The article considers two situations where group wisdom is tapped using a simple democratic process, and shows that it works in one instance and doesn't work in the other. A more interesting question, and one that is far broader in scope, is to try to identify which community structures amplify individual wisdom and which structures diminish it. Wikipedia and Slashdot would probably be extreme examples -- the former has built something remarkable by leveraging the collective wisdom of unremarkable individuals, whereas the latter has managed to take people who are smarter than average and reduce the discourse to juvenile banter. This should be interesting to people here, since Hacker News is an experiment in this vein. But I think the issue is important enough to merit rigorous study, so that people who are building various types of social sites can start with an idea of what kind of features to put in and what to leave out.
I speculate that the type of structure is the major factor causing the quality of social interaction to diverge. Wikipedia is people collaborating to make something, and only incidentally communicating with each other, on talk pages and user pages and so forth. Slashdot (and all its ilk, including Reddit and HN) is primarily people communicating with each other on comment pages. (I disregard story voting here because they don't have the same process, but also because voting is a relatively trivial input for a user, whereas writing a thoughtful comment/profane flame takes significantly more effort.)
In a nutshell, Wikipedia's product is not communication between users, so the Greater Internet Fuckwad Theory (http://www.penny-arcade.com/comic/2004/03/19/) finds less purchase. Slashdot/Reddit/HN do produce such communication, and then micro-optimizations determine the course such communication will take, like the moderation details/ability to downvote/etc. That's my hypothesis, anyway.
James Surowiecki, in his book "The Wisdom of Crowds", addresses some of this. Of the three conditions he says are necessary for the crowd to be wise, he says the following which applies to the holiday meal planning:
"Diversity and independence are important because the best collective decisions are the product of disagreement and contest, not consensus or compromise. An intelligent group, especially when confronted with cognition problems, does not ask its member to modify their positions in order to let the group reach a decision everyone can be happy with. Instead, it figures out how to use mechanisms -- like market prices, or intelligent voting systems -- to aggregate and produce collective judgements..."
The use of crowd wisdom to select a holiday meal is doomed as soon as the author says "eventually you realize there's only way to please everyone." That's not the way you use crowd wisdom. The better way is to give them some # of tokens and let them spend the tokens on possible meals. The final selections won't please everyone, but more people will be happy at the party. If there's limited meal choices and everyone has veto power as per the article, the solution will quickly go to null with the number of voters.
I watched Who Wants to be a Millionaire quite frequently, and it's interesting how the crowd fails when it fails to provide the right answer. The problem is always the same. People who don't really know do not vote randomly and they do not abstain. They vote for the cliche and therefore don't cancel each other out.
90% accuracy from hundereds of people is actually pretty bad. A few people in the crowd undoubtedly know the right answer but they have no way of telling the others that they do. The "ask a friend" life line has the advantage that the friend will tell them how certain he or she is. So the contestant can at least disregard that opinion.
So my conclusion is that wisdom of the crowds doesn't even work very well in those areas where it is supposed to work well. We rarely work in an environment that is similar to the cookie jar example. Most of the time we are surrounded by cliches and ideologies. That's why I'm usually very reluctant to believe a source that cannot tell me how it knows what it claims to know.
