The Throttled Singularity: Why Exponential Abundance Will Meet Its Match in Human Power

There is a story technologists love to tell about the next ten years, and it goes something like this. Artificial intelligence keeps doubling in capability on a cadence measured in months, not decades. Robots step out of the factory cage and into warehouses, kitchens, and hospitals. Self-driving trucks quietly take over the interstates at three in the morning, and robotaxis make car ownership feel as quaint as owning a horse. Reusable rockets drop the cost of reaching orbit until asteroid mining and orbital manufacturing shift from science fiction to line items in a quarterly report. Each of these curves feeds the others, and somewhere in the compounding, we cross a threshold. Machines become better than humans at nearly everything economically valuable, including the design of better machines. That is the singularity, and the exponential math says it could arrive within a decade.

The math is seductive because exponentials genuinely do behave this way. A technology improving at forty percent per year grows nearly thirtyfold in a decade. Human intuition, calibrated for a linear world, consistently underestimates this. We saw it with the genome, with solar panels, with the transformer models that went from writing plausible paragraphs to writing working software in a handful of years. If the curves hold, the endpoint is not hard to sketch.

The Abundance Scenario

Follow the exponentials to their logical extreme and you arrive at something economists have rarely had to think about: post-scarcity. If humanoid robots can build more humanoid robots, and those robots can mine, refine, manufacture, transport, and construct with negligible human labor, then the marginal cost of most physical goods collapses toward the cost of raw materials and energy. If space industry matures on schedule, even raw materials stop being a constraint, because the asteroid belt holds more accessible metal than humanity has ever mined. Energy, meanwhile, rides its own exponential as solar and storage costs fall.

In this world, the price of nearly everything trends toward zero. A house is a week of robot labor and some printed materials. Food is grown, harvested, and delivered by machines that never sleep. Medical diagnosis is a conversation with an intelligence that has read every paper ever published. And here is where the scenario turns radical: if goods and services approach infinite availability, money loses its function. Money is a rationing mechanism for scarcity. Remove the scarcity and you remove the point of the ration. Wealth, and the power that flows from wealth, would dissolve like salt in water. The optimists conclude that humanity would then drift, perhaps clumsily but inevitably, toward a flatter distribution of power. When no one can hoard what everyone can have, hierarchy loses its foundation. Some call it fully automated luxury for all. It is a genuinely beautiful vision.

It is also, I think, naive in one specific and fatal way.

What the Curves Cannot See

The abundance scenario treats power as a byproduct of scarcity, something that will evaporate once its economic substrate disappears. History suggests the arrow points the other way. Power is not a consequence of the economic system; the economic system is a consequence of power. Those who hold political, military, and financial control do not regard their position as an accounting artifact that technology may deprecate. They regard it as the point. And they have agency. They can see the same exponential curves the technologists see, and they can act on what they see.

Consider what abundance actually threatens. A currency that becomes worthless erases the leverage of central banks and the sovereign debt system that disciplines nations. Manufacturing that anyone can bootstrap erases the strategic advantage of industrial powers. Intelligence that anyone can download erases the moat around elite institutions, intelligence agencies, and professional classes. Energy independence erases the geopolitical weight of petrostates. A cheap ride to orbit erases the monopoly that a handful of governments hold over the ultimate high ground. Every single pillar of the current order is load-bearing, and the exponential wrecking ball swings at all of them simultaneously.

No incumbent in history has watched that swing passively, and there is no reason to expect this time to differ. What we should expect instead is suppression, and not the cartoonish kind with soldiers smashing servers. The suppression will be procedural, reasonable-sounding, and often genuinely wrapped in legitimate concerns. It has, in fact, already begun.

