Pod Insights Decoder Recode Decode: Rosetta Stone President Matt Hulett About this episode Matt Hulett, the president of education company Rosetta Stone, talks with Recode’s Kara Swisher about the company’s pivot away from language-learning CDs and into a more advanced and “approachable” mobile app. He also talks about his “promiscuous” background in the tech industry; Rosetta Stone’s literacy catch-up program for children, Lexia Learning; and the bigger picture of how language education is changing around the world. Plus: What’s the best way to make learning stick, and will Elon Musk’s Neuralink or some other moonshot technical development make learning languages unnecessary?
Featuring:
Matt Hulett, president of Rosetta Stone (@matt_hulett / @RosettaStone)
Hosts:
Kara Swisher (@karaswisher), Recode co-founder and editor-at-large
More to explore:
Subscribe for free to Pivot, Kara’s podcast with NYU Professor Scott Galloway that offer sharp, unfiltered insights into the biggest stories in tech, business, and politics.
About Recode by Vox:
Recode by Vox helps you understand how tech is changing the world — and changing us.
Follow Us:
Newsletter: Recode Daily
Twitter: @Recode and @voxdotcom
Learn more about your ad choices. Visit podcastchoices.com/adchoices
Transcript 0:00 Speaker 00 Real innovation is powered by real people. This is AMEX Corporate. Trusted experts, smarter technology, 24/7 dedicated support from day one to answer questions, help set up policies and approvals, and consult on spend capacity as your business needs evolve. Plus, on-demand resources help level up operations at every stage. That's the trusted service you've come to expect. Built to simplify, designed to scale. This is AMEX Corporate.
0:34 Speaker 01 You could try to find the perfect coffee to water ratio. Or you could press a button. With Keurig, each K-cup pod is precision measured for a perfectly delicious cup. No effort, no grind, no math required.
0:51 Speaker 01 Great coffee without the grind. Keurig.
0:57 Speaker 02 Hi, I'm Kara Swisher, editor-at-large of Recode. You may know me as someone who wants to be fluent in every language, but only swear words. But in my spare time, I'm a technology journalist, and you're listening to Recode Decode, a podcast about tech, politics, and media. We're part of the Vox Media Podcast Network. Today in the red chair is Matt Hewlett, the president of education company Rosetta Stone. After he got there in 2017, the company stopped selling CDs for learning a new language, and now focuses on reaching students on their phones. One of his previous jobs was as president of the Corporate Travel Division of Expedia under Barry Diller, who has been on this podcast many times before, so has Derek Kostroschahi. So we know lots of people in common. Welcome to Recode Decode, Matt.
1:39 Speaker 03 Very excited to be here, thank you.
1:40 Speaker 02 So I want to hear about your background, because I have a feeling it's rather interesting. How you got to doing Rosetta Stone is obviously a very storied brand for language, essentially. Initially was tapes, tapes, tapes, and then it was CDs. Maybe it wasn't tapes, but I called it CDs.
1:55 Speaker 03 CDs, yeah.
1:56 Speaker 02 Before that, I used tapes when I was learning German and stuff like that. And then, because I'm super old, and now it's sort of an online web focused, and then online.
2:05 Speaker 03 Yeah.
2:06 Speaker 02 App-focused. So, talk a little bit about your journey of getting here, because I find this area so interesting. I've been sort of touching on a lot of topics like ag tech or education tech that have been so profoundly affected by digital.
2:16 Speaker 03 Mm-hmm.
2:17 Speaker 02 And obviously, language is one of them, and it's an ongoing thing at all the search services to try to have translations that are perfect. And so, there's a lot of really interesting futuristic stuff there. But let's talk about your journey, how you got here.
2:29 Speaker 03 Yeah. I've always been in Washington State. I've always done technology. I've been professionally permissuous in terms of what I focused on.
2:36 Speaker 02 I like you already.
2:37 Speaker 03 I've jumped around between every software vertical imaginable: travel, games, marketplaces, business models like SaaS and microtransactions and stuff like that, freemium. I actually started out at a small software tech company in Seattle, got fascinated with the Internet.
2:51 Speaker 02 Which one was that?
2:52 Speaker 03 It was called WRQ. You've never heard of it. It sounds like a law firm.
2:55 Speaker 02 I've heard of a lot of those Seattle ones. I was up there a lot with Microsoft and Amazon in the beginning days.
2:59 Speaker 03 And then I got fascinated with the Internet when I downloaded the NCSA Mosaic browser.
3:05 Speaker 02 Sure.
3:05 Speaker 03 I think I compiled it on a Sun OS box. I got into Unix and X Windows at the time. Thought, wow, this is going to change the world, because I was using Telnet and VI, and I was somewhat techie, mildly techie, mildly geeky. And then I wanted to get involved with it, so I looked around for innovative Seattle companies that were doing things on the Internet, and I found one called Progressive Networks. Sent a letter, actually, to the CEO, Rob Glazer.
3:29 Speaker 02 Yep, Real Networks.
3:30 Speaker 03 Yep.
3:30 Speaker 02 He had just left Microsoft.
3:32 Speaker 03 He had just left Microsoft. I was employee 50. I was one of the first product managers there. I was the product manager for the Real Player.
3:37 Speaker 02 I'm so sorry. No kidding.
3:39 Speaker 03 It was a blast.
3:40 Speaker 02 Rob's a character.
3:41 Speaker 03 He's a character.
3:42 Speaker 02 The lead of a profile I wrote for the Wall Street Journal way back when, when he started Progressive, was, "He has a face like a fist."
3:48 Speaker 03 Hmm.
3:48 Speaker 02 Yeah, I thought it worked well. He liked it, of course.
3:51 Speaker 03 Yeah, he's one of the brightest folks, technically and business-wise, I've ever worked with.
3:56 Speaker 02 Explain Progressive. Progressive became Real Networks, but it was the first Player, essentially.
4:01 Speaker 03 Yeah, it really was. You know, at the time, you couldn't really do anything over the Internet because it was all dial-up, 288. And then the big innovation was 56K with the US robotics modem. This is dating myself, which is, I think, illegal in Washington State. And the only way you could actually enjoy anything was streaming at the time. So Real Networks, then Progressive Networks, invented audio streaming. And progressively, pun intended, got better and better at it, higher quality. Started with broadcasting voice, then we got into music. And then as bandwidth started opening up, we got into video and other formats.
