This has little to do with the media business, but it’s so cool that it’s worth sharing. Check out http://recaptcha.net.
This service supplies free captchas – those squiggly words supposedly readable only by humans and not computers, used to prevent automated submissions of forms. The captchas are taken from scanning books where the OCR software fails. After enough humans verify a captcha, the word is corrected as a part of the book scan.

This solves two real world problems at once – supplying unique captchas to Web sites, and deciphering unreadable words using human power. Brilliant!

(Read more from the site to see how the software knows that a captcha is correct before it’s actually verified.)