Duolingo Co-Founder, Severin Hacker: How AI Impacts the Future of Work and Education

Severin Hacker 0:00 You should always try to raise money from the best tier one VCs, even if the terms are slightly worse, because there's so much signaling. It's harder to raise three million than it is to raise 100 million. But I can guarantee you, Duolingo would not have been able to raise any money in Europe. You know, we had $0 revenue for the first five years.

Harry Stebbings 0:20 How much was the series A?

Severin Hacker 0:21 Three million at, uh, 15. Yeah. I think we only had one offer. It was ... or back to university.

Harry Stebbings 0:28 Ready to go? Severin, I'm so excited for this, dude. I had Luis on six or seven years ago, and I've heard so many great things before this show from Mayumi Bang, from KP, Brad and Albert from USV, many others. So thank you so much for doing this.

Severin Hacker 0:56 Thank you. Thank you for having me.

Harry Stebbings 0:58 Dude, I would love to start though with basically the news of the day. Um, I don't like the, you know, the whole context of, like, tell me your life story in normal podcast intros.

Harry Stebbings 1:07 Uh, Duolingo came out as being AI first the other day. And I really just wanted to start with, what does that mean and what does that not mean? So people have a clear understanding.

Severin Hacker 1:17 Okay. So maybe taking a step back. Why did we start Duolingo? Like, what is our mission? So Duolingo's mission is to provide the best education and make it universally available. That's why Luis and I started this company. Okay? And from day one, if you think about it, if you wanna build the best education, um, in the past, the best education was only available to the, the richest people, the kings, the heads, you know, the private tutors for their kids. And, and, you know, that was the, the way to learn, the most efficient way. And by the way, this is still true today. If you wanna become the best tennis player, you hire a one-on-one tennis coach. Okay? Now-

Harry Stebbings 2:01 You actually, you actually see this. I remember there was a US Open final, I think it was, and the two female players in the final were both daughters of billionaires.

Severin Hacker 2:08 Uh, yeah. (laughs)

Harry Stebbings 2:08 Yeah.

Severin Hacker 2:09 Yeah, exactly. So, so, you know, they get the one-on-one tutor. And when we started Duolingo, we saw the, the potential of technology to do this, but for everyone, right? So we can't all afford these one-on-one tutors. That's, it's too expensive. Um, but with technology, there's a way we believe that anyone can have access to a, the equivalent of a one-on-one tutor, this, with the same efficacy. And that's why we started Duolingo. So technol- ... Duolingo was, from day one, was technology first. And, you know, back then, it was g- you know, software, we called it software. Uh, now we call it AI. Um, so in a way, not much has changed, except AI, as we now call it, has become a lot better.

Harry Stebbings 2:58 Totally get you. And so take me to that discussion then, where you're inside the room with Luis and the board, and you decide, "Okay, we need to fundamentally change how we do everything and reshape the organization of this company towards being an AI-driven company."

Severin Hacker 3:14 This has been, you know, going on for the last two and a half years. So we were one of the launch partners of OpenAI when they first launched GPT-4, and we immediately saw the, the potential of this technology to help our mission, right?

Harry Stebbings 3:30 What was it that you saw that gave you such optimism?

Severin Hacker 3:32 Uh, so we had, we had early access, and we saw you could really use ... I mean, I don't know. Eh, when you first had access to it, it was im- impressive. It was like a, uh, an iPhone moment. It was like you just realize, "This is the future. We're gonna use this everywhere." That was my first personal impression. And then it's like, the second question is like, how do you use this at Duolingo? And the first thing we realized is this can really help us accelerate content production. Okay? So in, in the past, it would always take us a long time to produce new courses, right? So, uh, it requires a lot of, you know, effort to, to create these courses. But then with AI, there's, there's the potential that you can just pump out a lot of these courses all at once. In fact, we just did that. Uh, you know, it, it took us, I think, 12 years to, to, to build the first 100 courses or so. And now, within one year, we built another 148 courses. That was just not possible. It would have taken us decades to build this.

Harry Stebbings 4:38 Can I be really, but kind of weirdly granular? How are you using AI to then 12x content creation within Duolingo?

Severin Hacker 4:48 Yeah. I mean, there's still ... By the way, there's still a lot of, uh, human in the loop.

Harry Stebbings 4:52 Mm-hmm.

Severin Hacker 4:53 Um, so for example, the curriculum design, that's all human made. It's like h- you know, how, h- how do you structure the course? That's all human made. But then, you know, inside Duolingo, there's a lot of these, uh, short sentences, right? It's like, um, y- you know, the, the ones you see in, in, in the lessons. And those are now produced with, with AI. And, and there's also this, this, this concept of, uh, viability. You, you wanna introduce new words one at a time, and you can, you know, give these constraints to the AI, say, "Create a sentence that uses these words and doesn't use any of the other words, and only uses these grammar concepts." And it can do that. So that, that's kind of the magic, and, and that's how we've been able to generate the, kind of the, the sentence content within these courses. Okay.

Harry Stebbings 5:43 With respect on the curriculum design, when you look at completion rates of students and you have the data that you have, in six months, would it not be better to let AI do curriculum design knowing all of the completion rates, success rates, uh, satisfaction rates, NPS scores, to actually build that curriculum itself?

Severin Hacker 6:05 That's, that's a great question. I think the, the industry is definitely moving that way. Um, there is, you know, one of the most amazing things about Duolingo is we have this massive user base. Like we're by far the largest learning platform out there, by far. Uh, order of mag- orders of magnitude. And we see how people learn. It's like the largest school. You know, like we see, uh, usually in, in education, the, these audits were done with, you know, 20 to 30 students. We have hundreds of millions. And, and that's just, that's allows us to look at, okay, does this curriculum design work versus A, you know, this versus that. And, and we can use that data to improve the curriculum-

Harry Stebbings 6:45 Yeah.

Severin Hacker 6:46 ... design itself. But not only that, so, you know, right now, we still have these, these, these courses, and you know, your curriculum is the same as mine. But really, I think the future is gonna be quite different where we say like, "Hey, you know, you might be interested in, in rugby, and, uh, you have a trip planned to..." I don't know, France. I think France also is, is big into rugby. And, you know, you, you could, you could then custom design a course just for you that uses exactly the vocabulary you want to use when you go to France. And, you know, it could be completely different from mine. It's like, why do we even create this batch on the, on the back end? Like why don't we create it on the fly when you need it? Like when you open the app, it's like this sentence is custom designed for you, every single exercise.

Harry Stebbings 7:31 So personalization is the-

Severin Hacker 7:32 Personalization.

Harry Stebbings 7:32 ... future of education in your mind?

Severin Hacker 7:34 I believe, and I think that's part of the reason why these one-on-one tutors are so, um, so good, right? So because they, they, they know where you are. They're really good at assessment. Uh, you know, think of the tennis coach again. Uh, they're really good at assessment. They know where you are. They know where you need to go, and they're very good at giving you the feedback. And they also know kind of your interests and, you know, like maybe how you differ from other players.

Harry Stebbings 7:59 Totally get you and agree. I was actually watching Zuck last night on an interview, and he was saying the future of content is not passive consumption of videos like we have today. It's the interactivity where you will talk to the video, and suddenly, the fitness trainer will actually say, "Oh, no, you should have 12 grams of protein, not eight grams of protein." And it's that interactivity that will drive it. How do you think about personalization balanced with interactivity in the future of content, in the future of education? Like to what extent is it multimodal? You know, the thing I love about Duolingo that makes me sound weird as anything is I'm walking around London, and then I'm going, "Bruno es americano." (laughs) And yeah, it's, it's multimodal. It's me, my voice playing back. It's me typing. There's many different modes.

Severin Hacker 8:43 Yeah, yeah, yeah. I think it, it's gonna be multimodal, uh, 100%. There is, we have this feature inside the app called Video Call with Lilly, so you can, you know, call Lilly. Lilly's one of our characters. She's the one with the, the purple hair.

Harry Stebbings 8:55 Yeah.

Severin Hacker 8:56 And, and it, it is, she, she's not a, she's not a one-on-one tutor. She, she, she's more like a friend and, but she remembers stuff. So she remembers, for example, that, you know, that I like cooking. She knows where I live, et cetera, et cetera. So she, she personalizes in that way, but it's full conversation, right? So you learn by conversing with an AI friend. And I think that's gonna be super important. Um, now, there's always, like these, these modalities. Like sometimes, you know, there's, not in every moment, uh, you know, you, you can't talk in every moment, right? So sometimes you're, uh, in the subway, you don't wanna use, um, audio.

Harry Stebbings 9:34 Yeah. How has the way that you build internally in Duolingo changed with AI? What tools do you use? How has it changed cadence? How has that changed?

Severin Hacker 9:44 We use AI in, in three ways at, at Duolingo. Number one is content generation, and that's where we've seen, you know, the biggest success. It's, it's just completely changed it, like how we produce content. Um, this is the learning content. Number two is features, AI features that you previously couldn't build. So this is, again, this Video Call with Lilly. It's a interactive conversation with an AI bot, right? That, two years ago or three years ago, you couldn't build. Now you can do it. And it's actually the thing that, you know, was, was lacking in the product. If you ask our users like, "What's missing in Duolingo?" It's like this conversational piece, like speaking. I wanna get better at speaking. And, and now we, we can do this, and it's a really, really, really powerful feature, and it's, you know, we, we s- great adoption with this, uh, Video Call with Lilly feature. So that's number two. And I think there's more to, to be done there, so both making this feature better, but then also maybe there's, there's new AI features that, you know, we can now build. And the third is, is kind of overall productivity improvements across the company. So that's, uh, software engineers using Cursor, that is using AI for customer support.

