Resisting Computation as the Final Measure: Value, Tokenization, and the Limits of Calculability

This essay develops one of NVI’s central theoretical concerns: what happens when computability becomes a dominant criterion of value.

Beginning from token economics and AI infrastructure, the text asks how intelligence, judgment, meaning, and human activity are transformed when they are rendered into countable units. It positions tokenization as a new value regime, one that extends capital logic into cognition, language, and intelligence production.

The essay connects this concern to Jingyi Wang’s artistic practice: Post Capitalistic Auction addresses value in exchange, JUDGE ME addresses value in judgment, Voice to Voice addresses value in formation, and The Butterfly Entangled Effect extends the question into multi-agent conflict, ecological systems, and AI-mediated environments.

What increasingly concerns me in the current AI discourse is not only the technical development of large models, but the emergence of a new regime of value. In current discussions of token economics, the issue is not simply that tokens are a technical unit for processing language, images, or other data. The deeper issue is that tokenization is becoming a new way of measuring, pricing, optimizing, and legitimizing intelligence itself.

In this framework, complex cognitive activity is decomposed into countable units; intelligence becomes throughput, latency, cost per token, and revenue per token. What is being proposed is not merely a new technical language, but a new economic and epistemic infrastructure through which intelligence is rendered calculable.

This matters because it is not neutral. What is at stake is not only whether something can be tokenized, but what happens to it when it is. Tokenization does not simply capture intelligence, judgment, or meaning. It transforms them into what can be computed. It preserves what can be discretized, counted, compared, and optimized; at the same time, it risks flattening or excluding those dimensions of value that are slower, contradictory, context-dependent, unresolved, or irreducibly subjective.

In this sense, the central question is not: Can subjective value be computed? The more important question is: What happens to value when computability itself becomes the dominant criterion of legitimacy?

This is why I do not see token economics merely as a business model. I see it as a new value regime. It is a system that attempts to transform complex, layered, and heterogeneous forms of intelligence into a calculable plane. It is, in effect, a new operational form of capitalization in the AI era: not a departure from capital logic, but an extension of it into cognition, language, judgment, and intelligence production itself. Tokenization, then, is not the opposite of capitalization. It is one of its newest forms.

This is where the problem of computationalism becomes crucial. The deeper issue is not simply that many scientists, technologists, or mathematically minded thinkers believe that everything can eventually be formalized and modeled. The issue is that in a computation-centered regime, those things that cannot yet be stably formalized, operationalized, or optimized are increasingly treated as secondary, vague, soft, or not fully real. They may not be denied outright, but they are structurally downgraded.

In other words: The problem is not only the belief that everything can be computed. The problem is that computability itself is becoming a hegemonic criterion of value.

This is where my work enters. Across my projects, I have been approaching this problem from different directions. Post Capitalistic Auction examines how value is measured, converted, chosen, and exchanged. It asks whether value can be proposed and recognized beyond price alone. In that sense, PCA is already a critique of dominant value metrics and a test of alternative forms of justification.

JUDGE ME examines how social value systems become internalized as structures of self-judgment. It asks how external systems of valuation enter the subject and begin to function as inner adjudication. Here, the issue is no longer only exchange, but legitimacy, internalized conflict, and the violence of judgment.

Voice to Voice asks how value and value systems are formed. Through inquiry, testimony, and listening, it turns to the social and experiential conditions through which value becomes thinkable, speakable, and lived. It returns value to the domain of testimony and human formation rather than treating it as a fixed metric.

The Butterfly Entangled Effect extends this inquiry beyond the human social frame. It asks what happens when human, ecological, and artificial value systems are forced into the same dynamic field. Here, value is no longer only a matter of exchange, judgment, or listening within human institutions. It becomes a multi-agent problem.

Different agents -- human beings, ecological systems, non-human life forms, and AI-mediated systems -- do not necessarily share the same value functions. They may operate according to incompatible logics of survival, optimization, reproduction, adaptation, or meaning. The project therefore asks: What happens when incompatible value systems enter into conflict? What forms of negotiation or emergent system dynamics arise between heterogeneous agents? Can conflict itself become a condition for the emergence of new value logics?

This is why the AI dimension in The Butterfly Entangled Effect is essential. I am not interested in AI merely as a tool for representation or effect. I am interested in what happens when AI is positioned as a potential agent within a designed value field and exposed to contradictory or opposing value assignments. The question is no longer only what humans should encode into AI. The issue becomes whether conflict among heterogeneous value systems might produce unexpected behavior, emergent system dynamics, or even new ways of describing value functions.

This is a very different question from mainstream AI alignment discourse. Most alignment discourse still assumes that the task is to correctly define, translate, or stabilize human values so that AI can optimize them. But what if the problem is not that AI has not yet received the right values? What if the deeper problem is that there is no single coherent value system to begin with? What if value is always already plural, conflictual, situated, and contested across different agents and scales? And what if, under such conditions, emergence is not an error but a constitutive possibility?

This is why I am not simply asking whether AI can calculate value. I am asking whether the very attempt to render value calculable is already transforming what value is. And this, in turn, is why I am wary of computational sovereignty: the idea that what can be computed can eventually become the final arbiter of reality, intelligence, and worth.

My concern is not anti-technical, nor is it a romantic rejection of formal systems. I am deeply invested in structure, mechanism, protocol, and design. But I am equally concerned with the limits of formalization: with what gets lost, distorted, or violently reduced when calculability begins to function as the dominant standard of legitimacy.

So the real question is not whether everything can be tokenized, modeled, or computed. The real question is: What kinds of value are only recognized after they have been flattened into computable form, and what kinds of value disappear in that process?

This is the problem I am exploring across these works in different ways. PCA addresses value in exchange. JUDGE ME addresses value in judgment. Voice to Voice addresses value in formation. The Butterfly Entangled Effect addresses value in multi-agent conflict, game, and emergence. Taken together, these projects are not simply about value in the abstract. They are an inquiry into how value is produced, legitimized, internalized, contested, and transformed -- and into whether the future of value will be governed by calculation alone, or whether new forms of value might emerge precisely where calculability fails to exhaust the real.