>However, markets are socially constructed because they depend on property rights that are socially constructed, and there are many different ways that markets and property rights can be constructed, some of which are unexplored and potentially far better than what we have today.
This would be an interesting hypothesis to explore with swarm reinforcement learning. It seems to me that rights have a lot to do with optimal/stable cooperation strategies in certain types of games, and maybe RL can uncover better strategies, or maybe not.
The simplest well-studied problem of this sort is the iterated prisoner's dilemma. The strategies we have today are remarkably similar to the Axelrod's original strategies three decades ago.
This would be an interesting hypothesis to explore with swarm reinforcement learning. It seems to me that rights have a lot to do with optimal/stable cooperation strategies in certain types of games, and maybe RL can uncover better strategies, or maybe not.
The simplest well-studied problem of this sort is the iterated prisoner's dilemma. The strategies we have today are remarkably similar to the Axelrod's original strategies three decades ago.