Harry Stebbings 10:59 Uh, do you mandate Cursor? Do you say, "Hey, we prefer you to use this or this"? How do you do that?

Severin Hacker 11:05 We don't mandate. Duolingo is surprisingly, uh, egalitarian and, and, and bottom-up, and we just say, "Use AI tool or use whatever makes you more productive," and we pay for it. Uh, you can use Cursor. You can use, um, ChatGPT. You can use, uh, the Copilot. And, and then we see what people actually use.

Harry Stebbings 11:29 Where is AI... You said, kind of LMS are amazing at language. Absolutely. Where is AI much better than people think, and where is it not as good as people think and actually has quite a lot of room to improve?

Severin Hacker 11:41 You know, early on, uh, the, the, there were a lot of these, uh, hallucinations, the AI making stuff up. Now, I think they have gone down, A, number one. But then also number two, for us, it's-... not a problem if they hallucinate, because we teach languages. As long as the language is correct, grammatically correct, it is okay, you know, if it makes up a historical fact. It's not, we're, we don't teach history. (laughs) You know, it's like, the hallucinations are not a problem for our, our use case.

Harry Stebbings 12:13 I bet you have users that are like, "No, it was 1914, not 1916."

Severin Hacker 12:19 Yeah, yeah, we don't go that far into it. (laughs)

Harry Stebbings 12:22 (laughs) But when you have so many users, you're like, there's always one.

Severin Hacker 12:25 Yeah, yeah, yeah. Um, no, so, so the hallucinations are not, not a problem. I would say, the, I think there's a lot of Silicon Valley hyperbole around software engineering getting replaced, uh, very quickly. I'm less sure about that. I think it's, uh, I, I try all of these tools and I think they're still a lot they can't do well.

Harry Stebbings 12:53 What can they not do well?

Severin Hacker 12:55 So, okay, so what they, on the software engineering side, I think what they're really good at is going from zero to 80% for a simple app. You know, it's like, it's the, the Replit, the Lovable. And then the larger the code base, even the one that generates itself, the larger the code base gets, the worse they get. So, then, you know, it just kind of like, uh, adding the next 10% takes as long as the first 80%. And they kind of create their own tech debt, uh, (laughs) which they also cannot really, uh, solve. Um, that might get better. And then the other thing they're really good at is just kind of like isolated, uh, uh, transformations. So in a single file, um, you know, add all the comments or, um, change a function to use different input parameters or something like that. Really good at that. And then what they're not good at is, not yet, and again, maybe this, this will change, maybe I'm wrong, maybe, (laughs) you know, this, this will not age well, but what they're not good yet is large code bases and adding... well, it's like adding a feature, even a, a simple one, to a large code

Severin Hacker 14:05 base.

Harry Stebbings 14:05 Do you think you will have more or less software engineers in five years' time at Duolingo?

Severin Hacker 14:11 That's a great question, and, you know, if you had asked me a year ago, I would have said 100% more. Now I'm a little bit less certain. But I think two things will change. So in a way, what these AI tools do is they lower the, the entry level, like the entry bar, like to get into software engineering, right? So, um, you know, your mom can maybe build a, an app with Lovable or something like that.

Harry Stebbings 14:37 She has done. Yeah.

Severin Hacker 14:38 Um, right? Oh, yeah, right, so, so there you go. So, and then the other thing it does, it just lowers the cost of lines of code, of software, right? So production, producing software. And then there's the Jevons paradox which says, you know, if you, l- the cost of a good goes down, you will actually have more demand for it. Um, and, and I can, I, I also believe some of this will be true, so I think there will be a lot more people writing apps. Um, whether they will actually touch the code or not, I'm not sure. And then I think the second thing is like, related to this, is are we gonna call them software engineers or what are we gonna call them? You know? Uh, one thing we see is, inside Duolingo, we have t- the, the three biggest functions within Duolingo are product, engineering and design. Okay, tho- those compose probably about 80%, 70% to 80% of our, um, workforce. But with AI you can see that there's maybe like a, a, a, a new role, kind of a product engineer designer. Like one person who can do all of these roles to a certain, you know, degree. Like, they, they can build a, they can build a prototype and then maybe at some point they need to hand it off to an actual software engineer or an actual designer. But, but I think that's, that's gonna be part of the future. So I think it's gonna be more software, a lot more software, a lot more people creating software. Um, whether we're gonna call them software engineers or not, I'm not sure. And then I think there's gonna be this convergence of, of roles, of product engineering and design.

Harry Stebbings 16:16 So for everyone coming out of university today or contemplating university courses today and who are told, "Ah, CS in five years' time, it's not gonna be worth it," so do you still see immense value in CS as a principle and a mind- like, mindset?

Severin Hacker 16:30 I th- (sighs)

Harry Stebbings 16:31 If I was your little brother, and I'm 18, okay? And I'm like, you know, "Help me Severin, should I do CS at university?" What would you say?

Severin Hacker 16:40 So I did my CS, uh, degree, uh, yes, it's now quite a while ago, and we, they d- they don't actually teach you coding. You know, people think like, oh, CS means coding. Um, but CS is actually, uh, w- what, what they teach is kind of the fundamentals of, of not even software engineering but like how computers work, and then, uh, you know, like computer science is kind of applied mathematics. Okay, so I do believe there's still a lot of value in, uh, thinking logically, which is something that the best CS, um, courses teach you, right? So, so if, if you go, you know, studies at university, they, they don't teach you, you know, Java or Python, and they shouldn't. I think that that's actually, th- that, that is, that part is gonna go away. Like all the details of, of, you know, actual coding, but the, the fundamentals of, um, thinking about problems, it's problem-solving. I think that will still be valuable.

Harry Stebbings 17:48 And so should I do it?

Severin Hacker 17:51 I would say for the next five years, probably still a good, good investment.

Harry Stebbings 17:58 (laughs)

Severin Hacker 17:58 (laughs) Yeah. Uh, you know, like, there, there's a, there's like... Okay, so, so if you, if you extrapolate. Okay, so once we have full AGI-

Harry Stebbings 18:07 Mm-hmm.

Severin Hacker 18:07 ... okay, we, we no longer have jobs. Uh, nobody has to work anymore. Uh, we have UBI, et cetera. It's like kind of the, you know, the, the paradise. Um-

Harry Stebbings 18:15 Do you actually buy that?

Severin Hacker 18:17 Uh, I believe that's gonna happen. Um, again, the, the big question is when, when is it, it's gonna be... I don't think it's gonna be next five years. It's not gonna be. Maybe it's gonna be next 10 years if, if you really see a broad adoption of these technologies. More likely, maybe 20 years or so. Um, but in, in that world, it's like, what is gonna be valuable then, right? So in that world, what I think is like, you know, today, in the US, if you, if you run into a stranger at the Mm-hmm. ... conference or something, the first question is usually, "What do you do?"

Harry Stebbings 18:48 Always.

Severin Hacker 18:49 But in a, in, in this, this, uh, in this, in this future world, the, the first question might be like, "What are your hobbies?" (laughs) You know? Like, because we don't have jobs anymore. So, so all, everything is gonna be like, um, uh, about not work, right? So it's like-

Harry Stebbings 19:12 I, I, I do just question it though. I think we can, 'cause quite often get a little bit overexcited, and it's like, you know, when you had the agricultural revolution, it's like, "Oh, my God, we had a hundred people in the field, and now we have this combine harvester. We have one person driving it. We're all gonna have no jobs. Oh, my God, we have computers and calculators. What are accountants gonna do and people gonna do when we have..." We always find a next thing to do. The idea that society will live in this UBI lack of utility like state, I, I hope is not true.

Severin Hacker 19:44 Yeah. I, I, I don't, I don't know how, um, how it, how it's gonna play out, obviously. Uh, but I, uh, but I, I think the other thing that is interesting about this is, this is more on a, on a philosophical note, is, you know, if you, if you just, um, uh, ask people, you know, "Do you like your job?" Uh, everybody here, you know, you can ask anyone here, they all love their job. You love your job. I'm sure you love your job.

Harry Stebbings 20:07 Yeah.

Severin Hacker 20:07 Right? You would do it for less money, um, more or la- mostly-

Harry Stebbings 20:11 Dude, I did it for free for years.

Severin Hacker 20:12 You would do it for free, right?

Harry Stebbings 20:12 Yeah, I would. I literally-

Severin Hacker 20:13 But, but also, I love, I love my job.

Harry Stebbings 20:14 ... I get paid to hang out with amazing people.

Severin Hacker 20:16 It's like-

Harry Stebbings 20:16 Like, are you kidding me? Yeah.

Severin Hacker 20:17 So, so for, for, for, for, you know, for us, it's, it's, it's not just income, right? It's, it's purpose. We, we derive purpose from our job. Uh, but for a lot of people, the job is just, they do it for, for the income. They don't derive purpose from it. And, and then I think, you know, we are, in, in, in this, uh, the people that derive purpose from their job, who work the most, et cetera, you know, the, the tech people and, and, and the investors and all, all of these people, um, close to the AI revolution, I think we are most afraid of it, because we feel we might lose our purpose. And I think that, that's the- that's why, why we are sometimes not optimistic about, about this future world.

Harry Stebbings 21:02 It's funny. I had Jason Lenkin, who's a very famous SaaS investor, on the show recently, and he said, "The hard thing is with young people, they just don't wanna work. Like, fundamentally, they don't wanna work. And AI is as good as they are at sales, at marketing, whether you're an SDR, whether you're a social media content creator for memes or GIFs or, I don't know, whatever it is." Like, actually, uh, and so in 12 months, we're gonna see huge, huge unemployment in this kind of lower level tech employee, and it's go- it's, it's gonna happen. Do you think that's true?

Severin Hacker 21:34 I'm not sure. So, so here's another thing. Um, so, the, the, the general consensus is that these AI tools are really good for senior software engineers or staff level software engineers, right? So because they, they, um, you know, they have been trained, and now they can use AI and they become much more productive. They're now a 10X engineer. And, and then, uh, they can, you know, AI... The, the, the general consensus is that AI can do the, the job of an entry-level engineer or an intern or someone like that. Therefore, let's stop hiring entry-level engineers and only hire senior engineers. But I think that's a mistake, and, and we're actually not doing that at Duolingo. We still hire from universities! And one of the reasons is that we believe that people that grow up with these tools and, you know, start using these tools early will become much better at using them. Um, so in a way, yes, they're, they're, um-

Harry Stebbings 22:36 Why is that? Because young people have more plastic minds to new tools and new processes? What, what's the thinking behind that?

Severin Hacker 22:41 It, it's... I think it's just like, it's, it's like using these sorts of con- if you... I bet the person who spent, you know, a thousand hours using ChatGPT is a lot better at using it than I am, and, and I think that, that's the case for, for the new grads. Um, you know, another thing that's interesting is actually, you know, Duolingo wouldn't be where it is today in terms of, uh, you know, s- our social media presence if we never hired from Gen Z. It's like, you need to know your, your target demographic. And, you know, Zaria, who was on your show before, you know, she, she's Gen Z, and, and, and we've had-

Harry Stebbings 23:25 She just fundamentally understands it in a way that someone who's not Gen Z-

Severin Hacker 23:28 Exactly.

Harry Stebbings 23:28 ... could never understand it.

Severin Hacker 23:29 I, i- in fact, I, if I open, when I open TikTok and look at our own Duolingo, uh, stuff, as like, I don't get it. It's like, "Why is this funny? Why do people watch this?" But it's super, you know, it's, it's super popular.

Harry Stebbings 23:44 Yeah. You, you're not Gen Z anymore, Severin.

Severin Hacker 23:46 (laughs)

Harry Stebbings 23:46 I'm sorry to you. You don't, you don't fit in that demographic. (laughs)

Severin Hacker 23:49 (laughs)

Harry Stebbings 23:49 Um, we, we mentioned there how it changes, like engineering, product design. Another big area is customer support, and we chatted about it a little bit before. How does AI change customer support processes, workloads within Duolingo?

Severin Hacker 24:02 Yeah, so we, we started using, uh, an AI tool for, for customer support. I think it's another one of these early applications of AI, where AI can really do a, a, a, you know, a great job and, and transform an industry. And, um, we have found for a lot of the... you know, AI can do 70 to 80% of the, of the tickets that we get. Um, and then people are always like, "Oh, well, okay, you guys, you know, you, you don't need, uh, human customer support, um, agents anymore." That's not true. Uh, we actually... again, we... there's, there's, they're still there because you still have the, the other s- uh, 30% or so. But then not only that, now that, um, you know, you have this AI, kind of like an unlimited AI customer support agent, you can actually give customer support to a much broader base. So, so far, we only give it to our subscribers, the ones that, that pay for Duolingo. But if you can, you know, reduce the cost of customer support by 10X or 100X, then we can actually give it to everyone. So everyone benefits from this.

Harry Stebbings 25:08 Totally get you. Are you able to say which company?

Severin Hacker 25:10 And th- this is a classic, classic example of, uh, Turbin's paradox, right? So as cost goes down, demand goes up.

Harry Stebbings 25:18 Who do you use for customer support?

Severin Hacker 25:20 Uh, we use Decagon. Yeah.

Harry Stebbings 25:22 Got it. I would love to have that CEO on the show. I think it'd be a fascinating one on the future of customer support. You know what I'd love to? I'd love to do a panel with a couple of different providers.

Severin Hacker 25:30 (laughs)

Harry Stebbings 25:30 (laughs) Um, 70 to 80% is really very significant.

Severin Hacker 25:33 Mm-hmm.

Harry Stebbings 25:33 Can I ask you, when you think about offering it to more people and then also the implementation of, you mentioned, like the features that you have. Uh, implementing AI across the Duolingo product suite, it costs money. At the end of the day, you, you have to pay OpenAI for a lot of these features, and a lot of companies are. Do you see margin degradation by offering AI across the full product suite? Because you're not charging more, but you are having an additional cost out to a model provider.

Severin Hacker 26:01 Terms of margins, it's interesting. So as I explained before, there's these three use cases at Duolingo. Number one is the content generation. So c- content generation is, is just, um... it, it, it's, it's good for margins, right? So, so you, um-

Harry Stebbings 26:16 Phenomenal for margins.

Severin Hacker 26:17 Yeah. So... and, and it's also like you do... it's a one-off, um, thing. You generate the content, then it's done, right? And then number two is these AI features, like video call with Lilly, where you actually have a, a unit cost, right? So now you have... you directly interact with a LLM on the backend and you have to pay for that API call, right? So there's an actual unit cost there. And then there's, there's, uh, you know, you have to make that viable. So right now, video call with Lilly is, is in our... so we have three tiers. We have, um, you know, the free tier, there's super, and there's max. And it's mostly in our max, the higher tier subscription, uh, because of that reason, right? So, so we, we would like to give this to everyone, but the reality is today, there's a unit cost to, to AI, and therefore it's in our higher tier subscription.

Harry Stebbings 27:11 Got you. How-

Severin Hacker 27:12 We believe all of these costs will come down. So the one... you know, the, the, the batch, uh, processing cost has already come down. There's a lot of competition there. And then in the, um, the kind of the real-time models, multi-model, there's also gonna be... we hope there's gonna be a lot of competition. Cost will come down. Um, eventually, I think we'll be able to give this to a lot more people or almost everyone.

Harry Stebbings 27:35 Wait, when you think about like model relationships, you, you mentioned obviously being early access with OpenAI. How do you think about the, the flexibility of moving between models with different, uh, generations of model, with different specialties of model? Um, do you think you will have relationships with five model providers? Or do you think it's like, "Oh, no, we're tied to OpenAI?"

Severin Hacker 27:54 There, I think there's gonna be a lot of competition. And then in, in more specialized use cases, uh, like the, um, uh, like the visual models or, uh, for us, you know, audio is, is very interesting, multi-models. There's, there's less competition, and you can see that there could be companies that, that keep, uh, a lead. Um, and, you know, for example, for, for our voices we also work with ElevenLabs. And, and they're just amazing, uh, models, amazing voices. And, you know, I, I can see that, that might be harder to just, you know, uh, replace with one or the other model.

Harry Stebbings 28:32 When I hear you say about like ElevenLabs there, or Decagon there, or even think about Duolingo, the thing that I am struck with is, ah, but OpenAI are gonna do that. And I just love to go to the moment where I think Duolingo's stock actually took a hit, and I never talk about like public market prices, so don't worry, we're not going there. But I think it actually took a hit when there was the concern that OpenAI and ChatGPT would move into language learning very significantly. Can you take me to the internal discussion around, "Oh, shit, are ChatGPT and OpenAI going to move into the application layer for language learning?"

Severin Hacker 29:12 Okay. So when ChatGPT came out, one of the first killer use cases for it was homework cheating, for the lack of a better term.

Harry Stebbings 29:23 Mm-hmm.

Severin Hacker 29:23 Okay? So if you were in the homework cheating business, that was devastating. Uh, that was really bad. Um, so, so some of our... I'm not gonna name them, but like some of the other comp- companies in... (laughs) you know, th- they call themselves education but they're really more like in the homework cheating business, um-

Harry Stebbings 29:39 I love this homework cheating. I've never heard of this sector of the economy before.

Severin Hacker 29:43 (laughs)

Harry Stebbings 29:44 But-

Severin Hacker 29:44 So-

Harry Stebbings 29:44 ... okay, let's roll with it. The homework cheating business.

Severin Hacker 29:46 Yeah. So, so they, they s- they struggled with that. And then, and then I think there was actually a, a similar thing where people said, "Oh, well, you know, Duolingo is in education, um, therefore they're gonna be, you know, disrupted." And, and obviously the, the opposite, uh, is, is true because we're not in the homework cheating business. We're in the language learning business, or, you know, the, the general education, actually teaching you a new...... things, business. And that's completely different. So, and then, the other thing that, that I think you see happening in, in, in the public markets, and, and this, my only comment about this, is, like, there's, um, a bifurcation of the, of the market into, is this an AI winner or is this an AI loser? And I think, by mistake, we were first put into the, into the, the, the AI loser category, and now it changed to, like, "Oh, no, Duolingo's gonna be an AI winner." And there's lots of tailwind from AI. And, you know, I, I outlined some of this.

Harry Stebbings 30:44 So there was never a concern for you internally? You never had a call with Luis where you were like, "Oh, shit. Is Sam gonna do this?"

Severin Hacker 30:50 W- Well, you know, I, I, I think n- nobody is, is stopping them. So, but, so, so, so, the ... And I assume here you mean, like, OpenAI or ChatGPT building a really good language tutor? Right. So-

Harry Stebbings 31:04 100%. I'd be like, you know what? Yeah, like, can they take away a lot of your very, very casual users by doing a very simplistic, chat-with-a-friend, choose-your-language?

Severin Hacker 31:16 Yeah. So, so, th- that, that's definitely, uh, you know, somewhat of a, uh, of a concern. I think there's, there's a, a bunch of things to be said about that. Um, so first of all is, what is the hardest part a- about learning a language?

Harry Stebbings 31:31 Uh, I would say the discipline.

Severin Hacker 31:34 Yes. So it's, it's the actual motivation.

Harry Stebbings 31:37 Mm-hmm.

