Valuable work has two halves: deciding what to do, and doing it.
An organization succeeds by being great at whichever half is scarce.
For the last century, doing was the scarce half. Good work required skilled people, and skilled people were scarce. Talent went where the resources were, and resources pooled at scale. So the largest organizations gathered the most talent, and because talent was the binding constraint, scale won. Small organizations had one natural edge: the person who saw a problem sat near the person who could act on it. But the edge lay dormant. Doing the work still took more hands than they could afford.
That is now changing. Digital agents make execution cheap and abundant. The drafting, searching, reconciling, coding, and monitoring that once filled the working day now requires very little from humans. This does not abolish work. It ends the scarcity of doing. What is no longer scarce can no longer separate greatness from mediocrity. By elimination, the constraint moves to the other half: deciding.
A decision is not an opinion. It is a commitment to act under uncertainty. It takes judgment, timing, and the willingness to be wrong in public. Work divides among many hands. A judgment does not.
Two quantities determine the outcome.
The first is decision distance: the gap between knowing what to do and committing to do it. Distance grows with size. The more people an organization holds, the more damage one mistake can do, so it wraps its choices in reviews, approvals, and sign-offs. That is not a flaw; it is self-protection at scale. But every safeguard is a step between knowing and acting, and every step is another person to brief, persuade, and wait for. This is how a system everyone knows is broken stands for years inside a large company. Everyone can describe the problem. Everyone can name the fix. No one can make the call.
The second is decision capacity: how many good decisions an organization can make in a given time. Short distance is worthless if every choice funnels through one overloaded mind. A single person has no distance to speak of, but not much capacity either.
A large organization can raise capacity by adding deciders or by delegating. Every decider it adds lengthens the distance. And delegation asks it to tolerate the very mistakes its safeguards exist to prevent, so the safeguards follow the decisions down. Handing it agents changes little. The machinery that gates its people must gate its agents too. It can grow capital, headcount, and reach without limit, yet still lack the one kind of scale that now matters.
A small organization built on agents escapes the trade. It gives its agents the context, the tools, and the boundaries to make routine calls on their own and to stage the consequential ones for a person. Capacity rises. No layer is added. No distance returns. For the first time, an organization can hold both.
Size is only a proxy. A small organization can be slow and careless; a large one can shelter pockets of fast judgment. The variable is distance, not headcount. And agents supply capacity, not judgment. A system is only as good as the person who decides what counts and answers for it.
This is the Age of Decision. Its winners will not be a new set of giants but a multitude of small firms.
Scale does not disappear. It consolidates. The giants that thrive will run the infrastructure everything else depends on — the models, the networks, the compute, the logistics. The railroads and telecoms of the new economy, more essential than ever, supplying power to everyone and advantage to no one. And their customers, more and more, will be the multitude.
The multitude will not take the giants' work. It will do work that has never been done. A large organization can only afford problems big enough to pay for its own coordination; beneath that line, need went unmet or got a generic answer. Cheap execution and close decisions drop the line. The smallest serious firm shrinks toward one person. The smallest serious market shrinks toward one customer. Beneath the old line lies a vast territory of valuable work that no giant could see and no small firm could reach — audits for one industry, software for one trade, service for one town. Now someone can.
The winners will not announce themselves the way the last ones did. No campus, no platform, no logo on a stadium. From the street they will look like ordinary businesses — a supplier, a practice, a shop of five doing what once took five hundred. Inexplicably good. Impossible to dislodge. Thousands of them, close to their customers, answerable for their work, running more experiments in a year than the giants managed in a decade. The constraint has moved. The advantage has moved with it. What remains is to decide.