Tweeting for the machines

I’ve been having wonderful conversations these days with my temporary host, Rob Johnson of GNIP.

GNIP is doing well, and they’re doing in in a way that goes against normal internet wisdom. They are selling tweets.

A random Joe, somewhere in the world, sends out a little tweet like, “Enjoying a tasty burger with the gang at McDs.”

Does this interest you? Probably not. It’s the classic “Who cares?” criticism of Twitter. But the fact is that everyone has at least a few people for whom the smallest tweet matters. If you were part of “The gang” of Joe’s, you would take a little pleasure and interest knowing that he mentioned you. In science-speak, this is the “social grooming” that takes place in all human relations, similar to how apes will groom each other by off picking lice. (See all of Robin Dunbar’s work.) But if you’re not in Joe’s circle, this tweet boring drivel.


The genius of GNIP (and other emerging companies) is that when you take enough of these tweets together, you can find gold. Twitter supports lots of these little social grooming and social signaling messages, and then GNIP can roll them up and sell the aggregate data to a whole new class of interested parties.

The Wall Street Journal reports how Hedge Funds now use data from GNIP to directly influence trades. Enough people tweeting about their hamburger meal can be important information for beef futures, the performance of McDonalds, and a hundred other small signals that aren’t so obvious.

This calls to mind the classic essay by George Dyson about his visit to Google. His hosts explained, “We are not scanning all those books to be read by people, … we are scanning them to be read by an AI.” In the same way, these tweets are now a signal back into “the machine” that more and more directly effects the world.

These are fascinating times.