You are the computer
Way back when, computers were people. Problems were broken up into lots of little pieces, and shared across a room full of mathematicians armed with logarithmic tables, Napier's Bones, slide rules and the like. That way codes were analysed, paths plotted, and ballistic tables computed, ready to simplify the art of war. No one person knew the answer to the problem, they just had their little part of it, ready to deliver and add to the corpus that would become the final answer.
People still are computers, and the underlying concept carries across to ideas like Amazon's Mechanical Turk, where problems can be spread across many thousands of users, parallelising complex tasks. It's an approach that can help analyse air sea rescue photographs, or run art projects that simulate artificial intelligence. There's also something similar that happens when a problem goes viral across a social network. Names and addresses can be found as minor connections expand across the world, people identified, property returned.
This new networked society we're building is a strange creature, full of emergent properties that come from millions of eyes and millions of typing, touching and clicking fingers. We're still finding it hard to understand, believing that a literal camera that describes what it photographs to be the product of advanced artificial intelligence rather than the result of a person somewhere looking at a screen, typing a description and earning a handful of pennies. Sometimes the most advanced technology really is indistinguishable from the kind of magic we used to see on Saturday night variety shows that were broadcast from big tops on West Country beaches.
Perhaps we need to remain wary, ready to ask just why we're being asked the questions where being asked, why we're playing the games we're playing. Is that alternate reality game really marketing the latest Bond movie, or are you acting as the eyes and ears of an anonymous intelligence agency a thousand miles away or more? After all, there's no one less suspicious than someone who believes truly they're only playing a game…
Sure, you might say, that would never happen to me.
But it already is, in a way. Each time you fill in a reCAPTCHA form to access a website or post a blog comment you’re helping train the OCR engines at Google as they digitise books. The pair of words is there for a reason, one is a known, already recognised, piece of text that helps calibrate your response on the unknown text. But it's not just books you're helping digitise – you're also helping Google match the houses its streetside photography system has captured with the addresses in its mapping database.
Sometimes you'll get a reCAPTCHA form that looks like this:
Instead of text from a book, you're getting a section of a photograph with a number. Identify the number, and you're just part of the computer, helping match house and address, making it easier for Google to display the right picture when someone looks for a particular address. It’s a problems that computationally very hard, but now all Google needs to do is identify something that might be a number, deliver to one of the many millions using reCAPTCHA, and if identified, use the result. If its unidentifiable, the user will just refresh the form and a new image will be delivered, ready for identification.
You didn’t notice, and you weren't paid, but you were part of an algorithm, part of the machine. For the moment between looking at the image, filling in the form, and clicking submit, you were the computer.
Welcome to the machine. We’re so glad you’re here.
Simon Bisson