4:37 Speaker 02 Right, I remember he chose Jewel because she didn't move. One of the first, or something like that, one of the first video broadcasts. Because she moved very little.
4:46 Speaker 03 Yeah, back in the day...
4:47 Speaker 02 There was no activity with Jewel.
4:48 Speaker 03 You had, that's true. Also, we had a challenge when we had Spike Lee come in and do a, I remember, boy, I haven't thought about this in years.
4:56 Speaker 02 See?
4:57 Speaker 03 A live...
4:58 Speaker 02 I am your memory chest.
4:59 Speaker 03 You are. Savion Glover, he's a dancer.
5:01 Speaker 02 Yep.
5:02 Speaker 03 Tap dancer, quite talented. And we had a hell of a time streaming him because he was so fast. So we really had to go in and instruct Spike Lee on how to direct someone.
5:10 Speaker 02 Right, right.
5:11 Speaker 03 Which was somewhat awkward.
5:12 Speaker 02 Yeah, I bet, I bet.
5:13 Speaker 03 And so that was early in my career. Saw the RealPlayer download like 300 times a day. And then eventually, by the time I left, post-IPO, it was about 300,000 times a day.
5:23 Speaker 02 Right, right. And it was really interesting. Seattle this time, there was Real Networks, which was a big company, it was down by the water. We had Microsoft, which was sort of still pretending the Internet didn't exist for many, many years.
5:32 Speaker 03 Right.
5:33 Speaker 02 Even though they started MSN badly and most of their Internet efforts were bad, essentially. And then Amazon had its beginnings there. And I remember visiting him when he had five or six people.
5:42 Speaker 03 Yeah.
5:42 Speaker 02 You know, so, but there weren't, and then there were a lot of gaming companies, essentially.
5:45 Speaker 03 That's right. That's it, Big Fish Games and other, I worked for one of those as well. And so I left, actually, Real Networks with a peer of mine who started a company called Atom Films. Which was the idea of...
5:58 Speaker 02 Another big, early idea.
5:59 Speaker 03 Right, pre-Youtube, again, dating myself. A gentleman named Mika Solme and I worked together at Real Networks. He was the coolest guy I'll never be. You know, six foot four, flowing blond hair. I haven't seen him in a while, but he's always cool. If he's listening to this, he's gonna roll his eyes.
6:16 Speaker 02 That's true.
6:16 Speaker 03 And I joined him because we believed that short-form content was gonna be big on the Internet. We started licensing content. We eventually raised a bunch of money and we merged with then Shockwave and the geniuses that we were at the time combined the companies and called it Atom Shockwave.
6:35 Speaker 02 Right.
6:36 Speaker 03 And I think we were the last significant round raised in San Francisco at the time in the Bay Area. Sequoia Capital, Mike Moritz led it. It was around $20 million. And that was right as he sent that ominous slide deck out.
6:49 Speaker 02 Yeah, it's over. The death, the RIP.
6:51 Speaker 03 It's over. The tombstone, yeah, RIP.
6:53 Speaker 02 Yeah, but it was interesting because a lot of these ideas were super early and super correct, just wasn't the right company.
6:58 Speaker 03 Time, yeah.
6:59 Speaker 02 Time, absolutely. There weren't mobile phones. There weren't mobile phones changed and everything, obviously.
7:02 Speaker 03 That's right.
7:03 Speaker 02 But a lot of this stuff was, you know, bandwidth and bandwidth, which I don't care, you don't even say that word anymore, but an ability to access the internet and the device was critically important.
7:14 Speaker 03 Absolutely. And we were trying everything. We licensed content to HBO, to airplanes, anything with a screen. But eventually that, I left, went back to Seattle. Eventually that company was bought by Viacom. One of the board members of Atom Films at the time was Rich Barton. He hired me at Expedia. Rich, at that time CEO of Expedia, now CEO of Zillow.
7:34 Speaker 02 Right.
7:35 Speaker 03 And everything else, Glassdoor, AVO, anything with a directory. Rich hired me to Expedia and I started their corporate travel group. And then since then, I've been doing both startups and turnarounds.
7:45 Speaker 02 Right. And so you worked at, Expedia was bought by Microsoft, right? Who bought it?
7:50 Speaker 03 Spun out of Microsoft.
7:51 Speaker 02 Spun out of Microsoft, then bought by IAC.
7:53 Speaker 03 Yeah, standalone company for a while, then IAC a little bit, then all of it, and then not as much.
7:59 Speaker 02 And then, but it's been affiliated with IAC since then.
8:01 Speaker 03 Absolutely.
8:02 Speaker 02 And again, it was, that was one of the successful things pulled out of mine. There were several like that within Microsoft that spilled out. And so how did you get your Rosetta Stone? Like, you've got sort of, like, you're a typical journeyer of the internet companies.
8:15 Speaker 03 Yeah, yeah. A friend of mine is a recruiter called me. What actually got me most interested in the company was my personal story around literacy. I'm dyslexic. I have three kids. My youngest child is dyslexic. And most people don't know this, when they hear Rosetta Stone is, growing part of our revenue is literacy software. And that really spoke to me. And so that's really what got me excited.
8:38 Speaker 02 Alright, so to paint a picture today of the company, the company started, how many years? It was...
8:42 Speaker 03 A long time.
8:43 Speaker 02 Yeah.
8:43 Speaker 03 Twenty-seven years ago?
8:44 Speaker 02 Yeah, it's a long time, yeah. And it started with essentially, you, you used to learn languages was you went to class, essentially, or a night class if you're an adult or anything else. And then you would learn, and you would use flashcards and everything else, but you would not, and then they moved into CDs and tapes and then CDs where you listen to people speaking, essentially.
9:01 Speaker 03 Yeah. The company history is fascinating. The founder was unsuccessful at learning a language. Had learned German by being immersed.
9:11 Speaker 02 Yes. Yes. It should be nine, but it's not.
9:13 Speaker 03 Had a hell of a time with Russian. Decided that, twenty-seven years ago.
9:18 Speaker 02 Okay.
9:18 Speaker 03 Actually been a public company for eleven years. And the company was started by a gentleman who actually had difficulty learning a language, but had previously learned German because he was immersed in it. So he tried to replicate that experience in software.
9:31 Speaker 02 So he lived in Germany?
9:32 Speaker 03 He lived in Germany.
