ON AMAZON'S MECHANICAL TURK: DEFINING WORK 2.0

WORK 2.0:  MICROCHUNKING WORK 1.0

Who'd have thunk?  That potentially the next best use of the Internet since Google (via Bill Gross's GoTo.com/Overture efforts) could come from, not a bright-eyed, bushy-tailed, "web two'ed" start-up at 285 Hamilton Ave. in Palo Alto (via Om Malik), but from a scrappy web services team at "Web 1.0" Amazon.com?

At the admittedly very real risk of getting ahead of myself, this week's announcement of Amazon's Mechanical Turk (called AMT for the remainder of this post)  may mark the beginning of a whole new Internet gold rush, not seen since the Search 2.0 gold rush since Google's IPO.  AMT could be the shot that inspires the launch of a whole new stream of start-ups the world over, especially since this is about re-configuring outsourcing as we know it.

What is AMT?  As the announcement dryly and geekily describes it,

Amazon Mechanical Turk provides a web services API for computers to integrate Artificial Artificial Intelligence directly into their processing by making requests of humans.

Developers use the Amazon Mechanical Turk web services API to submit tasks to the Amazon Mechanical Turk web site, approve completed tasks, and incorporate the answers into their software applications.

To the application, the transaction looks very much like any remote procedure call - the application sends the request, and the service returns the results. In reality, a network of humans fuels this Artificial Artificial Intelligence by coming to the web site, searching for and completing tasks, and receiving payment for their work.

Here's what a typical work request looks like, for Amazon's A9 search engine.

TechCrunch explains AMT in simpler terms:

The “machine” is a web service that Amazon is calling “artificial artificial intelligence.” If you need a process completed that only humans can do given current technology (judgment calls, text drafting or editing, etc.), you can simply make a request to the service to complete the process. The machine will then complete the task with volunteers, and return the results to your software.

Volunteers are paid different amounts for each task, and money earned is deposited into their Amazon accounts. Amazon keeps a 10% margin on what the requester pays.

At first glance, one might go, huh?  What's the big deal about this?  You advertise for a task you want done on the internet, and get folks to respond to your job offer. 

For those of you around for the last Web 1.0 party, it might even sound similar to what a bright-eyed start-up backed by the Benchmark Boys of eBay, Respond.com was all about.  It's still around by the way.

But this is different this time, empowered by the technologies and the philosophy of "peer production" of Web 2.0. 

What gets me most about this idea, is that it can be applied not just for inexpensive software outsourcing as envisioned by AMT, but as a barter marketplace of services for service and/or content. 

In fact it's a whole new way to think about "payments for peer production", as I've mused about before.

VC Paul Kedrosky provides a good example of this is in a on-the-money post describing the potential precursor to AMT:

"I’m fond of the virtual sweatshop Amazon just launched via Mechanical Turk, but the progenitor of this stuff online is, as is almost always the case in technology, the porn industry. Here’s Cory at BoingBoing from January of 2004:

Someone told me about an ingenious way that spammers were cracking "captchas" -- the distorted graphic words that a human being has to key into a box before Yahoo and Hotmail and similar services will give her a free email account. The idea is to require a human being and so prevent spammers from automatically generating millions of free email accounts.

The ingenious crack is to offer a free porn site which requires that you key in the solution to a captcha -- which has been inlined from Yahoo or Hotmail -- before you can gain access. Free porn sites attract lots of users around the clock, and the spammers were able to generate captcha solutions fast enough to create as many throw-away email accounts as they wanted."

This is a whole new way to "microchunk" work, paraphrasing Umair Haque's great phrase (and work), just as many Web 2.0 companies are focusing on microchunking content (see pithy post on this by VC Fred Wilson today).

Susan Mernit dissects the economics of the current tasks being "microchunked" by Amazon in the AMT example above in this helpful post:

"A typical task right now is for A9 and reads as follows:

"You are presented with the name and address of a business as well as a set of photos taken along the street where the business is supposed to be located. Your task is to identify the best photo of the business that is listed."

For that, you get .03; if you do it 179 times, you get 5.37.
Amazon's estimates work so that if you have spent an hour to fulfil 1170 of these requests, you get $35.00, which means your time is worth $35 an hour if you can fulfill about 1 per second--if it takes you 5 seconds to fulfull each task, your time is only worth $6-7.00.

Amazon says that they take a slice of each transaction and that workers can transfer money to their U.S. personal bank account or to their Amazon.com account."

Microchunking work this way could remake parts of the outsourcing industry as we know it today.  One could see vertical work marketplaces that focus on microchunking tasks across a wide array of industries.  And this approach has potentially oodles of network effects of all types (Umair, any thoughts?).

While this may not be the next Industrial Revolution (maybe), it at the very least is at least as interesting as when Henry Ford took the century old ideas around Assembly Lines and "Googlefied" it into a "win-win-win" strategy for workers, manufacturers and consumers.  (Google of course did this for searchers, search providers and advertisers). 

Here's to the next, big Work Rush.

Comments

[Eli]

A very thoughtfull post,
Some ideas are clearly very well thought through.

-Eli, the turk

[Pete Cashmore]

If MTurk succeeds, I seriously believe it could revolutionise the web. But equally, it could be brought down by spammers, hackers and captcha breakers. A lot of these tasks could actually be achieved by a script, rendering the data gathered virtually useless. And if every site with a captcha can now be beaten by a single user and his army of human minds, we're going to see a *massive* increase in splogs and spam.

Either way, MTurk has already made its mark by simply putting this out into the world - I'm sure other people will pick up the idea and run with it, hopefully adding the necessary layers of security if Amazon has failed to do so.