Severin Hacker 31:37 And, you know, by the way, this, this is one of the key insights of Duolingo. It's like... And we've been sharing this with the entire world for the last 10 years and still nobody seems to believe us. And so the hardest part is motivation. Okay? And Duolingo is, because it's so gamified, it, it, it's really good at this. It's, it's, it's a motivation engine. I call it the motivation engine. And it really gets you to do this. It really gets you to come back every day and, and do Duolingo. And if you stick to it and you, you know, do, do the hard material, you will actually learn a language. That is the, the power of Duolingo. And, in general, people underestimate how important that is. So if you wanna build something that is, you know, as good as Duolingo, you have to solve this somehow. Okay? And then also the, the, the, uh, the, you know, Duolingo is, if you ask our users, "What would you do if Duolingo went away?" They never mention any of our language learning competitors. They say, "Oh, I'd sp- spend more time on social media."

Harry Stebbings 32:35 Mm.

Severin Hacker 32:36 Okay? So the, the hyper-casual users, they love Duolingo because it's so motivating. Okay? So that, that's one thing. And then the second thing is, like, if I was Sam Altman, you know, (laughs) uh, I would go after the, the big trillion-dollar companies. Like, why go, go into, you know, language learning when you can go into search or social media? Like, those are the, you know, the, the big, big markets.

Harry Stebbings 33:05 Well, if you were sitting there as Sam today, you're thinking there's kind of a couple of cool areas which are winners. There's chat, which we already own, right? Chat interface. Uh, there's two, which is customer support, which I think they'll buy someone very soon. There's coding, which they've obviously just bought someone. And then I'd say language learning is the fourth. As you said, it's so perfectly... I mean, LLMs, it's kind of in the name. It is so perfectly aligned to the use case of what they are best at. It's not a stretch.

Severin Hacker 33:34 Yeah. Well, we'll see. (laughs)

Harry Stebbings 33:37 We shall see. Uh, o- one thing that I'd love your thoughts.

Severin Hacker 33:40 The, you know, the other, the other thing is like, it's, it's, um... Uh, Luis always says this. Like, there's a difference between being a first-class citizen and a second-class citizen, uh, as a, as a, as a feature, right? So, so if, if, uh, you know, language learning... You know, Duolingo works because we have these gamification mechanics. We have the streak, we have the XP, we have the, the leagues, and, uh, and all that. And you can build this all inside, um, you know, a, a different app. But, but if it's just kind of a, a second-class citizen buried somewhere, again, it's like, it's hard to, uh, make that experience as sticky as it... if you build, you know, a- an app just for this purpose.

Harry Stebbings 34:22 Is there anything within Duolingo that's a second-class citizen that you wish you could make a first-class citizen?

Severin Hacker 34:28 Uh, there's one thing maybe. So, so we have these stories inside, uh, Duolingo. So, so every now and then you get a story. I think they're actually becoming really good, again, with AI. And there's kind of an entertainment aspect with this. Like, so, uh, uh, y- you know, maybe you could actually make them more prominent and go into more, like, reading, including, like, f- like, uh, your, your first language, L1. Um, or maybe something for, for kids, like learning how to read. Um, but we haven't, haven't done that yet.

Harry Stebbings 35:03 You know, when you, when you-

Severin Hacker 35:04 It's not, it's not at the, the, the level of, you know, Harry Potter. (laughs)

Harry Stebbings 35:07 (laughs) You know, one, one thing that strikes me when I, you know, I use Duolingo every day, uh, it, it's exactly what you said though, which is, like, the little things which make the product experience delightful. And it reminds me of Gustav Soderstrom from Spotify who always told me, "The details are not the details. They are the product." And there's so many little things which make me feel that when I use Duolingo. To what extent do you believe that and advise that to founders who call you, versus, "Jesus, don't spend too long on the way that something pops up. Just get it done in the early days"?

Severin Hacker 35:40 Uh, it's a fine balance, right? Um, I think ultimately, product quality is details. And my... Uh, Luis, he's obsessed with details, you know. He can, he can spot something, like, being not properly aligned within half a second. Uh, i- i- the details really, really matter. If you, if you look at, um, y- you know, retention, which is the, the main m- uh, metric we, we, you know, we care about, is like-... the details really matter. It really matters. It also matters in terms of, uh, how likely are you to recommend it to a friend, right? So, uh, y- j- y- a- and you're not gonna say, "Oh, you know, it's like, oh, it's super, like, you know, details are really good." But it's like, "It's a great app. Yeah, it's, it's a great app." Or it's like, "Yeah, i- i- it's fun," and, and all of that. That, that's all details. That, that's the difference between a, a good app and a great app, is details.

Harry Stebbings 36:37 Huh. I love that. Um, I also love the first class attention to details.

Severin Hacker 36:42 Attention to details, yeah, yeah.

Harry Stebbings 36:43 Uh, can I ask you, where did you spend time that with the benefit of hindsight you're like, "Ah, we shouldn't have spent time there?"

Severin Hacker 36:50 Okay. Uh, so one thing that comes to mind is, so when we ... So we started with languages, and, but the mission was always all of education, right? So, so at some point we said, "Okay, let's, let's build new subjects. Let's build ..." Um, you know, we wanna add new subjects and then someone said, "Let's add math." Math is another one of these, like, fundamental skills that we'll probably need to learn for the next 100 years or so. Um, a- uh, that's our belief, or my belief. So, let's add math. And then the, the first, um, idea was to build a new app. In fact, we built a new app. We built a new math app. And-

Harry Stebbings 37:30 Like completely standalone?

Severin Hacker 37:31 Com- completely standalone.

Harry Stebbings 37:32 Wow.

Severin Hacker 37:32 So there was Duolingo, the main app, and there was a entirely new math app. We built it.

Harry Stebbings 37:37 Did it, did it have a name?

Severin Hacker 37:38 Well, yeah. Uh, it was, uh, it's, it was in the App Store. I'm, I'm actually not sure if it's still in the App Store. It was the Duolingo math app. (laughs) Again-

Harry Stebbings 37:44 Creative.

Severin Hacker 37:45 (laughs)

Harry Stebbings 37:45 Well done, Stefan. Wonderful.

Severin Hacker 37:47 Uh, and then, and then we realized, you know, this is looking more, uh, more like the, the main Duolingo app, and it's like, uh, we were about to add a leaderboard, and, you know, a league system, and a streak.

Harry Stebbings 38:05 (exhales)

Severin Hacker 38:05 And it was like, "Why are we doing this? Like, why are we replicating this whole functionality?" That was, you know, um, insight number one. And then insight number two is, for the user it was also not a great experience. Because A, you first had to find ... Well, first you had to know that Duolingo had a math app. Then you had to find it in the App Store. Then you had to download it. Then you had to install it. Then you had to grant it, uh, permissions again to send you push notifications. And then-

Harry Stebbings 38:30 (exhales)

Severin Hacker 38:30 ... it was like, "Wait, is this streak different from my Duolingo streak?" And, you know, it, it looks more, uh, feels very much like Duolingo but it's not quite Duolingo. Um, and then we said, "No, no, no. No more new apps. If it's education, it goes into the main app." So internally, we called this, um, uh, well, but it's kind of like a super app strategy, right? So, so that was, uh ... And w- w- we probably spent two hours working on, on this separate a- . Now the good thing is like, all the content and all that, we could just transfer in. So we didn't, you know, throw it out. But we, you know, merged it with the, the

Severin Hacker 39:10 main app.

Harry Stebbings 39:11 Can you talk to me about chess? How did the chess come about? We mentioned it outside, and it was a great story. How did chess come about? And, and talk to me about that.

Severin Hacker 39:20 Yeah. Chess, chess is a, is a interesting one because, you know, is it, is it really education? Uh, you know, it helps you think. I mean, we, we had this discussion-

Harry Stebbings 39:28 Mm-hmm.

Severin Hacker 39:28 ... before, uh, about, you know, should you still study computer science? I think chess is similar in the way that it helps you think, helps you think logically. Um, and so we believe it, it, it fits the, within the mission of Duolingo. That's number one. And then it was, uh, two people who were just super passionate about chess. And it's like, "Hey, let's, let's try to build a chess course." And we were already-

Harry Stebbings 39:51 Two people internally?

Severin Hacker 39:52 Internally, yeah, yeah. And we were not convinced, uh, at first. But this is another thing-

Harry Stebbings 39:57 Why were you not convinced?

Severin Hacker 39:57 We, well, we're like, well, you know, this is, this is just a game. We, we don't wanna build just a game, right? Like, there needs to be some educational aspect. Be- that, that's al- always a, a, uh, y- you know, a, a trade-off between, okay, are you just a game or are you, uh, an educational game? Or are you just, you know, a textbook, boring experience? So you have to ha- you know, find, find the balance between the two. And we don't want it just a game. So with chess, it's different. Like, so chess is, is not, you know, like a, a racing game or something. Chess is like, you know, it helps you think. So that's why, why we're good with it. But then, um, it's like, well, you know, how does this fit into the tree? Like, how do, uh, you know, the, sorry, the, the path? How does this look inside Duolingo? Can we make it fit into the Duolingo design language and all that? So that, that's another concern. Um, but then there's again, that's super interesting and, um, uh, uh, I wanna share a story. So, so the two of them, they built this prototype with AI. Okay? So they built it, the, the, there's a, a PM, and a designer. And the PM, he has some engineering background, so he knows how to think logically. (laughs)

Harry Stebbings 41:11 (laughs)

Severin Hacker 41:11 Um, but, but he hadn't written much code before, uh, you know, before Duolingo or at Duolingo. So, but he was able to build this first version just using Cursor, and, and, and a bunch of AI tools. And, and the prototype is what convinced us that we should do this. You know, if it had just been a, a Figma design, we probably was like, "Ah, you know, uh, maybe, maybe not." But the prototype was so good that it convinced everyone. It's like, "Yeah, let's build this."

