That's a good point and reminds of a professor I had in grad school. I remember having this discussion with him over beers once and while he could explain how any algorithm worked (or figure out how to explain), he didn't really care. What got him excited was taking all these low level pieces and using them to build things that never existed before and solving higher order problems.