Ok, so I've dipped out for a while on this thread, but this argument comes up time and time again, and honestly, it's not that much of a barrier for science to cross. People who allow this to be their stumbling block should think of time in a less linear fashion, especially when you're talking on a universal scale like this. It's useful for us to think about time as a linear dimension in Cartesian space, but it's almost certainly not an inherent property of the universe. If you have to think about time like that, forwards and backwards, and in a straight line, then the best thing I can tell you is that there was no beginning, there was no nothingness and then somethingness. Our own big bang may be attributable to a date 15 billion years ago, but to try to keep going back to "what came before that" is reductionist. Because we haven't filled in the gaps yet about the precursors or prerequisites to the big bang (materially speaking), this is a facile argument for creationists, but it shouldn't be the backbone for any arguments.
Also, I have no idea what you mean by "all empirical science rests on the premise that the future will resemble the past." Entropy and enthalpy, man. Shit changes and the universe is always different for it.
Great post, Townie. Love the bolded part.