If I have to follow Occam's Razor (which I consider a tool for human scale, not for the degree of complexity of the whole cosmos)... I can also realize that, when trying to reduce the number of agents in the whole system, if I discard the possibility of an universe that was detonated by some force (call it whatever you like, but the concept of intelligence that usually appears in these debates seems terribly anthropocentric to me) I will be left between the alternatives of a universe collapsing out of nothing, or an eternal universe, cyclic or not. Those alternatives do not simplify our human understanding of it either. Seems that we have to work beyond the classic reason tools to make some progress, don't we?

And the discussion of randomness and unpredictabilty, there's probably no point in trying to discuss it if you are a firm believer in the strong AI. Sincerely, good luck with your reverse-engineering quest. You are aiming high, I just hope that if you ever come across good ideas and tools that do not seem to fit your rules for admission, you are able to reconsider them.