← All writing Final · Dec 2025 · Updated Sep 2026

Economics in the Intelligence Age

A layered model of economics, built from the ground up: scarcity, prices, externalities, public goods, political power, and what AI changes about each.

Introduction

Economics often feels like a subject you’re supposed to already understand. People argue about markets, regulation, capitalism, and government as if everyone shares the same definitions. Most of us are really reacting to fragments: prices at the grocery store, headlines about corporate greed, stories about inequality or pollution. We rarely see how the pieces fit together.

This essay starts from a simpler place. Economics is about how people coordinate when they can’t have everything they want. You already know that problem. Any time you’ve chosen how to spend limited time, money, or attention, you’ve done economics.

My goal isn’t to push a political position. It’s to build an explanation of how modern economies work, one layer at a time: why markets exist, where they fail, why government steps in, why that step is always risky, and how artificial intelligence changes the problem. Each layer solves a real problem while depending on the layers beneath it. When we forget those dependencies, well-intentioned fixes can collapse the structure they rely on.

David Deutsch argues that progress comes from the search for good explanations, ideas that account for reality and are hard to vary without breaking:

In this book, I argue that all progress, both theoretical and practical, has resulted from a single human activity: the quest for what I call good explanations. — David Deutsch, The Beginning of Infinity

What follows is an attempt at such an explanation for economics.

1. Scarcity: The Ground Floor

Scarcity is the condition that human wants exceed available resources. It covers food and energy, but also time, attention, skilled labor, and trust. Scarcity isn’t a moral failure or a political choice. It’s a constraint we inherit by being human.

In wealthy societies it’s easy to forget. Shelves are full and entertainment is endless. That comfort tempts us to treat economics as a debate about values alone. But every stocked shelf represents millions of coordinated decisions, and that coordination doesn’t happen automatically or for free.

Because resources have alternative uses, choosing one outcome excludes another. Land used for housing can’t also be farmland. An hour spent building software can’t also be spent caring for a child. So some way of allocating resources is unavoidable. Refusing to choose doesn’t eliminate scarcity. It just lets allocation happen through power, chaos, or collapse.

Technology can make particular things abundant, but as long as wants outrun resources, a proposal that ignores this constraint isn’t compassionate or ambitious. It’s incoherent. Scarcity is the ground floor. Everything else sits on top of it.

2. Prices: Coordinating Dispersed Knowledge

A free market lets people produce, trade, and consume voluntarily, with prices emerging from competition rather than decree. The obvious alternative is central planning, and it’s tempting to think that with smart enough people, or good enough computers, planning would run more smoothly.

The main obstacle isn’t a shortage of intelligence. It’s information and causal complexity. The knowledge needed to allocate resources well doesn’t sit in one place, and its effects don’t move in straight lines. An intervention meant to lower prices in one area can create shortages elsewhere. A subsidy meant to help one group can distort incentives far downstream.

Most of this knowledge is local and tacit, a point Friedrich Hayek made famous. A shop owner senses when customers hesitate. A farmer notices changes in soil or weather. A factory manager adjusts schedules around bottlenecks that no report captures. On top of that, no one knows how millions of these local adjustments will interact as they ripple through the system.

Prices compress both that knowledge and its causes and effects into a single number. You don’t need to know whether a price rose because of drought, regulation, war, fashion, or a shipping bottleneck. You only need to respond. That response, repeated by millions, is how coordination happens without anyone understanding the whole causal web.

When prices are suppressed or overridden, people lose more than information; they lose feedback. Actions no longer reliably signal consequences. The result isn’t moral failure but systemic misalignment: well-intentioned decisions that compound into shortages, waste, or instability.

Could AI change this? If AI systems could one day model these causal chains, anticipating ripple effects better than price signals do, markets might stop being the best coordination mechanism in some domains, most plausibly those where the relevant knowledge is stable. That remains an open conjecture. One limit won’t yield to computation: prices don’t only transmit existing knowledge; they call forth knowledge no one has yet created, and no model can include what hasn’t been discovered. For now, prices are the most robust tool we’ve found for an economy more complex than any mind can grasp.

3. Externalities: When Prices Stop Telling the Truth

This is usually where patience with markets wears thin. You look at pollution or exploitation and think: this can’t be working. That reaction is reasonable. Something real has gone wrong.

Markets rely on prices to reflect costs. An externality is a cost (or benefit) that falls on people who didn’t choose the transaction. When costs are pushed onto others, prices stop telling the truth. A company can look efficient while imposing harm on others. The system hasn’t become immoral. It has become inaccurate.