9:32 Speaker 02 Yeah, I tried to speak German too, and I'm terrible at languages, but go ahead.
9:36 Speaker 03 Yeah, and so, well, it worked for him in terms of immersion, and so he decided to build software around it. And so, CDs came out, we became known for the iconic yellow box, the CDs, airplane, you know, kiosks.
9:47 Speaker 02 Right.
9:48 Speaker 03 And built a huge brand around language learning.
9:50 Speaker 02 Right. And so this was going along, CDs, obviously, everyone had first CDs at home and then in their computers, and then in Walkman and things like that. But that changed really drastically, like your, people didn't learn this way anymore.
10:03 Speaker 03 Yeah, and I think, like most companies, the arc of Rosetta Stone has changed significantly over the last five years. We bought several companies years ago, one of them was Lexia, which is really interesting company that focused on literacy in USK 12 schools. And also, the arc of the consumer business with Rosetta Stone took a different turn. They were later to innovate in terms of digital technologies than I think they would have liked. They focused on the Lexia software, and our last CD was sold in December of 2017, right as I was getting into the company.
10:34 Speaker 02 Right. And so, why was that? Because there were so many, like, you can think of a million companies like this, and there's others that catch up really quickly. Facebook, others, there's lots of people that shift and turn. But this is something that, you know, the era of the CD was a very brief one. It was a relatively brief one.
10:49 Speaker 03 Yeah.
10:50 Speaker 02 In terms of media.
10:51 Speaker 03 Mm-hmm.
10:52 Speaker 02 So what was the problem?
10:53 Speaker 03 Yeah, it's always hard to think through, you know, because I wasn't there, the playbook as to what got them there. But I do think it's very difficult to cannibalize yourself as a company. You think about like, which company has been successful turning themselves around from business models? Like Adobe comes to mind, moving to a subscription model. Apple obviously did a major turnaround, but not many brands have been able to do that. And I think Rosetta Stone is one of those brands that, one, is well-known. It's well-liked. It almost has a nostalgia to it. My biggest competition is actually the fact that everyone knows who I am. They just don't know I have an app.
11:26 Speaker 02 Right, right, right.
11:27 Speaker 03 And I think that a lot of companies get stuck in this rut of lack of innovation. Because it's very hard to move from a thousand-dollar CD to a product that's $9.99 a month. And I think those are hard decisions to make, and most companies don't survive that. So we went through that. I was hired to bring on a team and a product and really make this product really mobile-native, which we've started to do.
11:48 Speaker 02 So started to do. The mobile's been around since, what, 2000?
11:52 Speaker 03 Quite a while, yeah.
11:54 Speaker 02 Ten years, I'd say.
11:55 Speaker 03 And so, you know, we're obviously playing a little bit of catch-up. And the space is about a $50 billion language learning space. About 6% of that's digital. There's a lot of competition in the space. But we're starting to make some strides there.
12:08 Speaker 02 So when you're thinking about when you come into a company, you've been working at all these various companies, some of which, you know, Real Network got caught up in that too. Like everything, you know, things became in the phone. You can think of a dozen companies like that where that happened. Why did you decide to come here, to the company? What was your thinking on where it could go?
12:25 Speaker 03 Yeah, I usually am handed issues and problems that I almost feel like MacGyver at times. You know, like, a lot of times...
12:33 Speaker 02 Let's reference that for the young.
12:34 Speaker 03 Yes.
12:34 Speaker 02 MacGyver could do anything with a rubber band.
12:36 Speaker 03 That's right.
12:37 Speaker 02 And not dirty.
12:38 Speaker 03 Bailing wire and Elmer's glue.
12:39 Speaker 02 Right.
12:40 Speaker 03 But in all seriousness, I've been in situations before where a company's about ready to run out of money, investors aren't happy. You know, those are stressful situations. I haven't been in a situation where there's a world-class brand. And customers want the product. They like the current digital product. It just needs some restructuring and some innovation. That's a different type of problem.
13:00 Speaker 02 Right. How much of your business is still CDs and things like that?
13:03 Speaker 03 Zero.
13:03 Speaker 02 Zero. You don't sell any. Your last one was 2017.
13:06 Speaker 03 Yeah, and basically how our revenue works is-
13:09 Speaker 02 And no retail element, like you had those-
13:11 Speaker 03 We have a little bit in Amazon, but you know, we sell through Amazon, but no CDs at all. About half our business is selling to USK12 schools.
13:20 Speaker 02 Mm-hmm. Which is dyslexia, which is literacy.
13:23 Speaker 03 Which is literacy. And the other half is consumer and enterprise. And so those components are what I'm restructuring right now. The Lexia business is growing pretty nicely, 20% year over year.
13:34 Speaker 02 All right. What is that? That's to allow kids to... English as a foreign language, or what's the...
13:40 Speaker 03 No, it's... Well, you know, again, one of my passions is literacy. Two-thirds of kids are reading below grade level by the time they get to grade three. And it's... This is nothing new, and that trend continues through 8th grade. Unfortunately, it's been like that for 30 years. I'm sure you've had...
13:57 Speaker 02 Yes. And there's been a lot of startups where people tried. There were a whole bunch... A lot of ex-AOL people tried a whole bunch of literacy efforts. There's been a lot.
14:04 Speaker 03 There's been a lot.
14:05 Speaker 02 And then much of it charitable, I would say, as businesses.
14:08 Speaker 03 That's right, and I think the-
14:10 Speaker 02 Jim Barstool was another person.
14:11 Speaker 03 Yep, the Barstool Institute. There's been a lot of tech investment in it. This product's interesting because we probably have, quantitatively, more PhDs that know the science of reading than most companies have in terms of software developers. And there's like 17 evidence-based academic research papers that show that this software works. It's software that's worked in conjunction with a teacher in the classroom. It's not a self-serve product. It's not what you typically have read about, like the flip model where the student does all the work and the efficacy is a little bit lacking. This is a product that's used within the environment of the classroom. The teacher gets personalized-
14:48 Speaker 02 So you're using digital means to help people with literacy?
14:50 Speaker 03 Yeah, about a third of the time, the kids' allocated time in the class is using the product. That gets data, keystrokes, results back to the teacher. They can personalize what the issues are with the student and provide very personalized training plans for the student. There are six elements to the science of reading. We need to be teaching that across the board in the United States. I learned about those six elements. I have one called phonemic awareness. I'm dyslexic. The others are phonics, structural analysis, comprehension, vocabulary.
