Lunch with BS: Luis Von Ahn
Source: Anjuli Bhargava
He's a MacArthur Genius. Popular Science has declared him one of the world's 10 most brilliant scientists. Discover has adjudged him one of the 50 best brains in science. He's one of the most innovative people under 35, according to the MIT Technology Review. At 22, he sold a company to Google and earned enough to retire; only he didn't.
Sitting at Machan, The Taj Mahal Hotel, New Delhi, I think I am definitely the wrong person to be meeting Luis Von Ahn, the creator of Captcha, reCaptcha and now Duolingo. What will a dyed-in-the-wool techie find to say to possibly the most technology-challenged person around?
Ahn walks in looking impossibly young (he's 36), relaxed and unassuming; he doesn't look like a mad genius or a geek - much more the professor at Carnegie Mellon University, a position he holds. He's in India to spread the word about Duolingo, his latest idea and venture, and to meet students at Indian Institute of Technology Delhi.
We order - a minestrone soup for him and a soup and salad for me - quickly and start talking. A Guatemalan by birth, Ahn headed to college in the United States (Duke University) to become a mathematician - a short-lived ambition, as he moved to computer science subsequently.
Ahn was in his first year of PhD at Carnegie Mellon when the chief scientist of Yahoo! (Yahoo! then was like the Google of today) came to speak at his college and listed 10 challenges that the company did not know how to solve. Ahn - then 21 - listened, went home and tried to solve all 10. While he couldn't find answers to all, he did think he had one answer and that was Captcha - the code one enters into websites to distinguish a human from a computer programme or a robot. Yahoo! was overjoyed and started using Captcha almost immediately.
"Did Yahoo! pay you for it," I ask. Ahn says it didn't but that he and his PhD advisor (who had helped him with it) were so overjoyed that Yahoo! was using it that the thought of getting paid didn't occur.
Soon after, almost every website started using it. That's when Ahn started thinking about his first discovery a bit more. He calculated that on an average 200 million people were typing Captcha in a day. And it took about 10 seconds for a person to do it each time, which, taken together, was a mammoth waste of time. Moreover, he knew that most people found it irritating. He began to feel bad - and a bit guilty - about it.
He started thinking of ways to make use of that time. "There is something special about that time - your brain is doing something that computers cannot do," he says. He wanted to use this time better - for humanity.
So he came up with reCaptcha - a second round of sorts for Captcha. The idea was that every time someone types a Captcha, it can be used to help companies digitise books. For example, Google was trying to digitise all books that ever existed and Captcha could help, he figured.
He goes into a complex explanation of how it works. I smile politely and nod at the right junctures so he thinks I am with him, while I focus on my salad. When I tune in again, he has just finished explaining that while humanity is still spending the time, now it's been put to good use. I, too, feel good at the thought.
Again, I ask whether Google - someone, anyone - ever paid him for his effort. reCaptcha, the company which produced the eponymous product, I learn much to my relief, was bought by Google in 2009, which left Ahn several million dollars richer.
By that time, Ahn had finished his PhD (he was a professor of computer science at Carnegie Mellon at the time) and had made enough money to retire but instead set up Duolingo, a website that allows one to learn several languages online and free.
Ahn wanted to do something in the field of education. Coming from Guatemala - a country where approximately 75 per cent of the population live below the poverty line, according to a 2015 World Bank report - Ahn argued that education, instead of bringing equality to social classes, deepens and widens the divide between the rich and the poor. "The rich can buy themselves the best education in the world and that helps them get richer, whereas the poor barely learn how to read and write so they never make much money."
I nod vigorously, adding that this sounds more like my country than his.
Why languages, I ask. Education being quite general - and rather vast - he's decided to start with language. Knowledge of English, he says, can make all the difference between having a good job and not having any. "In many countries - yours included - people earn twice as much if they are familiar with English than if they are not." Languages are also immediately valuable, he says.
Ahn says the language market is also a strange one. About 1.2 billion people worldwide are currently trying to learn a foreign language - a majority of them are trying to learn English primarily for a job and they belong to the lower socio-economic classes. Yet most of the ways to learn the English language -through online portals - on offer in the United States are expensive. In Latin America, apps and sites offer it at similar prices. "Most of the people who were trying to learn English were doing it to get out of poverty but the cost of learning it would push them further into poverty," he says. The irony of it struck him.
So Duolingo, which he co-founded in 2012 with one of his PhD students, is free. He has raised $83 million in funding and has 110 million users, of whom 25 million are active learners in any given month, learning 17 languages.
But if it is free, can it be sustainable, I ask. Why would venture capitalists be so eager to part with their money if there is no hope of any return?
Ahn maintains that learning will always be free - and no advertisements come up on the site/app to irritate users either - but once you have learnt something, if you want a certificate from Duolingo, there's a small $20-fee to be paid. Two years ago, users started writing in thanking Duolingo for teaching them but also saying they needed a certificate after completion of a programme.
That's when Ahn started looking at the TOEFL and IELTS models. These cost around $200 per head and test takers need to reach the testing centre. About $10 billion is spent globally every year on such testing. "While in the US, this may just be annoying, in a country like India, it's more than annoying. This $200 is a lot of money, plus test centres are not easily accessible." Ahn himself spent a tidy sum travelling to El Salvador to take the TOEFL as his own country had run out of seats for the year when he took the examination. So a separate app, the Duolingo test centre, allows one to obtain a certificate for $20 after taking a standardised test from the app.
What credence would a Duolingo certificate hold over a heavyweight TOEFL, I ask. That, he admits, is the tough part, but now 12 US universities (and some companies) have started accepting it and the number is growing. So far, they have earned just half a million dollars through certificates but he hopes Duolingo at some point will become the de facto standard. He says he has a good chance unless rivals stop charging the monopoly prices they do.
Language may be where he's begun but it will not be where he plans to end. He says a billion adults around the world don't know how to read and write, but 50 million of them have smartphones. "If we can reach even five per cent of them, it would mean a five per cent decline in world illiteracy rates."
At some point he wants to teach everything that can be taught better by a computer for free. "The standard education model has many good things but reach is not one of its strengths."
Reach is where Ahn hopes to find his strength. Time will tell if he will manage to reach his goal, but this Guatemalan - the first I have ever met - makes me feel he might just manage.
| }
|