I think MS has created a fair amount of PR damage through the rather careless use of the term "legacy" in reference to desktop applications. Though it may not be the "cool new thing", using the term legacy brings lots of bad connotations and an expectation of impending deprication. Other than maybe Sinofsky, I don't think that is really the intent. This is just like the "Win8 apps will be programmed in HTML" fiasco that preceeded Build. A few misplaced comments cause a stir, and then complete radio-silence from MS allows fears to fester. MS may not want to share plans (or may not even know them yet), but it wouldn't take much for a vague hint at a product roadmap. If you don't have the confidence of your developers, they aren't going to provide the software that is critical to the success of the new platform.
No doubt there are a lot of things that will be changed along the way. I wouldn't be too surprised to see a lot of changes even for Win8RC. They certainly are hearing a lot of feedback, whether they are publicly responding or not.
It will be interesting to see what comes next for the desktop. The obvious thing would be to see WinRT extended to desktop applications, but it's not clear if that really makes sense. The Metro/contract/sandboxing parts of WinRT wouldn't apply. Efficiency in the runtime would be nice, but the main issue is creating a modern windowing API for the desktop (that works with native code as well as managed). That might be a very different animal from WinRT. All we can do is wait and see.