15:18 Speaker 03 And automaticity, which is sometimes lacking for me called fluency. And that's built into the product and it's used in conjunction with schools. We're in about 10% of schools.
15:28 Speaker 02 Okay, we're here with Matt Hewlett. He's the president of Rosetta Stone, which is well known as a language company. When we go back, we're going to talk about the landscape and where language learning is going.
15:41 Speaker 04 It's me, the
15:44 Speaker 04 We don't need the jingle twice.
15:47 Speaker 04 Stop it. This is about vacation inflation and how Priceline negotiates amazing deals on hotels, flights and rental cars.
15:55 Speaker 04 Yeah, but I didn't mention that you can save up to 60% off hotels in the Priceline app. Fine, no one deals more deals than Please stop. Touche. Priceline.
16:09 Speaker 02 We're here with Matt Hewlett. He's the president of Rosetta Stone, which is, you know, pretty well known for having sold CDs and teaching people to do languages. And now he's been focusing it on their phones. The landscape is pretty competitive. There's like lots of little apps. I was just looking at them all last night. There's zillions of them. Duolingo is probably the biggest online. I don't know if that's your biggest competitor, but map the landscape of where language learning is going.
16:33 Speaker 03 Yeah, it's, we mentioned before, $50 billion worldwide. 80% of that's schools and offline tutoring. Like we talked about before, it's schools and tutoring is still a big part of that. Digital is smaller. It's single digits, but growing quickly, not unlike a dynamic you've seen in other markets. In the United States, we're still number one, you know, in terms of overall dollars. All of my competitors are private, so I have no idea what their real revenue is. I would say in terms of share, you could definitely, you know, look at Duolingo and say they've done a fantastic job. They've built a very engaging product, a very addictive product. We like to think of ourselves as a product that actually gets efficacy and results. We built speech recognition of the technology to get people producing the
17:17 Speaker 03 the actual language, so you become more proficient, sound like a native speaker. But that landscape has changed significantly over time. It's interesting that as players like Duolingo have entered the market, we've actually seen our brand awareness go up. And one of the things about language learning that I'm excited about is it's still growing at pretty strong clips. There's people like Duolingo, there's Babbel, those are some of the larger players than ourselves. And there's a ton of other people in this space. And I find this market, as being relatively new to it, relatively disloyal. So we find that like 56% of our customers have already used another product. So we think there's some opportunity there to go deeper in the stack to get different types of experiences for customers.
18:00 Speaker 02 All right, let's talk a little bit about the science of what's going on now, because one of the, you know, I was playing around with Europe, it's very different than before. Before, it's essentially a lot of book learning and listening, essentially. What's changed in language, from your perspective? Like, obviously, delivery method over mobile is one.
18:18 Speaker 03 Yeah, I think language learning, if you break down, and I'm not an academic in language learning, and they roll their eyes, the people on my team, when they hear me speak about it, but it's really about usage, frequency, and time. And we have a very specific pedagogy called immersion. So you're, we're not teaching you anything. Procedurally, you're, we don't conjugate things for you, you're just learning as you're seeing objects and video and pictures. What's changed over time is that, thanks to the 3 billion worldwide market of smartphones, everyone's ready to have digestible content. And you can chunk out the lessons to make them more approachable versus feeling like you're behind your desktop and you're going to allocate, you know, a couple hours a day, which no one wants to do.
19:03 Speaker 03 And that's what fundamentally changes, the bite-size aspect of the learning. Also, you can take advantage of cool things that make it much more personalized. You know, we grab through ARKit in the Apple iPhone environment, different images that you can collect, like Pokemon Go for learning, so you can actually personalize objects around you and incorporate that in a conversation. Those are things and experiences you haven't been able to do in the past.
19:24 Speaker 02 Right. So when you think about the, but in terms of how people learn, obviously, people are twitchier than ever, they're doing, since you're talking about, when you talk about digestible, it's not really the way language has been taught in the past. It's been in a very, you know, long-term classroom.
19:42 Speaker 02 Oriented where effort is required rather than instant learning.
19:45 Speaker 03 That's right. I think about things as US versus everyone else. US, it's like saying you're in a rock band in Seattle, everyone says they're in a rock band. They're not really in a rock band.
19:56 Speaker 02 Right.
19:57 Speaker 03 Meaning world languages is kind of a nice to have, not a must have. So when you're going to, let's say, Mexico, everyone wants to learn Spanish, everyone decides to spend the prerequisite amount of time, two weeks. Takes a little bit longer than two weeks to learn a language. Outside the United States, the hunger element is way high. So there's about a billion and a half people around the world that are wanting, that are learning English right now. There is no issue around hunger because they know they have to get to the next tranche of educational academic achievement and/or the next job. And so the hunger is already there and the appetite to learn is very high. So the two different learners types in the United States are casual learners that really don't need to learn
20:41 Speaker 03 a language. They want to do it for experiential reasons and outside the United States where they have to do it for other reasons. The digestible component makes it much more in terms of a form factor, approachable, more economic. And for example, Asia, there's a trade off between, especially for parents and their kids, learning a language and food. They're more willing to spend education, money on education than food. And so I would say that it varies by where you're wanting to learn a language. The fact that you have technology and basically enables anyone with a smartphone to have access to teachers everywhere. In fact, there's a big language learning trend to educate folks using American teachers to learn English. So it's opening up a whole new world. And I think that the
21:26 Speaker 03 digestible component and the communications aspect of it is completely changed.
21:29 Speaker 02 Let's talk about the technology elements of it right now. Obviously, again, it was rote learning, I think. I don't know what, I'm not a teacher, but it was essentially repetitive. What's changed in terms of how language is taught? And people's brains are starting to rewire themselves, I think, in the digital age. We'll see that in 20 years' time, but we won't know that until...
21:48 Speaker 03 It's becoming more personalized. And it's also becoming personalized because of machine learning and LP.
21:56 Speaker 03 Speech recognition is also changing, so we have our own speech recognition technology. And so getting all these inputs from real, live input or from what you're doing is actually catering very personalized experiences for learning. That's the fundamental thing that's changed. Some people learn differently than others, most people do. And the technology is able to pick up, whether it's our literacy products or language, and I think other companies are doing this, very specific ways and pathways that are curated to basically map how you want to learn. And I think that's what's changed fundamentally versus being in a class. I was just in South Korean, South Korean Hong Kong two weeks ago. I was amazed to still see offline classes, they call them cram schools. Where you see kids in a class, literally just kind of cramming for tests.
