> This could be of interest compared to known heuristic algorithms. The problem is that it isn't. Yes, a big enough cellular automaton can be used to solve pro

> This could be of interest compared to known heuristic algorithms. The problem is that it isn't. Yes, a big enough cellular automaton can be used to solve problems, but the ratio of "processors" to "problem" isn't practical and we are not learning anything to improve heuristics because the "solution" here is only practical on a 2d surface.