The Discontinuity Thesis: The Only Coherent Economic Model for the AI Era

Every major AI discussion today suffers the same flaw:
people keep trying to fit an unprecedented technological event into familiar economic narratives.

We hear comforting lines like:

• “AI will eliminate some jobs, but new jobs will emerge.”
• “We’ve survived automation before.”
• “Regulation will ensure humans remain central.”
• “Society will adapt just like in the Industrial Revolution.”

These are soothing stories, not serious analysis.

AI isn’t another tool.
AI is cognition at scale.
And cognition is the foundation of every white-collar industry, every decision-making process, and every economic function we associate with “work.”

This is why only one framework actually matches the incentives, the math, and the reality of competitive global markets:

The Discontinuity Thesis (DT).

It’s the first—and so far only—coherent economic model that explains what happens when cognitive automation reaches cost dominance across entire sectors.

Below, I explain what the Discontinuity Thesis says, why its core logic is airtight, and how it compares to the “next best” explanation mainstream economists cling to. By the end, you’ll understand why almost every other theory is simply delaying the inevitable conclusion:
capitalism as we know it does not survive the arrival of broadly capable AI.

1. Unit Cost Dominance: The Engine of the Discontinuity Thesis

The DT is built on one brutally simple observation:

When AI can produce cognitive output cheaper, faster, and at higher quality than humans, firms will defect toward AI in every domain.

This is not philosophy.
This is microeconomics 101.

A rational firm always chooses the production method with the lowest marginal cost.
And AI—particularly when paired with a single human verifier—drives marginal costs of cognition toward zero.

This shift is what the DT calls Unit Cost Dominance (UCD).

Here’s why UCD matters:

• AI does not need breaks.
• AI does not need salaries, healthcare, or pensions.
• AI does not plateau in skill.
• AI gets better with every model upgrade.
• AI can scale infinitely and instantly.
• AI has near-zero replication cost.

When a production method hits these characteristics, substitution is not optional—it is forced.

If one firm in an industry automates its cognitive labour, the rest must follow or die.
This is competitive defection.
It is unavoidable.

Once UCD is triggered in a domain, human labour cannot stand on the production frontier.

This is the first pillar of the Discontinuity Thesis, and nobody has produced a remotely credible rebuttal to it.

2. The DT’s Most Important Insight: Capitalism Requires Productive Participation

Most economists debating AI fixate on jobs.
They argue about displacement, new categories, reskilling, and labour market churn.

But the DT doesn’t focus on jobs.
It focuses on productive participation—the actual engine of the modern economic system.

Here’s the distinction people miss:

• System survival means the majority of adults perform economically valuable labour.
• System replacement means the majority of adults do not produce, but continue to consume via redistribution.

These two states are completely different systems.
If productive participation collapses, capitalism collapses.
You can keep everyone fed with dividends or UBI, but you have not preserved capitalism—you’ve replaced it with something else.

The Discontinuity Thesis makes the point clearly:

Post-WWII capitalism is defined by mass participation in production, not just consumption.
Once AI outcompetes human cognition, mass participation dies.

Attempts to preserve consumption through redistributive mechanisms do not save the system; they simply mask its death.

This redefinition is key, because it exposes the central flaw in the “everything will be fine” stories: they implicitly downgrade capitalism to mere distribution, ignoring the structural necessity of productive human labour.

3. Why Coordination, Regulation, or “Human-Only Zones” Can’t Save Us

Every optimistic counter-argument relies on some version of:

• “We’ll regulate AI.”
• “We’ll preserve human-only tasks.”
• “We’ll tax AI usage.”
• “We’ll slow down deployment to protect jobs.”

This is all fantasy, and here’s why:

3.1 Competitive Defection Always Wins

If any one country decides to fully deploy AI, it gains:

• higher productivity
• lower costs
• military advantage
• faster innovation cycles
• cheaper services and exports

Every other country must follow or become irrelevant.

There is no global coordination mechanism—not the UN, not the G7, not the OECD—that can prevent defection.

Governments haven’t solved tax competition or climate coordination.
They won’t solve AI coordination, which moves faster and has higher stakes.