22:39 Speaker 02 For tests.
22:39 Speaker 03 Not understanding what they're actually learning, understand the vocabulary, but not truly comprehending what they're doing.
22:44 Speaker 02 Does that mean, like, you know, language lessons have always, again, been analog. Does that mean it's over for that, in that regard? That people shouldn't, this is the kind of thing you can do utterly? One of the things I always say is everything can be digitized, will be digitized. And language seems perfect for that. Because you'll have a personalized, it reacts to your voice, there's voice recognition, there's an ability with AI to learn even more about your experience while you're doing it and to correct you. Is that over for the analog business?
23:13 Speaker 03 Yeah, I do. I do think that's going to happen. I mean, it's, it's, if you think about the biggest sectors that haven't really been digitized yet, healthcare is making a pretty good progress. And since 70s, no one would actually invest in healthcare. Now it's actually
23:28 Speaker 02 Right.
23:29 Speaker 03 A very investable category. Education. And there's, education's like a $6 trillion category. Very small digital penetration. So I think that's one of the last great sectors that's going to be digitized. I think the language aspect of that most certainly is going to be innovated through digital means, absolutely. And that's happening. It's happening slower than I would have guessed, being relatively new to this space.
23:50 Speaker 02 And when you think about education in general, you can think how courses could, they're just, you're just somewhere else, you're just not, it's analog but not. This is a very different thing. Do you apply, so when you're applying AI, what are the uses of that when you're doing language learning?
24:05 Speaker 03 For us, it's about speech production and sounding like a native speaker. And again, the use case is changed by where you're learning. For us, what we do is digitize the waveform of when you're speaking into a speech model that gives you basically a histogram of how you're speaking against kind of other native speakers.
24:25 Speaker 02 Right.
24:26 Speaker 03 And we'll give you kind of very basic audio cues on how you're doing against that. I've argued that I'd actually like more data. That's just me. But we give those speakers visual cues as to how they're sounding against the target language they're learning. That speech model just gets better and better over time as we get more people in the product.
24:45 Speaker 02 Right.
24:46 Speaker 03 The component of that's super important. Outside the United States, especially in Asia, where I was spending a lot of time, speaking with confidence, speaking English in particular, is a major hangup. One of the gentlemen that is building a product for me, he spends an hour and a half working on small talk before he goes to a party. And just because of the speech component of this. And I think the machine learning aspect of that is going to be even better over time. I don't think you're going to have an AI-only product. I think you're going to have AI mixed with some HI, some human intelligence.
25:19 Speaker 02 Right.
25:19 Speaker 03 But yeah, I think, you know, if we can actually build products, all of us, I don't care which product you use.
25:26 Speaker 02 Right.
25:27 Speaker 03 That really develops and goes after the underserved, it's only going to be better for the planet.
25:31 Speaker 02 Mm-hmm. And then you use them, you use people's voices, what they're saying as they learn a language, correct?
25:39 Speaker 03 That's right. And we map that against our speech modules. We also build in tutoring in some of our products as well.
25:44 Speaker 02 With real people?
25:44 Speaker 03 With real people. We've been doing tutoring, I think we're the first ones since 2009 introducing tutoring in the software.
25:51 Speaker 02 Right. Versus anything else. Are there statistics on how people learn these things and retain them versus an analog teacher?
25:58 Speaker 03 Yeah, there's a bunch of studies out there. The net of the studies is they learn faster. If it's more personalized and it's more pervasive and more addictive, you're going to learn faster. Absolutely.
26:07 Speaker 02 So right now, the cost of this right now is on a recurring revenue basis?
26:11 Speaker 03 That's right.
26:12 Speaker 02 That people use it. Do people stop using it? Is it like a dating app? I'm trying to like...
26:16 Speaker 03 Trying to get a sense of it, yeah.
26:17 Speaker 02 Yeah, like what's the business line?
26:18 Speaker 03 Yeah, it depends on where you're coming from. If you're in Asia, it's, for example, we're testing a product in South Korea right now, which is fascinating to look at the data. We're introducing tutoring and content. Voracious appetite for consumption of that product.
26:33 Speaker 02 Because?
26:34 Speaker 03 Because they need to understand how to, you know, they want a quarter office. They want to get into the next school.
26:38 Speaker 02 They have to speak English.
26:39 Speaker 03 They have to speak English and pass a bunch of tests like the TOEFL tests. A lot of employers actually require some level of English proficiency. We have a corporate business where we sell software to Accenture and Vodafone and others. It's absolutely important to get it, to get a better job, to speak better English. But in general, I think most US learners are a little bit lazier than most. Not that, well, we could say that about other things, but they're a little bit lazier in terms of their conviction around the product. In general, you'd have to spend about 25, 26 hours in our product to get to the next level of proficiency. And that's more than what most people would want to commit to. But if you commit the time and the usage, you can get there.
27:24 Speaker 02 All right, when we get back, we're talking to Matt Hewlett, the president of Rosetta Stone. When we get back, we're going to talk about where technology is going, including Neural Links. Why even bother studying that when we get back after this?
28:06 Speaker 02 We're here with Matt Hewlett, the president of Rosetta Stone. Rosetta Stone is obviously a well-known language learning company, but you, half your, half your business is literacy or more?
28:16 Speaker 03 About a third is literacy.
28:17 Speaker 02 Literacy.
28:18 Speaker 03 About half USK 12.
28:19 Speaker 02 Let me ask you, what is the most popular language right now?
28:22 Speaker 03 English.
28:22 Speaker 02 English.
28:23 Speaker 03 Absolutely.
28:23 Speaker 02 English.
28:24 Speaker 03 Yeah.
28:24 Speaker 02 Why doesn't anyone just speak English? No, I'm teasing. Don't I sound like Donald Trump? What other languages are popular now? Spanish?
28:31 Speaker 03 Spanish, yeah. The population, the EL learners in terms of trends. English as a second language is one of the fastest growing cohorts.
28:40 Speaker 02 Right.
28:40 Speaker 03 So it's about 10% of the population now, should be about 25% by the next three years.
28:45 Speaker 02 Has English become the world language? I know everyone will go crazy at me saying this, but it seems like that's the trend.
