Skip to content
Who's in the Video
Luis von Ahn is a Guatemalan computer scientist and tech entrepreneur. At the beginning of the century, he worked with Manuel Blum to invent CAPTCHAs, computer-generated tests that humans are[…]

It’s been proven that students taught by a personal tutor perform two standard deviations higher than those in a classroom. Luis von Ahn wants to replicate this effect with technology. He explains how Duolingo, the free language-learning app he helped create, continues to evolve in pursuit of that idea.


“We would like to make it so that everybody has access to the absolute best type of education, regardless of whether you’re wealthy or not. So we want to be as good as a one-on-one tutor.”

Luis von Ahn: We know one thing about education; well we know many things about education, but one thing we know is that one-on-one tutor is actually much better than a lecture hall. There's a famous study called the 2 Sigma Study and it's been replicated in many different disciplines, but the idea is the following. You compare people who are taking a class from a classroom, you know, 30 people, 50 people or whatever in a classroom; you compare their performance against somebody who is being tutored one-on-one. And it is always the case that the people that are tutored one-on-one perform two standard deviations higher than the people that are in a classroom. That's why 2 Sigma, two standard deviations.

What that means is they are in the 98th percentile of the performance. So people who are in a one-on-one tutor perform in the 98th percentile. That means it really is better to have a one-on-one tutor than a teacher. And it makes sense because if you're in a classroom size 30 or whatever, the teacher can teach — they have multiple choices. And one of them is to teach to the lowest student, but then it's very inefficient to all others. Or the teacher can teach to the middle for example, and then what happens is it's inefficient to everybody above the middle and the ones below the middle don't understand anything. So there's no great way to do it in a classroom and this is why one-on one-tutors are much better because they completely adapt to your personal needs. However, we can't have one-on-one tutors for everybody because then that would require having as many teachers as there are students and that's just not scalable.

So, what I would like to do, and this is one of the stated goals of the Duolingo, is we want to be your one-on-one tutor for learning a language. We want to be as good as a one-on-one tutor. I don't believe we are yet. Duolingo is as good as a classroom. We have studies that show that we are as good as a classroom. That's great. But I think in the future we're going to be as good as a one-on-one tutor and that's what I would like to do. And there's a good reason for that and it has to do also with the divide between people who have a lot of money and people who don't have a lot of money. If you have a lot of money and you want to learn something, the best thing you can do is find the best tutor in the world and have them teach you. For example, in language learning there was a very famous guy named Michel Thomas who was a very famous language teacher. The idea with him is if money was no issue you would hire him and he would teach you. The Princess of Monaco hired him and all kinds of famous people from Hollywood would hire him to learn a language and he would teach you a language in a few weeks. He was amazing. He would charge of course tens of thousands of dollars to do this. So if you had a lot of money, you had access to this type of education.

If you don't have a lot of money, the best thing you can have is either a classroom or maybe not even a classroom. So what I would like to do, and this is the thing with Duolingo, we would like to make it so that everybody has access to the absolute best type of education, regardless of whether you're wealthy or not. So we want to be as good as a one-on-one tutor and that's the whole idea.

Directed/Produced by Jonathan Fowler, Elizabeth Rodd, and Dillon Fitton


Related