WEBVTT - Duolingo CEO Luis von Ahn Talks Earnings, AI

0:00:02.520 --> 0:00:08.920
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news. Shares of dual Lingo

0:00:09.000 --> 0:00:12.400
<v Speaker 1>are tanking today. The stock in fact right now down

0:00:12.600 --> 0:00:17.079
<v Speaker 1>about seventeen percent. It did though, hit its lowest intra

0:00:17.160 --> 0:00:19.920
<v Speaker 1>day since February of twenty twenty three, now down more

0:00:19.960 --> 0:00:23.599
<v Speaker 1>than eighty percent since May of twenty twenty five, and

0:00:23.640 --> 0:00:25.919
<v Speaker 1>about twenty two percent of the flow is short, so

0:00:25.960 --> 0:00:28.800
<v Speaker 1>there's some real negative sentiment on the name. Shares sliding

0:00:28.840 --> 0:00:31.920
<v Speaker 1>after the companies said it's drive to gain subscribers would

0:00:31.960 --> 0:00:34.920
<v Speaker 1>mean slower earnings growth and narrower profit margins in the

0:00:34.960 --> 0:00:37.040
<v Speaker 1>short term. Tim, you keep reminding us that there's a

0:00:37.080 --> 0:00:39.400
<v Speaker 1>lot going on with the company, some transformations here. Yeah.

0:00:39.400 --> 0:00:41.360
<v Speaker 2>The company said it would step up investment in AI

0:00:41.520 --> 0:00:44.360
<v Speaker 2>and sacrifice some degree of monetization in order to accelerate

0:00:44.479 --> 0:00:47.920
<v Speaker 2>user growth and engagement. The goal double the current number

0:00:47.920 --> 0:00:50.600
<v Speaker 2>of daily active users to one hundred million in twenty

0:00:50.720 --> 0:00:53.080
<v Speaker 2>twenty eight. Great to be talking again with a member

0:00:53.120 --> 0:00:56.120
<v Speaker 2>of the Dulingo c suite, back with us. Luis von Ahn,

0:00:56.200 --> 0:00:59.040
<v Speaker 2>co founder, chairman, president and CEO of Duelingo. He joins

0:00:59.120 --> 0:01:01.360
<v Speaker 2>us from New York. Lou He's always good to see you.

0:01:02.280 --> 0:01:04.280
<v Speaker 2>We're gonna talk about the quarter and the outlook in

0:01:04.319 --> 0:01:06.120
<v Speaker 2>a minute, But I want to start big picture with

0:01:06.160 --> 0:01:09.800
<v Speaker 2>AI because I mean, you've been thinking and researching AI

0:01:09.880 --> 0:01:12.080
<v Speaker 2>for decades. At this point, you're a recipient of a

0:01:12.120 --> 0:01:16.080
<v Speaker 2>MacArthur Genius Foundation grant back in the year two thousand

0:01:16.120 --> 0:01:19.160
<v Speaker 2>and six. We're working on AI. This is nothing, nothing

0:01:19.200 --> 0:01:21.640
<v Speaker 2>new to you, what's happening right now. But answer the

0:01:21.720 --> 0:01:25.120
<v Speaker 2>question for investors, is AI a threat to your business?

0:01:26.280 --> 0:01:27.640
<v Speaker 3>Well, first of all, thank you for having me.

0:01:29.600 --> 0:01:32.360
<v Speaker 4>You know, I think that we're in a unique point

0:01:32.360 --> 0:01:35.319
<v Speaker 4>in time where because of AI, we're going to be

0:01:35.319 --> 0:01:40.040
<v Speaker 4>able to teach significantly better. And you know, Dueling was

0:01:40.240 --> 0:01:43.880
<v Speaker 4>is by a wide margin, the largest app in education

0:01:44.040 --> 0:01:47.000
<v Speaker 4>in the world, and I think because of that, we're

0:01:47.000 --> 0:01:49.480
<v Speaker 4>going to be able to really take advantage of this,

0:01:49.840 --> 0:01:53.440
<v Speaker 4>you know, AI boon to be able to teach significantly better.

