I don’t know why it’s taken me so long to post this. Several weeks ago I was playing a turn-based fighting game that I love, and I got to thinking “what would an AI that plays this game well look like?” And as I thought about it, it became clear to me that the challenge of playing a complex game is the same as reaching goals in any sufficiently understood environment.
Let me elaborate on that last point, because I think it’s important. Most of what artificial intelligence, as a field, focuses on (and rightly so) is building a model of an environment. This learning involves taking raw sensory data and turning (or rather, integrating) it into a set of concepts and relationships. But with the simple case of a game, we don’t really need to learn what the rules are, or learn how to take actions; we can focus entirely on how to play well.
This post is in many ways a response to 
Imagine a system, any system. It could be Super Mario, the solar system, facebook, a cell, whatever. Now, this system is almost surely filled with various quantities, be they distances, velocities, hidden variables, or just countable things. I will call each quantity an x. (One x might be how fast Mario is moving along the left-right axis, while another might be how many lives Mario has.)