The reality is that apps are the new native programs. You wouldn't take it seriously if it were suggested that your desktop programs be replaced by your web browser. (Although more and more of them are seeing just that happen). By the same token, apps are the new native code that can be faster, more efficient, and provide an overall better experience than a general web app that must dynamically fit a wide range of devices.
What Paul is saying in that article is that Apple has, with iOS, an operating system that is already designed from the ground up to be simple and efficient. It is much easier to extend that system to have more powerful capabilities built upon the foundation and rules of that system than it is for Microsoft to strip away the legacy stuff from Windows. Microsoft built Windows 8 with this duality because they felt that it was the only way to get their users to accept an operating system that was re-built in the modern sense. In 8 the duality was too stark, and the advantages of the Metro side of things were not yet developed to the point to tempt many users. In 10 they have blurred the lines between the two sides a bit. But make no mistake, Microsoft is not giving up on the 'metro' side of their OS. They continue to develop software for it, and in my opinion, what will happen is that slowly, over time, that side of the OS will come to dominate, with the 'desktop' side left behind for legacy compatibility. In this way Microsoft gets to build a new operating system that grows up along side their old one.
This is what I think Paul is missing in his analysis. Apple is much more willing to jettison legacy hardware and software in order to advance more quickly. Microsoft doesn't have that flexibility. So while I agree that an iPad pro could, potentially, take a lot of sales from the Surface market, I also believe that Microsoft is on the right track to do what they need to stay competitive in the long run.