The author's biggest miss seems to be the realization that the range of computational devices and their application is growing exponentially and that Microsoft doesn't necessarily need to be in every game.
The iPhone and iPad are examples of computational expansion not necessarily a complete paradigm shift.
Besides, some of the stuff MS has shipped in the past few years (Win7, Office20##, WinPhone, etc.) seems to run pretty well. They might not be quite so bleeding edge anymore but it sure looks like a bunch of people are working hard on their products.