The Miniaturization of the Firm

For centuries, building a company meant gathering people before you even knew whether the idea deserved to exist.

You had to hire before you could validate. Coordinate before you could simplify. Build structure before you had density. In practice, vision was rarely enough; you often had to assemble an entire organization just to execute one serious intuition. The firm was, to a large extent, a response to the cost of production. Because producing was expensive, capital, functions, hierarchies, and human talent had to be concentrated inside the same machine.

That logic does not disappear overnight. But it is beginning to loosen.

Artificial intelligence introduces a new entrepreneurial idea: capacity no longer grows only by adding people. It also grows by adding systems.

That sounds technical. In reality, it is economic. For much of modern history, the growth of a company was tightly tied to the growth of its human structure. More sales required more operations. More operations required more support. More support required more coordination. More coordination required more management. Scaling meant adding layers. And over time, those layers became part of the weight the company needed just to keep moving.

A different architecture is now becoming possible.

A founder, an operator, or a very small team can leverage agents, automations, generated software, semi-autonomous systems, orchestrated workflows, and tools that no longer merely assist, but increasingly execute. The consequence is not that people stop mattering. The consequence is that the relationship between people and productive capacity begins to change in a radical way.

The firm does not disappear. It becomes smaller.

And that miniaturization may become one of the defining economic stories of this decade. Not because every company will become tiny, but because the minimum viable size required to build something serious is falling fast. We are entering a phase in which software can be launched, processes can be operated, systems can be maintained, sales can be supported, documentation can be generated, customer work can be handled, analysis can be performed, and coordination can happen with far fewer people than before. Not perfectly. Not without judgment. Not without effort. But with a density that would have looked unusual only a few years ago.

That changes a boundary we took for granted for a very long time: the boundary between the individual and the firm.

In the twentieth century, the company was the institution that assembled capital and labor and turned them into production at scale. The individual, with some exceptions, was a component inside that structure. Now another possibility is beginning to emerge: that a single person, or a very small cell, can operate with a level of practical force that previously required a much larger organization. That serious ambition may no longer require a proportionally serious structure from day one. That the firm may start looking less like an organizational factory and more like a design intelligence connected to a distributed operating layer.

That also changes the kind of founder this moment demands.

For years, the business world and the technical world divided the work between them. Some people thought about capital, market, growth, and narrative. Others thought about architecture, product, and technical execution. That split worked when the distance between having an idea and deploying a real system was still large. But when AI sharply lowers the cost of building, that division starts to feel incomplete.

Technical people need to think more like entrepreneurs. Entrepreneurs need to think more like system architects.

Not every founder needs to become a deep engineer. But the literacy required is changing. You need to understand systems. Define requirements. Design operations. Know what should be automated and what should remain human. Understand how an AI-assisted workflow is structured. Know where the advantage lies. Know where the real friction still is. Know how to turn a business thesis into an executable architecture.

That shift matters because it also redefines the value of judgment.

Pure technical ability will still matter. But the bottleneck is moving. It matters less who writes every line from scratch and more who knows what system should exist, what part should be automated, where the moat is, which layer of the problem contains the edge, and how to design an operation that captures that value. When building stops being so expensive, strategic judgment becomes more expensive.

That is why I do not think the dominant figure of this phase will be just the programmer, nor the traditional CEO who delegates technical reality to others. I think the most powerful figure will be hybrid: someone able to think in markets, design systems, understand enough technology, allocate capital intelligently, coordinate automation, and maintain a high operating standard.

An entrepreneur with an architect’s logic.

That is, in large part, the kind of structure I am interested in building. Not heavy companies that grow by adding organizational friction, but compact, vertical, dense ones. Structures able to operate with more power than their visible size suggests. Software that captures judgment. Agents that reduce manual work. Systems that work in front of and behind the business. Operations that until recently were too expensive to exist and are now becoming entirely rational.

The business history of the twentieth century was, in large part, the history of concentration. Concentrated capital, people, infrastructure, and capability inside large organizations. I believe the twenty-first century will also become the history of compression. More capacity per person. More software per unit of capital. More company per unit of talent. More output without bloating the structure.

And when that happens, ambition depends less on the initial size of the apparatus and more on the clarity of the design.

That is a profound change.

Because it means the future will not belong only to those who can gather large teams, but also — and perhaps above all — to those who know how to gather systems, judgment, and will at a higher density.

The firm is not becoming irrelevant.

It is becoming smaller, faster, and more dangerous.