Guys, this guy isn't quite as incoherent as you're making him out to be. He's addressing a very well-known philosophical concept call the "Problem of Induction." See:
http://en.wikipedia.org/wiki/Problem_of_induction
The basic jist of the argument is that using evidence-based reasoning (induction, more formally,) to arrive at truth assumes that evidence has some relationship to truth, and moreover that patterns that are observed in the evidence will remain constant. A very simple illustration is that you can't prove the statement "All swans are white" simply by observing white swans, since there's no guarantee that you've observed all the swans in existence, or that the swans you've observed remain white after you've observed them.
At a more 'meta' level, the problem of induction gets thornier since you can't resolve it inductively; That is, pointing out cases where patterns in evidence lead to a correct answer as justification for the concept of "evidence" assumes what you're trying to prove.
Add in some of the "presuppositionalist" apologetics of Cornelius Van Til, (who stated that only the existence of God allows for a rational/logical universe,) and you realize that SyeTenb's argument isn't quite as philosophically incomprehensible as you're all making it out to be.
HOWEVER, just because the argument is marginally coherent doesn't make it less stupid: There's light-years of rational distance between suggesting "The existence of a God who wants there to be a comprehensible universe would explain the existence of a comprehensible universe," and asserting "The existence of God is the ONLY way in which a comprehensible universe can exist."
"I axiomatically assume a God who created a universe in which induction is a valid path to knowledge" and "I axiomatically assume that induction is a valid path to knowledge" are both ways to resolve the problem.