Harry Stebbings 41:39 What was it about the prototype that was so good? It was fun? It was progressive in terms of mindset? What it w-

Severin Hacker 41:45 Y- e- e- you could play it and it, it just felt like this, this belongs inside the app. You know, it's like when, when, when you just see a screen it's like, "Uh, yeah, maybe." You have to kind of imagine it. Uh, versus if you have the prototype, it's like, "Well, this feels, looks and feels like Duolingo." Um, and, and then the, the other thing is like because of AI, again, we were able to build this within nine months. This is the fastest we ever created a, a new course.

Harry Stebbings 42:11 What do you not have ... Sorry, nine months from start to finish?

Severin Hacker 42:15 Yeah.

Harry Stebbings 42:16 Wow. And normally it would be-

Severin Hacker 42:18 ... well, you know, uh, years.

Harry Stebbings 42:21 What do you not have that you would love to have in terms of activity?

Severin Hacker 42:25 On-

Harry Stebbings 42:26 Like, and you... Uh, so you, you said that you've added languages, you've added maths, you've added music, and now we have chess. What's the next one that you'd love to have?

Severin Hacker 42:34 So for me right now, I think the, the... What I'm most passionate about is the, the social aspect of the app. You know, I think we have this leaderboard mechanic, um, but I think we have this... We now have this ma- massive user base, and we have all this brand love. People love Duolingo. And I feel like we... there's a- we're- there's an untapped potential of connecting the people that are on Duolingo. Um, I don't know how to do it (laughs) so, um, but I'm, I'm very What would that look like?

Harry Stebbings 43:03 ... interested in that.

Severin Hacker 43:03 Sorry, Duo- Duo Social? Duo Date? Uh, uh (sighs) um, f-

Harry Stebbings 43:08 'Cause Duo Date's not a terrible idea.

Severin Hacker 43:10 (laughs)

Harry Stebbings 43:11 It's not.

Severin Hacker 43:12 Well, uh, yeah. I don't know. It, it, it could be, could be something like that. It could be, you know, learning together. Um, but it could also be... Uh, I don't know. I, I think there's just a lot of untapped potential that there's not that many apps that

Severin Hacker 43:28 have our size of user base.

Harry Stebbings 43:28 Would you ever open a physical school? And I know that sounds completely counter what Duolingo's all about, but having, in New York, a branded amazing experience where people could go for a week, meet the owl, sit with people in their leaderboard, and build that community play that is really special and different.

Severin Hacker 43:50 It's a good question. So not sure about that, uh, directly. But one thing that is interesting about education is... Okay, so there's K-12 and then there is higher ed, universities, college, et cetera. And one thing that is interesting when, when people talk about education, we always assume it's the instruction. It's like, you know, here's skill, I learned this skill, and now I go from A to B, right? That, that's kind of like how we think of education. But higher ed, for example, is actually more than that. It's like, it's... Going to college, it, it's three things. It is the actual instruction. It's the actual teaching. It's actually the, the learning of new skills. Okay, that's number one. But then there is also the degree, uh, the cre- credentials, right? So, uh, university says, "Harry, you got a degree. You, um, you know, successfully finished this, this, uh, this course, and you can put this on your, on your LinkedIn or resume." And, and then an employer can say, "Oh, yeah, I guess, you know, um, Harry did, uh, take this computer science class, and therefore we can hire them as a engineer." So that's credentials, number two. Very important, because otherwise, you would always have to prove again that you know something, right? But that's what it... That's the, the purpose of the degree. And number three is the social aspect. So a lot of people, uh, meet their, you know, their, their partner at the... at, at, at the university. And, you know, it's their, their friendship network. So most companies that look inside at tech, they just focus on number one. But I think if you really wanna... let's say, if you really wanna disrupt the Harvards of the world, you have to solve all three. Um, that's one thing. By the way, w-

Harry Stebbings 45:44 Do you want to disrupt-

Severin Hacker 45:44 On the number two-

Harry Stebbings 45:45 ... the Harvard?

Severin Hacker 45:46 No. So we... Well, we wanna achieve our mission. That, that's, that's why we started this company, okay?

Harry Stebbings 45:53 But by achieving your mission, are you not disrupting Harvard?

Severin Hacker 45:56 Well, yeah.

Harry Stebbings 45:57 (laughs)

Severin Hacker 45:57 (laughs) We would be. Um, but one thing, uh, uh, on, on the second thing. So we have not solved the, the social thing, right? Okay. This is what higher ed does. We have not solved, solved the social thing. But on the number two, the credentials, one thing we're also very passionate about is, is the Duolingo score. I'm not sure if you, if you've noticed it within Duolingo, but we now give you a score, like how good you are at, at this language. And, and we want this to become the way people talk about language proficiency. So it's like, "Hey, my Duolingo score is 61 in French." And then people know, "Okay, I guess you can have a... this kind of conversation," et cetera, et cetera. So, so we're getting into the credentials or the degree aspect of it with this score, so.

Harry Stebbings 46:38 It's actually pretty helpful the amount of times where people say, "Oh, do you speak French?" And you're like, "Uh, a l- a little bit." And it'd be helpful to say like, "Oh, I'm actually a 40." And they're like, "Okay, not really. He's stupid," (laughs) you know?

Severin Hacker 46:50 (laughs)

Harry Stebbings 46:50 Or whatever it is. Do you know what I mean? It's like quantifies it a lot more easily. Can I ask you? You men- and this is kind of jumping around, but you mentioned like the push notifications element. Does that make a difference still? We have so many push notifications. With the oversupply of push notifications, are you like, "Ah, that was ruined, it was good," or not?

Severin Hacker 47:08 Still makes a huge difference. And, and we know because whenever the service goes down that sends the push notifications, you can see the dip.

Harry Stebbings 47:16 Really?

Severin Hacker 47:16 Yeah, yeah, yeah, 100%. It's, it's, it's huge. It's massive, yeah.

Harry Stebbings 47:19 Wow.

Severin Hacker 47:20 Yeah. Um, going back to the other, the other-

Harry Stebbings 47:23 Yeah.

Severin Hacker 47:23 So, so there was the higher ed, three things. For K-12, there's also, you know, people also, "Okay, you know, like, my kids, um, they, they, they go to school to learn things." But again, it's not just the learning. In fact, you know, a lot of K-12 is childcare (laughs) you know? And, and so even if you solve one aspect of it, like let's say you have an AI tutor that is really, really good, you still need to solve the, the, the childcare, um, problem. And, you know, it, it, it... It's unlikely that people will just go to... or to a school without any humans, without any teachers, and just use apps and then, you know, go home. That just... Uh, or that we would have some sort of Optimus, you know, in, in a school. I think there's... a lot of it, uh, needs to be done by humans still. And, and by the way, the other thing is like... So, I do believe that the AI tutors will become really, really, really good. And it's, it's gonna be the way most people learn most things in the future. I, I do believe that.... however-

Harry Stebbings 48:25 What does the future of schools look like then, with that? Is it, like, 10 children sitting in a class with AI tutors?

Severin Hacker 48:31 Uh, I, I think that there's still gonna be this, this human element of, "What do I need to learn?" Kind of guidance and coaching and motivation. So guidance, coaching, and also, "Why should I learn this?" It's like, an AI will never tell you, like, why you should learn this. It's like, the, they'll help you learn it, but it's not gonna tell you, you know, "Why should you learn this? Why is this important?"

Harry Stebbings 48:53 Do you know what I wish you had? I'm sorry. You have leaderboards, which is like social status or social mechanics that drive motivation and incentives. I, I'm sure you've studied this. I have 'cause I'm a loser. Um, it's financial incentives and it's actually when you offer someone money, they are much less likely to do something. When you take money away from them, they are much more likely to do it. Basically, the fear of losing is much more than the gain of winning. And I wish that Duolingo had this, for every day you don't do your streak, we'll take a committed amount off you. Say I commit to 100 pounds a day. Fuck, dude, I would do that.

Severin Hacker 49:30 (laughs) You would do that?

Harry Stebbings 49:30 Like, uh, yeah, 100%.

Severin Hacker 49:33 (laughs)

Harry Stebbings 49:33 100%.

Severin Hacker 49:34 You know, you, you, you could, you can, you can just give me the money whenever you lose your streak. (laughs) Um, no, no-

Harry Stebbings 49:41 That would be the best sidekick,

Severin Hacker 49:42 (laughs)

Harry Stebbings 49:42 ... to use another 100. Harry's just giving me money. (laughs) Turn off his push notifications, guys. No more. (laughs)

Severin Hacker 49:49 Yeah, yeah, I mean, you know, you know, the, the, the streak is exactly the, this loss aversion mechanic, right? So it's like, it's like, it's motivating to build up your streak, but it's like I absolutely do not want to lose my streak, right? It's like, I will do a lot of things to not lose.

Harry Stebbings 50:05 Have you had crazy stories of people who do crazy things to-

Severin Hacker 50:07 Oh, yeah, yeah, yeah, it's cra- And, and we also sometimes, um, you know, get emails like, you know, "I was, uh, on this, uh, trip to climb Mount Everest and I couldn't extend my streak and-"

Harry Stebbings 50:17 (laughs)

Severin Hacker 50:18 "... can you please, you know, uh, restore it or something? And I had no internet access and..." Um, yeah, always.

Harry Stebbings 50:25 Do you go, "No." S-

Severin Hacker 50:26 No, no, no. We, we, uh, it depends. Oh, by the way, also when there was, you know, there was a power outage in, in, uh, Portugal and Spain, we, we proactively, uh, extended their streak. So, so they don't lose it.

