A hacker is just a satisficer that places little value on a norm or norms. A munchkin is an optimiser.

Removing one constraint allows a satisficer to achieve better results on all the other constraints; by contrast, an optimiser will violate as many constraints as it takes to get the best result on the optimised criterion.

I came back to ask a similar question. I would not call the issue of choosing cryonics more then a hack. I think it is the difference is that a hacker is often some one who has optimized well in a narrow area while a munchkin will look at the whole system and optimize it and constantly look for new rules to exploit. The difference I see EY drawing is one of local optimization verses global optimization(or at least an attempting to).