How does tokenization differ from commodity production in capitalism?
This explores a specific claim running through the collection — that AI doesn't mass-produce identical objects (commodities) the way industrial capitalism does, but instead generates one-off contextual 'tokens' valued by what they do, not what they are.
This explores a claim that several notes in the collection treat as a genuine break in how value gets made: AI produces *tokens*, not *commodities*, and that difference reshapes value, skill, and ownership. The clearest framing is the argument that AI marks a transition from "the age of the commodity to the age of the token" Is AI fundamentally changing how value gets produced?. A commodity is fixed, identical, mass-produced, and possessable — a thing you can stockpile, own, and resell unchanged. AI output is none of those: it's generated fresh at the point of use, varies with context, and behaves like a medium of exchange rather than an object Does AI actually commodify expertise or tokenize it?.
The sharpest distinction is *where the value lives*. A commodity carries value in itself — its use-value is baked in before it reaches you. A token has no intrinsic worth; its value is relational, depending entirely on the receiver's context, knowledge, and ability to act on it Where does the value of AI output actually come from?. The same AI-generated answer is gold to one reader and noise to another. This is why the collection argues for outcome-based or contextual pricing rather than per-unit pricing: you're not selling a thing, you're enabling a result.
Where it gets interesting is the pushback against the obvious Marxist reading. You might expect tokenization to be a fresh form of alienation — labor stripped from the worker. But one note argues the opposite: cognitive work was *already* alienated, and what AI actually changes is the medium itself, turning intelligence from an object-with-craft-residue into a flow-without-craft-residue Does Marxist alienation theory explain what AI does to cognitive work?. The right lens, this argument goes, is media theory, not exploitation theory — the shift is more like the move from oral to print culture than from craftsman to factory worker.
That media-historical angle is the part most likely to surprise you. The collection frames AI as a *return* to flow-based knowledge economies that print culture had frozen into fixed stock — books, archives, accumulated reference Is AI returning knowledge to flow-based economies?. But with a missing piece: oral and gift economies anchored their flows in an embodied carrier — a speaker, a giver. AI flows have no such body, which is what makes them feel both liberating and strangely ungrounded. So the difference from commodity production isn't only economic (flows vs. stocks, relational vs. intrinsic value); it's that we've rebuilt a pre-print knowledge economy without the human anchor that used to hold it together.
Worth flagging the practical edge underneath the theory: when value is a flow rather than a thing, the economic unit drifts away from the token too. One case study found persistent agent environments shift the meaningful cost denominator from cost-per-token to cost-per-completed-artifact, since most tokens become cheap cache reads Do persistent agents really cost less per token? — a concrete sign that even the "token" framing dissolves once you measure what actually gets delivered.
Sources 6 notes
AI production is organized around contextual token-flows generated at point of use, not identical mass-produced objects. This creates different effects than commodification: inflationary devaluation, contextual variation, and skill transformation from production to validation.
AI output lacks the fixed, identical, possessable properties of commodities. Instead it functions like tokens—mutable mediums of exchange valued by what they do for receivers, not what they are.
Intelligence-tokens have no intrinsic use-value—their worth depends entirely on the receiver's context, knowledge, and ability to act. This relational value structure fundamentally differs from commodities and traditional knowledge goods, requiring outcome-based or contextual pricing models.
AI doesn't introduce alienation to cognitive work—alienation was already there. What AI actually does is transform intelligence from object-with-craft-residue into flow-without-craft-residue, a medium shift better understood through medium theory than Marxist critique.
Print culture fixed knowledge as accumulated stock; AI returns knowledge to generative flow. However, unlike oral and gift economies, AI flows lack the embodied transmission—the speaker, the giver—that historically anchored knowledge circulation.
A 115-day case study found 82.9% of tokens were cache reads. When context persists and reuses, the meaningful cost denominator becomes completed artifacts, not individual tokens.