There are 2 billion people in the world learning a foreign language. Most of them are learning it in schools or something. We only teach a little more than 100 million. So there's a lot more there. I also want to do for math the same thing we did for languages. And then after math I probably want to do physics. I mean so generally we just want to do as many subjects as possible and really have hundreds of millions of people learning on their phone.
It's similar. I don't think it's that bad, but I do think that generally when you say that you want to have positive impact on the world, investors a lot of times think that you're not going to make a lot of money. Fortunately, we did not have difficulty for a number of reasons. I think the first is that I had made a previous company that got sold for a lot of money and so investors thought that I was going to do that again and then pretty soon after we launched Duolingo started getting a lot of users. So even though we weren't making any money at first, we just had a lot of users and I think investors in general thought we were going to find a way to make money. I mean, I think what has happened with Duolingo is that we just have such a large audience that even though only 8 to 10% of our users pay us, that's still a very large number.
It was not marketing at the beginning. It was really just the product was so good. I mean, when we launched, it was a fun way to learn a language that was also free. And it was in 2012 and it was the first good app to learn languages. It was not the first app to learn a language but it was the first good one and I think that that just made it so that people downloaded and told their friends. So for the first several years the entirety of the growth was through word of mouth where people just told their friends. We didn't really do any marketing up until about 2019.
I think that it was free and fun and I think people just saw it when they tried it. They saw it and they're like, 'Oh, wow. This is free and it's fun.' And they told two of their friends and then their friends told two of their friends, etc., etc.
We do, we have a lot of studies that have shown what we do is we take people that just start using Duolingo. We let them use Duolingo for sometimes a year or two and then we measure again how well they know the language. So whenever they're starting we measure how well they know the language. Usually they're beginners and then after 2 years we measure how well they know the language. And it turns out people really do learn. For example, you can learn about the equivalent of five semesters of university level education in that language on Duolingo in about half the time.
Yeah, mobile based is because we really wanted to give access to education to as many people as possible and the one device that people have throughout the world is a mobile device. I mean, you can even everybody has one. It's not true that everybody has one, but a lot of people have one much more so than laptops or anything else. So it's really the cheapest way to reach a lot of people is to make a mobile app. So that was why mobile. In terms of AI, the other possibility for teaching people is to have humans teach them and that's very expensive to have humans teach them. And in addition to that especially for practicing conversation most people don't want to practice conversation in a language that they're not very good at with another human. Most people are either they get too nervous for that. So this is why we chose AI because you can actually practice with a computer and you don't get nervous about it.
The first one is all the content that we used to have to teach a language. So all the sentences and everything like that used to be done in part by humans and in part by computers. So some of the sentences were written by humans or at least verified by humans. Nowadays, the entirety of the content is done by AI. The second thing is that with the latest improvements in AI over the last 2 years, we're finally able to really practice conversation. So we added a new feature called video call with Lily, which is a way to practice your conversation with an AI character that was not possible 2 years ago.
I mean, we get a lot of emails or letters every single day. So there's a lot of things that have been impactful for me. The fact that we help refugees in different countries learn the language of the country that they had to relocate to, particularly Syrian refugees in Europe. That was impactful for me. But we get all kinds of things. I mean people very old people who say they are using Duolingo to keep their brain healthy or people who say oh thank you because of Duolingo they were able to get a job. That's all very impactful to me.
I mean we try to, you know, there's a couple of other things. We have this thing called a Duolingo English test which is trying to make English testing much more accessible to everyone. Before the Duolingo English test, most English tests around the world to try to get either university admissions or something were very expensive and they were only in certain places where you could take the test because you have to go to a testing center. Now with the Duolingo English test, not only is it cheaper, but you can take the test from home. So there are entire countries where, for example, there was no way to take an English test, whereas now you can take it on the internet. So I think that's a lot of social impact. And then we're of course as a company we do all kinds of things like we give scholarships to people to go to universities and things like that. But I think in general by far the most impact that we're having is just helping to educate more than 100 million people.
I mean generally the things that I am most interested in is countries where people who don't have a lot of money don't have good access to education. There's a lot of countries like that. The poorer the country the more this happens. So I'm very interested in how much we can help people be educated in the poorest countries.
It's very natural. I don't think we have a very specific system. I mean, in general, yes, we're going to help things like, you know, we try to help with Ukraine relief and things like that, but in general, again, we spend most of our effort on just being able to provide the best possible education with our app.
Yeah. I mean, I really I'm a big fan of the premium model that we have. It turns out that the people who pay, so only about 10% of our users pay us, the people who pay usually have stable jobs or live in wealthy countries and the people who don't pay usually don't have stable jobs or don't live in wealthy countries and they're the beneficiaries of the free education. So I like that basically the wealthy people are paying for the education for everyone. I like that very much. It's compatible with what we wanted to do.
We're going to do something similar that we do for whenever we enter new markets. We increase the course offering. So we're teaching people more languages here. And then after that we try to do social media marketing. And the good news is we generally don't have to spend a lot of money on marketing to grow. So with a little bit of money and a lot of creativity we see that markets grow a lot. So that's our hope that with a little bit of that plus all the word of mouth that we're going to grow a lot in Korea.
I think in general for most businesses the most important thing is to have a really good product. Whatever it is you're making, make it really good. I think people get distracted a lot of times by spending most of their efforts on the wrong thing. For example, if you're making a restaurant, you should make sure that you can make the best food possible as opposed to having a really nice sign outside. I think people sometimes get distracted on things. So, the thing that you're doing, you should spend most of your effort on. Again, the restaurant analogy is a really good one. You can if you concentrate on having really really good food, it may take a long time for people to discover you, but once they discover you, they really will tell their friends and then over a long period of time, you will be the most popular restaurant. Whereas if you make a really good sign outside, maybe a lot of people will discover you very fast, but then nobody will come back. So I think if you actually want long-term success, you should make sure that you have a really good product. And in our case, that is the language learning app.
Yes. I mean, as a company, what we want to do is we really want to be able to teach as many people as possible the things that are important to learn. You know, we want screen time on the phone to be valuable as opposed to just scrolling on TikTok all day long. And for me personally, I mean, I want to continue working on Duolingo. We have a lot to accomplish. I think we're only scratching the surface so far.
I mean, look, there are two billion people in the world learning a foreign language. Most of them are learning it in schools or something. We only teach, you know, a little more than 100 million. So there's a lot more there. I also want to do for math the same thing we did for languages. And then after math, I probably want to do physics. I mean, so generally we just want to do as many subjects as possible and really have hundreds of millions of people learning on their phone. Duolingo for the last 14 years has just been continuously improving and I think we're just going to see it continuously improve. I mean, we really believe in continuous improvement and we can do that.
