The Two-Speed Economy

For most of human history, the central economic problem was making enough stuff. Everyone needed food, clothing, shelter, light, heat, transport, medicine — and there wasn’t enough of any of it. Solve that problem, and you’ve solved most of human suffering.

Over the past two centuries, technology did exactly this, and the cost compressions have been almost beyond comprehension. A candle-hour of light in 1800 cost something like an hour of labor to earn; today it takes a fraction of a second. Quality-adjusted television prices have fallen roughly 97% since the late 1990s. Long-distance phone calls that once cost a daily wage are now effectively free. Food took 40% of an American household’s income in 1900; today it takes about 10%. Computing, clothing, transportation, communication — every category of replicable good has gotten dramatically cheaper, often by orders of magnitude. Increasingly, technology creates value not by raising output but by collapsing costs.

The result was the largest improvement in human welfare in recorded history. Global extreme poverty fell from nearly two billion people in 1990 to under 700 million today. Life expectancy in poor countries rose by fifteen years in two generations. Mobile phones reached villages that had never seen a landline. The miracle is real, and it’s still working — anywhere consumption is dominated by replicable goods, technology continues delivering extraordinary gains in lives lived.

So here’s the puzzle. The same decades that lifted over a billion people out of extreme poverty also produced rich-country middle classes who feel poorer, more squeezed, and less optimistic than their parents. Why does life in San Francisco or London or Tokyo feel harder, not easier, even as technology in rich countries kept advancing — AI, biotech, automation, the internet, all accelerating? Why does housing seem impossible? Why does healthcare keep eating more of every paycheck? Why does the next generation expect to do worse than the last? These two stories shouldn’t coexist. They do, and they’re the same story.

The squeeze isn’t about absolute deprivation. By historical and global standards, today’s rich-country middle classes are extraordinarily well-off in material terms, with access to medicine, technology, food, and entertainment their grandparents couldn’t have imagined. The tension is about relative position: the things that mark having “made it” — a home in a good neighborhood, top schools for one’s children, status security, a foothold in the right institutions — have moved further out of reach even as televisions, food, communication, and most consumer goods have become abundant. Absolute welfare keeps rising; relative advancement keeps narrowing.

The answer follows an arc, and it runs on basic microeconomics. Goods differ in how much people want of them as they get richer. You only eat so much bread no matter how wealthy you get, but housing, healthcare, education, experiences, and status all expand sharply with income. So as technology compresses costs in the basics, freed-up income flows toward more income-elastic categories — and the dynamics of those categories define what each phase of the economy feels like.

Phase 1: Goods. When the basics are scarce, the economy revolves around producing more of them. Technology drives prices down, abundance up, and welfare follows directly. This was the story from roughly 1800 to 1980 in the wealthy world, and it’s still the story across much of Asia, Latin America, and Africa, where household budgets remain dominated by food, clothing, energy, and basic manufactured goods.

Phase 2: Services. As goods become cheap, spending shifts to services — healthcare, education, hands-on care, professional and information work. Many of these sectors historically resisted scaling because the value of the service was inseparable from human time delivered at irreducible ratios. A string quartet still requires four musicians and thirty minutes to play a Beethoven piece. A childcare worker can only physically supervise so many children. A barber still cuts one head of hair at a time. As farming and manufacturing got cheaper by orders of magnitude, these sectors mechanically absorbed rising shares of GDP and household budgets — which is why labor’s share of income increasingly concentrates in domains that resist productivity gains.

But the line between scalable and non-scalable is now actively moving. AI is rapidly automating administrative work, information processing, content creation, customer service, parts of legal and medical analysis, and pieces of education — categories of cognitive labor that previously defined white-collar employment. What looked like the permanent character of services is being scaled in ways previously reserved for manufacturing. As that happens, the income freed from now-scaling services doesn’t disappear — but where it goes depends on distribution. Where AI gains diffuse broadly through wages, lower prices, and consumer surplus, they flow further down the arc, intensifying demand for what remains genuinely irreducible: hands-on care, in-person work, the trades, live performance, and everything in Phase 3. Where gains concentrate among capital owners or a narrow class of beneficiaries, the same demand intensification plays out for fewer people, while displaced workers face wage compression or unemployment rather than upgraded consumption.

Phase 3: Comparatives. When goods are abundant and services increasingly scaled, the marginal dollar flows toward goods with even higher income elasticity and even more constrained supply. Many of these are absolutely scarce — fixed regardless of demand. Land in desirable locations is bounded by geography and policy. Time is capped at twenty-four hours per person per day. Attention is biologically finite — and as information became abundant, human attention itself became one of the scarcest goods in the economy, fought over by platforms, advertisers, and creators in what now constitutes a substantial share of digital activity. Status is inherently comparative. Authentic relationships require presence and can’t be mass-produced. Institutional access — to top schools, top firms, networks of consequence — is constrained by capacity and gatekeeping. As rising income meets this fixed or slow-growing supply, prices rise and competition intensifies. This is straightforward microeconomics: more demand chasing supply that can’t expand.

The arc explains the puzzle. The same technological progress that lifted billions out of poverty — by making goods cheap and abundant — has, in already-wealthy societies, exposed the constraints that lie beyond goods and services. Those constraints aren’t manufactured inefficiencies; they’re absolute or near-absolute. You cannot scale time. You cannot scale attention. You cannot replicate a particular location or a particular relationship. And as more income chases unscalable goods, their prices and the energy spent competing for them rise mechanically.

Two stories are happening at once, and both are true. Globally, technology is producing the greatest welfare gains in human history, especially where consumption is dominated by goods that scale. In advanced economies, technology has revealed what was always going to happen once material scarcity was solved: economic life increasingly revolves around what technology can’t replicate — time, attention, location, status, institutional access, trust, and authentic connection. And as AI now scales the cognitive services that once seemed safe, that revelation is only accelerating.

The progress is real. The squeeze is real. They’re the same story, told from different points along the arc.