Micro-computers enabled the GUI, they enabled personal computers on every work desk and home office. They put into everyone's hands capabilities and tools that were previously the purview of a select few. There are a great many more experts in typography today because of the power of the PC. The same goes for movie makers, photographers, etc.
Mobile computing isn't just about smartphones or computing on a handset, it's a platform and a paradigm. It's the appstore + sandbox model. It's more intuitive interfaces (e.g. multi-touch). Etc.
If you look at the iPhone or the iPad and say to yourself "these won't replace Windows" you will be fooling yourself. This is not a static industry, indeed it's one of the most dynamic, falling back to static conceptions of the world is a recipe for being left behind. You need to imagine what the possibilities are for the future of these paradigms. iOS has been in users' hands for less than 4 years, the appstore has only been around for 2 years. What will iOS and the application landscape look like after it's evolved for 4x or 5x that time? What will android and its applications look like? And Chrome OS? And WebOS? And whatever anyone else comes up with?
Apple knows what's up, they're morphing the OS X experience to be more like iOS, they can see a future where the experience of using a Macbook is more like that of using the iPhone, where installing applications over the internet is the rule rather than the exception.
Here's a simple hypothetical. Imagine the iPad all grown up, it's faster, it has terabytes of internal solid state storage, it has a larger screen. Plug it into a docking station on a desk and it becomes a monitor with a keyboard and mouse attached and connected to a wired network. It runs all of the grown up office applications anyone ever uses anymore, installed via the corporate intranet app portal or preloaded in the base install. Disconnect it and take it home or to lunch or to a meeting, it has 4G/wifi connectivity and connects to the corporate VPN and internet facing secure services (corporate email and web-based collaboration software) seamlessly. Could such a beast compete with a Windows based system? Certainly. And the real devices of this sort from 2015 and 2020 will blow this simple hypothetical out of the water.
"Desktop computing" won't go away, but it may well be redefined and subsumed by new platforms that come along. Indeed, I think it's far more likely that desktop computing in 2020 will share more in common with iOS and android than it will with Windows as we know it today.
