As a developer, you have to move forward.
Sure. But I don't think that time is now. Maybe by Win 9. The fact that Microsoft doesn't have a version of Office for WinRT is the sort of thing that makes experienced developers pause; if Microsoft believed their own marketing they would have shipped some version of Office RT by now, a preview, a lite version, something. They've had years to work on it after all. And yet they haven't.
Remember, Microsoft said the same things to developers when Windows first came out. Windows 1 sucked, and developers that bought in to Microsoft's line tended to go out of business. They said the same things when Windows 286 came out. It sucked, and developers that targeted it tended to go out of business. They said the same things when Windows 3 came out, but this time it didn't suck, and coincidentally Microsoft came out with Word and Excel for Windows. More recently, WPF was unusable for serious work until Microsoft used it to write Visual Studio, and suddenly it became acceptably fast and stable for serious apps.
WinRT is too new and unfinished, and the market just isn't there yet. It needs time to cook a bit.
I'm afraid you don't know what you're talking about. That sentence possibly describes the .Net BCL, but not WinRT. WinRT is very new, and though object-oriented, it is almost - but not quite - completely unlike Win32 and the .Net BCL.
Windows Runtime provides ZERO friction for targeting all three architectures!!!! Why would you NOT do that?
Because it means incurring the cost of writing new software for very little incremental gain. If the market changes and I have a good chance of recouping the development costs for a WinRT version then I'll reconsider. Especially if the WinRT market turns out to be better than iOS or Android. Which could happen, especially for business apps. Though it isn't likely until Windows 9 comes out.