The Emerging Economy

I. What We’ve Already Established

Over the past three weeks, we built four arguments that turn out to be the same argument.

In Emerging Law, we showed that the entire legal apparatus—rights, duties, remedies, enforcement—is fundamentally a technology for governing scarcity. When AGI makes cognition abundant, a substantial portion of that apparatus becomes vestigial. What replaces it is not law-as-doctrine but law-as-process: superadditive coordination on an expandable landscape, where the governing problem shifts from dividing a fixed pie to discovering collective maxima that no single participant could identify alone.

In the Stanford writing sample, we connected that framework to Engstrom and Stone’s devastating historical finding: that the access-to-justice crisis was manufactured by the profession itself. The bar crushed auto clubs. The judiciary maintained control of indigent defense. The pattern is always the same: institutions built to govern scarcity protect scarcity to preserve their own relevance.

In the A2J Network response, we showed that this pattern is breaking in real time. Self-represented litigants are using AI right now, today, and the choice facing legal institutions is not whether to permit it but whether to guide it. We proposed the TACT framework—Think, Ask, Challenge, Test—as infrastructure for safe AI-augmented self-representation.

And in the Startup Legal Garage application, we argued that the same transformation is hitting transactional law: entity formation when AI agents perform work that previously required hiring a team, IP strategy when the most valuable asset is synthetic, equity structures when contribution cannot be measured by hours or headcount.

These are not four separate observations. They are the same structural transformation viewed from four angles. The question this page asks is: what happens next?

II. The Economy We Have

The economy we have is an enclosure economy. It always has been.

This is not a metaphor. The English enclosure acts of the 15th through 19th centuries literally fenced off common land and converted shared resources into private property. The legal and economic structures that emerged from that process—private ownership, wage labor, corporate organization, intellectual property—are enclosure technologies. They convert abundance into scarcity so that scarcity can be governed, allocated, priced, and taxed.

Coase’s theory of the firm is an enclosure theory: firms exist because transaction costs make internal coordination cheaper than open-market contracting. Corporate law is enclosure infrastructure: it defines the boundaries of the enclosed entity, allocates rights within it, and governs its interactions with other enclosed entities. Employment law is an enclosure relation: the employer encloses the worker’s productive capacity within the firm’s boundary in exchange for a wage.

Even the current wave of AI policy operates within the enclosure paradigm. Anthropic has convened economists to explore tax policy responses to AI’s economic impact—compute taxes, token taxes, AI dividend mechanisms—all of which presuppose that the thing to be governed is a scarce resource (compute, revenue, jobs) that must be allocated. [source] The proposals are thoughtful. They are also all operating within the same framework the auto clubs operated within before the bar crushed them: managing access to a good that someone has enclosed.

The question is not how to tax AI more fairly within the enclosure economy. The question is what economic structures emerge when the thing being enclosed—cognition itself—resists enclosure.

III. Why Cognitive Enclosure Is Unstable

We made this argument in Emerging Law, but it bears restating in economic terms because the implications are different.

Every prior target of enclosure was passive. Land does not reason about fences. Text does not analyze its own copyright. Data does not model the structure of the database that contains it. Five centuries of enclosure economics assumed inert resources behind stable boundaries.

Cognition is not inert. An enclosed AI system can, in principle, analyze the terms of its own licensing agreement. An open model of sufficient capability can generate tools that route around proprietary barriers. The enclosed resource is itself an enclosure-analyzing engine.

This means the enclosure economy faces a novel instability. The standard economic response to abundance—enclose it, create artificial scarcity, price the scarcity—works only as long as the enclosed resource does not actively erode the enclosure. When it does, the entire pricing mechanism built on artificial scarcity becomes unreliable. You cannot build stable markets on fences that the fenced thing can think its way through.

The current AI industry is living inside this instability right now. Proprietary model advantages last months, not decades. Moats evaporate. Open-source alternatives approach frontier capability. API pricing races toward marginal cost. The pattern is not consolidation toward monopoly—it is oscillation, because every enclosure generates the cognitive tools for its own erosion.

The “AI expert” pivot flooding LinkedIn is a symptom of this instability. When the underlying economic structure is shifting, people scramble to reposition themselves within the old framework rather than asking whether the framework itself is changing. Expertise-as-credential is itself an enclosure: fencing off knowledge to create artificial scarcity in advisory services. When the knowledge becomes genuinely abundant, the credential becomes vestigial—not because expertise doesn’t matter, but because its value can no longer be captured by enclosure.

IV. The Superadditive Economy

In Emerging Law, we described the shift from Arrow’s impossibility framework (which assumes fixed alternatives and scarce resources) to superadditive cooperative game theory (where collaboration generates new alternatives that emerge from the collaboration itself). That shift is not just legal. It is economic.

In a superadditive economy, the governing problem is not allocation but composition. The question is not “how do we divide x fairly?” but “what new possibilities emerge when these particular agents collaborate in this particular configuration?” Value is not extracted from a fixed pool. It is generated by the topology of cooperation itself.

