I think at least one major flaw in that argument is in it's refutation to the point of an infinite god existing "outside" the universe. The counterargument was that since the universe is everything, there is no "outside the universe" where god could be.
This to me seems to be wordplay - using the definition of "universe" as "all there is", when, infact we cannot know whether the thing we've traditionally called the universe, i.e. the area of space-time where the laws of nature we observe apply (which true for the universe as far as we can observe), really IS in fact "all there is". For all we know, what we've been calling the universe that is bound by the familiar laws of nature may be a tiny corner of reality, perhaps even unrepresentative of the much greater whole.
There are, for example, many multiverse theories which postulate that the result of the big bang, the universe we live in, is just one local expansion event in vastly larger, perhaps infinite, "bubble bath" of expanding local "universes".
Some cry foul at the use of the word "universe" thus, but I contend that when a term that has been traditionally used to denote something is revealed to be a missaplication, but the term has stuck, then perhaps it is right to bend the definition of the word, and not stick rigidly to the original meaning.
For an example of this, take the word atom - which originally meant indivisible. Atoms were thought to be the smallest particles, and indivisible, as their names suggest - yet later it was found that this was not the case - atoms are composed protons and neutrons, and protons and neutrons are composed of of still smaller particles, quarks.
Yet we do not stop calling atoms atoms, and pass the name on to what we call quarks, because quarks really do seem to be indivisible.
Similarly, the original use of the word "the Universe" would indeed refer to everything there is, but then again, we thought that everything there is is limited to the outcome of the Big Bang. What if it's not? Do we then relable what we now call the multiverse "the universe", and invent some new term to describe the particular area of the whole which we inhabit and can observe? That would be profoundly confusing - as confusing as starting to call quarks atoms, and inventing a new name for atoms.
It seems much simpler to simply allow the language to evolve new meanings to words, and go with the flow - "the universe" in my mind, may be just a locality of the multiverse, and I wouldn't have trouble talking about "universes" - a plural that would not have made sense under the traditional usage of the word.
As for "proof" of god's non-existence, I doubt very much if that would be possible. Yet I do think that there exists a good evidentiary case to be made for disembodied, infinitely capable, conscious beings being far too implausible to take seriously, if the matter is given careful scrutiny, in the light of what we know about cognition, it's mechanisms, and the world we live in.
I'm quickly posting this because I really should go to bed.
http://www.youtube.com/watch?v=2KxHD6o259I
Inside I was laughing with glee while watching this video. I am pretty sure I had worked out a similar argument to the one in this video myself before (something like that if God is something specified, He has limitations, but if He isn't, He is a mystery), but I'm not certain whether the argument actually is correct.
It seems obvious that the only way 'God' can be disproved is by means of scrutinizing its definition. And that leaves me wondering whether this is not just some clever play with definitions of words (just like the 'ontological argument for God' is).
Any thoughts?