The Coming Age of Formulation Economics

For most of modern economic history, intelligence has been understood as a productive resource. A more intelligent worker could solve problems more effectively. A more intelligent manager could allocate resources more efficiently. A more intelligent scientist could discover better explanations for observed phenomena. The underlying assumption remained the same: intelligence creates value by navigating a world whose structure already exists.

Artificial intelligence has largely inherited this framework. Discussions about economic impact center on productivity. How many tasks can be automated? How much labor can be replaced? How many reports can be generated, diagnoses performed, contracts reviewed, or lines of code written? These are important questions, but they all concern the same underlying activity: reasoning within an established space.

Reasoning, however, may not be the deepest source of economic value.

The most transformative developments in history were not simply new solutions. They were new formulations. They altered the space in which future solutions could exist.

Double-entry bookkeeping did not merely improve accounting. It created modern accounting. The scientific method did not merely improve research. It created a new framework for producing knowledge. The transistor did not merely improve electronics. It created an entirely new technological landscape. These developments were not valuable because they solved a particular problem. They were valuable because they generated thousands of new problems worth solving.

The distinction is subtle but profound.

Reasoning operates within a landscape.

Formulation creates the landscape.

A society can possess enormous reasoning capacity and still produce relatively little conceptual novelty. Thousands of brilliant individuals may spend lifetimes optimizing existing systems. By contrast, a single powerful formulation can redirect the efforts of entire civilizations for generations. The economic consequences of the formulation often dwarf the effort required to produce it.

This suggests that there are actually two forms of cognitive capital.

The first is reasoning capital. Reasoning capital consists of the ability to generate answers inside existing frameworks. It includes analysis, optimization, prediction, planning, and problem-solving. Modern AI has already begun dramatically increasing the supply of reasoning capital. As a result, activities that once required specialized expertise are becoming increasingly available at negligible cost.

The second is formulation capital. Formulation capital consists of the ability to generate new frameworks. It creates distinctions that did not previously exist, reveals assumptions that had gone unnoticed, unifies previously separate domains, and opens entirely new spaces for exploration. Unlike reasoning capital, formulation capital does not merely improve activity within a market. It often creates new markets.

The difference between these forms of capital resembles the difference between roads and destinations. Reasoning helps us travel more efficiently from one place to another. Formulation determines which destinations become reachable in the first place.

From this perspective, much of contemporary discussion about AI appears focused on the wrong horizon.

The dominant concern is that machines are becoming increasingly capable reasoners. Yet reasoning has historically exhibited diminishing returns. Once a problem can be solved quickly enough, making it solvable even faster often produces modest gains. The transition from a ten-second computation to a one-second computation is useful. The transition from one second to one millisecond is less revolutionary than it initially appears.

Formulation behaves differently.

A successful formulation can produce returns for decades.

Einstein’s equivalence principle generated more than a new theory. It generated a century of scientific work. Shannon’s formulation of information created vast industries that continue to evolve. The invention of computation as a conceptual framework transformed not one profession but nearly every profession.

The value of these developments did not come from immediate productivity gains. Their value emerged from the intellectual ecosystems they enabled.

This observation suggests a different way of thinking about economic growth.

Traditional models often focus on labor, capital, technology, and productivity. Growth occurs when output increases more rapidly than inputs. Such models describe much of economic reality, but they often treat innovation as an external phenomenon. New ideas simply appear and shift the production function.

Formulation economics places the production of new possibility spaces at the center of the analysis.

The fundamental economic resource is not information. It is not labor. It is not even knowledge.

It is the ability to generate fertile conceptual territory.

A fertile formulation behaves like intellectual infrastructure. Once constructed, millions of individuals can operate within it simultaneously. Entrepreneurs can identify opportunities within the new landscape. Scientists can explore consequences. Engineers can construct applications. Educational systems can transmit its principles. Institutions can reorganize around its assumptions.

The original formulation functions much like a runway. Its value is not measured by the concrete beneath it. Its value is measured by the number, diversity, and productivity of the flights it enables.

This perspective leads to a provocative possibility. Future AI systems may ultimately prove more valuable as producers of formulation capital than as producers of reasoning capital.

Today, AI primarily expands reasoning capacity. It helps people think faster, search more broadly, and execute more efficiently. These are significant achievements. Yet they remain fundamentally bounded by existing conceptual frameworks.

Suppose, however, that AI systems begin generating formulations that humans would not have produced independently. Not merely better answers, but new categories. New abstractions. New intellectual architectures. New ways of organizing knowledge itself.

The economic implications would be extraordinary.

An AI that writes software faster improves software development.

An AI that invents a conceptual framework as important as object-oriented programming creates new industries.

An AI that performs scientific analysis more efficiently accelerates research.

An AI that introduces a framework comparable to information theory transforms the nature of research itself.

At that point, economic value would no longer arise primarily from labor substitution. It would arise from possibility generation.

This distinction matters because possibility generation tends to expand economic activity rather than reduce it. Every major formulation in history has created more intellectual work than it eliminated. The scientific method generated entire disciplines. Information theory generated entire industries. Computing generated entire economies.

The most valuable future AI may therefore be the system that repeatedly creates questions worth pursuing rather than merely answering existing ones.

This suggests a future benchmark for intelligence that differs radically from today’s evaluations.

Instead of asking whether a system can pass an examination, solve a theorem, write software, or outperform experts in existing domains, we might ask a different question:

Can the system generate a formulation whose consequences occupy researchers, entrepreneurs, and institutions for the next fifty years?

The benchmark is demanding because it shifts attention from performance to generativity. Great formulations are not measured by the correctness of a single answer. They are measured by the richness of the worlds they create.

The long-term significance of artificial intelligence may therefore have little to do with automation. Automation concerns efficiency. Formulation concerns possibility. Efficiency improves existing civilization. Possibility reshapes civilization.

The coming age of formulation economics, if it arrives, will not be defined by machines doing humanity’s work. It will be defined by the emergence of systems capable of generating new conceptual frontiers for humanity to explore. In such a world, the ultimate scarce resource will not be computation, labor, or information.

It will be the creation of new spaces in which entirely new forms of value can emerge.

This is just an idea, a formulation itself. The questions it raises should be obvious now that the distinction is made:

1. What is formulation capital?

2. How should intelligence be measured?

3. What is the economic value of formulation?

4. Why do some intellectual breakthroughs matter more than others?

5. Why does modern work often feel purposeless?

6. Are human categories becoming stagnant?

7. Could AI become a generator of conceptual frontiers?…

Kenneth Myers