> Who's to say there isn't something out there we haven't discovered yet

Occam's razor? We should work with as few assumptions as possible to get a model with the largest scope. Otherwise we get stuck with a hard to falsify mess.