Unit Cost Dominance is the economic condition where an AI, often with minimal human oversight, can produce a unit of cognitive work at a significantly lower marginal cost than a human. This isn't about the cost to train the AI, but the cost to run it for a task—just as you don't care about Microsoft's R&D budget for Excel, only your cost to make a spreadsheet.
An AI drafts a 20-page contract in 30 seconds. A senior lawyer (the "verifier") spends 10 minutes reviewing it for accuracy. The core creation work is replaced, leading to a 50-100x reduction in unit cost. One elite verifier can oversee the output of a system that replaces dozens of junior associates. This is replacement.
A plumber uses an AI-powered diagnostic tool that identifies the source of a leak in seconds. The plumber still performs all the physical labor of fixing the pipe. The AI makes the human more efficient, perhaps 1.5-2x faster. The human's skills are enhanced, not replaced. This is augmentation.
Now, you can choose.
AI + human verification is now 20–50x cheaper than human-only cognitive work. In a competitive market, adoption is not a choice—it's an inevitability.
*Illustrative costs based on median cognitive wage vs. API/compute costs + fractional verifier time.
In the world of bits (code, reports, analysis), the AI performs 100% of the creation. The human becomes a mere verifier, overseeing vast AI output. One expert can replace an entire team, creating an exponential cost advantage. This is mass replacement.
Some domains resist full automation because verifying *is* the job. A surgeon still cuts, a judge still rules. AI removes friction—processing scans, surfacing precedent—making the expert 2-5x faster. The core act remains human, but the throughput is amplified.
In the world of atoms (plumbing, construction), the human still performs 100% of the physical labor. An AI is just a powerful tool—diagnosing a pipe, overlaying a schematic. It augments the plumber, making them perhaps 50% more efficient. This is an incremental improvement.
This hybrid model is key because it neutralizes the two biggest objections to AI: hallucinations and impersonality. The human verifier's role is to catch errors and add the necessary nuance, empathy, and human touch. This creates a system that is both ruthlessly efficient and reliably human-centric, powering all three tiers of dominance.
To refute Unit Cost Dominance, you must falsify the following:
Coordination is impossible. This is not one prisoner's dilemma; it is billions of them played out daily at the international, corporate, and individual level. Each actor's rational choice to adopt AI leads to collective ruin.
The boundary problem is a Sorites paradox. At what point does helpful AI assistance become total human replacement? Like removing single grains of sand from a heap, no individual step feels decisive, yet the endpoint is absolute. You cannot regulate billions of these daily dilemmas when the line between cooperation and defection is literally undefinable. There is no boundary to police, only a gradient into obsolescence.
The wage-demand circuit breaks. Production without wages means products without consumers. This is not a recession; it is a structural failure.
The collapse of the wage circuit leads to the collapse of productive participation. The majority of humans can no longer contribute economically valuable labor.
Consumption-supporting mechanisms like Universal Basic Income (UBI) do not solve this. They formalize it. You are not liberated; you are managed. When your economic function is eliminated, you become a cost to be optimized, a pet to be fed.
Without productive participation, we become zoo animals in a comfortable, algorithmically-curated enclosure.
Unit Cost Dominance does not democratize; it stratifies. Value concentrates at the apex of judgment, leaving the majority behind.
This pyramid itself is a temporary structure. The verifier's role is a trap: every correction they make is training data that teaches the next-generation AI to make fewer mistakes. The machine learns from its auditor, eventually consuming the verifier's function. The pyramid doesn't just shrink; it collapses into a single spike where even the elite are made obsolete.
The psychological impact of accurate perception. Society victim-blames, telling you to "just retrain"—as if it's your fault. But retraining against an exponential curve is futile. Your anxiety is not a disorder; it is a correct response to the data.
"I was told to get a degree. Now my degree is worth less than a $20/month AI subscription."
"My entire career path was just a temporary lag between two versions of an algorithm."
"I played by the rules, but the game ended before I could cash in my chips."
"Why can't they just work hard like we did? The system worked for us."
The political system cannot address the invisible force of Unit Cost Dominance. It must invent a visible enemy. Your justified rage is harvested and redirected.
Knowing you are in a trap does not grant you the tools to escape it. Our political and social tools were designed for the 20th-century conflict of Labor vs. Capital. Now, Player 3—the machine—has entered the game. It does not respond to protests, votes, or legislation. It only responds to the logic of Unit Cost Dominance.
The Discontinuity Thesis is internally airtight. P1 guarantees that AI+human workflows—whether via verification or augmentation—will always undercut pure human labor, collapsing costs by orders of magnitude. P2 guarantees that no coordination regime can hold the line, because the payoff matrix makes defection survival and cooperation death, while the Sorites paradox makes definition not just fuzzy but logically undefinable, even before enforcement. Together, they don't just shut the escape hatches; they seal them with concrete from the outside.
The only uncertainties are external shocks—compute limits, geopolitical fracture, or systemic collapse—that can delay but never reverse the mechanics. The wage-demand circuit of post-WWII capitalism is not fragile, it is finished.
Unit Cost Dominance doesn't just end jobs. It structurally dismantles the core pillars of post-WWII capitalism.
The system was defined by mass employment, not just mass consumption. UBI and dividends are system replacement, not survival. They are welfare for a population that no longer has economic function.
Political power for the masses was a downstream effect of their economic leverage. When over 50% of the population depends on transfers from a tiny AI-owning elite, their votes become performative. You cannot vote your way out of dependency.
The promise of "Cognitive Ladder Restoration" is a fantasy. The cognitive ladder itself is gone. Previous revolutions automated muscle, pushing humans into mind-work. AI automates mind-work. There is no next rung for the masses to climb.
The market for mass human cognitive labor ceases to exist. AI's marginal cost becomes the new wage floor—a floor no human can survive on. Competitive wage-setting, the engine of the labor market, dies.
Horses weren't exterminated. They were just no longer worth feeding.