Intelligence develops like an economy

We often think of ourselves as having some degree of intelligence, which we then apply to whatever problems we’re trying to solve. It’s as if intelligence is a reservoir of water than can be applied to various uses, like drinking, cooling, or firefighting. Another perspective is that intelligence is never stored in a system but always constructed in the moment as the organization of a body to meet a task in an environment. Calling a system intelligent, in this view, is a way of categorizing how well it self-assembles solutions to problems.

The constructed nature of intelligence becomes particularly clear when you consider how intelligence develops over time. Infants don’t grow in intelligence by enlarging their thinking-reservoir, or by growing improved internal distribution mechanisms for transporting thinking-stuff to the problem site. Instead, they get better at using internal processes, environmental affordances, and task constraints to assemble suitable forms.

This can still be hard to see in infants because we don’t view their internal processes very well. The economy is a useful model system for studying the development of intelligence because the economy doesn’t have a traditional skin, a boundary making it opaque to much of the environment.

The economy is a general intelligence—a highly powerful optimization system that can solve complex problems across a wide and diverse range of problem spaces, efficiently navigating around many obstacles and correcting all kinds of perturbations, across massive ranges of space and time. The economy’s intelligence can be seen in the creative, indirect solutions it produces to many different problems. The economy even exhibits recursive self-improvement in the form of economic growth.

But the economy doesn’t have an “intelligence part”. There’s nothing in the economy analogous to the traditional view of the brain as the place where smartness happens. Instead, the economy’s ability to construct solutions to problems is a function of the resources, markets, technologies, industries, and divisions of labor that it can draw on. As economic development occurs, the economy gains the ability to construct solutions to problems via enhanced abilities to anticipate, mitigate, exploit, reallocate, and transform events relevant to the economy’s allostatic processes.

The division of labor helps enable this problem-solving capability. As tasks are broken into specialized roles, the economy’s coordinating structures allow them to be combined into larger functional assemblies. Solutions to complex problems can therefore be constructed by linking together many specialized processes. Qualitative and nonlocal effects can result.

Economic development thus increases intelligence not by enlarging a central stock of reasoning power, but by expanding the repertoire of specialized capabilities and the pathways through which they can be recombined. The division of labor allows the economy to assemble solutions that are prescribed or preordained nowhere in the system.

From this perspective, the development of intelligence is not the accumulation of a substance but the growth of a system’s capacity for organized assembly. New processes, structures, and connections enlarge the repertoire of components that can participate in solving problems, while new forms of coordination allow those components to be brought together in increasingly indirect and flexible ways. The result is a system that can construct responses to a wider range of situations. Infants, artificial intelligences, and other adaptive systems become more intelligent in the same basic way: by developing richer sets of parts and more powerful means of organizing them into temporary structures suited to the demands of the moment.