The basic idea is that you can make a computer more affordable to people by eliminating the high-up front cost of a PC and replacing it with a sort of leasing arrangement. This, of course, has been done for years, and it's never been all that popular because even dumb people are able to figure out that they'll end up paying $600 for a $300 computer. In almost every case, you're better off just putting it on a credit card and making payments. But the innovation that Microsoft is trying to patent is that the computer will have a "scalable" capability, and that the user will be able to easily unlock both software-based functionality and hardware-based power as they need it.
In the cloud computing paradigm, this is workable and desirable, because many businesses find that in an early stage of their online app or web site they have long periods of time in which they're in development and have very modest resource needs, but after they launch, they hope to have rapid uptake and will need to scale up very quickly. With the cloud, they don't have a buy or lease a lot of powerful servers up front that will mostly sit idle, and don't have to worry about the pitfalls of bringing new hardware online quickly once they're successful. The service provider can use new visualization technologies to efficiently share their overall computing resources among all their clients, and everyone's happy.
At first glance, this Microsoft idea looks very similar. But then you realize that the users of these Microsoft PCs aren't going to be accessing hardware and software resources over the network and sharing them efficiently with all of the provider's other customers. What it looks like is they're going to be getting a powerful gaming PC that's been crippled and throttled back to allow only basic performance, and if they find they want more power, they'll be able to unlock it by paying more. It will presumably also have a lot of pre-loaded software that will be able to be accessed as they need it (which isn't that bad of an idea, considering how cheap disk space is).
But the draw of cloud computing is that you're not paying for hardware that's left idle. But in Microsoft's scenario, you are. For this program to work, all of these PCs need to be powerful enough for the users to grow into them, so they need to be expensive PCs. They're expensive PCs that in many cases will be allowed to rot and go obsolete without ever being used to their full potential. Where does that make sense? So somebody is going to be paying for those expensive, crippled computers up-front, and guess who will eventually foot the bill? Whatever poor suckers sign up for this program.
The other thing that's wrong with this idea is that it's a perversion of the inherent magic of the personal computer. You know, the central idea that made Microsoft the superpower that it is today? The thing that makes PCs great is that they are so flexible that they've been put to uses that their makers have never imagined. Hardware and software can be combined and recombined in so many different ways that mind-boggling innovation has brought us new advances in computing that surprise us all every day. These pay-per-use computers, for this program to function as designed, will not be PCs. They won't be flexible, or open. As a user, you won't be able to exert any real control over the hardware or software that's on them. So you'll be left behind when other PC users are enjoying the fruits of the latest innovations.
Ultimately, most attempts by the computer industry of simplifying or "cheapifying" the computing experience beyond a certain threshold have failed. Dumb terminals and network computers ended up being neither more useful nor cheaper than the PCs they were supposed to replace. I predict that Microsoft's pay-per-use idea is also doomed for the scrapheap. And good riddance.