Does this make sense to anyone?
http://www.wired.com/science/discoveries/magazine/16-07/pb_theoryI'm trying hard to grok how exactly massive data analysis can replace models and theory. Is he just using hyperbole?
Although I've now gotten a good idea from this whole Petabyte Era series of articles they've run... outsourced massively parallel pattern recognition.
You have massive sets of data to parse, and your hardware is crazy fast and awesome; but your machine intelligence and algorithms just aren't good enough to find what you're looking for - so you need humans to find your patterns. You go somewhere in Asia, start up or rent a data/call center, and hire a bunch of people really, really cheaply to eyeball sets of data.
For example, there's this project called HiRISE which is searching for the crashed Mars polar lander. They are relying on volunteers to scan through many tiny sections of massively high resolution images to find the impact site - which would likely appear in something like 30 x 30 pixels of the picture. Admittedly they have no money for this, but if they did, they could easily hire a room full of very low paid people to scan, once they learned what to look for. Such a thing could even be useful in searches for lost planes or for people lost at sea, assuming you had good enough satellite imagery.
I'm sure you could get a lot of people power for a relatively small amount of dough. And I think it's a business that will thrive in the age of MASSIVE DATA.