On Nvidia seemingly being in perfect position to ride successive hype cycles of technology: their product category is one of the few types of hardware with significant technical advancement in the last decade, making it only natural that people would seek to leverage that category of hardware with new software efforts. So less that they have been brilliant strategically and more that they have been brilliant technically and that has been creating the market so to speak. What is not clear is whether, as an industry, we are constrained by this technology into seeking solutions that can be massively parallelized.<p>More simply, when every leap in compute is a massively parallel architecture, every problem seems like it needs to be solved by a massively parallel system. But I'm guessing before this cycle is over we start to see the limitations of that.