The Duolingo Journey

Chris Luebkeman, leader of the Strategic Foresight Hub in the Office of the President at ETH Zurich, sat down with three young entrepreneurs for a discussion about how ETH students can turn and have turned their ideas into a solid business, and how the options for doing so have changed over the years.

The most renowned participant, Swiss-born Severin Hacker, has returned for the discussion to ETH Zurich, where he received a BS in Computer Science in 2009 before moving on to graduate studies and the development of Duolingo at Carnegie Mellon in Pittsburgh. His journey towards entrepreneurship was different than that of the other panellists.

“My career would not have been possible in Switzerland [at the time] for sure,” Hacker explains. “One of the things many people don’t know is that we made no money in the first five years, but we were always able to raise it – and that was only possible in the US. Nobody would give you the money [in Switzerland].” For Hacker, the journey towards his own business was very hands-on, from his studies to idea development to mentorship from his PhD advisor (and Duolingo co-founder, Luis von Ahn), learning as he went along.

Some 15 years after Hacker worked his way toward successful entrepreneurship in the US, post-graduate students Julia Carpenter and Nicole Kleger, co-founders of the ETH spin-offs Apheros (cooling systems) and sallea (cultivation of cell-based meat and fish) “caught the bug within the lab,” as Kleger agrees. Which means that the inspiration for a start-up can develop from a student’s studies. “I was really lucky to be in a [research] group where we had several start-ups coming out of that group previously. And I think only because of that did I even get the idea that entrepreneurship is a thing,” says Kleger, who wasn’t interested in an academic or industrial career. “I started with a small [start-up] course just to pressure-test a bit. I got to know people in the game. That also shows how important it is to have people around you.”

In this way, both were able to take advantage of ETH start-up courses and the culture of entrepreneurship that already existed within the university.

All three entrepreneurs agree that success comes only with in-depth development and organizational skills.

“You need to have a good idea,” says Hacker. “You need to have a great team, and then you need to have the funding. If any one of these is missing, it doesn’t work.”

Carpenter agrees. “Building that first team … finding the people that want to work in a start-up … that’s really a big decision and it’s something that I’m working on right now.”

It was hard finding those initial team members, says Hacker. “They were first afraid they’d get outsourced to India, so they were hard to recruit.” But von Ahn knew which students would be good for the job, so they picked them out from Carnegie Mellon’s research pool.

While it’s good to “look at students doing master theses to match positions to fill, you also need to look for people with start-up experience. It’s a matter of balance,” explains Kleger.

Questions to the Severin Hacker also focus on the Duolingo concept itself and how it developed since the first public version was released in 2012.

With studies showing that the most effective way to educate someone is through 1:1 lesson with a personal tutor, Hacker and von Anh decided to develop a personal language-learning app. They compensated for the lack of a human teacher by focusing on motivation and turning the learning experience into a game. Which is why Duolingo hands out points and rewards, and gives users the option of sharing streaks with friends, while its green owl mascot quietly cries at you from your inbox to guilt you into coming back if you haven’t opened the app in a while.

On the technical side, Duolingo’s product development is experiment-driven. This basically means that the app learns from the user as they learn the language. The findings are analysed, then adapted for the app and implemented back into the product. In doing so, the developers make sure each language level is neither too hard nor too easy, for example, again to keep the focus on motivation.

We got an example of this in our own household a few years back, when the Italian course we were taking repeatedly made sure we were familiar with the term “pizza”. Which, honestly, has to be one of the most internationally understood words in the world. This glitch (?) eventually stopped, so we can assume we experienced Duolingo’s experiment-driven development process first-hand.

Duolingo currently offers a list of over 40 – and counting – language courses, which include constructed ones like Esperanto, High Valyrian and Klingon.

And so the inevitable question comes from the audience: “Why is there no Swiss German?”

Hacker laughs. “Should I be honest? I don’t want to add it because I want Swiss German to remain that secret language that you can’t learn anywhere.” The local audience applauds loudly.

Watch the full recording:

Find out more about our ETH Global Lecture Series here.