I honestly don't know what they were thinking with the Zune. What's amusing is that by all accounts it's a really nice portable media player. Microsoft basically finally made a nice portable media player just after the point where growth stopped in portable media players.
Courier I find interesting. I was one of the people who thought the demo looked really interesting. It totally spoke to me as a device that if it operated the way it did in the video (which is a big if), I'd have had a real use for. But I don't know if that means it would have actually been successful as a product. I think if Microsoft thought they could build it, and that it would be successful, they'd have done so. The fact that they scrapped it (or did whatever it is you do with a product concept video that you never actually said was going to be a product) shows me that they don't understand that market, or don't think it exists sufficiently.
I think part of the problem is the way Microsoft Research is structured. According to Bill Buxton (who is kind of my dream "wish he was my uncle"), Microsoft Research is basically tasked with incubating long-term technologies.
In this video, he talks about how some of the stuff he's been working on over the past decade is now starting to get used in shipping projects: http://channel9.msdn.com/posts/LarryLarsen/CES-2010-NUI-with...
While I think that's awesome for Microsoft to be that forward-looking, I think they also need a more short-term skunkworks.
Maybe not as radical as Google's "20% time", but I think something that shortens the amount of time between "concept and prototype".
Windows Phone 7 seems like a pretty good idea. It's a fairly novel approach to interface design; but it didn't just spring up. Microsoft basically plodded through ten years of bad mobile strategy and pissed away their technology lead until even they couldn't pretend that they were headed in the right direction.
If you contrast that with Palm, who also developed and launched a completely new platform in a relatively short period of time, I think it's obvious that Microsoft works in glacial time. Then again, Palm's gamble hasn't really paid off either.
I think Microsoft couldn't possibly bet too big on mobile. I know it's trendy, and I'm certainly not being profound, but I don't think it's ludicrous to consider it the largest growth market in computing for the next decade. I think the problem (which the original article points out, is that Google has pretty much screwed their existing business model by giving away Android).