That is why calls for regulation arise so reliably here. When profit no longer tracks the public good, intervention feels necessary. The danger is in how we intervene. Replacing markets entirely throws away the information system we still need, because weighing a harm against the benefits of the activity causing it requires knowing what things cost.

The more promising approach is to push hidden costs back into prices through liability, property rights, or targeted taxes. The difficulty isn’t motivation; it’s measurement. When a harm can be measured well, it can be priced, and the market adjusts around it. When it can’t, fixes become blunt and risky.

4. Public Goods and Contested Values

Some things matter deeply yet don’t show up cleanly in markets: social trust, mental health, stable families, a sense of belonging. Markets don’t neglect these out of cruelty. They undersupply them because much of the benefit spills over to people who don’t pay, so no single actor captures enough of it to justify the cost.

This is where government intervention feels most compelling. If we want a society that supports human flourishing, surely someone must step in. But this is also where disagreement explodes, because people don’t agree on what flourishing means or which values deserve protection.

The more morally charged the issue, the greater the risk that intervention shifts from coordination to coercion. What one group sees as protection, another experiences as domination. Legislation doesn’t dissolve that tension. It moves it into politics.

Restraint here isn’t indifference. It’s an acknowledgment that power used to enforce contested values carries real costs, even when intentions are good.

5. Political Gravity: Why Power Accumulates

Once intervention moves problems into politics, a new force takes over. I call it political gravity: concentrated power tends to attract more power. Markets concentrate resources through success and scale. Governments concentrate authority through law and enforcement. Where the two meet, regulatory capture, where the regulated come to steer their regulators, becomes likely.

Rules written to restrain large firms often entrench them by raising compliance costs that only incumbents can afford. Protections meant for the public end up freezing competition and innovation. None of this requires cinematic corruption. Ordinary incentives are enough.

Once power concentrates, it resists constraint. Political gravity doesn’t care about intentions, only structure.

6. Structural Constraints: Designing for Human Failure

If power accumulates regardless of intentions, good intentions can’t be the safeguard. Because humans are fallible, systems should assume misuse will occur. Karl Popper reframed the old question “who should rule?” as “how can we arrange institutions so that bad rulers do the least damage?” Constitutional limits, separation of powers, competition law, and transparency are answers to that question. They aren’t expressions of trust in human goodness. They’re defenses against predictable failure.

Every new lever of control should be treated like a drug with side effects. It may relieve one symptom while creating others. The question isn’t only whether an intervention helps, but what it breaks, and whether the damage can be undone.

Constraints feel frustrating when you’re trying to solve an urgent problem. They slow things down. That friction is the point.

7. AI and the New Scarcity

Artificial intelligence introduces a new factor of production: scalable cognitive labor. Tasks once limited by human attention, expertise, or endurance can be done cheaply and repeatedly. The immediate result is efficiency. The deeper result is time, as hours that went to that work are freed for other things.

As some scarcities recede, others move to the foreground. Attention becomes more valuable. Meaning may become harder to find, and mental health may become a binding constraint. Belonging and meaning may become the scarce goods, which is why work built on relationships with particular people gains value, the “Hearts” work described in State of AI. These are exactly the goods from section 4, where markets struggle and centralized control is most dangerous.

AI may eventually coordinate resources better than prices in some domains, where the relevant knowledge is stable enough to model. If it does, the case for markets weakens there. Where knowledge is still being created, prices keep a job no model can take over, for the reason given in section 2. But that shift wouldn’t remove the need for pluralism, autonomy, and constraint. It would make them more important. A system that coordinates for everyone is concentrated power, subject to political gravity, and no amount of computation settles what people disagree about valuing.

Conclusion: Economics Is Layered

We started with a question: how do people coordinate when they can’t have everything they want? The answer is a stack of imperfect solutions, each resting on the one below. Scarcity makes coordination unavoidable. Prices use dispersed knowledge without central control. Where prices miss real costs, we patch them, and where values conflict, we turn to government. That brings power, which needs constraints. AI doesn’t end this; it shifts which constraints matter most.

The dependencies are the lesson. You can’t remove prices and keep coordination. You can’t add power and assume restraint. And you can’t solve moral disagreement with optimization alone.

None of this is clean, and none of it was centrally designed. That food arrives, knowledge spreads, and conflicts are often resolved without violence is not obvious or guaranteed. The modern world isn’t a failure of intelligence or morality. It’s a fragile, ongoing achievement. We should be critical where explanations fail, cautious where power accumulates, and ambitious about improvement, but also a little impressed that, given our nature and our constraints, things work as well as they do.