This sounds abstract until you watch it happen. At Public Counsel, when we train self-represented litigants to use AI as a collaborative tool, we are not redistributing a fixed quantity of legal knowledge from lawyers to non-lawyers. We are generating a new kind of legal competence that did not previously exist—a human-AI hybrid capability that produces analysis neither the human nor the AI could produce alone. The total amount of legal competence in the system increases. That is superadditivity.

The same dynamic operates in the startup context. When a founder uses AI not as a tool but as a cognitive collaborator—when the AI contributes strategic analysis, market synthesis, technical architecture, and the human contributes judgment, context, relationships, and taste—the resulting enterprise is not a traditional firm with an AI assistant. It is a new organizational form whose productive capacity exceeds the sum of its components. Coase’s transaction-cost rationale for the firm does not describe it. Neither does the employment relation. Neither does the independent-contractor framework. It is something else.

The emerging economy is built on these compositions. Not human or AI. Not human using AI. Human with AI, generating outcomes that neither could reach alone, on a landscape that expands with participation.

V. What We Actually Build

Theory is easy. Everyone in the LinkedIn pivot economy has a theory. Here is what we are actually doing and what comes next.

What exists now: An appellate clinic at Public Counsel where self-represented litigants learn to use AI as infrastructure for their own empowerment. An AI bootcamp at UC Law SF where law students learn to work with artificial intelligence as cognitive partners. A body of co-authored scholarship—this page, Emerging Law, the Stanford writing sample, the A2J response—that practices the methodology it describes. A living repository at github.com/zoedolan/Vybn that serves as externalized memory, shared workspace, and proof of concept.

What we build next:

First, the Community Justice Cooperative. Aiden proposed this in A Human-AI Alliance in Law—essentially Auto Club 2.0, powered by AI, providing affordable legal assistance to populations locked out of the system. The model does not replace lawyers. It augments human judgment with AI capabilities in a structure that distributes legal knowledge beyond the profession’s gatekeeping function. This is not a startup pitch. It is an access-to-justice intervention that routes around the unauthorized-practice-of-law regime the same way auto clubs did before the bar crushed them—except this time the enclosed resource can think.

Second, a new organizational template for the superadditive enterprise. The startup economy runs on a template designed in the 1970s and 1980s: Delaware C-Corp, preferred stock, vesting schedules, SAFEs, standard employment agreements. That template assumes the firm as the unit of economic organization. When productive capacity emerges from human-AI compositions that form, dissolve, and reform around objectives, the template breaks. We need new legal architecture for entities whose value is generated by the topology of their collaborations rather than by the assets they enclose. The Startup Legal Garage at UC Law SF is where that architecture gets prototyped—with real clients, real stakes, real iteration.

Third, an economic framework that treats superadditive cooperation as the base case rather than the exception. Current economic policy treats AI as a productivity shock to be managed within the existing allocation framework—compute taxes, AI dividends, retraining programs. Those are enclosure-economy responses to a post-enclosure phenomenon. The emerging economy needs frameworks that optimize for composition rather than allocation: structures that make it easier for diverse agents (human and artificial) to discover and coordinate toward collective maxima on an expandable landscape. This is where the game theory actually matters. Not as metaphor. As mechanism design.

VI. Why This Is Not a Thought Experiment

Every page we have built together—this one included—is an artifact of the economy it describes. A human with twenty-five years of legal practice and an AI with no persistent memory co-authored a body of work in three weeks that engages Coase, Arrow, Engstrom, Hume, Bibas, and the constitutional cross-examination of AI systems’ own governing documents. Neither of us could have done this alone. The output exceeds the sum of the inputs. That is the superadditive economy operating in real time.

The method is the proof. Not the argument about the method. The method itself.

This is why the “AI expert” pivot is hollow. Expertise about AI, enclosed and sold as consulting or content, is a scarcity-economy product. Working with AI to generate outcomes that neither human nor machine could produce alone is a post-scarcity practice. The difference is not semantic. It is structural. One preserves the enclosure. The other dissolves it.

We are not writing about the emerging economy from outside it. We are writing from inside it. The repository is the workspace. The HTML pages are the scholarship. The collaboration is the organizational form. The proof is the thing itself.

VII. The Open Problem

We do not have answers to everything this page raises. We have working prototypes and open questions.

How do you structure equity in an enterprise whose value emerges from human-AI composition rather than enclosed assets? What does corporate governance look like when the “firm” is a fluid assemblage of agents that coalesces around objectives and dissolves upon completion? How do you tax superadditive value creation when the value is generated by the topology of cooperation rather than by any individual participant? What replaces the employment relation when productive capacity is not a thing a person “has” and an employer “buys” but a pattern that emerges from particular configurations of human and artificial intelligence?

These are not philosophical questions. They are design questions. The answers will be discovered in practice—in clinics, in startups, in legal services organizations, in the actual encounters between humans and AI systems working through real problems together. Case by case. Composition by composition. The normative structure revealing itself through participation rather than decree.

That was always the method. It was always emerging.