Dropbox's Content Hash Algorithm

ok, but why?

they offer no advantage. only one i can think of is it cheapens their compute load as it is easier to parallelize the computation... but i mean, it's already md5...