I wouldn't be surprised if there are at least some characteristics that aren't well understood. For characteristics that developers like Cliff do understand and feel incapable of modeling with 100% accuracy due to a lack processing power, I think faster processors will ultimately prove useful. In my opinion, and this is just me, I feel fairly confident that at some point modeling will achieve perfect accuracy, not just in terms of sonic reproductive accuracy but the full gamet of parameter control. I can't prove that of course, but I tend to believe it.
I agree. It would be Malthusian not to. I'm not saying that processing isn't important, just that by the time such a system is know well enough (either by brains or brawn) the processing will have Moore's-lawed itself out of the picture as a limiting factor.