On Edge, Martin Rees – a noted cosmologist, someone who definitely knows what he’s doing – has an article about multiverse cosmology. (This is a general-public article, BTW, not a technical one)
Now, I’ve got one primary objection to this article, which is that it makes multiverse cosmology sound like something much more agreed upon in the scientific community than it actually is; its current status is more on the lines of “a definite possibility, one whose details haven’t been worked out yet, but has some really good features as a prospective model.”
But that aside, he has some interesting things to say on the subject. One intriguing possibility he brought up, which I hadn’t heard before, was that in an infinite universe, someone has very likely developed extraordinarily powerful computers, ones powerful enough to simulate not just individual life forms but even large sectors of the universe itself, and that for all we know, we may be living in such a simulation.
The philosophical implications of this sort of thought are rather intriguing. On the one hand, we couldn’t necessarily intrinsically tell whether we’re in such a simulation or not, since we don’t have any outside reference points to work by; but it may be possible under such circumstances to communicate with the simulators, by some means or another, and indicate that yes, there’s life inside here.
The good/bad thing with this line of thought is that it seems that one could use it as a basis for virtually any argument; it’s like theology, but even less constricting. Can’t explain some property of the universe? Maybe the simulators programmed it to be that way. (Like the “well, that’s just how God planned it” excuse – a great way to avoid answering any inconvenient questions)
What’s interesting, here, is that this seems to arise naturally as a consequence of certain kinds of cosmology: If the universe really is large enough to contain all things, one must expect such simulations to exist therein. But this gives us no information whatsoever about whether we’re elements of such a simulation.
Anyway, this is just sleep dep and a great deal of physics (from the past few days) talking at this point. There are some interesting issues brought up here, but they’re about on a par with solipsism as far as really arguable points. Perhaps it’s time for bed.
Yonatan Zunger’s Blog 
I say, if there’s no way to tell the difference, does it matter?
for a different cosmological take on things check this out:
http://www.electric-cosmos.org/
I think the plasma physicists are on to something. It’s a strong argument, at any rate.
Any sufficiently detailed simulation is indistinguishable from reality.
Discuss.
Or in this case, would that be “any sufficiently advanced simulator is indistinguishable from God?”
Anyway, the statement is obviously true; the simulation just has to be self-consistent. Someone living purely in the sim would have nothing else to compare it to, and therefore trivially conclude that it was reality.