Sheb raised the point I wanted to. Organisms are filled with redundant stuff. I've said before, nature trends towards two things: Hack jobs and efficiency, and somehow the latter comes from the former.
It's all a hack job. But it's also all (at least anything that's been around for a handful of generations and not causing you problems) a hack-job that generally works. Which is obviously more efficient than a hack job that doesn't, or an efficiency that falls over due to a critical loss of redundancy.
I mean, why even have two sets of genes? Surely once the whole hot, steamy gamete-on-gamete mix'n'matching action is concluded, everyone could get away with one of each. Except for the XX/XY pairing, where us men, at least, need the pair... but women can use just the single X though, right?
(We should all know the answer to those questions. Consider them rhetorical. Polyploidy (>2N) also seems to help greatly with adaptability of organisms (up to 12N, last I heard!), but it looks like we manage well enough at the level of diploid (2N), whether or not we could somehow reduce our functional genome to the haploid level if we
really tried, by deliberate hacking and patching back up with the molecular equivalent of duct-tape.)