As I promised, I was reading Kathy Harris last night and found a good topic “Dumbness of Crowds” – in it she separates collective intelligence against “dumbness of crowds”. Using a Flickr example, she mentions how the Flickr userbase organically selecting the coolest pictures is an example of collective intelligence while a community asked to collectively create an original picture is demonstrative of the potential “dumbness” of crowds. Despite art’s subjective nature, I agree, the thing won’t have a guiding vision, which some think all great art needs to have. (Kathy Harris may also appreciate filmmaking’s ‘auteur theory’). As always, she enlightens us with a nice disclaimer about relying too much on a central web 2.0 tenet. I have to admit I was just a bit disappointed she didn’t veer the conversation into one of several hybrid approaches for community involvement.

Hybrid Approach # 1: The Benevolent Dictator

Benevolent dictators don’t exactly let the community decide (what self respecting dictator lets the community decide?), but like Google and Mayor Bloomberg and other powerful community leaders in a dominant position, they have the community’s interest at heart and generally heed the advice (or rage) of an advisory board made of community representatives, but they really don’t have to, and often depart from the script at will whenever they please. I used to think Craig Newmark was a Benevolent Dictator, but his flagging system has become so automated that its much more of a collective intelligence approach to cleaning up spam.

Approach #2: The Moderated Community.
This is alot like a community run by a benevolent dictator, but in this case, its more like a country run by Congress, instead of the President, or a city run by the Assembly instead of the Mayor. Many smart forum communities have established “wranglers” that steer the conversation, kick out users, and generally run the community under a mostly silent leader/publisher. This is a great approach, its only downfall is it takes resources to have alot of wranglers and moderators around, and for a bootstrap start up, I had to come up with something else.

Approach #3: collective intelligence meets historical standards.

This is our approach, which utilizes collective intelligence and community empowerment but also sets pre-determined thresholds based on a community’s historical leanings on the issue. For example, while community intelligence determines that a ’sex service’ ad on craigslist is deemed Offensive, who determines the number (or threshold) at which point these ads should be purged? My research on craigslist suggests that its somewhat artibrarily determined by Craig & Co. and is often randomized in a bid to foil spammers. On hykoo, we will take a city’s historical tolerance when setting these thresholds, and therefore each community will have its own threshold standards for exactly when ads are purged after being flagged. It may take 50 flags to be removed from the New York market, but the same salacious ad may only be flagged 20 times in Minneapolis before disappearing. While we’re still working on a way to quantify historical decency and other standards for every community, lets not get too crazy, we all know Vegas is not Omaha and vice versa.