28:51 Speaker 03 Yeah, you know, if you break down the population, just math, no emotion.
28:54 Speaker 02 Right.
28:55 Speaker 03 Billion and a half people in China, billion and a half people in India, 365 million or so million people in the United States. But the predominant language that's being taught around the world is English.
29:04 Speaker 02 Mm-hmm.
29:05 Speaker 03 Yeah.
29:05 Speaker 02 How did that happen? What is that? What, because, just because?
29:09 Speaker 03 Well, world, you know, hegemonic forces.
29:11 Speaker 02 Yes, of that, yes.
29:12 Speaker 03 Yes, and we can get into that. But there's clear historical reasons why that's happened. But as we've dealt with like large corporate entities, what are you going to standardize your language on?
29:24 Speaker 02 Right.
29:24 Speaker 03 English has become the standard for a variety of different historical reasons. Maybe Mandarin. Mandarin is being taught more. That is becoming more and more popular, but I don't see that.
29:34 Speaker 02 Is that popular on your service? Mandarin, learning people?
29:37 Speaker 03 It's popular, not as popular as English and Spanish.
29:40 Speaker 02 It was for kids for a while.
29:41 Speaker 03 Yeah.
29:42 Speaker 02 My kid took it 14 seconds.
29:44 Speaker 03 Yeah.
29:44 Speaker 02 Not under my pressure, it was his own.
29:46 Speaker 03 Very difficult language.
29:47 Speaker 02 Yeah, he was terrible at it.
29:48 Speaker 03 Yeah, I started Vietnamese.
29:49 Speaker 02 He can barely make Spanish words together.
29:52 Speaker 03 No, it's very difficult. Yeah, I failed at Vietnamese.
29:55 Speaker 02 Vietnamese?
29:56 Speaker 03 It's very difficult.
29:56 Speaker 02 What isn't, would you get, you dump languages, like?
30:00 Speaker 03 Me?
30:00 Speaker 02 Yeah, not you, just the company. Like, okay, we're not teaching this one anymore.
30:04 Speaker 03 Not really.
30:05 Speaker 02 How many languages do you have?
30:06 Speaker 03 26.
30:06 Speaker 02 26.
30:07 Speaker 03 Yeah, and then we teach.
30:08 Speaker 02 What's 26?
30:10 Speaker 03 What's?
30:10 Speaker 02 What's number 26?
30:11 Speaker 03 Gosh, Gaelic, I don't know.
30:13 Speaker 02 Gaelic? You have Gaelic?
30:15 Speaker 03 Somewhere around the, yeah, I mean, we've got so many languages. We've been around forever. We also teach advanced.
30:19 Speaker 02 I get that, but who's learning?
30:21 Speaker 03 We also teach advanced. There are, you'd be surprised.
30:23 Speaker 02 Yeah.
30:23 Speaker 03 We also have done a lot of work with Native American tribes to preserve their language. Pachanga, for example, is a language that we're preserving. We work with a lot of Native American tribes to preserve their language.
30:34 Speaker 02 Well, that seems worthy.
30:35 Speaker 03 Yeah, you know, it's crazy when we bring in an elder to record their speech patterns into our sound recording system. And literally, you're talking to, interacting with one of the last speakers of a language.
30:48 Speaker 02 Right, a live speaker.
30:50 Speaker 03 A live speaker, that language.
30:51 Speaker 02 Because some people speak Sumerian, but like three people essentially.
30:54 Speaker 03 Right, right, or Hivalarian.
30:56 Speaker 02 Yeah, yeah.
30:57 Speaker 03 For fun, but you know, it's poignant, you know, and it's not the. Obviously, there's not a huge population for it, but it is more profound to our speech recognition technology folks when they are preserving a language.
31:08 Speaker 02 Right. So trending languages are, I would imagine, Chinese and English, Spanish.
31:14 Speaker 03 Mm-hmm.
31:14 Speaker 02 What, is French downward trending, German?
31:16 Speaker 03 No, it's, they're kind of like, holding their own.
31:18 Speaker 02 Everyone used to take French here.
31:20 Speaker 03 They're holding their own. I think world languages being taught in United States schools is...
31:25 Speaker 02 Spanish.
31:26 Speaker 03 Is Spanish, but I think the funding for these world languages is definitely drying up.
31:31 Speaker 02 Mm-hmm.
31:32 Speaker 03 I know like in Washington State where I live, I live in Seattle, there is a big focus on bilingual education. Not every state has the same philosophy.
31:41 Speaker 02 No, no. So when you're thinking about, like, because you can get easily dragged into political things. When you make these decisions on what you're teaching, you don't dump languages, right?
31:50 Speaker 03 No, no. I've thought back to my career in tech. I've worked for some very sharp-elbowed people that I admire, but not the easiest people to work for. And I definitely wanted to have a no A-hole policy. And the core mission of the business is to really enable anyone to read, write, and speak with confidence. And it would be like ripping our dead hands off, you know, a speech module to actually dump a language. We wouldn't do it.
32:21 Speaker 02 Wouldn't do it. So you're doing some preservation of languages in terms of collecting people talking about them. When you think about where language is going, everyone thought we'd be talking in ones and Os by now and not speaking languages. But to that point, with the Google Translate, Microsoft was working on translation. I know such an Adela debuted one of ours. They're still glitchy, but it seems like the trend line is why even have learn languages when you could speak it? It would be translated. I mean, that was the dream, right? That we don't need to, we speak whatever and we're have, we have a universal translator. I remember Sergey Brin talking about that extensively, just on his own. The idea that you say one thing and then... It comes out a different thing.
33:03 Speaker 03 Yeah, you know, like in Marvel Comics. You know, they have the universal decoder.
33:09 Speaker 02 Yes, of course, yeah.
33:09 Speaker 03 You know, Captain Marvel, she had one built into her wrist. Part of it is, for me, kind of comes down to basic. What kind of populace do we want to have? Do we want to have a population where we outsource our learning, our encoding and decoding to AI and computer software and science?
33:27 Speaker 02 Why, yes, we do, but go ahead.
33:28 Speaker 03 I think some aspects will be automated. I think some won't. Not to get too metaphysical about it. I do worry that if you outsource too much of your cerebral cortex, there's not going to be much left of us. If we're just the limbic system, we're pleasure zones and motivation zones. I do worry a little bit about that.
