Growth expectations, page-38

That is not the point. Are you saying you'd turn your back on Microsoft? That video explains where they are at when it comes to server farms. From listening to that video the conclusion for me is that FPGA's aren't disruptive in any real sense. They can't revive Moore's law. They might provide a factor of 100 x equivalence scaling by accelerating the functions of parts of the cloud software. FGPA's also have similar fundamental limits to most CMOS devices as they are made from three terminal devices (transistors) which have reached their limit on a size/cost basis. They can't jam anymore transistor onto a chip without upping the cost which destroys the economic models that the semiconductor companies have been trading on for decades. Some scaling will probably be achieved using FGPA's but not enough to keep up with Moore's law which is the Holy Grail now that the end is near/here.

As mentioned IMO stackable ReRAM is the true disruptive technology that will change computing and all associated devices and infrastructure forever.

I've asked before how many gates would you need to achieve a relatively simple learning task (like voice or face recognition which is currently achieved in software)? No answer was forthcoming from this forum. Forget about true intelligence. Learning for me is just the beginning of intelligence and when translated to silicone, unless it has enough connections, will only produce a sophisticated sensor which may be useful and may be commercialised but won't be disruptive.

Please explain how Peter's patents are disruptive because I can't see why yet.

Eshmun