Harry Stebbings 50:38 It is amazing-

Severin Hacker 50:39 Yeah.

Harry Stebbings 50:39 ... isn't it?

Severin Hacker 50:40 Yeah.

Harry Stebbings 50:40 I mean, that is so cool. Um, listen, I, I do want to talk a little bit about how you work 'cause you're fascinating in how you work. I, I saw in, in some of your posts, you said you're not good at finishing the final 20% of a project or a bit of work. Is that... Ta- (laughs) That seems problematic. Is it? And how do you reflect on that?

Severin Hacker 51:02 It would be problematic if it was just me (laughs) and everybo- or everybody was like me, right? So I think ultimately, you need to have complementary skills, uh, within the founders, but then also within the whole exec team, within the whole company, right? So, so if everybody was really bad at finishing or, like, the last 20... By the way, the last 20% are basically details, right? So you need to have someone who's obsessed with details. In, in Duolingo, that is, that is Luis, uh, to this day, um, but also, you know, from day one, he was just obsessed with details and he really cares. And, and it's not, by the way... I'm, I'm like... I, I a- I also care about details. (laughs) But he is an order of magnitude better at that. I think-

Harry Stebbings 51:43 Has he ever been obsessed on details to the detriment of the company? "Dude, we don't need to obsess over it that much. Come on."

Severin Hacker 51:56 (laughs) Uh...

Harry Stebbings 51:57 (laughs) That's a yes, but I'm not sure whether I should say it. (laughs)

Severin Hacker 52:04 (laughs) Uh...

Harry Stebbings 52:06 (laughs)

Severin Hacker 52:06 Uh, uh, he, you know, he, he mostly... So he, he's, uh, he mostly works with the PM org, the product org, and they would probably say something like that when, uh, you know, we still have this product review process where every change goes through, um, a review, a- and Luis is a- sits in most of those. So he sees-

Harry Stebbings 52:25 He still does?

Severin Hacker 52:26 He still does. So that's a founder mode for, for, for, (laughs) for him, is like he's still in most of the product reviews. Um, and I think that's rare for a company CEO of, you know, a company like, like the, like, our size, uh, where, where the CEO sits, sits in most of the product reviews.

Harry Stebbings 52:44 Wow, so you decided to delegate that to the product team-

Severin Hacker 52:48 (laughs)

Harry Stebbings 52:48 ... on whether he obsesses or not. I love that.

Severin Hacker 52:50 (laughs)

Harry Stebbings 52:52 Uh, how do you... You said founder mode there-

Severin Hacker 52:54 Mm-hmm.

Harry Stebbings 52:54 It drove so much engagement and discussion. How do you reflect on it? And when you look at your own style of management and leadership, how do you think about that?

Severin Hacker 53:02 My role has changed practically every year since the, the beginning, since when we started, so-

Harry Stebbings 53:08 Yeah, Tanzeen told me this. Uh, I spoke to him from General Atlantic. I stalked you, I told you. Um, and he said your role has changed so significantly.

Severin Hacker 53:15 Yeah, yeah, yeah. So, so early on I was, you know, implementing the thing. The thing I was building, Duolingo. And then it, it morphed more into, um, more, like, into, into code re- architecture, uh, how to build, uh, the backend system and, and all that, system design. And then it morphed more into management and hiring. I think I'm actually quite good at hiring, uh, and then hiring engineers. Uh, but then I, you know, there's also other stuff that you, people wouldn't expect (laughs) from a CTO.

Harry Stebbings 53:46 Cool.

Severin Hacker 53:47 Um, so one of the... You know, I was the one who pushed for hiring our first product manager. Uh, it was actually interesting because Luis, at the time, he's like, "Oh, we don't need product managers." He's... He had seen some of them at, at Google. Uh, his, his previous, uh, company was acquired by Google, so he had see- uh, seen some of them at, at, at, at Google. And he had worked with some of them and was like, "Hey, you know, I don't think they provide that much value and we don't need this. You know, I can d- do this myself and, and we just need a bunch of designers." I go, "No, no, we really should..." You know, like if, if every single company has a product manager our size, maybe, maybe we are, maybe we are wrong. You know, let's, let's hire one and if it doesn't work out, then, um-... you know, we can, we can still, uh, go with that. But obviously, that was the right decision. (laughs) Obviously, that was the right decision, so.

Harry Stebbings 54:33 Hiring. What do you think makes you good at hiring that is non-obvious or not typical?

Severin Hacker 54:40 We have a really strong consumer brand. So literally, you know, everyone, everyone knows Duolingo at this point, uh, worldwide.

Harry Stebbings 54:48 Given the strength of Duolingo's brand-

Severin Hacker 54:50 Yeah.

Harry Stebbings 54:50 ... do you think Duolingo is underpriced today? I think we dramatically underpriced brands. And when you think about global brands, which billions of people know.

Severin Hacker 55:00 Uh, underprized in, in terms of?

Harry Stebbings 55:03 Monetary value. Like-

Severin Hacker 55:04 Oh, uh, the stock price?

Harry Stebbings 55:06 Yeah.

Severin Hacker 55:06 Yeah. (laughs) I c- I cannot talk about that.

Harry Stebbings 55:08 (laughs)

Severin Hacker 55:09 (laughs) Uh, but one, w- I mean, it, it is just amazing-

Harry Stebbings 55:12 Do you see what I mean by that though? Like, you know, people often think about like Manchester United, it is a, a, a global brand, for example. And I think it's drastically underpriced given the fact that you have, like, children in India and Nepal, Brazil wearing my new shirt. And it's a $5 billion company?

Severin Hacker 55:28 Yeah. I, I, I don't, I don't know. I, I, I think the, the, the, there's like, it's a massive brand, uh, Duolingo. So it's, it's... I think at some point we were bigger on TikTok than Nike, which is, you know, a, a brand that has existed for a very long time. We all obviously know Nike. Um, so it, it is a very strong brand. And the other thing that is like, I always say is like, we, we run the world's most efficient marketing organization. It's like people think, "Oh, yeah. You know, they're, they're big because they, you know, they spend a lot of money on advertising and, uh, social media influence and stuff." It's actually, by impact, we must be the, the, the most efficient marketing org. It's a small team. It's a small team.

Harry Stebbings 56:16 How big is the team?

Severin Hacker 56:17 And, uh-

Harry Stebbings 56:18 Give or take.

Severin Hacker 56:18 ... I, I have, I think, 40 or so. 40 people or so. Um, and, and, and the reason for it is, for a very long time, we had no marketing budget. So we, we didn't advertise nothing, like $0 marketing budget. Because we're also not profitable, we also didn't have any revenue. So it's like, "No, we can't spend money that we don't have." (laughs) So, so there was, like, the constraint. But then the constraint, um, was really good in, in enforcing us to come up with good ideas. So we always say that Duolingo marketing it's, it's not powered by money, it's powered by ideas, good ideas. It's, uh, you know, you know, what you see on our TikTok channel. So it's, it's, you know, having the, the, the green owl, our mascot, having the costume, and also, you know, kind of some- sometimes, um, doing unhinged things. Uh, and, and, and kind of, like, also making fun of, of, of, of ourselves in a way. Uh, that is, like, why it, it is just so, so effective. Yeah.

Harry Stebbings 57:25 I, I, I love that in terms of the creativity that comes from the constraints. Can I ask you, you mentioned about the not having PMs, and actually, if all the biggest companies have them, maybe there's something to it. And it makes me think of this, like, conventional rules are conventional for a reason sometimes. What conventional rules do you find are conventional for a reason, one? And then what one conventional rule do you find, "What the fuck? That, that's just the way it's been, but it's not right."

Severin Hacker 57:54 Well, first of all, I, I would say you should always, not always, but every now and then, especially when there's a big technology shift like now, you should question all of these assumptions, right? Like, uh, you know, how you build companies. I think it, it just we're in the middle of this massive transformation. You can build companies with probably 10 times fewer, uh, em- employees. Um, so, so that, that, that convention is, is, is, uh, you know, uh, completely overthrown, uh, uh. Uh, you know, that you need to hire a lot of people. Um, that's one thing. Uh, I would say, you know, (laughs) early on there was like a f- uh, this is, this is going back. It's not a, not a great answer. But, um, there was like this, this belief that you can run a completely flat organization, right? No, no managers. Um, no hierarchies. No levels. And I think that's not really possible. Uh, at a certain scale, you need to have managers or layers, or you need to have some sort of hierarchy. You need to have some sort of, um, career progression and stuff like that. I think that, uh, you know, the, the oldest organization in the world, the, the Catholic Church that is still around, it's very hierarchical. And, and I think there's probably some, some reason for it. Yeah.

Harry Stebbings 59:20 Does that not go against founder mode and Louis sitting in the product reviews? And when you think about managers or managers and layers, and the diffusion of kind of knowledge and power that comes with those layers, does that not go against the man- founder mode?

Severin Hacker 59:34 Uh, yeah. Uh, uh, I think the, it's always this, this w- how do you strike the balance, right, between the two?

Harry Stebbings 59:39 What do you still do today that you don't think you should do?

Severin Hacker 59:43 Uh... (laughs)

Harry Stebbings 59:47 So like, for me, I li- you know, I listen to every single show we do. And I... yeah, yeah, 10 years in. Why? It's 'cause it's my product. I'm not shipping product that is not perfect.

