Dave Naffziger’s Blog

That’s how many captchas are filled out by people proving themselves to be human.

Enter reCaptcha - they have figured out a way to make all of this captcha solving useful - digitizing books.

Each ReCaptcha captcha has two words:

1. An unidentified word from a scanned book
2. A known word

So, you get to digitize the world’s information by doing something that machines can’t. Assume an average book is 80K words (200 pgs by 400 words per page) and that OCR is 95% accurate: 4K captchas mean 1 book is digitized. That creates a potential of 15K books a day.

Cool. It kind of reminds me of the first SETI screensavers.

I’ve installed it on my comment form.