TED: Luis von Ahn on CAPTCHAs and Massive-Scale Online Collaboration

January 11, 2012

I can’t say enough good things about this TED talk. Well worth your time, and incredibly inspiring. Feel free to ignore anything I’ve written here, but do yourself a favor and take a few minutes to hear von Ahn’s presentation. Absolutely fascinating.

Whenever you’re filling out a CAPTCHA (those blurry, wavy words they sometimes force you to type on web pages), you should think of Luis von Ahn. He was involved in the early days of CAPTCHAs, but ended up starting a company called reCAPTCHA that revolutionized this seemingly mundane task.

CAPTCHAs are a necessary evil – there to determine who is a real person, and who is an automated script trying to “trick” the system. While humans have no problems deciphering the altered text, today’s computers (2012) are still not sophisticated enough to pass these basic texts.

On average, users spend 10 seconds filling out CAPTCHAs. When von Ahn did the math, factoring in the millions of people filling out CAPTCHAs on a daily basis – the numbers were staggering. He realized that he could put this collective effort to better use, and his reCAPTCHA project revolutionized how one task can serve more than one purpose. For every user who goes through the reCAPTCHA process, their work helps to digitize books… one small word at a time.

I haven’t real Malcolm Gladwell’s Outliers, but in his book he talks about “the 10,000 rule” – a magical number of hours one must spend in order to master a specific skill. In his talk, von Ahn talks about a magical 100,000 number – the number of people it took to build the pyramids of Egypt, the number of people it took to get a man on the moon.

In the past, organizing and paying this large number of people was incredibly difficult; with the advent of the Internet, these constraints become less problematic.

Von Ahn is looking to radically change how we go about learning languages with his next project, Duolingo. Similar to reCAPTCHA, users will help translate information on the web into other languages, as they learn a new language. It’s a stellar idea that accomplishes two goals with one task: