Competitive environments like wild ecosystems and markets, where most of what constitutes the relevant environment for fitness is the behavior of other agents, are often described with the metaphor of the Red Queen’s Race — named after the character in Through the Looking Glass who says “here, you see, it takes all the running you can do, to keep in the same place. If you want to get somewhere else, you must run at least twice as fast as that!”
This metaphor is lossy because it does not sufficiently describe power asymmetry.
The accumulation of power asymmetry is at the core of what makes environments competitive in this way. Within a single iteration, we may easily justify this asymmetry: successful agents gain power because they have won it by succeeding; this power incentivizes them to try harder; they most successful are also the most worthy of this power because they are the most likely to do something useful with it. However, looking at a larger iterative game, we see a different pattern: power is a results-multiplier, so early successes quickly compound. It’s not that everyone needs to run as fast as they can to stay in place — the Red Queen herself is so far ahead that she can sit on her throne and rest, while everyone else runs as fast as they can only to inevitably fall behind.
Early successes compound even if they are accidental or extremely environment-specific: a short string of good luck produces circumstances under which any action, so long as it’s not a really terrible choice, will usually produce better results than the best possible choice available to a less-powerful person.
Nevertheless, in a rapidly-changing environment, what constitutes a terrible choice is quite variable. If an agent has lucked into power, the most sensible thing to do with that power is to create a…