It's doing about what I expected. As I've been saying for months, WinRT is far too immature and unpleasant to be taken seriously by developers at this point (notice that Microsoft can't even write a decent email client with it, much less something significant like a WinRT version of Office). Although I've been hammered for making saying this, I'll repeat it here - there just isn't enough incremental market share to be gained from suffering through a major WinRT effort, even had Windows 8 really taken off. For WinRT to be worth developing for, there needed to be a huge movement of *ARM* tablets - Clover Trail does nothing for WinRT marketshare since it can also run traditional Vista and Win7 apps. We're in the process of writing a touch app for Windows 7, and it's been as easy as pie, it needed a new UI with a lower-density touch-friendly design (easy with WPF), and make sure the app is lean and battery-friendly. Due to the nature of the app it made a good test-case for WinRT evaluation. Firstly the fact that WinRT isn't supported on Win7 and Vista was immediately a problem, and will be until the Win8 adoption levels get much higher. The WinRT API itself was awesome if you're trying to write a calculator app, or a RSS scraper, or the sort of games that used to be written in Flash, or a few other things that it has extensive support for. Anything beyond that and you'll spend all your time fighting the API instead of getting useful work done.
In some ways the WinRT thing reminds me of the early days with the PS3 and Xbox 360. Like the PS2, the PS3 was difficult to develop for but developers put up with it because that's where the market share was. The transition to the PS3/X360 changed that dynamic, suddenly the Xbox offered a comparable market and since developers were starting from scratch anyway they opted for the easier and more mature development system and the PS3 tended to get the lower-quality ports. Like Sony, Microsoft has a habit of aggravating-to-use APIs (Win16, Win32, OLE2 - the .NET BCL was a delightful exception). Like Sony, Microsoft continued this tradition at a time of transition when the market share argument wasn't sufficiently in their favor.
Philosophically, WinRT reminds me of Newspeak from Orwell's 1984, which was deliberately crippled language who's primary goal was to make it impossible to say or think bad things (from the government's perspective). WinRT is a deliberately crippled API who's primary goal is to make it impossible to write bad programs (from Microsoft's point of view). But Microsoft's point of view is not our point of view, nor our customers' point of view. They have real problems that need real software solutions, and if WinRT makes it impossible (or even difficult) to accomplish that, then Apple and Google are happy to offer viable alternatives.