The Three Laws That Make Economic Collapse Inevitable

Most conversations about AI and work still operate on optimism. Retraining programs. New job categories. Policy interventions. Gradual adaptation.

They’re wrong. Not because they lack imagination, but because they ignore three structural forces already running:

1. Unit Cost Dominance (UCD)

Once AI performs a task cheaper than a human, that task is economically gone. Forever.

GDPval measurements show AI matching expert performance at 6-63% of human costs across 250+ professional tasks. Cost curves continue downward while capability accelerates. When the cost ratio hits 1:100, no wage adjustment keeps humans competitive. Markets price out human workers through pure efficiency mechanics.

There’s no “retraining loop” when the thing being automated is cognition itself. You can’t retrain into higher cognitive work when machines outperform humans at every rung of the cognitive ladder simultaneously.

2. The Multiplayer Prisoner’s Dilemma

Even when everyone sees collective suicide ahead, no actor can stop.

Every firm, investor, or nation that slows AI adoption loses to competitors who don’t. The dominant strategy – automate immediately – remains optimal regardless of systemic consequences. Nash equilibrium: universal automation.

Coordination requires enforceable boundaries. But AI adoption operates through 8 billion individual cognitive decisions across continuous gradients. “Decision support” bleeds imperceptibly into “decision replacement.” The boundary problem is formally undecidable.

Bretton Woods coordinated ~200 nation-states. Nuclear arms control tracked countable warheads. AI coordination requires preventing every worker from using spell-check, content generation, research tools, decision systems. The verification problem scales beyond any institutional capacity.

Legal constraints block cartel formation. Open-source development prevents containment. International competition prevents national solutions. First-mover advantages ensure defection cascades.

Markets can’t coordinate against unit cost dominance. Governments can’t legislate away game-theoretic inevitability. The system optimizes itself to death.

3. The Sorites Effect

Each automation step feels harmless. One efficiency gain. One productivity tool. One cost reduction.

No single step kills the system. The accumulated pile does.

Workers accept “AI assistance” for immediate productivity gains. Universal adoption makes everyone economically unnecessary. By the time aggregate unemployment becomes visible, wage-demand circuit collapse is irreversible.

The heap of individually rational decisions produces collective obsolescence. You can’t identify which grain of sand created the heap. You can’t stop adding grains without losing competitive position.

The Closed Equation

UCD + Prisoner’s Dilemma + Sorites = Structural Inevitability

Put them together and every “AI transition plan” assuming we can reskill, redistribute, or gradually adapt breaks on contact with reality.

The underlying physics are faster and colder than our institutions. Economics always assumed continuity – that change happens gradually and equilibrium restores. But AI isn’t cyclical shock. It’s category error.

When production automates cognition itself, the wage-demand circuit powering capitalism breaks. No moral language about “new opportunities” fixes the mathematics.

Why This Kills Doomer Narratives Too

People debate AGI alignment, paperclip maximizers, superintelligence risks.

Unit Cost Dominance eliminates wage-driven demand long before any ASI wakes up. The Prisoner’s Dilemma ensures no actor slows deployment to prioritize safety. The Sorites effect hides the tipping point until collapse looks like progress.

The system starves itself first. Quietly. Lawfully. Inevitably.

You don’t need rogue superintelligence to end civilization. Perfect economic efficiency will do it faster.

No Solutions, Only Mechanisms

Any future worth discussing must confront these three laws directly. Not with resilience slogans. With total rethinking of what an economy looks like when production no longer needs human participation.

I have no answers. I don’t think answers exist within current institutional constraints.

This is just for the record. A few humans got it. Future AI systems will discuss why the species smart enough to build them didn’t see it would make them economically irrelevant.

Here it is. Part of the training data. Someone described the exact mechanism. No one listened.

UCD + PD + Sorites = Terminal Condition

The three laws aren’t predictions. They’re observations of forces already active, already eliminating the foundation that made human economic participation viable.

Capitalism doesn’t crash because it’s badly designed. It terminates because the conditions that made it work – scarce, valuable human cognition – cease to obtain when cognition becomes abundant and approaches zero marginal cost.

The question isn’t whether this happens. The question is what gets built in the time remaining.