Cognitive Capital and the Economics of AI

Most of the conversation around AI still revolves around the same questions. Which jobs will disappear, which skills will remain, how quickly companies will automate. These are valid questions, but they stay close to the surface. They assume that what is changing is primarily the content of work.

A more interesting question is what happens when the underlying structure of value creation shifts. Some recent discussions have started to move beyond this, asking what happens when human time is no longer the foundation on which labor, contracts and economic value are built. For a long time, value has been closely tied to time. People are paid for hours worked, and productivity is often understood as output per hour. But AI weakens that link. A relatively small amount of human input can now generate a much larger amount of output, often continuing without direct involvement. If time is no longer the main constraint, the system built around it starts to change as well.

That framing already goes further than most discussions. But it still assumes that access to AI, or control over systems, will be the main source of advantage. In practice, access is becoming close to universal. What is not evenly distributed is the ability to use these systems effectively.

This shifts the focus more fundamentally. Not from labor to capital, but from capital to something less tangible. The scarce resource is no longer time, and increasingly not even access to technology, but the ability to think with it: cognitive capital. By cognitive capital, I refer to the ability to structure, interpret and apply knowledge in ways that create economic value. This matters because AI does not just increase productivity, but changes what type of human input becomes economically valuable.

~

If cognitive capital determines how effectively AI can be used, then differences in outcomes are no longer mainly driven by access to technology, but by how people work with it.

In practice, the same tools can lead to very different results. Much of that comes down to how a problem is approached in the first place, how instructions are formulated, and how output is interpreted and adjusted. Using AI effectively is not just a technical skill. It requires a way of thinking often described as systems thinking: being able to structure ambiguity, question assumptions, and work with ideas that are not fully formed yet.

Cognitive capital builds on existing forms of human capital, such as education and experience, but extends them in a specific direction. It shapes how much value someone can extract from these systems. Two people with access to the same tools can therefore end up with outcomes that differ not marginally, but by orders of magnitude.

~

To understand how cognitive capital translates into economic advantage, it helps to separate three mechanisms that in practice are closely connected. Each of them compounds over time, but in different ways.

The first is skill compounding. As people work with AI systems, they gradually develop a better sense of how to use them. They learn how to formulate prompts, how to guide outputs, and how to recognize when something is off. This is not a one-time skill. It improves through use. Someone who has spent months working with these systems will approach problems differently from someone who has just started, even if both have access to the same tools.

The second is cognitive compounding. Working with AI does not just change how ideas are processed, but how they develop. Thoughts no longer need to be complete before they can be explored. They can be tested early, challenged, and extended in real time. Used this way, AI becomes a constant sparring partner. Over time, this does more than improve individual outputs. It improves the underlying process of thinking itself. Ideas are generated more quickly, connections become easier to see, and creativity increases as partial thoughts can be developed rather than discarded. Each interaction builds on the last. In that sense, cognitive capital compounds through use: the more it is applied, the stronger it becomes.

The third is output compounding. AI makes it possible to generate and build on output at a scale that is not limited in the same way by human attention, working hours or physical space. A factory cannot expand indefinitely. A building has boundaries. Human labor runs into natural limits of time and energy. Digital output does not work that way. Once generated, it can be replicated, extended and built on continuously, forming the basis for further output without the same ceiling that applies to physical capital.

~

These three mechanisms do not operate in isolation. They reinforce each other. As people develop better skills in working with AI, they are able to engage with it more effectively as a thinking partner. That, in turn, strengthens cognitive compounding. As thinking improves, the quality of input increases, which leads to better output. And as output accumulates, it creates more material to build on, further accelerating both skill development and cognitive capacity.

The result is not a linear increase in productivity, but a self-reinforcing process. Small differences in how effectively someone uses these systems can widen over time, as each layer of compounding builds on the others. This creates a form of trajectory divergence. Those who are able to develop cognitive capital early benefit not just from higher initial output, but from faster improvement over time. Those who do not face a different dynamic. Catching up becomes increasingly difficult as the gap is reinforced through repeated interaction.

~

Taken together, this changes how value is determined. If output increasingly depends on how effectively AI systems are used, then value is no longer primarily tied to time, and not solely to ownership of technology. It is shaped by the ability to translate these systems into meaningful results. In that sense, cognitive capital becomes a key driver of value creation.

This also changes how inequality emerges. When access to technology is widespread, differences in outcomes are no longer explained by who has access, but by how effectively that access is used. Because the mechanisms behind this are compounding, these differences do not remain static. They widen over time.

Scarcity shifts accordingly. Human time becomes less binding, and access to tools becomes less differentiating. What remains scarce is the ability to think with these systems in a structured and effective way.

Returns follow that distribution. As cognitive capital becomes more decisive, those who are able to apply it effectively are likely to capture a disproportionate share of the value created. Not because they work more hours, or own more assets, but because they are better positioned to direct increasingly powerful systems.

What is less clear is how quickly this shift will become visible, and how easily it can be addressed once it does. If the underlying dynamics are compounding, the effects may emerge gradually, but accumulate in ways that are difficult to reverse.