If you are a tourist interested in seeing a baseball game while in New York, you can Find out which of its teams are in town simply by sending a message to AskForCents. com. In a few minutes, the answer comes back, apparently supplied by a machine, but actually composed by a human. Using humans to process information in a machine-like way is not new- it was pioneered by the Mechanical Turk, a famed 18th-century chess-playing machine that was operated by a hidden chessmaster. But while computers have since surpassed the human brain at chess, many tasks still baffle even the most powerful electronic brain.
For instance, computers can find you a baseball schedule, but they cannot tell you directly if the Yankees are in town. Nor can they tell you whether sitting in the bleachers is a good idea on a first date. AskForCents can, because its answers come from people. "Whatever question you can come up with, there’s a person that can provide the answer-- you don’t have the inflexibility of an algorithm-driven system," says Jesse Heitler, who developed AskForCents. Mr. Heitler was able to do this thanks to a new software tool developed by Amazon, the online retailer, that allows computing tasks to be farmed out to people over the internet. Aptly enough, Amazon’s system is called Mechanical Turk.
Amazon’s Turk is part toolkit for software developers, and part online bazaar: anyone with internet access can register as a Turk user and start performing the Human Intelligence Tasks (HITs) listed on the Turk website (mturk.com ). Companies can become "requesters" by setting up a separate account, tied to a bank account that will pay out fees, and then posting their HITs. Most HITs pay between one cent and $ 5. So far, people from more than 100 countries have performed HITs, though only those with American bank accounts can receive money for their work; others are paid in Amazon gift certificates.
Mr. Heitler says he had previously tried to build a similar tool, but concluded that the infrastructure would be difficult to operate profitably. Amazon already has an extensive software infrastructure designed for linking buyers with sellers, however, and the Turk simply extends that existing model. Last November Amazon unveiled a prototype of the system, which it calls "artificial artificial intelligence". The premise is that humans are vastly superior to computers at tasks such as pattern recognition, says Peter Cohen, director of the project at Amazon, so why not let software take advantage of human strengths
Mr. Cohen credits Amazon’s boss, Jeff Bezos, with the concept for the Turk. Other people have had similar ideas. Eric Bonabeau of Icosystem, an American firm that builds software tools modeled on natural systems, has built what he calls the "Hunch Engine" to combine human intelligence with computer analysis. The French postal service, for example, has used it to help its workers choose the best delivery routes, and pharmaceutical researchers are using it to determine molecular structures by combining their gut instincts with known results stored in a database. And a firm called Seriosity hopes to tap the collective brainpower of the legions of obsessive players of multiplayer online games such as "World of Warcraft", by getting them to perform small real-world tasks (such as sorting photographs) while playing, and paying them in the game’s own currency.
Amazon is successful with Turk is probably because()
A. Turk has very p software infrastructure developing potentials
B. its previous model has laid a solid foundation for an extension
C. its system is based on artificial artificial intelligentce
D. Turk is profitably adaptive to almost any software infrastructure
参考答案:B
解析:
[解题思路] 原因细节题。文章第四段第二句说:亚马逊已经有了非常广泛的用于连接买主和卖主的软件基础,Turk只不过是把现有的系统延伸罢了。那么亚马逊使用Turk成功的原因就可以从另外一个角度来说,即:其本身的系统为进一步延伸奠定了基础。