Obsolescence of the Entrepreneurial Function
The most unsettling prediction in Capitalism, Socialism, and Democracy was not that capitalism would destroy the working class, or the artisan class. It was that capitalism would destroy the entrepreneur — not through failure, but through success. The progressive rationalization of economic life would eventually bureaucratize innovation itself. The heroic individual entrepreneur would be replaced by teams of specialists operating within corporate research departments. Schumpeter's prediction seemed wrong for the second half of the twentieth century, which produced a resurgence of individual entrepreneurship. But the prediction was about mechanism, not timing. The mechanism has been operating continuously, and AI has brought it to its moment of maximum tension.
In the AI Story
Schumpeter devoted an entire section of Chapter XII of Capitalism, Socialism, and Democracy to this prediction. His argument was structural: the same progress of rationality that bureaucratizes the managerial function will eventually reach the entrepreneurial function. When innovation is systematized — when teams of specialists in research departments explore possibility spaces methodically — the lone visionary becomes redundant.
The prediction was partially confirmed for the mid-twentieth century. Corporate R&D departments at Bell Labs, GE, IBM, and DuPont did routinize innovation to a remarkable degree. But the personal computer revolution, the internet, and the biotech era produced a resurgence of individual entrepreneurship that seemed to refute Schumpeter.
AI brings the prediction back to center stage. Claude Code and its peers can generate thousands of product variations in the time a human entrepreneur can sketch one. They can explore combinations of features, architectures, and designs that no individual team could enumerate. If the entrepreneurial function is the introduction of new combinations, and the machine introduces them faster and more comprehensively than any human, then the function appears to be undergoing the rationalization Schumpeter predicted.
The open question is whether machines can perform the selection under uncertainty that constitutes the entrepreneur's irreducible contribution. Generating combinations may be rationalizable. Choosing among them — caring enough about a specific outcome to bear the risk of failure — may not be.
Key Ideas
Prediction by mechanism. Schumpeter's claim was not about when the entrepreneurial function would obsolesce but about the mechanism: progressive rationalization eventually reaches every economic function.
Mid-century resurgence was partial. Individual entrepreneurship returned in the PC, internet, and biotech eras, but corporate R&D continued to routinize innovation in parallel.
AI as rationalization completed. The machine that generates thousands of combinations is the rationalization of the innovation process Schumpeter predicted, in a form more complete than corporate R&D ever achieved.
The selection question. Whether machines can approximate selection under uncertainty — the willingness to stake resources on a specific vision — is the open question the AI decade will answer.
Debates & Critiques
The debate splits between those who believe selection under uncertainty is a computation that can eventually be replicated and those who believe it requires the specific kind of caring that only mortal creatures with stakes in the world can possess. Schumpeter's framework cannot resolve this but forces the question into the open.
Appears in the Orange Pill Cycle
Further reading
- Joseph Schumpeter, Capitalism, Socialism, and Democracy (1942), ch. XII
- Peter Thiel, Zero to One (2014) — contemporary defense of entrepreneurial irreducibility
- Edo Segal, The Orange Pill (2026), epilogue