I'd imagine it's solvable by just about everything. This is one of those cases where the core problem isn't that the problem is hard, it is that it is so easy that it is a challenge to get humans to even perceive it as a problem on such a small scale. Any planning system will hit something like this all but immediately.

I'd say a lot of the earliest AI optimism was a failure to understand just how hard things that in the simplest case are so simple we don't even perceive them as problems in the first place. For a much more complicated example, I doubt anyone in 1970 would imagine that in 2018 we're still trying to get robots to walk from here to there. (Yes, lots of progress has been made, but it's still very much cutting edge, a research project, and a bespoke one-off effort each time, not a "Oh, I'll just swing on down to the robot store and pick up a walking chassis".) We don't see it as a challenge because for our conscious minds, by the time we're old enough to be doing robotics research, it's zero cognitive effort for us.