Luis Von Ahn

And we've done these research studies that are in some of the most amazing things that I've seen. When you, you know, you talk to a user, you ask them, uh, what do you think could make Duolingo better? And, you know, historically in the past they have said, well, more practice conversation with a real person. They have said that.

And we've done these research studies that are in some of the most amazing things that I've seen. When you, you know, you talk to a user, you ask them, uh, what do you think could make Duolingo better? And, you know, historically in the past they have said, well, more practice conversation with a real person. They have said that.

And we've done these research studies that are in some of the most amazing things that I've seen. When you, you know, you talk to a user, you ask them, uh, what do you think could make Duolingo better? And, you know, historically in the past they have said, well, more practice conversation with a real person. They have said that.

And then you asked the user, okay, so you're telling me if I put a human on Duolingo, you would do that? And then they say, yes, yes, I would. And you can even ask them, would you pay for it? And then they'll say, yes, I'll pay for it. And then you tell them, okay, do you want to do it right now? And the answer invariably is no, no, no, not right now. People just don't want to do that.

And then you asked the user, okay, so you're telling me if I put a human on Duolingo, you would do that? And then they say, yes, yes, I would. And you can even ask them, would you pay for it? And then they'll say, yes, I'll pay for it. And then you tell them, okay, do you want to do it right now? And the answer invariably is no, no, no, not right now. People just don't want to do that.

And then you asked the user, okay, so you're telling me if I put a human on Duolingo, you would do that? And then they say, yes, yes, I would. And you can even ask them, would you pay for it? And then they'll say, yes, I'll pay for it. And then you tell them, okay, do you want to do it right now? And the answer invariably is no, no, no, not right now. People just don't want to do that.

So that's why we haven't put the humans there. And I think it's been a good decision, especially now that with large language models, we can do a pretty good job getting you to practice conversation without a human.

So that's why we haven't put the humans there. And I think it's been a good decision, especially now that with large language models, we can do a pretty good job getting you to practice conversation without a human.

So that's why we haven't put the humans there. And I think it's been a good decision, especially now that with large language models, we can do a pretty good job getting you to practice conversation without a human.

That's exactly right. It is a really, really good application. I mean, there's a number of things. You said it right. It doesn't matter if it says something that is a little wrong. because you're just practicing your language. Also, it doesn't matter if it makes a small mistake. Sometimes it makes a small grammatical mistake.

That's exactly right. It is a really, really good application. I mean, there's a number of things. You said it right. It doesn't matter if it says something that is a little wrong. because you're just practicing your language. Also, it doesn't matter if it makes a small mistake. Sometimes it makes a small grammatical mistake.

That's exactly right. It is a really, really good application. I mean, there's a number of things. You said it right. It doesn't matter if it says something that is a little wrong. because you're just practicing your language. Also, it doesn't matter if it makes a small mistake. Sometimes it makes a small grammatical mistake.

People can't even notice because they're usually kind of beginners at Spanish or beginners at French. You don't know. And it also can adapt to your level really well. Large language models are really good at adapting to your level. So we tell it, OK, adapt to a beginner Spanish.

People can't even notice because they're usually kind of beginners at Spanish or beginners at French. You don't know. And it also can adapt to your level really well. Large language models are really good at adapting to your level. So we tell it, OK, adapt to a beginner Spanish.

People can't even notice because they're usually kind of beginners at Spanish or beginners at French. You don't know. And it also can adapt to your level really well. Large language models are really good at adapting to your level. So we tell it, OK, adapt to a beginner Spanish.

And by the way, we even tell it, hey, because we've seen this person learn on Duolingo, we actually know all the words they know. So we tell the language model, this person only knows these 200 words. So please, mostly use only these 200 words. So it works really well for that.

And by the way, we even tell it, hey, because we've seen this person learn on Duolingo, we actually know all the words they know. So we tell the language model, this person only knows these 200 words. So please, mostly use only these 200 words. So it works really well for that.

And by the way, we even tell it, hey, because we've seen this person learn on Duolingo, we actually know all the words they know. So we tell the language model, this person only knows these 200 words. So please, mostly use only these 200 words. So it works really well for that.

We're investing a lot. Fortunately, it's actually been good for us in terms of profitability for two reasons. So there's two places where we invest in AI. The first is generating data that is going to be used in our lessons. That data used to be generated online. partly with humans, and now it's mostly generated with AI. And it turns out it's a lot cheaper to generate with AI than with humans.

We're investing a lot. Fortunately, it's actually been good for us in terms of profitability for two reasons. So there's two places where we invest in AI. The first is generating data that is going to be used in our lessons. That data used to be generated online. partly with humans, and now it's mostly generated with AI. And it turns out it's a lot cheaper to generate with AI than with humans.