With Ageia's card having already debuted and the graphics maker going nuts to save face, I have one question. In this multi-core processing world.. Why do we need a card to do physics? I hope it's not a silly question, but wouldn't it makes sense to develop code to run on it's on thread on a cpu core it can reserve? After all, we're going to Quad core very soon... then 8 cores after that. (K8L on 4x4+) So why not just hand the work off to 1 or more of the CPU cores?