33:46 Speaker 02 And universal basic income. It's all the Andrew Yang dream.
33:49 Speaker 03 Yep, no, I listened to that episode.
33:51 Speaker 02 Thinking of it, why even bother to learn certain things if you can? I mean, I don't mean to get all matrix-y on you.
33:56 Speaker 03 Sure, yeah, yeah.
33:57 Speaker 02 You know, Elon Musk has talked about the Neuralink. If you can actually, you don't even need a neuralink. He wants to embed a chip in your brain. And it's like on the Matrix. You suddenly can do karate or you can suddenly know 10 languages and things like that. So that's one way to going, is that we chip our brains or we make our brains smarter in some fashion, using digital means. The other is just have a translator. Like, you would speak and then it would come out. Like, it seems like that is more inevitable than anything else.
34:27 Speaker 03 Yeah, I think there's, if you think about all the things that are going to be automated, to your point, it'd be ludicrous for me to say that's not going to be automated over time. I think that most of the translation software is pretty darn good. When it started out like four or five years ago, you know, you get your face slapped in a restaurant.
34:42 Speaker 02 Sure.
34:43 Speaker 03 Based on what the output. Now it's pretty darn good. The subjectiveness of language, though, is still not being captured in the software. Subjectiveness meaning when you know how to, the social, cultural norms of that language. Isn't necessarily a one-to-one mapping right now.
35:00 Speaker 02 Mm-hmm.
35:01 Speaker 03 Language is very subjective.
35:03 Speaker 02 Right, sure.
35:03 Speaker 03 Machine learning is very-
35:04 Speaker 02 Like idioms and everything.
35:05 Speaker 03 Yeah, exactly, like, you know, when I'm in the Netherlands and they say a monkey's going up my sleeve, you know, that's a hard one to kind of interpret. And so I think that there will be a role of learning a language.
35:16 Speaker 02 What does that mean? I'm sorry, I'm completely out.
35:18 Speaker 03 I have no idea what that means.
35:19 Speaker 02 But it's-
35:19 Speaker 03 When they say it, I think it's like, are you sneaking one by me?
35:23 Speaker 02 Oh, I see.
35:24 Speaker 03 You know, I have no idea, like, there's a monkey going up my sleeve.
35:27 Speaker 02 Okay, okay.
35:27 Speaker 03 But that's, you know, a very subjective comment that, you know, will AI figure that out?
35:33 Speaker 02 Mm-hmm.
35:33 Speaker 03 Probably over time.
35:35 Speaker 02 Right.
35:35 Speaker 03 You know, there's been some interesting machine translation things like the GPT-2.
35:40 Speaker 02 Explain that.
35:41 Speaker 03 That's the text extraction that OpenAI introduced that made pretty darn good poems.
35:48 Speaker 02 Mm-hmm.
35:48 Speaker 03 And you put the data set, there's a huge data set called the internet, you extract it and try to make some reasonably good poems.
35:55 Speaker 02 Mm-hmm.
35:55 Speaker 03 AI2, Paul Allen's Institute in Seattle, did a product called Grover
36:00 Speaker 02 Mm-hmm.
36:01 Speaker 03 that actually did a pretty reasonable job at replicating fake news.
36:04 Speaker 02 Mm-hmm.
36:04 Speaker 03 They actually built it to identify fake news. Those types of things are just gonna get better over time. So, the long-term answer to your question is, will you need to learn a language? Yes, but in the long, long, long term, probably not.
36:19 Speaker 02 Right.
36:19 Speaker 03 And then you have to decide what aspects of your brain do you want to outsource and what's important to you as a human.
36:25 Speaker 02 So, when you're thinking that as a, I mean, obviously, you're not gonna be running this company at that moment, but what, or maybe you have a plan to live for 150 years, but what do you-
36:33 Speaker 03 I'm intermittent fasting for the last two decades.
36:35 Speaker 02 Oh, no, you really?
36:35 Speaker 03 No, I'm just, I'm, sorry, I-
36:39 Speaker 02 I am so tired of talking about intermittent fasting with tech dudes.
36:42 Speaker 03 I know.
36:42 Speaker 02 Literally, I sit down to a dinner and they start yammering at my ear and I'm like, I have no interest, pass me that roll over there, which I'm sure you want to eat right now. So, when you think about running a business, like, look, Rosetta Stone almost got run over by not, by being in CDs, obviously Netflix shifted, like, any business. How do you run a business that is based on information that is so digitizable?
37:01 Speaker 03 Mm-hmm.
37:02 Speaker 02 How do you run it, when, you know, Google's working over here, Untranslate, Microsoft is doing stuff, there's all kinds of efforts and so, how do you think about the business going forward?
37:11 Speaker 03 Yeah.
37:11 Speaker 02 Five, 10, 15 years?
37:13 Speaker 03 Yeah, you know, it's a little bit like how, I hate to say it, Uber's thinking about their business. You know, driverless cars and there's going to be other things that happen. I think for us, we think about AI and HI together, the blended learning aspects of learning. Eighty percent of the world's digital language learning, or sorry, overall language learning is still offline. Gosh, you know, we have many, many years until that actually becomes something that's digitizable. But I still think there's going to be a human element, component to it. There's so many aspects to language learning, reading, everything else that requires a human tutor. I don't believe that we will have complete AI learning for students.
37:51 Speaker 02 For students.
37:52 Speaker 03 And I definitely see a role for us integrating personalized learning with the human being. Because I happen to like human beings. So as us as a company, as us as a company, that's kind of what we focus on is how do you marry the learner, the teacher, and the data in the middle? Whether it's speech recognition, whether it's machine learning, supervised or unsupervised, how do you make that work together really, really well?
38:16 Speaker 02 All right, do I finish up talking about, you know, I have your app I'm going to start using. I've been wanting to learn Spanish for years and I was a terrible Spanish student. I was a terrible German student. I was a terrible French student. And I have a feeling I'm just a terrible language student. And there's lots of reasons for that. Different people's brains are wired differently. Mathematical people, or my brother speaks like 16 languages. Whatever that is, eventually, I was hoping there would be a pill someday and then I would just take it. That was my hope and I'm down with Elon Musk's idea of neural net. What are some of the mistakes, given current technologies? What are some of the mistakes people make? Or what's important to do when you're using any of these language apps? What's the critical part? Besides moving to France or something.
