Comment by fc417fc802

> I don't see how you could have as a goal to explore alternative path without a good metric of what is a good alternative path (like rating a chess position), which to me, seems unlikely to exist.

As a counterpoint, we can algorithmically estimate chess positions and clearly expert humans are capable of demonstrating useful heuristics for exploring the state space of mathematical problems thus it would appear to follow that such a thing is possible for AI in principle even if we don't currently know how to go about building it.