Moravec's Paradox

My regular work in Robotics, and my interest in King and Pawn endings meets at the name Moravec. In the field of robotics, Hans Moravec is a leading researcher, and the creator of Moravec's Paradox. Basically he (and Marvin Minsky et al) argue that seemingly intelligent tasks such as playing chess aren't particularly difficult for computers, while seemingly simple tasks such as walking or picking up an object are quite difficult to get robots to do.
In the field of chess studies, Josef Moravec (1882-1969) composed a number of wonderful K+P endings. I have even thought of bringing them closer together by using Hans Morovec's work on repulsion fields in robot navigation as a way of finding the solution to various K+P endings, in a kind of "Moravec solves Moravec" way. But I haven't progressed much beyond thinking it's a good idea.
For now I'll show you a Moravec study (White to play and win), which could also be titled "Moravec's Paradox", as the obvious idea for White doesn't quite work.