It started with CAPTCHAs. You know those things, the crazy little scripts you have to transcribe before you can download a file or sign up for a mailing list on the Web? They are used to make sure you are the human you claim to be, and worthy of the service you are requesting online. Luis von Ahn is the guy who came up with that bright idea, which takes 10 seconds of time, but when multiplied by 200 million CAPTCHAs, totals to about a half million hours a day. And he felt bad about that. So he created reCAPTCHA, which adds a second word, pulled from a scan of some old book where an optical character reader can’t make out a specific word. But you can. Complete that second word in reCAPTCHA and add to the sum of global human knowledge too (Not to mention some fraction of a penny to von Ahn’s wallet. My description is the exec summary. Guy Gugliotta provides the details in the NYT if you are interested in an actual explanation.) So a really good idea for the age of the internet. Everybody gets something for nothing. So how does Ahn top that?
He’s going to get language students to translate the Web for free. And the premise is so amazing that I was able to sit through about 10 minutes of one of those TED presentations before his talk outran my attention span.
Ahn argues that translating the content of the Web is simply too costly to ever occur. The number of professional translators is too vanishingly small to make the tiniest dent in the process. Even bilinguals are in short supply, less than 100 million among billions, according to Ahn. And like us professionals, most of those amateurs are quite sensible about their attitude towards working for free. They don’t care to do it, thank goodness. And that’s where Ahn and his group saw their opportunity.
In von Ahn’s words: “The solution was to transform language translation into something that millions of people WANT to do, and that helps with the problem of lack of bilinguals: language education. . . . It is estimated that there are over 1 billion people learning a foreign language. So, the site that we’ve been working on, Duolingo, will be a 100% free language learning site in which people learn by helping to translate the Web. That is, they learn by doing.”
Nowadays, learning a language takes a lot of time and money. Ahn wants to trade the time of language learners for translations they provide. The user interface is still not public yet, but the principal will be similar to that behind CAPTCHA. Purchasers of the translator won’t be buying a snippet from a single bad translator, but a statistical analysis of multiple bad translations, prepared by multiple bad translators while feedback from other users allows the crème of the translations to rise to the top of the translation memory. Quantity, as the saying goes, has a quality all its own.
Really amazing. It will be interesting to see how it works for translation, but my gut feeling is that it’s sure-fire for language study, and it’s easy to envision what a great community platform this will be for language learners. And free all around, except at the very edges, so perfect for the Web. Here’s the sign up for the alert, with the go-live date about three weeks out.