0:01:53.480 --> 0:01:55.360
<v Speaker 4>And our sense is that within the next few years

0:01:55.400 --> 0:01:57.880
<v Speaker 4>we're going to have an app that teaches us well

0:01:58.120 --> 0:02:00.520
<v Speaker 4>as a one on one tutor, but also so is

0:02:00.720 --> 0:02:03.120
<v Speaker 4>as fun as a mobile game. And if we're really

0:02:03.120 --> 0:02:05.240
<v Speaker 4>able to do that, I think there's just so much

0:02:05.240 --> 0:02:06.160
<v Speaker 4>more that we encapture.

0:02:06.200 --> 0:02:08.960
<v Speaker 3>So that's what that's what we're doing. We're really shooting

0:02:08.960 --> 0:02:09.280
<v Speaker 3>for that.

0:02:10.200 --> 0:02:13.320
<v Speaker 1>So jumping on the AI bandwagon, I mean, I guess

0:02:13.360 --> 0:02:15.320
<v Speaker 1>what I guess what investors want to know is what

0:02:15.480 --> 0:02:20.639
<v Speaker 1>is dual Lingo's advantage over potential new AI learning apps.

0:02:22.360 --> 0:02:25.240
<v Speaker 4>Well, you know, at the moment, we're not we're not

0:02:25.280 --> 0:02:28.800
<v Speaker 4>particularly concerned with competition in terms of you know, other

0:02:29.120 --> 0:02:30.200
<v Speaker 4>other education apps.

0:02:30.200 --> 0:02:30.400
<v Speaker 2>We have.

0:02:30.800 --> 0:02:32.880
<v Speaker 4>For example, for language learning, we have about eighty five

0:02:32.919 --> 0:02:35.359
<v Speaker 4>percent of the market share of all daily active usies

0:02:35.360 --> 0:02:36.720
<v Speaker 4>of people who are learning a language.

0:02:37.080 --> 0:02:38.480
<v Speaker 3>So we have huge scale.

0:02:38.840 --> 0:02:42.440
<v Speaker 4>And what differentiates us is that, because of that scale,

0:02:42.480 --> 0:02:45.120
<v Speaker 4>we just have more data into how people learn than

0:02:45.160 --> 0:02:47.960
<v Speaker 4>anybody else and we can use that to train our

0:02:48.000 --> 0:02:50.600
<v Speaker 4>own specific models. So we just have the advantage of scale,

0:02:50.639 --> 0:02:54.160
<v Speaker 4>of being able to watch literally billions of exercises every

0:02:54.240 --> 0:02:57.280
<v Speaker 4>day by people who are learning different things.

0:02:57.480 --> 0:02:59.560
<v Speaker 1>But as you know, this is going and this is

0:02:59.560 --> 0:03:03.120
<v Speaker 1>moving quickly, the AI impact it seems, and to be fair,

0:03:03.200 --> 0:03:05.960
<v Speaker 1>you are right, like we'll see many say this is

0:03:06.040 --> 0:03:08.000
<v Speaker 1>you know, we're kind of early in on this process

0:03:08.040 --> 0:03:11.080
<v Speaker 1>and the impact. But what does dual lingo offer that

0:03:11.080 --> 0:03:14.400
<v Speaker 1>can't easily be replicated by a large language model And

0:03:14.960 --> 0:03:18.079
<v Speaker 1>how do you kind of view that competitive landscape evolving.

0:03:19.800 --> 0:03:21.520
<v Speaker 3>Yeah, I mean there's a there's a there's a number

0:03:21.560 --> 0:03:21.880
<v Speaker 3>of things.