38:54 Speaker 03 Frequently.
38:55 Speaker 02 Frequently.
38:56 Speaker 03 It's as simple as that. You have to do it for a little bit at a time, about every day. Five, six days a week.
39:04 Speaker 02 Right.
39:04 Speaker 03 To get to the next level of, there's like six levels of what's called CEVR, the Common European Framework for Learning Languages. Starts at A, A1, A2, B1, B2, C1, C2. If you want to be an advanced level speaker, you know, it's quite a big commitment. We have software that we sell to corporations. It's advanced learner's software that gets you there. But it's just frequency and little bits at a time.
39:26 Speaker 02 And when you think about, sort of when you're leaning into the future, when you think about where language is going, do you imagine a time where you'll have a VR kind of?
39:39 Speaker 02 Element to it, like you were in France, you were in Germany. Because immersion is actually, that's the one time I learned, like when I was in Germany, I lived there and I did speak okay. Like not great, but okay.
39:49 Speaker 03 Yeah, and I get confronted, well, long-winded answer, which I'm known for, is yeah, I think immersive experiences, I think AR is a very interesting one. VR, we'll see about headsets. Augmented reality. My nine-year-old saved up his money and bought a Oculus Quest. And I was totally blown away, by the way.
40:07 Speaker 02 Right.
40:08 Speaker 03 I do think that those are going to be interesting experiences to unlock. We're starting to do a little bit of that with that Seek and Speak feature I mentioned, kind of Pokemon Go for language learning. But how cool would it be that you'd have like AR glasses on, you're interacting with the world, you're getting feedback about how you're speaking, interacting, maybe have like a Find Home feature where like a live tutor can actually provide, you know, a walk-around tour of where you're at. That'd be really cool.
40:30 Speaker 02 Right, but not a VR immersive, like I'm in France.
40:32 Speaker 03 Potentially.
40:33 Speaker 02 Yeah. Are you guys working on that?
40:35 Speaker 03 Not on VR. It's been a headset issue for us, there just hasn't been enough of them yet. It's been a specialty. But I'd say like the technology's getting pretty good right now. We've started playing around with some things. But AR is kind of where we focus.
40:47 Speaker 02 AR, which would be like you're in France, you're looking at the thing, it was speaking to you, saying you should say this here.
40:52 Speaker 03 Yeah, and very, very, even in France, it taught me about the subjectivity of French. It was interesting that, as an example, we had a blog post about the whole new set of more hip sayings in France. So, language is evolving all the time, obviously, but wouldn't that be cool if that was reflected in, you know, a geolocation AR device?
41:14 Speaker 02 Yeah, they just don't like it when you speak French to them. They just, they still don't.
41:18 Speaker 03 They're not so bad.
41:19 Speaker 02 I literally, I thought that. I thought, oh, I'm being the cliche of French, but literally, they're like, stop speaking French immediately. I was like, fine, I'll just speak in English.
41:28 Speaker 03 Public transit is another issue, you know, trying to get around, but...
41:31 Speaker 02 Well, to me, AR would be really interesting, is like when I was in Japan, all those signs are in, very few people speak English in Japan, really. Even in Tokyo, it's really interesting. It was much, many fewer people speak English than you think in certain countries.
41:44 Speaker 03 Yeah, that's true.
41:45 Speaker 02 And if I could have looked up at a sign and it would translate it, it would have been enormously helpful.
41:50 Speaker 03 That's right.
41:51 Speaker 02 It would be really interesting. And then you start to, you start to recognize exit and things like that, as you, you know, it's just an interesting time. All right, what languages do you speak? That's my final question.
42:00 Speaker 03 I'm, I am kind of a...
42:02 Speaker 02 Fluent.
42:03 Speaker 03 Not fluent, I'm learning Spanish.
42:05 Speaker 02 Okay.
42:05 Speaker 03 And I'm actually getting back into using my tutor.
42:08 Speaker 02 Okay.
42:09 Speaker 03 Through our software.
42:10 Speaker 02 Uh huh.
42:10 Speaker 03 But Olga, who's my teacher, my Spanish teacher.
42:13 Speaker 02 Olga?
42:13 Speaker 03 Yeah, and she's listening, I'm committed to coming back to you, Olga. But Spanish is the one that I'm learning right now.
42:20 Speaker 02 All right, and how are you doing?
42:22 Speaker 03 Uh, I've been better.
42:23 Speaker 02 I've been better?
42:24 Speaker 03 First six months, I was pretty darn good.
42:26 Speaker 02 Were you?
42:27 Speaker 03 Yeah.
42:27 Speaker 02 And then?
42:28 Speaker 03 And then, you know, life happens.
42:29 Speaker 02 Yeah, exactly.
42:30 Speaker 03 And so this is kind of me being an American.
42:31 Speaker 02 Right, exactly. All right, thank you, Matt, so much, it's so interesting. Rosetta Stone is fascinating. The whole idea of where language is going to me is gonna be really interesting. But I do think eventually it will all be AR-oriented, like it'll be translated for you, or you will have a chip, one or the other.
42:46 Speaker 03 Just, I hope no one hacks into our heads.
42:48 Speaker 02 And then finally, I can achieve my long-held dream of speaking ancient Aramaic.
42:53 Speaker 03 Gosh, I wish I'd just learned Kung-Fu.
42:56 Speaker 02 Anyway, Matt, thank you so much for coming on the show. You can follow me on Twitter at Kara Swisher. My executive producer, Eric Anderson, is at Eric America. My producer, Eric Johnson, is at HeyHeyESJ. Matt, where can people find you and Rosetta Stone online?
43:08 Speaker 03 Yeah, my Twitter handle, Matt underscore Hewlett.
43:11 Speaker 02 Okay, what about Rosetta Stone, at Rosetta?
43:12 Speaker 03 RosettaStone.com.
43:13 Speaker 02 Okay, great. If you liked this episode, we really appreciate it if you shared it with a friend. And make sure to check out our newest podcast, Reset. Just search for it in your podcasting app of choice. Thanks also to our editor, Joel Rabe. Thank you for listening to this episode of Recode Decode. I'll be back here on Wednesday, tune in then.
Previous Recode Decode: OpenDoor CEO Eric Wu / Evite CEO Victor Cho Next Recode Decode: The inside story of the Cambridge Analytica scandal Page updated October 13, 2026 · transcript and summary are machine-generated