Massive-scale online collaboration | Luis von Ahn

TL;DR

In this content, the speaker discusses the invention of CAPTCHA and how it has evolved to help digitize books and provide free language education through Duolingo.

Transcript

How many of you had to fill out a web form where you've been asked to read a distorted sequence of characters like this? How many of you found it really annoying? (Laughter) OK, outstanding. So I invented that. (Laughter) Or I was one of the people who did it. That thing is called a CAPTCHA. And it is there to make sure you, the entity filling out ... Read More

Key Insights

• 🤖 CAPTCHAs are used to ensure that web forms are filled out by humans and not automated programs. They are effective because visually-impaired humans can read them, but programs cannot.
• 💡 CAPTCHAs can sometimes show random characters that form a word. This can lead to humorous or confusing situations, such as users thinking they have to wait or getting offensive combinations of words.
• 📚 The CAPTCHA Project and reCAPTCHA, developed by Carnegie Melllon and later acquired by Google, have found a way to use CAPTCHA efforts to help digitize books. When users type CAPTCHAs, they are also transcribing words from old books that computers can't recognize.
• 🌐 Many websites, including Ticketmaster, Facebook, and Twitter, use reCAPTCHA to help digitize books. This has led to digitizing about 100 million words per day, equivalent to approximately 2.5 million books per year.
• 🎨 The use of two randomly chosen words in reCAPTCHA has sparked the creation of CAPTCHA art, where users create drawings or illustrations based on the word combinations they see. This has become a popular internet meme.
• 🔒 The reCAPTCHA system has successfully engaged over 750 million people to help digitize human knowledge, which is approximately 10% of the world's population. This highlights the potential of large-scale collaboration on the internet.
• 💬 Duolingo is an upcoming language learning website that aims to provide free language education by combining it with web translation. Users learn a language while simultaneously translating web content, creating a fair business model that doesn't discriminate against the poor.
• 🚀 With the potential for 100 million people to contribute to projects facilitated by the internet, the question arises: what can we achieve with this large-scale collaboration? Projects like Duolingo and digitizing books highlight the possibilities of harnessing the collective efforts of millions of people.

Questions & Answers

Q: What is a CAPTCHA and why is it used on web forms?

A CAPTCHA is a distorted sequence of characters that is used on web forms to verify that the user is a human and not a computer program. It is used to prevent scalpers or spam bots from submitting forms multiple times.

Q: Why do humans find it easier to read CAPTCHAs compared to computer programs?

Humans, especially those who are not visually impaired, have no trouble reading distorted characters in CAPTCHAs. Computer programs, on the other hand, have difficulties in accurately recognizing and deciphering these characters.

Q: How does the CAPTCHA Project at Carnegie Mellon University help digitize books?

The CAPTCHA Project at Carnegie Mellon University has evolved into a new project called reCAPTCHA. When users type a CAPTCHA, they are also helping to digitize books. The words they type are actually words from books that computer OCR technology cannot recognize. By presenting two words, one known and one unknown, users not only authenticate themselves as human but also help digitize an unknown word accurately.

Q: What is Duolingo and how does it combine language education with web translation?

Duolingo is a website that offers free language education while simultaneously translating the web. It utilizes real content and sentences from the web to help users learn a new language. While users learn, they also provide translations, which are as accurate as those done by professional translators. This innovative approach combines language education with valuable web translation, creating a fair business model that doesn't discriminate against those who can't afford expensive language learning programs.

Summary & Key Takeaways

• CAPTCHA is a security measure used on web forms to ensure that the person filling out the form is human, not a computer program. It works by presenting distorted characters that computers have difficulty reading.

• CAPTCHAs are used to prevent scalpers and ticket-buying bots from buying up large amounts of tickets. However, CAPTCHAs can sometimes be annoying and difficult to read for humans as well.

• The speaker's project, reCAPTCHA, takes the wasted time spent on CAPTCHAs and puts it to good use by having users help digitize books. This has resulted in 750 million people helping to digitize human knowledge.