Who Killed Progress? Charles Frey on AI, Power, and the Next Collapse

by: Kianga J Moore + Amanda Smith

Carl Frey doesn’t predict the future. He studies how it unravels. In How Progress Ends: Technology, Innovation, and the Fate of Nations (Princeton University Press, 2025), the Oxford economist and professor argues that innovation doesn’t die from exhaustion but from control. Frey gives concrete examples of how prosperity fades when systems built to protect growth become the structures that prevent it.

Early reviews have called the book “a masterwork of economic history and foresight” and “a wake-up call for the digital age.” Scholars at the University of Chicago and MIT have described it as “a rigorous and unsettling examination of how institutions, not technology, decide whether progress continues or stops.”

When my colleague Amanda Smith and I interviewed Frey, he spoke with the calm precision of someone describing a pattern rather than an opinion. “Innovation suffocates when it becomes the work of too few,” he said. Frey’s concern wasn’t technology’s speed or complexity—it was the narrowing of ownership, the slow concentration of decision-making into fewer hands. “Every era imagines its advances are inevitable,” he told us. “But innovation is political—it depends on how power is distributed.”

Smith asked how earlier societies managed that balance between growth and control. Frey pointed to the Dutch Republic—once the most inventive economy in Europe—whose decline, he said, began “not with collapse, but with caution.” Rules meant to preserve stability hardened into protection for established guilds and merchants. Experimentation slowed, and risk migrated elsewhere. He traced similar patterns throughout nineteenth-century Prussia, where bureaucratic precision delivered early industrial power but later turned into restraint. “Planning creates order,” Frey said, “but it also creates dependency. When institutions prize control over exploration, progress narrows.”

No example captured the tension more sharply than the Soviet Union. Zelenograd, designed as the USSR’s “Silicon Valley,” embodied the contradiction of centralized ambition (Computer History Museum; Wikipedia). Ministries could summon scientists, funding, and equipment overnight, yet bureaucracy strangled flexibility. “You can’t plan discovery from a ministry,” Frey said. “It’s not a production quota—it’s trial and error.”

“Soviet / Russian Silicon Valley” 

These examples, he said, weren’t distant history but evidence of the same cycle—institutions tightening control in pursuit of order as innovation turns procedural and risk is pushed out of the system.

That pattern defines today’s economy. Frey’s argument lands hardest in industries once celebrated for openness and disruption. Amazon now dominates logistics, health care, and cloud computing. Google controls search, maps, email, and increasingly, AI infrastructure. Meta owns the channels through which billions communicate. Microsoft, having defined personal computing, is embedding itself across artificial intelligence, data, and enterprise systems.

His warning recalls Alfred Chandler’s Visible Hand, which chronicled how twentieth-century corporations replaced Adam Smith’s Invisible Hand of markets with hierarchical management. Efficiency became justification for dominance. “Every breakthrough begins with uncertainty,” Frey told us. “When exploration gives way to administration, discovery stops.”

He sees the same logic shaping digital capitalism today. Ownership defines competition, not innovation. “When a handful of firms control both the infrastructure and the story of progress,” he said, “it’s not competition—it’s coordination.”

Frey and Smith discussed how this pattern extends beyond technology to the structure of economies themselves. He pointed to Japan’s Zaibatsu Conglomerates of the Meiji era—family-controlled industrial networks like Mitsubishi and Sumitomo. They accelerated modernization but left little space for challengers. “They show how a system designed to catch up can become too efficient to evolve,” Frey said. “Dominance becomes its own purpose.”

In How Progress Ends, Chapter 3, The Rise of Europe, Frey attributes England’s early technological advantage to decentralization—multiple centers of experimentation competing without a single authority able to suppress them. By contrast, in Chapter 8, The Age of Planning, he describes how twentieth-century bureaucracies achieved rapid catch-up growth before stagnating. “Centralized management,” he writes, “is most useful for exploiting what is already known. It halts renewal once discovery requires uncertainty.”

Those observations connect directly to the present.

During the 1980s, Japan’s electronics boom seemed unstoppable. “Thomas Johnson once said the Cold War was over and Japan had won,” Frey recalled. But Japan missed the shift to software and the internet. “It wasn’t bad luck—it was structure,” he said. “Weak antitrust meant incumbents could hold the frontier indefinitely.”

The United States, by contrast, dismantled its bottlenecks. Regulators broke up AT&T and forced IBM to open its architecture, clearing ground for Microsoft, Apple, and the next generation of firms. “That willingness to disrupt your own giants,” Frey said, “is what keeps a system alive.”

Today, that willingness is fading.

Progress, in Frey’s view, depends less on invention than on diffusion—the ability of ideas to spread beyond their origin. “Progress cannot be taken for granted,” he said. “It is constant work in progress.”

Data from OECD and Brookings confirm the trend. Business formation has fallen to historic lows. Startups are fewer. Productivity growth lags even as profits climb. “Market concentration used to reflect efficiency,” Frey said. “Now it reflects stagnation.”

When asked whether China represents an alternative model, Frey said it illustrates the same problem in reverse. “China channels innovation through political stability,” he said. “It’s efficient but brittle. The United States has the opposite imbalance—too much power concentrated in firms, not the state.” Both models, he argued, lead to the same result: exploration replaced by control.

His concern is structural, not ideological. “The question isn’t capitalism or socialism,” he said. “It’s whether any system can stay open enough to keep learning.”

Artificial intelligence, he believes, is the clearest test. Smith asked whether AI marks a new frontier or another consolidation of power. “Large models are remarkable tools,” Frey said, “but they don’t remove the need for decentralized exploration. They can accelerate, but they can’t substitute.”

Los Angeles Examiner/ USC Libraries/ Corbis 

His view aligns with MIT economist Daron Acemoglu, who argues that AI’s productivity potential will remain limited unless institutions share its benefits rather than enclose them. For Frey, the question is no longer how advanced the tools become, but how open their systems remain. “Who owns the compute,” he said, “will determine who defines progress.”

The implications extend beyond technology. Frey pointed to Detroit as an early metaphor for modern innovation. “Early Detroit looked like Silicon Valley,” he said. “Ford’s assembly line was planning at its best, but it worked because independent suppliers and workers kept improving it.” Scale produced efficiency, but networks sustained renewal. When those networks thinned, collapse followed. “Scale without feedback,” he said, “is fragility disguised as strength.”

In Chapter 5, The Age of Exploration, Frey discusses how Spain and Portugal turned discovery into empire. Royal charters dictated who could sail and where, turning exploration into administration. “Explorers became contractors of the crown,” he writes. “Innovation continued but diffusion stopped.” England and the Dutch Republic allowed privateers and merchants to compete, spreading navigation techniques and ship designs through rival ports. Frey describes that period as the first instance of progress ending not from failure, but from over-coordination.

Later, in Chapter 10, Diffusion and Decline, he examines postwar Europe. Welfare systems built to protect workers and industries reduced mobility between firms and disciplines. “The safety net became a hammock,” Frey writes. Economic security expanded, but experimentation slowed. That logic now repeats in modern corporations, where benefits and control over data create internal stability at the cost of innovation.

That fragility is visible again. Global productivity has slowed to its weakest pace in three decades, according to the World Bank. The Stigler Center describes dominant digital platforms as “gatekeepers of innovation.” Both reinforce Frey’s argument that progress rarely ends with a crash—it ends through concentration, quietly, until change feels impossible.

Even so, Frey resists fatalism. “The danger isn’t inevitability—it’s complacency,” he said. “We still control how our systems are built.” His prescription is pragmatic: enforce antitrust laws that restore competition, fund public research that diversifies experimentation, and design institutions that prioritize diffusion over dominance. “Our problem isn’t invention,” he said. “It’s permission—who gets to explore, and who decides when enough is enough.”

How Progress Ends contends that technology alone will not shape what comes next. Whether innovation leads toward openness or control depends on how societies govern their choices. The next frontier rests on those decisions— but somehow history rarely grants second chances.

Further Reading

How Progress Ends: Technology, Innovation, and the Fate of Nations
By Carl Benedikt Frey
Published by Princeton University Press

Purchase How Progress Ends from Princeton University Press →