The Machinery of Deceleration

Look at the tools already on the table. Export controls on advanced chips are framed as national security, and partly they are, but they also establish the principle that compute is a controlled substance, like enriched uranium, to be allocated by states. Licensing regimes for frontier models are framed as safety, and partly they are, but they also raise the cost of entry until only a handful of firms, each deeply entangled with government contracts, can play. Autonomous vehicle deployment is throttled city by city, not because the technology fails but because the technology succeeds against an industry of drivers, insurers, and municipal revenue streams that vote. Drug discovery by AI still funnels through a regulatory apparatus built for a slower century. Space launch is licensed, orbital slots are allocated by treaty, and any serious attempt at asteroid retrieval will collide with international law written when such a thing was fantasy.

None of these measures stops the exponential. Each one merely adds friction, and friction is enough. An exponential with a damping term is no longer an exponential; it is a logistic curve, an S-curve that bends over and flattens into a plateau. The genius of institutional suppression is that it never has to win outright. It only has to slow the doubling time from twelve months to thirty-six, again and again, until the revolutionary becomes the merely impressive, and the merely impressive can be absorbed, taxed, licensed, and owned.

There is a second, subtler mechanism at work, which is capture rather than suppression. The most likely custodians of transformative AI are not garage tinkerers but a few enormous firms whose interests align comfortably with the status quo. Their incentive is not to make intelligence free; it is to meter it. Abundance delivered through a subscription is not abundance at all. It is scarcity with better marketing. The same robots that could make goods nearly free can just as easily make goods for whoever holds the platform, with pricing set not by cost but by what the market will bear. In that world, money does not become worthless. It becomes more important than ever, because it is the only key to the machines.

The Decade of Contested Exponentials

So what does the next ten years actually look like? My speculation is neither the utopia nor a simple freeze, but a contested middle that is stranger than either. The technology will keep advancing, because exponentials are hard to kill and because great powers are locked in a competitive dynamic where unilateral restraint feels like surrender. China and the United States will each accelerate the capabilities that serve state power, namely military autonomy, surveillance, cyber operations, and strategic industry, while both throttle the capabilities that diffuse power downward, namely open models, cheap general-purpose robotics, and permissionless biotech. We will get astonishing machines and astonishing rules about who may use them.

Expect self-driving freight to succeed first, because trucks threaten a labor force with waning political weight and deliver savings to corporations with growing political weight. Expect humanoid robots in factories before homes, because factory robots concentrate returns to capital while home robots diffuse them to households. Expect space commercialization to flourish precisely because it extends rather than threatens existing power, giving militaries new terrain and billionaires new frontiers. And expect artificial general intelligence, if it arrives, to be announced not with an open release but with a classification decision.

The deepest question of the decade will not be technical. It will be whether the suppression holds. Exponentials have a way of routing around obstacles, and the same open research culture that built this wave keeps leaking capability to the edges. Models shrink and run on laptops. Techniques publish faster than laws draft. A serious possibility exists that control fails not through revolution but through evaporation, as the marginal cost of intelligence drops below the cost of policing it. If that happens, the abundance scenario returns through the back door, messier and more chaotic than the optimists imagined, arriving not as a shared decision of a wise global community but as a fait accompli that institutions must scramble to survive.

Betting on the Collision

If I had to place a wager, it would be this. The next ten years deliver perhaps a third of what the pure exponentials promise, which will still be more change than the previous fifty years combined. Goods get dramatically cheaper but never free, because pricing power migrates to platforms. Money does not die, but its meaning shifts as governments experiment with distribution schemes, from expanded transfers to sovereign AI dividends, less out of generosity than out of the need to keep consumption, and consent, alive. Power does not equalize, but it does fragment, contested between old states, new platforms, and a public with unprecedented access to intelligence if not to its fruits.

The singularity, in other words, will not be canceled. It will be negotiated, and the negotiating table is not level. The exponential is a force of nature, but its expression in human affairs is a force of politics. Anyone forecasting the next decade by extrapolating the curves alone is modeling the engine and ignoring the brakes, along with the hands that grip them. The future will be fast. Just not as fast as the math allows, and not nearly as fair as the dreamers hope, unless people notice, while the negotiation is still open, who is doing the throttling and why.