I hate this sort of thing really. Saying "what should Microsoft do with...loadsoftechnologies" is basically the wrong way round of thinking about it. It should be, "what would customers want to do which uses this
sort of technology?".
Seriously, I'd suggest for techie people to look up a "Computers & Technology" type hour on a TV channel like QVC. Ultimately, it really doesn't matter what the miniscule percentage of the types on Channel 9 and other technology sites think, its the sort of people the likes of QVC market to.
Look at XP. In terms of ordinary users, one of the features that I always see that impresses people is the Camera & Scanner wizard. More precisely, the fact that a user can plug in a camera and a little box will come up with a bunch of very-simple-to-choose-from tasks. Ordinary folk really like it. I fear that too many techies are almost inflicted with the Clippy thing and see that helpful box as a bad thing.
So, what you really need to ask is this, what sort of things do users do or want to do with their media that this
sort of technology encompasses? This is where the biggest problem lies in the proposed idea. Robert suggests "we need to build Internet Content Sharing into everything we do". I can tell you something FOR SURE: what you think is Internet Content Sharing is very different to the everyday user who would be mortified that something they shared is available for all to see - purely because they don't know, and frankly don't care, what the Internet is.
The biggest oddity with this whole area is that there seems to be very little that really needs to be done. There must be a key ideology to all this - its THE USER'S media and thus, you should give them as much control but make it as simple and as clear as possible, just like with the Digital Camera dialog, so the user doesn't at any point feel out of control of the media. Apart from that, open up APIs or whatever it takes so that Web Services/Smart Clients can actively and seamlessly work with the media via Explorer, create a standard and robust and simple security model for publication and encourage other essential services. Simple!
