I still haven’t decided if FriendFeed, which aggregates activities from other websites like YouTube, Twitter, your blog, Flickr, etc., is the ultimate walled garden of our personal data or possibly the answer to the Centralized Me problem.
I keep pushing FriendFeed to take the lead in the DataPortability wars by allowing users to export all that stuff they gather. For now, their feeds and APIs allow access to that data, but it all links back to FriendFeed in one way or another. Over time I’m hoping they start to release the actual data in a responsible way. From my discussions with them, they seem to want to do that, and I believe them (I generally don’t believe the big social networks when they say the same things).
In the meantime though, the service just keeps getting more useful. Tonight they released yet another feature called personalized recommendations. You can now view the items posted by your friends based on how interesting the network thinks each item is. You can view results by the last day, week or month. No RSS feeds for now, but co-founder Bret Taylor says they’re coming soon.
This essentially slices data two ways. The first filter is people you’ve subscribed to, so presumably you’re somewhat interested in what they have to say. The second filter takes a look at how many people are commenting or bookmarking the items to determine if it should be highlighted.
A natural next step is for them to release the feature without the first filter, so results are shown across the network. Allow people to tag items, and you’ve got yourself a breaking news engine that may be incredibly useful. Taylor says this is something they might do in future.