Severin Hacker 59:57 You know, I'm actually quite good at, at, at delegating things and, and kind of, like, uh, self-awareness of what I'm good at, what I'm bad at, what I like, what I don't like, what is beneficial to the company and what is not beneficial. So that's also why my role has always changed. And one of my, one of my principles is, reduce, automate, delegate. But the-

Harry Stebbings 1:00:21 Can you walk me through that? So-

Severin Hacker 1:00:22 Yeah. So the, the, the main idea is, uh, really, you know, maybe like once a month, once a quarter, so really think about, like-... does this need to be done, at all? You know, what happens if you don't do it, if you just don't do it? Is it the end of the world or is it, is there a actual, uh, negative effect? Um, so that's number one. A- and particular, what happens if I don't do this? What happens? Okay. And number two is, what you cannot reduce, what you cannot remove. Uh, let's say you decide, "Okay, this action needs to be done." Then the second question is like, can I automate it? Like, now again with AI, it's a lot easier. Like, can I automate this part of my job? Uh, can I use, uh, ChatGPT to, to do this, um, report, or answer this question, or... So that's the second... And, uh, that's, you know, self-injury, that's very natural. It's like, "Okay, can we build a tool that does this for me or the company or whatever?" Um, so, and then what you can't automate, uh, you delegate. Okay. So that I, I try to delegate to, um, other people in Dork. And (laughs) you know, I've handed off most of the day-to-day engineering to our head of engineering. She's really good with this. I'm, I'm now a little bit out of the, uh, out of the weeds, and I'm thinking more... Right now actually, I'm only focused on two things. One is like, this, this question is with AI. Like, what's gonna happen? Like, what, what is the future look like in, you know, uh, in our space? What is the implications for Duolingo? What actions do we need to take? That's number one. I probably spend 80% of my day thinking and acting on this, this AI question. And number two is, is, is more M&A. Um, and, and it's, it's r- relates to the, the first question. It's M&A, like, you know, what companies should we buy? Um, where do we invest, uh, outs- e- externally? And, and that's, that's basically all, all I do right now.

Harry Stebbings 1:02:22 Where have you not invested that you would like to invest?

Severin Hacker 1:02:26 Uh, we did make a bunch of acquisitions. Um...

Harry Stebbings 1:02:29 How many have you made?

Severin Hacker 1:02:30 Not that many. Four or five.

Harry Stebbings 1:02:32 Any lessons from doing that? It's hard. I remember having ............................ on from Lightspeed, the payments provider, um, in .......................... And I... they've done like 18, and there was a lot of lessons there. (laughs)

Severin Hacker 1:02:43 You know, you know, it's actually m- very interesting. So one insight, and i- in hindsight it's, it's kind of obvious. But, you know, they say, like, "Wh- what is the best predictor of, uh, two founders working out?" Like, you know, they're actually, uh, they're not gonna get into a fight and, uh, founder conflict. And the number one, uh, thing to look for is have the two founders worked before they started this, this company. Like, have they worked together? Um, for, for example, Luis and I, we had been working, uh, together on, you know, research for two years prior to starting Duolingo. So... And it... By the way, it's not the same a- as friendship. So it's like, you can know someone for a very long time, but they might be very different at work. So, you know, prior work experience, um, of working together. That is number one. And with M&A it's the same thing, I actually think. It's... Do you have prior experience working with this company? Yes or no? So I would say that's the number one advice is like, "Have you worked with them before?"

Harry Stebbings 1:03:57 Do you incorporate with the-

Severin Hacker 1:03:57 And, and it doesn't have to be like... So this was a, you know, customer, uh, vendor relationship. It can also be a, an investment relationship.

Harry Stebbings 1:04:05 Do you do Duolingo ventures? Do you invest early?

Severin Hacker 1:04:07 We just made some minority investments, but it, it's, it's, it's not a lot.

Harry Stebbings 1:04:13 It's always with a view to buy, essentially.

Severin Hacker 1:04:16 Uh, we have to figure this out. (laughs)

Harry Stebbings 1:04:19 (laughs) Uh, uh, it's a competitive space. Um, you know, I've, I've seen many competitors. I remember, I think you came pretty close doing a round in Practica. Um, how do you view the competitive landscape today and the distribution of value across it? Is it like Uber and Lyft where Uber takes 90% and everyone else takes 10%, and that's Duo and everyone else? Is it a much more distributed value kind of dispersion? What does that look like?

Severin Hacker 1:04:46 In consumer it's much more, uh, winner take all, kind of, like, um, mechanic. So we've been now running this for 12 years. I would say every single year there was one or two companies that grow really fast in language learning. One or two. And inevitably, they go down again. And what happened is they, they used venture money or some other funding to grow through paid acquisition. So they spent it on, you know, Facebook ads, uh, Google ads, and, and they go really fast. But ultimately, the only thing that matters in our space is retention. So if you... You can spend as much money as you want, if you don't have retention it's a leaky bucket. All these people are gonna leave. And so we're not... Now when you see one of those, we're not scared, because we've seen this play out a million times. So the one company I'm scared of is one that has higher retention than Duolingo.

Harry Stebbings 1:06:01 What's the day retention number that matters to you?

Severin Hacker 1:06:05 Of the, you know, millions and millions of people that use Duolingo daily, how many come back tomorrow?

Harry Stebbings 1:06:09 Got you. How-

Severin Hacker 1:06:10 And, and, and... You know, all of these correlate with each other. So if you're, uh, current user retention is very high, you also have very high D7 retention, you also have very high D30 retention. But you get a data point every single day.

Harry Stebbings 1:06:24 That was absolutely fascinating. It reminds me of Alex Schultz who says exactly the same thing on like, "Retention, retention, retention." So I totally get

Harry Stebbings 1:06:30 you there. I, I, I do wanna ask before we move into a quick fire, um, you're European. Uh, you are also based in New York. When you look at Europe today-I sit here and I'm very aware of our fading into irrelevance, which is why we do things like Project Europe. How do you think about Europe? Are we fading into irrelevance? And how do you see that having the perspective of being a European in the US?

Severin Hacker 1:06:55 Uh, I- I'm very conflicted internally. Y- you know, I- I want Europe not to fade into irrelevance, of course. Um, however, if you asked me, you know, if- if- if a young founder, European founder came to me and said, "Hey, I have this brilliant idea of how to use AI," or, like, uh, a new model or something like that, "What should I do?" Uh, I'd say, "100% go to Silicon Valley. 100%. Uh, you- you absolutely maximize your chance of..." And by the way, you know, this is, this is-

Harry Stebbings 1:07:26 Why? Why?

Severin Hacker 1:07:28 I- it's just the environment is so much stronger. It's just like, it- it just... And I- I can say this because we- we built Duolingo in Pittsburgh. We're not a Silicon Valley company. We don't have an office in Silicon Valley. And we- we succeeded despite not being in Sil- Silicon Valley. And- and by the way, there's also pros of not being in Silicon Valley and- and we can talk about this. But I can guarantee you Duolingo would not have been able to raise any money in Europe. You know, we had $0 revenue for the first five years. Zero. There's- there's no European investor who would have invested. Not in Switzerland, not anywhere in Europe. In fact, in the US-

Harry Stebbings 1:08:13 Do you think that's true today? And I'm not being rude, dude. You're a really smart guy. You came out of great institutions. Lewis, I know, is a super smart guy too, out of great institutions. There's so much money in Europe now. Dude, I'll give you a term sheet today-

Severin Hacker 1:08:28 (laughs)

Harry Stebbings 1:08:28 ... when you were, but just the two of you coming in with-

Severin Hacker 1:08:31 I- I think it has changed to a degree, but it's still... You need to look at what it is the most ambitious people and- and- and, you know, the- the founders, where- where did they go? And- and, you know, it's- it's a little bit sad because in- in Europe I see a lot of European founders and I often think, like, "How successful would they be if they had stayed in Europe?" And I think for most of them, I would say, "Mm, best maybe half as successful," or maybe sometimes 0%. Uh, you know, like, I- and I- I would like to change that. I would like to change that. I mean, that's why, you know-

Harry Stebbings 1:09:09 (laughing)

Severin Hacker 1:09:09 ... I- I invest in- in European VCs. I invest, uh, you know, I'm part of the Project Europe. Uh, we need to change that because... (laughs)

Harry Stebbings 1:09:16 How do you think then about Synthesia, about Eleven Labs, about some incredi- lovable in Sweden being built from Europe?

Severin Hacker 1:09:24 That is great. I think for... And, you know, uh, they- they- they should deserve all of our support.

Harry Stebbings 1:09:32 (laughs)

Severin Hacker 1:09:32 And, you know, the- the- the- the- the- the sad part is also I think in, you know, in Silicon Valley is like being a tech founder is- is- is really, you know, you're- you're mostly a good person. Now, of course, there's also... (laughs) Uh, it's- it's- it's aspiring, right? So- so people want to be that. I think a lot of- a lot of people in Europe you're just, like, you're kind of like, "Why are you doing this?" Like, "Why? Why?" You know like... Or- or- or they- they- they look at it as like, "Oh, yeah, you know, they're just doing it for the money," or, uh, there's this suspicion, uh, almost of- of, uh, you- you can't be too successful or you can't be... You shouldn't be too ambitious. And- and I think that- that is very, uh, detrimental to the- the European startup ecosystem.

Harry Stebbings 1:10:21 You shared an insight with me before that I'd never thought about, about having offices in San Francisco as a non-SF HQed company. What was that insight and how did it lead to your not wanting to have an office in SF?