0:03:21.919 --> 0:03:25.600
<v Speaker 4>I mean, for one, with do a Lingo, what we

0:03:25.639 --> 0:03:29.080
<v Speaker 4>do is we try to build a habit of learning.

0:03:29.280 --> 0:03:31.760
<v Speaker 4>I mean, for example, fifteen million of our daily active

0:03:31.840 --> 0:03:34.760
<v Speaker 4>users have a streak of longer than three hundred and

0:03:34.760 --> 0:03:37.040
<v Speaker 4>sixty five days, meaning they have been using du Lingo

0:03:37.080 --> 0:03:39.200
<v Speaker 4>for a year or longer and haven't missed the day.

0:03:39.640 --> 0:03:42.160
<v Speaker 4>So we're really trying to build a habit for that.

0:03:43.160 --> 0:03:45.720
<v Speaker 4>And this is why do a linguist part game, part education,

0:03:46.520 --> 0:03:49.240
<v Speaker 4>And that's you know, I think that's just the right thing.

0:03:49.280 --> 0:03:52.240
<v Speaker 4>It just turns out the hardest thing about learning something

0:03:52.240 --> 0:03:56.160
<v Speaker 4>by yourself is staying motivated. The actually the content to

0:03:56.280 --> 0:03:58.320
<v Speaker 4>learn has been there for a long time. So for example,

0:03:58.640 --> 0:04:00.640
<v Speaker 4>it's you know, you've been able to learn a language

0:04:00.680 --> 0:04:03.080
<v Speaker 4>by just reading a book for hundreds of years.

0:04:03.160 --> 0:04:04.160
<v Speaker 3>That has been there.

0:04:04.360 --> 0:04:06.840
<v Speaker 4>It's just what we do differently is we motivate people

0:04:06.920 --> 0:04:07.880
<v Speaker 4>to continue doing it.

0:04:08.280 --> 0:04:09.839
<v Speaker 3>And that's hard to get right.

0:04:10.480 --> 0:04:12.640
<v Speaker 2>So what are what our investor is not getting right?

0:04:12.680 --> 0:04:16.520
<v Speaker 2>In your view, Carol? Carol mentioned some of the stock moves,

0:04:16.520 --> 0:04:18.599
<v Speaker 2>and now you know down at a point that you

0:04:18.600 --> 0:04:21.000
<v Speaker 2>haven't seen since February of twenty twenty three. What are

0:04:21.000 --> 0:04:22.039
<v Speaker 2>investors missing here?

0:04:23.800 --> 0:04:27.000
<v Speaker 4>I think it's a matter of timing. You know, on

0:04:27.000 --> 0:04:29.560
<v Speaker 4>our end, we are trying to build a company for

0:04:29.600 --> 0:04:31.760
<v Speaker 4>the very long term. We're trying to do something that

0:04:32.320 --> 0:04:35.880
<v Speaker 4>you know, for example, in our shareholder letter, we said

0:04:35.920 --> 0:04:38.760
<v Speaker 4>that we are we're shooting to have a double the.

0:04:38.720 --> 0:04:40.480
<v Speaker 3>Daily active uses that we have now. I mean, at

0:04:40.480 --> 0:04:41.520
<v Speaker 3>the moment, we have a little.

0:04:41.320 --> 0:04:44.240
<v Speaker 4>North of fifty million daily active users worldwide. We're trying

0:04:44.240 --> 0:04:46.880
<v Speaker 4>to have more than one hundred million. But we're saying

0:04:46.880 --> 0:04:49.479
<v Speaker 4>that that's going to take some time, and so we're

0:04:49.520 --> 0:04:53.280
<v Speaker 4>cons concentrated in the long term. And you know what

0:04:53.320 --> 0:04:56.880
<v Speaker 4>we're doing is we're trading off this year's you know,

0:04:56.920 --> 0:05:00.280
<v Speaker 4>financial metrics for a much larger thing in the long term.

