Amazon launched Mechanical Turk this week, the product I worked on for the past year I was at Amazon. Here is the official blurb from the Amazon Web Services site on Mechanical Turk:
"Today, humans still significantly outperform the most powerful computers at completing such simple tasks as identifying objects in photographs – something children can do even before they learn to speak. However, when we think of interfaces between human beings and computers, we usually assume that the human being is the one requesting that a task be completed, and the computer is completing the task and providing the results. What if this process were reversed and a computer program could ask a human being to perform a task and return the results? What if it could coordinate many human beings to perform a task?
Amazon Mechanical Turk does this, providing a web services API for computers to integrate Artificial Artificial Intelligence directly into their processing."
I was responsible for Product & Program Management for Mechanical Turk, and I can honestly say this is the most interesting concept I've ever worked on. It is fun to work on something that is a bit "out there."
Over the past year, one of the things I often what would be the reaction to the MechTurk concept. I waited anxiously for news of MechTurk to spread into the blogosphere. Here are some posts and comments from others.
From Greg Yardley's Blog:
"Did I mention how much I like this initiative? Because let me tell you - outsourcing the sort of grunt work the Turk could potentially do is a giant throbbing pain in parts that shouldn’t be hurting.
Another inspired initiative by a market leader looking to become a market maker."
From David Flanagan's Blog:
"At first I was intrigued by this. Interesting use of technology. (There's even an API for it: hey, I could use Mechanical Turk tasks as a captcha for blog comments, and readers would earn money for me...) Free market goodness...
I don't know if I'm more disturbed by the philosophical implications of this Matrix-lite inversion of control, by the commodification of intelligence it represents, or by the implications for workers rights. A system like this will just accelerate the race to the bottom: even offshore worker's jobs won't be safe. Heck, they could probably train pigeons to do some of the image recognition tasks they're currently paying 3 cents for. 3 cents probably buys a lot of pigeon pellets."
There are also a bunch of great comments on digg.com. They range from some folks thinking this is a scam, others thinking it is 'slave labor', and other admitting to how strangely addictive it can be. I can tell you that I and many of the folks I worked with on this project, were surprised by the addictive nature of MechTurk.
BlogoScoped seems to have really picked up the developer angle:
"This service opens up great new possibilites if you're writing web software. The Mechanical Turk is still in Beta and I'm sure there are many issues to settle once it runs full-speed, but the potential is immense."
I think TechDirt really missed the boat by claiming MechTurk is a copy of Google Answers. You could build an app like Google Answer's using MechTurk's API, but Answers falls well should of MechTurk. One, Answers doesn't offer programmatic access. Two, Answers doesn't provide a mechanism to ensure the quality of the results (e.g. Qualifications). Three, Answers doesn't provide a way to render the question format in highly specialized manner.
