Guy Gugliotta has an article in the Times on the use of Captchas to interpret scanned text through a special software program. Apparently when you see 2 words in the Captcha, one is a true word which the software knows, the other is scanned text which the software isn't sure of. So if you get the true word right, you're a human and the software will consider your answer to the other. Very interesting. They claim 500,000 hours of brain effort are being spent on replying to Captchas, so their software converts that to useful work.
It raises the question to me: how are economists capturing these gains to utility (or however they'd word it)? It's unpaid work, but it's very useful, converting the poorly scanned texts of old NY Times and 19th century books into readable, accurate English. Come to that, how do they account for the improved research which historians can now do using Google Books and Google Scholar?
No comments:
Post a Comment