Chris Brogan and Clarence Smith Jr. (am I supposed to just say “@chrisbrogan and @dykc” to appear more ‘linked-in’ and whatnot?) just posted a collaborative piece about managing the (many, many) streams of information from sources like Twitter (or any aggregator, like Google Reader, I’d add).
My immediate thought is that this is a perfect task for adding intelligence on the client side. For instance, I would love to have more options in Twitterific for controlling the incoming tweetstream.
With a ‘pause’ feature, if @InterestingDude is temporary at ReallyBoringConfFest07, and will be until Thursday, I’d love to be able to put him on “pause” until then, so that I don’t see all his excited tweets about impromptu hallway meetings with ReallyBoringPeople and ReallyBoringProjects.
But I don’t want to quit following @InterestingDude permanently; I’ve just lost interest for the next few days. I don’t want to have to remember to re-follow him, either; I just want to skip the next couple of days (or hours or whatever) of tweets from him.
The problem Chris and Clarence are examining is how to get more of the info that’s interesting and less of what’s not, when the “interestingness” isn’t always determined by the person or the topic. The ability to follow tweets or blog posts by “interesting people” is a start, but only a very primitive one. There are lots of fascinating people doing fascinating work, but who have hobbies that bore me to tears (and vice versa). There are a very few people who write so well that it hardly matters what they’re doing, I just enjoy reading their writing. And there are many, many people who normally would not be of interest to me at all, except when they happen to have seen/talked to/read/bought/tripped over the very piece of information which is extremely relevant to something I’m working on or discussing with other people.
Perhaps I should just quote Chris and Clarence a bit:
The problem arises when the people you follow are initiating and participating in conversations that you do not find interesting at all. [...] Said another way: I might like YOU, but not be into every little thing you are into. [...] How can we catch your tweets about social computing but skip the tweets about being stuck at JFK for a 3 hour delay?
It’s not that YOU should have to twitter differently, but rather, we should have a way to adjust the lens on what comes through our “interestingness” gate. And of course, this is all relative to whatever you’re interested in, who, and often times where. For instance, if we’re visiting Seattle, we might want to get MORE about the area around us than less, in case something newsworthy is happening (like avoiding a traffic jam).
They also (briefly) consider the personal/social implications of such filtering:
How do we on/off the conversational flow of people in such a way that we receive more of what’s interesting to us (again, very relative), without it signifying anything negative about a person?
This I don’t find as compelling, partly (perhaps) because so far I only have an extremely limited scope of friends (in anything resembling the traditional sense) who are Twittr/Whatevr users. But I feel pretty confident that the notion of any social implications of being filtered or un-followed (is that a word?) will adjust with the tools. As people spend enough time swimming in the flow of social network infostreams, it will become more apparent that social standing with a person is largely disconnected from how much of your socnet output they’re consuming at any given time (perhaps it will always have more to do with which parts of our output they consume than how much).
Why can’t we have a system that’s partly like Flickr’s “interesting” and “favorites” system, that helps train Twitter (and other networks) to predict which conversations will matter to us? Something more than keywords. How do we apply this same thinking to the people we currently “follow?” What if Clarence loves when Chris talks about data centers, but doesn’t care about Chris’s current trip to New York City? How could we “gate-on” based on information, and then “gate-off” when the interestingness vanishes? [...] How could we build tools that turn on and off our view of someone’s Twitter stream based on things like: location, context, content?
This is where I think there’s a big ol’ gaping opportunity for an interesting next-gen aggregator. How many different ways are there to arrange the panes in a blog reader? Well, OK, there’s a zillion, but the point is, how many of them are enough better to compel me to change to them? Just like with the mail client, the things that really are going to motivate people to change en masse will be the addition of “intelligence” to make managing the burgeoning information therein easier.
The “pause” thing would be cool; but man, a client that could determine what I was interested in at the moment, and help me find “more of this and less of that”, would really pique my interest. Whether it did it by reacting to explicit actions like tagging or even rating (3 stars or 4?) tweets or people, or by implicit observation of my behavior (in the same vein as how Google Reader’s “Select by Auto” works), it would be a great step in the right direction.
Plus, you know, a pony. And an aeroplane. And a perpetual motion machine. I mean, it can’t hurt to ask, right?