I think that there are 3 important pieces to the functionality these guys can provide:
1) First is the basic aggregation across news/activity feeds. This is on its way to being accomplished, at least for activity streams that are exposed through APIs or in a structured fashion.
2) Second is smart filtering/alerting on that aggregated feed. This is in terms of both what you allow others to see about yourself and what you choose to see about others. This is where privacy comes in. This seems like a relatively short step from where we are today, but it will be hard to get exactly right.
3) Third is re-packaging the aggregated feed for consumption by services elsewhere. In a way, this is like a FeedBurner for your personal activity stream. It may format the feed in a certain way for certain services, it may restrict what gets sent in that particular feed (although this is more part 2) and it may add new information as necessary.
As an example of part 3), you could allow Amazon to receive your feed. You would send them the feed in a format they can process (standardized or handled automatically by your personal FeedBurner), you would only send them information that is useful and that you are comfortable with (e.g., what movies I've seen, where I am going on vacation, what I've bought at Buy.com) and you would bundle it with other useful info (e.g., Amazon Associate #).
Amazon could use this information to:
- Get to know you better and make timely offers based on your interests and activities (e.g., 10% off a guidebook for your vacation destination, a lens for the camera you bought elsewhere)
- Make offers to your friends/people you influence and give you some credit (hence the associate #). This would work best in combination with a Beacon-like service (see this past post for some discussion on this).
Anyway, this example is a bit weak and contrived. But to Josh's point, there is a lot of potential in taking all these feeds and aggregating/filtering/repackaging them. FriendFeed and the other 1st generation aggregators may get there, but there are probably a bunch of other startups with similar ideas.