Severin Hacker 1:10:34 Yeah. So Duolingo does not have an office in Silicon Valley or anywhere in California and the reason is (laughs) when- when we thought about opening one, and of course all of our investors they said, "Oh, you've got to open a SF office." And we in fact did go to SF and we looked at office space and we were this close to opening an office in SF. But before we pulled the trigger, we asked a bunch of founders who had companies that were headquartered outside of the S- outside of Silicon Valley and then opened an office in- in Silicon Valley and we asked them, "What do you think? It was good decision, bad decision?" And all of them said, "Worst single decision ever." And I said, "Why?" And they said, "Well, you know what happens is it- it creates this internal, uh, funnel of your best employees moving from your HQ to Silicon Valley," which by itself is not a problem. The problem is then they get recruited away by the, whatever the hot Silicon Valley company is at the time. You know, like it could be Airbnb or Uber at the time, maybe now it's OpenAI or Anthropic, and you basically create a hiring funnel for Sil- Silicon Valley and you- you lose your- your talent. And that was also, like, new to me and then it was like, "No, no, let's not do that."

Harry Stebbings 1:11:56 I love that. I hadn't thought about that at all. And so you will not do an office in SF?

Severin Hacker 1:12:01 Nope.

Harry Stebbings 1:12:02 Do you think I would be more successful if I was in SF? I know that's an unfair question to ask, but it's something that I've always been told, Severin, for years and years. "You have to. You have to. Why are you in London?"

Severin Hacker 1:12:14 Uh, let's see. As a founder, I would say 100%. As, um...

Harry Stebbings 1:12:20 Let me tell you my rationale and you can see if you think I'm right and you can call bullshit on me. Like, in SF, I'm competing against Peter Fenton, Marc Andreessen, Vinod Khosla, Eli Gill, the Collisons, Daniel and Nat Friedman. I- I rate myself. I buy my product. But that is a fucking hard market. And Marc Andreessen getting you round to his house for breakfast on a Saturday morning and brol...In Europe, there's probably two good VCs.

Severin Hacker 1:12:49 Yeah. Y- you know, by the way, so I (laughs) I, uh, I started angel investing with very similar thesis in Pittsburgh because it was like, there's no one other than me and Luis. (laughs) You know? It's like, I, I... But, but then, you know, a lot of those, uh, companies eventually moved to Silicon Valley, um, so... But yeah, that was, that was exactly my, my investment thesis in, in Pittsburgh. But I... So I, as... I think as an investor, it's not clear. As an investor and, you know, doing this show, I think it's, it's not clear because you... Again, yeah, I think the... this is much less crowded. You can really become the, the, the top, you know, investor that founders wanna work with in Europe. I think, I think that, that... I agree. But I was, I was explaining this as a, as a founder, if you wanna start a business, an AI business, and you have this crazy idea, would you stay or not? I would, I would move.

Harry Stebbings 1:13:42 You know, when someone's-

Severin Hacker 1:13:43 But, but that's exactly the problem. Like, how can we change that? You know, part of the problem... This, this is a little, uh, um... Not, not, not obvious, not obvious, is part of the problem is that it's so easy to move to Silicon Valley. It's like, it's actually... If you, if you have, you know, a good degree and, um, y- you know, you can raise money, etc., the US is actually really good at, you know, getting its brains, uh, to come to Silicon Valley. And in a way, the best thing that could happen to, to Europe is if the US, US made it really hard to immigrate for, for founders. You know, if they, if they made it really hard, if they restricted immigration, such that you, you know, people like me couldn't, couldn't move there anymore, then, well, I, I would build it here. And what I would do first, it would... Let's first change the environment here to make it easier for startups to succeed, right? Because right now, there, there's also n- no pressure from, from, you know, the industry to change the regulation to, uh, y- you know, make it easier to start companies, because I can just go, I can just... I can vote with my feet.

Harry Stebbings 1:14:49 What would you change about Europe? Is it the regulation? Is it the level of capital? Is it the mindset that we tell young people? All of the above, obviously.

Severin Hacker 1:14:58 All of, all of the above. All of the above.

Harry Stebbings 1:14:58 But is there one?

Severin Hacker 1:14:59 Uh, regulation is, is definitely hurting. Uh, the regulation, um, is definitely hurting. I think there is... And then there's country-specific things.

Harry Stebbings 1:15:08 I always fall back to no great entrepreneur was stopped by regulation, though. Which is like, Travis wasn't stopped by regulation with Uber, uh, Elon with SpaceX wasn't stopped by... Regulation does not inhibit the greatest entrepreneurs.

Severin Hacker 1:15:21 Um...

Harry Stebbings 1:15:23 Discuss. (laughs)

Severin Hacker 1:15:25 I, I... Again, I think it depends if there's an alternative or not, right? So again, as a founder, I can be like, "Okay, I can work with the EU regulation on AI," which is, I think some of the dumbest regulation ever written. It's literally like, you know, the, the Ottoman Empire, they had, um... When, when, you know, they invented the printing press, it's like, "Oh, this is not good. It's not good for our community. It's not... Let's not do the printing press." I feel like the A- EU AI Act is that. It's like, let's not do AI, you know, because there could be negative repercussions or negative side effects or something. It's like, it's that dumb. It's like, it really... Like, yeah. I, I... You know, if, if there, if there was no alternative, then I would agree. Um, by the way, in, in... For example, in, in, in Uber's case and, and Airbnb, there's a lot of local regulations. There's actually lots of alternatives. So e- even if like, say, one c- one city doesn't work, like, it's like New York, this is the regulation, you have 200 other cities in, in the US.

Harry Stebbings 1:16:26 If you could go back to the early days of Duolingo and do something differently, what would you do?

Severin Hacker 1:16:31 Oh, yeah. So I would say, uh,, and we're pretty open about this, I think that the two mistakes we've made at Duolingo, they were not fatal, but... (laughs)

Harry Stebbings 1:16:42 Obviously. (laughs) 20 million monthly pay-

Severin Hacker 1:16:44 But, but I would say the two big, big ones were, one, we, we waited too long with monetization. So we just didn't take it seriously enough early on, and we were, like, operating under this business model of let's just raise more money. Like, our business model, people, people ask, "What's your business model?" "Oh, it's, like, venture capital." (laughs)

Harry Stebbings 1:17:05 (laughs)

Severin Hacker 1:17:06 So, uh... And eventually, we did take it seriously. And, you know, we're very good at, at this now. Um, you know, we have this internal, uh, concept of the Green Machine. It's like you, you run a lot of experiments, see what works, and then double down. And that's what we did with monetization. It's, it's, you know, super good now. But we waited, we just waited for too long.

Harry Stebbings 1:17:26 Was that because of naivety? Was that because of a push for user growth and engagement?

Severin Hacker 1:17:31 A combination. So it was also like the, the investors, they didn't push us to monetize early, which is... You know, it was always counterintuitive to me. It's like, "Why don't they ask for monetization?"

Harry Stebbings 1:17:41 But I never would because I'd want engagement and I'd want user behavior and I'd want habit forming. Don't fucking put a paywall in front of them.

Severin Hacker 1:17:48 Yeah. And then, you know, our investors, they also invested in, in Twitter and, and, and some of these... And, and for, for those platforms it was always the right decision just grow users first, grow users first. So that's what we did. So number one is we knew we could always grow users first. I mean, we could always grow users more. Number two is we knew we could always raise more money. (laughs) And, and I think the other one was our mission, right? So, so our mission is to provide the best education and make it universally available, and for a long time, we didn't know how to build a business model that is compatible with this mission, right? So, uh, for a long time it's like, well, we just have to... You know, everything has to be free. But then we realized, no, no, no, what really matters is that we don't exclude people. Like, we want to be inclusive. We, we don't, um... There's... You should be able to learn everything that is on Duolingo without a bank account. And it's still true today. In fact, you know, most of our users do not pay Duolingo. It's... They use the free tier. And that's actually another thing I'm, I'm very proud of. It's like, it's not just, you know, a great business, but it's, it's a great business with a great purpose. Um, so... But that, that's also, like, why we probably waited for a little bit too long. So, you know, we should have started two years earlier.Number two is wha- we, uh, (laughs) we also waited for too long hiring more senior managers. So we just ... For, you know, up, up, up until 30 employees or so, it was completely flat. It was, uh, uh, (laughs) a, a lot of people from new grad. It was pretty chaotic.

Harry Stebbings 1:19:24 I, that goes against everything everyone says to me on the show. Everyone always eschews or, like, you know, says, "Don't hire the senior manager. If they've been there, done it before, the playbook's not transferrable. You want youngsters that you bring up through the ranks, um, they wanna be involved more in the weeds and hands-on. Beware of the manager."

Severin Hacker 1:19:44 Y- you know, again, there's like a, you have to strike the right balance. You cannot run a thousand-person company with all new grads.

Harry Stebbings 1:19:53 (laughs)

Severin Hacker 1:19:54 (laughs) You need, you need a little bit of both. By the way, today, it's still like, Dueling is a very, uh, young organization. So we, we, like, we still hire a lot of people directly from, from university.

Harry Stebbings 1:20:07 You mentioned you always knew that you could raise more money. I love that. What was the business model? Venture capital. Harry's like, "Thank God I still have a job." Um, but when you think about the fundraising that you did, what was the hardest?

Severin Hacker 1:20:21 Of, uh, series A?

Harry Stebbings 1:20:22 Talk to me about that.

Severin Hacker 1:20:23 Uh, so, so we, we started as a research project at, uh, Carnegie Me- Mellon University, and the belief was, oh, this, it's these two academics. Mi- Luis was a professor, and I was his PhD student. It's these two academics and, and, and (laughs) they're gonna build something that's not gonna make any money, and they're gonna sell to Google. Okay? And then that, that was kind of the belief. And by the way, that, that belief lasted until we actually made money. (laughs)

Harry Stebbings 1:20:50 (laughs)

Severin Hacker 1:20:50 Except for like five years, the first five years.

Harry Stebbings 1:20:52 (laughs)

Severin Hacker 1:20:52 It's like, "Oh yeah, they're, they're, you know, these like, these, these two researchers, and they're gonna build this, this thing, and uh, they're eventually gonna sell, you know, they're not, not,