Regarding the architecture, I was thinking specifically of the mixed delivery of a whole new one (Metro/touch-driven) and the old/current one (classic Windows/mouse-driven). I've read quite a few articles about this but one I find myself agreeing with the most is Gruber's: http://daringfireball.net/2011/06/windows_8_fundamentally_fl... -- the main point is Microsoft failed to make the choice to potentially come up with two products, one designed for Touch and this new Metro architecture (ARM, HTML5 apps, etc…) and one geared for the desktop computer. They're so caught up on their perceived value of the "Windows" brand, they're failing to make the hard decisions…
As to your second question, I think my above point also covers this one in part. Microsoft is in a position to make some radical changes and improve on a product but avoids making those choices. Even the Surface will have a version with a full-blown CPU and ability to run desktop apps. Metro is caught in the middle of this "identity crisis" and, as such, fails to deliver a good experience in either camp. Furthermore, I think the core design model of "colored boxes" fails to deliver the unique experience that each app can offer: my experience with the "Music" app can't be the same as the "Control Panel". It becomes forgettable and everything blends together way too much. Yes, this particular point maybe a bit subjective but I'm sure some UX specialists could pitch in and give the hard evidence. My point is a "gut feeling" but informed by knowledge of what a good experience is… :)
Thank you for your comment! :)