Charbax.com

What will be a killer feature is when they filter filtered searches using an algorithm on the amount of likes users get for their items. This will filter out the good stuff from all the noise automatically.

For example:

- If I add some intelligent, precise, relevant items on my feeds that other users like on friendfeed, then my friendfeed activity should be rated more relevant than the random tweet that is going through all the feeds.

- Friendfeed should implement liking of comments on items, so a user can gain “points” by posting intelligent, relevant, precise comments in all sorts of real-time conversations on friendfeed.

- Real-time conversations on friendfeed should be filterable by relevancy of the comments. The interface should make it possible for thousands of users to be in the same chat room, yet still every user by default only sees the most relevant chat comment items, while the rest of the noize stays in layers further down which are hidden by default but can be shown as well.

- This will make it possible to type any topic of your interest on the friendfeed global search, then view a real-time stream of items and comments being posted about that topic. And even with thousands of users interested about the same topics and items, frienfeed should filter not only by friends, but also using those algorithms on general likes but also on likes by friends, friends of friends, friends of friends of friends, and stuff like that, including also based on likes of liked users, likes of liked liked users. And so on. To determine in fact a personal relevancy on the stream of news to each user and not only a general relevancy.

I posted this at http://www.techcrunch.com/2009/05/13/friendfeed-enables-peoplegroup-tracking/#comment-2747669