But science and the acquisition of knowledge is all about pushing deeper into the implementation. We can't look at something like gravity and say, 'yep, that seems to work fine, adheres to its interface, we don't really need to know how it works, just that it passes its tests (apples fall, helium balloons float above oxygen, etc)'. We can't leave it there, because we have to know the implementation. We'll never improve if we don't. And we still haven't nailed it, but that's no excuse to stop.
It's like a beginning programmer looking at a hash table object and hitting 'view source'. What kind of programmer would he be if he didn't want to see behind the curtain at least once? How could he be expected to excel?