3.2 You Cannot Define “Human-Only Zones” in Cognitive Work

This is the Discontinuity Thesis’s strongest—and most devastating—argument:

AI dissolves task boundaries continuously.
What begins as support becomes substitution.

Examples:

• Spell-check → writing assistance → drafting → full composition → decision-making
• Coding suggestions → code generation → full systems design
• Image editing → full image creation → full creative direction

Unlike nuclear arms treaties, which tracked discrete objects, AI cannot be counted, bounded, or frozen.
Cognitive tasks blend together; you cannot regulate gradients.

There is no stable way to prevent AI from leaking into every cognitive domain because cognition itself is continuous, not discrete.

This kills every serious regulatory proposal at the root.

4. The Best Counter-Explanation—and Why It Fails

The strongest non-DT explanation is the “Historical Analogy” model:
AI is like every previous technological revolution. It destroys some jobs, creates new ones, and humans move up the value chain.

This is a comforting narrative, but it suffers one fatal flaw:

AI occupies the entire value chain simultaneously.

In every past revolution:

• Electricity automated muscle.
• Computers automated arithmetic.
• The internet automated distribution.

But humans remained the bottleneck for cognition.

With AI, cognition itself becomes cheap, fast, and infinitely scalable.
There is no “upward” rung of the ladder for humans to climb.
AI sits above the entire ladder.

The usual economist line “new jobs will emerge” is not a theory—it’s an incantation.

No historical analogy applies because for the first time, humans are competing not with tools, but with general intelligence at scale.

This is why the Discontinuity Thesis beats every other model:

• It doesn’t rely on historical parallels that no longer apply.
• It doesn’t assume infinite new job rungs appear.
• It doesn’t assume global coordination.
• It doesn’t assume humans magically retain a comparative advantage in cognition.

It matches the reality that AI is a general-purpose cognitive engine—not an industry-specific tool.

5. What Happens After the Wage Economy Dies?

The Discontinuity Thesis does not claim that society ends.
It claims something more precise:

The system we’ve known for 80 years ends.
What replaces it is not capitalism, but something else.

This could take many forms:

• AI-funded UBI
• National AI dividends
• Sovereign AI ownership models
• Techno-socialism
• AI-powered neo-feudalism
• Shared compute cooperatives
• State-run AI wealth engines

But whatever the political form, the economic substance remains:

Humans become consumers, not producers.
Value creation is automated.

This is not dystopian.
It could lead to immense prosperity.
But it is a discontinuity—a break, not a continuation.

Even if every citizen receives €60,000 a year from a national compute fund, it does not mean capitalism survived.
It means capitalism ended and was replaced by a post-labour system.

The Discontinuity Thesis forces us to stop pretending redistributive fixes preserve the old order.
They don’t.

They simply keep humans comfortable while the underlying economic engine switches species.

6. Why the Discontinuity Thesis Wins

Here is the simplest way to compare the DT to every other model:

The DT explains everything the mainstream model cannot:

• why white-collar wages stall
• why productivity spikes don’t lead to hiring
• why “AI safety regulation” is political theatre
• why new “future-proof” jobs collapse after 6–18 months
• why every major AI advantage becomes universal within weeks
• why AI adoption is accelerating even in conservative sectors
• why capital owners are gaining power at unprecedented speed

The DT’s predictions have matched reality with brutal consistency.
The optimists’ predictions have not.

7. Final Verdict: The Only Model Worth Taking Seriously

The Discontinuity Thesis is not a prediction of doom.
It’s a map of incentives, cost curves, and strategic behaviour under technological change.

It says:

• Once AI becomes the cheapest producer of cognition,
• competitive defection forces universal adoption,
• productive participation collapses,
• capitalism ends,
• and a successor system emerges.

There is no hedging in this logic.
It’s the cleanest, most coherent framework for understanding where AI-driven economies are heading.

Every other explanation relies on wishful thinking, bad analogies, or a refusal to think past the next election cycle.

If you want a theory that matches the actual behaviour of real firms, real markets, and real political systems under pressure—

the Discontinuity Thesis is the only one that fits the facts.