jan 4


In the comments of this post, my friend Rico makes an interesting comparison: lexapro on Twitter Search vs. lexapro on Google. Very different results! It suggests an interesting question: could conversational knowledge eventually usurp search knowledge -- and doesn't it already, in many ways?


Isn't it just "information" versus "opinion"?

Therefore it's not really surprising, that there are differences. You could argue, that Google tends towards "opinion" but the top results are still information-based and they will stay this way.

posted by KMB at 4:28 PM on January 4, 2009

Look again. It's not really "opinion."

posted by Rex at 4:39 PM on January 4, 2009

Definitely for things that happened 5 minutes ago, Twitter is more effective. But eventually, maybe for long-term overall information, conversational knowledge will provide better results for humans, by humans?

Remember: Google's results are only a guess as to what is relevant to a query - a guess based on a set of algorithms that must be comparably effective among the universe of (web-spider-accessible) known information, constrained by what a very large network cluster can compute hundreds of thousands times per second. I think a couple thousand human conversations could be a much better distiller of information, even if it's entire content is a distillation of what is on Google. It would be at least as effective as Google, and in many cases more effective as there are more inputs to human discourse beyond a web spider. As long as the relevant population size is sufficiently large - for most topics, there's always enough of a percentage of 7 billion people to have a large enough relevant group - the results will be a better guess than what Google can spit out.

posted by Brian Van at 4:49 PM on January 4, 2009

OCD grammar addendum: "even if its entire content" (delete apostrophe)

posted by Brian Van at 5:02 PM on January 4, 2009

I didn't know what Lexapro was before I clicked the two links, but had assumed it was some sort of information service for lawyers, maybe like Lexis-Nexis. As it turns out, it is not.

posted by Rachel Sklar at 5:15 PM on January 4, 2009

All this tells me is that a lot of Twitter users take (or know) about Lexapro. Draw your own conclusions. (I have.) So what other search terms do you wanna test this theory with? Prozac and Absinthe?

posted by krucoff at 7:39 PM on January 4, 2009

Having conversational knowledge supplement spidered results would be useful and I'd prefer to see Google snatch up Twitter than Facebook.

posted by rico at 11:42 AM on January 5, 2009

NOTE: The commenting window has expired for this post.