0:05:01.400 --> 0:05:03.760
<v Speaker 1>Go ahead, Well, I get that right. You know, companies

0:05:03.800 --> 0:05:06.480
<v Speaker 1>have to invest in order to provide growth going forward.

0:05:06.520 --> 0:05:08.839
<v Speaker 1>And I think it's fair to say that we understand

0:05:08.880 --> 0:05:12.640
<v Speaker 1>why investors. I mean, they're not loving your forecast for

0:05:12.720 --> 0:05:15.520
<v Speaker 1>first quarter adjusted EBITA that's missing the mark from what

0:05:15.520 --> 0:05:19.000
<v Speaker 1>the street was expecting twenty twenty six revenue. Also that

0:05:19.080 --> 0:05:24.240
<v Speaker 1>forecast missing same story for twenty twenty six adjusted EBITA overall.

0:05:24.440 --> 0:05:29.680
<v Speaker 1>So you talk about building this to double your daily

0:05:29.720 --> 0:05:32.800
<v Speaker 1>active users to one hundred million, how much will how

0:05:32.839 --> 0:05:36.200
<v Speaker 1>long will that build actually take and what kind of

0:05:36.279 --> 0:05:38.679
<v Speaker 1>visibility do you have on that to kind of maybe

0:05:38.720 --> 0:05:42.599
<v Speaker 1>reassure investors at this isn't something that takes longer while

0:05:43.000 --> 0:05:46.760
<v Speaker 1>the AI world continues to challenge established players.

0:05:48.320 --> 0:05:50.960
<v Speaker 4>Yeah, I mean we're the way we're seeing it is

0:05:50.960 --> 0:05:55.760
<v Speaker 4>We're going to spend this year really working on three things.

0:05:55.800 --> 0:05:59.800
<v Speaker 4>Really teaching better, improving the free user experience, and also

0:06:00.000 --> 0:06:02.520
<v Speaker 4>spanning to other subjects. I mean, by now we are

0:06:02.600 --> 0:06:05.960
<v Speaker 4>we already teach math, music, and chess, and you know,

0:06:06.000 --> 0:06:08.799
<v Speaker 4>we've been growing quite a bit. About fifteen percent of

0:06:08.440 --> 0:06:11.680
<v Speaker 4>our active users are learning things that are you know,

0:06:11.720 --> 0:06:15.040
<v Speaker 4>other than languages. Our chess course has grown quite significantly.

0:06:15.240 --> 0:06:17.000
<v Speaker 4>So these are the three things that we're working on,

0:06:17.040 --> 0:06:18.760
<v Speaker 4>and we're going to be working on that for the

0:06:18.800 --> 0:06:22.440
<v Speaker 4>rest of the year, and we're expecting that towards the

0:06:22.560 --> 0:06:26.040
<v Speaker 4>end of the year we're going to see some improvements

0:06:25.640 --> 0:06:27.920
<v Speaker 4>in our year over year growth rate.

0:06:28.520 --> 0:06:29.200
<v Speaker 3>We're still growing.

0:06:29.200 --> 0:06:31.880
<v Speaker 4>I mean, we've had five years of really phenomenal growth.

0:06:31.880 --> 0:06:35.080
<v Speaker 4>We iPod in twenty twenty one and since then, we've

0:06:35.080 --> 0:06:38.440
<v Speaker 4>more than five x our active users, so you know,

0:06:38.640 --> 0:06:41.240
<v Speaker 4>we expect that will continue growing pretty pretty strongly.

0:06:41.320 --> 0:06:44.120
<v Speaker 2>Luis, is the idea in making changes to the free

0:06:44.160 --> 0:06:47.040
<v Speaker 2>tier is it to get more people who don't use

0:06:47.200 --> 0:06:50.279
<v Speaker 2>du lingo now to use the free tier, or is

0:06:50.360 --> 0:06:53.960
<v Speaker 2>to get those free free tier users to upgrade to

0:06:54.040 --> 0:06:54.719
<v Speaker 2>a paid tier.

