Twitter will need filters in order to thrive longtern

New communication channels often get praised because signal:noise ratio is way higher than mature channels. Email was fantastic - then came spam and 10,000 people having my address. Part of the love for Twitter is that it still has a pretty high signal:noise ratio. But this changes over time as you add more followers - people will ruin their own experiences by introducing noise due to social pressure and design failures (see my posts "Psychology of the Salad Bar" and also how this is impacting FriendFeed).

So what does this mean for Twitter? IMHO their continued growth and longterm utility will be based upon filters - both explicit user controls (such as groups) and implicit prioritization (order by "interesting" not chronology). If Twitter doesn't do it, certainly the ecosystem of developer apps will nail it. Ev has talked about designing with constraints - he famously fantasized about a social network which limits you to only 10 friends. Well, that's one way to keep high signal, low noise. Another is to treat data intelligently, and my bet is that's the way Twitter (or the ecosystem) will go.

Here's some of the filters i'd find potentially useful:

Explicit (ie dependent upon me)

• groups/folders
• deprecate/mute tweets containing kw= or #=
• Google Reader integration - if i already subscribe to someone's feed, i don't also need to see their tweet announcing they just posted

Implicit (ie do it for me)

• sort by interestingness (based upon # of retweets, how often i tweet @ that person, etc)
• collapse dupes - turn retweets among those i'm following into a single group
• conversation view a la TweetTree

btw, if you want to follow me: http://twitter.com/hunterwalk