In Part 1, I discussed how capital markets behave like an evolutionary system. Capital is energy. Corporations are organisms. Technologies, business models and knowledge all act like genes.

Capital markets are the environment. Return and profit are proof of survival and reproduction.

The conclusion was an uncomfortable one. Natural selection has no morality. It does not reward the good, the safe, or the wise — only those fit enough to generate return and profit. And because the system rewards return and profit, every organism is pushed to capture more energy than it may sustainably need. That is how overshoot begins.

We will return to overshoot, and the bust it leads to. But first, a question that has to come before it:

This is where genetic variation enters the story.

Before evolution can act, new genes need to enter the pool.

Neither environments nor organisms are static. Both change constantly, through the interplay of changing ecological conditions, genetics, and learned, culturally transmitted traits. Before evolution can act, there must be genetic differences within a population. New genes have to enter the pool somehow.

In nature, this happens through three main processes, and each one has a useful corporate analogue.

Mutations: the spontaneous change in the genetic sequence. In capital markets, this is novelty arising from within a firm: R&D, a founder’s idiosyncratic bet, a side project, an accident. Like biological mutations, they can be beneficial, neutral or harmful, and the vast majority of new technologies, products, ideas, strategies, features, and pivots die quietly. But every so often, one provides a real advantage. Slack began as the wreckage of a failed game. AWS began as Amazon’s internal plumbing. The mutation that survives looks, in hindsight, like genius; at the time it was mostly persistence meeting a willing environment that enabled it to reproduce — what we usually call luck.

Gene Flow: the transfer of genetic material between different populations. In capital markets, this is the transfer of knowledge, ideas, intellectual property, skills, technology, talent, code, research, and, yes, sometimes less legitimate forms of transfer such as espionage. A gene that took one organism years to evolve can jump to another in a single hire.

Recombination: the shuffling of two genetic sequences into a unique combination. In the capital markets environment, this is recombination at the corporate level: mergers, acquisitions, joint ventures, partnerships. Two corporations’ genes — one technology, the other distribution — combine to produce a new business model, product, service or technology that neither could have produced alone.

Slowly, the advantageous trait spreads.

Over time, the environment and food chain change as their various components act on one another. To persist, organisms must adapt. During this process of adaptation, genetic variation occurs through mutation, gene flow, or recombination.

If a gene’s traits provide an advantage that helps the organism survive and reproduce in the environment, that gene is more likely to be passed on. Slowly, the advantageous trait spreads.

That process is known as natural selection. And this is where AI enters the story.

The finches were not inferior. The environment changed against them.

To illustrate this concept, let us consider the small finches inhabiting the island of Daphne Major. These finches possessed beaks that enabled them to feed on seeds. The smaller-beaked finches were limited to cracking open the smallest seeds, while those with larger beaks favoured larger seeds.

In 1977, a severe drought sharply reduced the availability of small seeds. During this period, the small-beaked finches starved, leading to an approximately 85 percent collapse of the finch population, with the small-beaked finch dying at the highest rate. This reduction continued until rains resumed in 1978.

With the small seeds gone, the average beak size in the population shifted larger: beak size is a heritable trait, and the larger-beaked birds were the ones that survived to breed. Small-beaked finches became rare or disappeared. The population became less varied in its beaks and with it, less genetically diverse.

Over successive generations, as each new generation inherited and reproduced the advantageous gene, it spread until the population became different from the generations before it.

AI is a gene, a trait and an environmental shock.

Using the framework I set out before, AI can be understood as a gene — specifically, a technology — but also as an adaptive trait and environmental shock.

As a gene, its traits and characteristics trace back to mathematics, computer science, engineering, statistics, and the sciences. It is the product of mutation, gene flow, and recombination among the technologies that came before it.

As a trait, its advantage is its ability to consume, process, generate, and automate work at a scale humans cannot physically match. This gives it an edge in its current environment.

At the same time, AI is also an environmental shock. Its introduction and use changes the conditions of the environment around it. For some firms, it creates advantages: cost reduction, automation, accelerated research, personalised products, code generation, data processing, and greater output. For others, it makes existing advantages less valuable.

The moment the environment shifts, selection begins.

Like the finches, firms and their genes can coexist as long as the environment stays constant — as long as nothing removes the “small seeds” they depend on. The moment the environment shifts, selection begins. This is where the forces of nature come into play.

Nature is not stable. What looks like randomness is often the visible result of the net forces acting on a system at a given moment; it is simply hard to predict, because so many variables are in play at once.

Back to AI. In this economic food chain, AI is a gene, a trait and an environmental shock.

As with the finches, the AI-enabled firm does not become dominant simply because AI is inherently superior. It becomes dominant if the environment changes in a way that rewards AI-enabled traits and punishes the older traits firms previously depended on.

The smaller-beaked finches were not inferior; the environment changed against them.

The same can happen in markets.

“The Most Fascinating Profile You’ll Ever Read About a Guy and His Boring Startup,” Wired, 2014.

B. Rosemary Grant & Peter R. Grant, “Evolution of Darwin’s Finches,” Ernst Mayr Lecture, 4 November 2004.