0:06:57.160 --> 0:07:01.200
<v Speaker 4>Historically, what we have done is we have, right, you know,

0:07:01.240 --> 0:07:04.560
<v Speaker 4>the way we generate value on dualingo is there's two

0:07:04.560 --> 0:07:07.280
<v Speaker 4>parts of it. One is how many active uses we have.

0:07:07.360 --> 0:07:08.960
<v Speaker 4>That's kind of like the size of the pie that

0:07:09.000 --> 0:07:12.800
<v Speaker 4>has grown quite significantly. And then there's how many of

0:07:12.840 --> 0:07:16.160
<v Speaker 4>these people are paying us because we have this freemium model,

0:07:16.200 --> 0:07:17.840
<v Speaker 4>that's you know, like the fraction of the pie that

0:07:17.880 --> 0:07:21.400
<v Speaker 4>our payers. Historically we have worked on both. We have

0:07:21.480 --> 0:07:24.120
<v Speaker 4>worked on growing the pie and also growing the fraction

0:07:25.200 --> 0:07:27.120
<v Speaker 4>of the pie that is paying, which is why our

0:07:27.160 --> 0:07:30.200
<v Speaker 4>bookings have grown really significantly over the last five years.

0:07:30.440 --> 0:07:33.160
<v Speaker 4>This year, we're mainly concentrating on just growing the pie,

0:07:33.200 --> 0:07:35.480
<v Speaker 4>that is the number of people that are actively using

0:07:35.560 --> 0:07:38.400
<v Speaker 4>dual ingo, because it is our belief that.

0:07:39.720 --> 0:07:41.120
<v Speaker 3>Because of the.

0:07:41.200 --> 0:07:43.120
<v Speaker 4>Moment that we're in, we really want to have as

0:07:43.160 --> 0:07:47.200
<v Speaker 4>many people as possible actually learning something meaningful and dual lingo.

0:07:47.680 --> 0:07:47.800
<v Speaker 1>Uh.

0:07:48.160 --> 0:07:49.720
<v Speaker 4>And once we're able to do that, I mean, one

0:07:49.720 --> 0:07:52.120
<v Speaker 4>of the main advantages that we have, certainly against new

0:07:52.240 --> 0:07:56.239
<v Speaker 4>entrants is scale. They're just there's there's no other education

0:07:56.320 --> 0:07:57.800
<v Speaker 4>app that has the scale that we have.

0:07:59.200 --> 0:08:02.400
<v Speaker 1>Well, great to away get a check on what you

0:08:02.440 --> 0:08:04.960
<v Speaker 1>guys are up to and really appreciate it. Love to

0:08:05.000 --> 0:08:06.960
<v Speaker 1>hear more too as we go out throughout the year.

0:08:07.320 --> 0:08:08.160
<v Speaker 1>How things are going.

0:08:08.280 --> 0:08:11.480
<v Speaker 2>I'm a three fifty five, I think, Lingo, how is this?

0:08:12.280 --> 0:08:12.960
<v Speaker 2>I'm almost at a.

0:08:13.000 --> 0:08:15.600
<v Speaker 1>Year I think I think Sebastian actually said he's on

0:08:16.200 --> 0:08:18.080
<v Speaker 1>his streak is about seven hundred and seventy. That's one

0:08:18.120 --> 0:08:20.480
<v Speaker 1>of our producers. Hey, Louise, thank you so much. Luis

0:08:20.560 --> 0:08:23.360
<v Speaker 1>fan On, co founder, chairman, president, CEO of Duel Lingo.

0:08:23.400 --> 0:08:25.520
<v Speaker 1>When we come back and check on trading and stocks

0:08:25.560 --> 0:08:27.760
<v Speaker 1>on the move. This is Bloomberg Business Week Daily