creation in fragmented industries. Within that frame, his analysis is correct. The management strip works. The platform play is sound. The exit math is compelling.

What his frame excludes is the variable that the Coordination cluster placed at the center of its argument: what the structure does to the people inside it, measured by their own standards.

Dale wants good routes and staged parts and nobody between him and the work. The AI gives him that. Kevin wants a stable income and a role that uses his skills. The management strip eliminates both. Charlene wants a paycheck that matches what she was making before the plant closed and a place for the skill in her hands. The cooperative gives her that. Nina wants to belong to something. The assembled workforce gives her work but not a workplace.

The Capital View has been honest about what capital sees. It has also been honest, I hope, about what capital does not look at. The blue mug. Kevin’s mortgage. The difference between Charlene’s tolerance and the system’s tolerance, between adequate and good. The photograph on the refrigerator. The meetings with plastic chairs. The arguing that is theirs to do.

Capital does not model these things because its instruments cannot measure them. The investment memo has no line item for belonging. The hundred-day plan has no section on what happens to the operations manager’s mortgage after the severance runs out. The exit analysis has no variable for whether the people inside the transformed company have reason to care about what they are doing.

These are not soft concerns appended to the real analysis. They are the variables most likely to determine which model produces the better outcome over time. Alignment, belonging, care, the willingness to flag the weld defect that is within the system’s tolerance but not within yours: these are the inputs that the cooperative’s structure is designed to produce and the capital structure is designed to ignore.

Whether the cooperative’s structural advantage in producing these inputs is large enough to overcome its structural disadvantage in speed and capital formation is the question this arc cannot answer. It is the question that the next decade will answer, deal by deal, cooperative by cooperative, line item by line item.

The Trawler

#

Marcus bought the trawler in Portugal twenty years ago. He has never explained why. He has moved it through four offices and recently moved it from his desk to the windowsill, where the afternoon light catches it and where he would have to turn around to see it.

I have not asked him about the trawler. It did not seem like a question he would welcome. But at the end of our last conversation, the one where he said “I am building the one someone owns” and then paused and looked at the windowsill, I asked him something different.

I asked whether he had ever considered building the other one. The one nobody owns. The cooperative version. The coordination infrastructure as a commons rather than a product.

He was quiet for longer than usual.

“I don’t know how,” he said. “Everything I know how to do is organized around ownership. Capital structure. Governance rights. Exit. The entire vocabulary of what I do assumes that someone owns the thing and someone else uses it.”

He paused again.

“If I built the commons version, I wouldn’t know how to finance it. I wouldn’t know how to govern it. I wouldn’t know how to make it work at the scale the problem requires. I would be starting from nothing in a domain I don’t understand, competing against people who are building the private version with every advantage I just described.”

He looked at the trawler.

“But I notice I keep thinking about it.”

He did not say more. He picked up his phone. The conversation was over. I left his office and walked to the elevator and thought about what it means that the person best positioned to build the capital version of the coordination economy keeps thinking about the other one.

It does not mean he will build it. It means the question is alive in the room where the decisions are being made, even when the decisions do not reflect it. It means the counter-thesis is not external to capital. It is internal to it, held by the people inside the structure, visible in the moments when the frame relaxes and the thing behind the frame becomes briefly available.

Whether that is enough to change the outcome is not something I can know. The race is underway. The window is finite. The people inside both structures are building what they know how to build, as fast as they can, with the tools and the vocabularies and the institutional architectures they have.

Charlene is on the factory floor. Her ear for defective welds has not atrophied. The cooperative argues about surplus distribution and the things that people argue about when the arguing is theirs to do.

Marcus is in his office. The platform is being deployed across the portfolio. The returns are on track.

The trawler is on the windowsill, between them, catching the light.

This is the twelfth and final essay in The Capital View, a twelve-essay arc examining the AI transition from the position of capital. The original nine essays (TAM-CV.01 through TAM-CV.09) examined how capital organizes the transition from human-delivered to AI-orchestrated services. The extension (TAM-CV.10 through TAM-CV.12) examined how capital responds when the coordination function itself becomes automatable and ownable. This essay holds the counter-thesis: the cooperative model’s structural advantages that capital cannot replicate, and the genuine uncertainty about which model prevails. It does not resolve the question because the question is not yet resolved. The race between private and collective ownership of the coordination economy is the most consequential structural question in the AI transition, and its outcome depends on speed, governance, policy, public infrastructure, and whether the people building the capital version keep thinking about the other one. This essay connects to the owned factory in TAM-RIM.6-04; to the lock and the unlock in TAM-RIM.6-SYN; to the government question in TAM-RIM.6-08; to the assembled workforce in TAM-RIM.6-06; to the blue mug in TAM-CV.05; to the enclosure of coordination in TAM-CV.07; to the toll booth economy in TAM-033 and TAM-051; and to the distillation thesis in TAM-072.

References

#

Cooperatives: Theory, Evidence, and History

Cheney, George. Values at Work: Employee Participation Meets Market Pressure at Mondragon. Cornell University Press, 1999.

Dow, Gregory K. Governing the Firm: Workers’ Control in Theory and Practice. Cambridge University Press, 2003.

Pencavel, John. “Worker Cooperatives and Democratic Governance.” Handbook of Economic Organization, edited by Anna Grandori, Edward Elgar, 2013.

Whyte, William Foote, and Kathleen King Whyte. Making Mondragon: The Growth and Dynamics of the Worker Cooperative Complex. Cornell University Press, 1988.

Capital, Ownership, and Distribution

Mazzucato, Mariana. The Value of Everything: Making and Taking in the Global Economy. PublicAffairs, 2018.

Piketty, Thomas. Capital in the Twenty-First Century. Translated by Arthur Goldhammer, Harvard University Press, 2014.

Stout, Lynn A. The Shareholder Value Myth: How Putting Shareholders First Harms Investors, Corporations, and the Public. Berrett-Koehler, 2012.

Platform Economics and Enclosure

Boyle, James. The Public Domain: Enclosing the Commons of the Mind. Yale University Press, 2008.

Ostrom, Elinor. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press, 1990.

Scholz, Trebor. Platform Cooperativism: Challenging the Corporate Sharing Economy. Rosa Luxemburg Stiftung, 2016.

AI, Labor, and Institutional Change

Acemoglu, Daron, and Simon Johnson. Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity. PublicAffairs, 2023.

Autor, David H. “Work of the Past, Work of the Future.” AEA Papers and Proceedings, vol. 109, 2019, pp. 1-32.

Sen, Amartya. Development as Freedom. Knopf, 1999.