Anunț publicat de Severin Hacker

I'm super excited about Duolingo's second acquisition: Hobbes

People know Duolingo for its personalized lessons, but we use AI in many other places across our products. Our in-house experts spend a lot of time thinking about how AI can support and scale their work so that they can get new content to learners faster than ever before. Here is a non-exhaustive list of where AI enhances the Duolingo experience: - Assembling personalized lessons - Determining when learner’s should review old content - Generating interactive exercises from expert-created raw content - Auto-suggested text in freeform exercises - Generating a range of possible accepted translations - Grading exercises - Creating character voices - Generating DuoRadio scripts - Generating real-time responses in Role Play - Providing context on mistakes with Explain My Answer - Triggering character animations with Rive - The Duolingo English Test question generation and scoring  - Deciding when to send push notifications - A few more exciting features I can’t talk about just yet 🤐

Today we announced that Mario Schlosser is joining the Duolingo board. Mario is the CTO and co-founder of Oscar Health. He has a ton of valuable experience in building consumer products that use the latest, most powerful technologies. He champions AI innovation and his guidance will be invaluable in helping Duolingo achieve its mission to develop the best education in the world and make it universally available. Welcome, Mario!

Last week, we hosted our first-ever Engineering Summit at Duolingo. This was entirely the brainchild of Brian Dewey, who pitched this event to Natalie Glance and me, and then helped bring it to life. The day was organized into different "tracks," and engineers could build a schedule and attend different talks and sessions that fit their interests. This is exactly what it means to "take the long view" on hiring—Brian joined us 5+ years ago, and now he's shaping the engineering experience at Duolingo. He gave an excellent talk on communication for engineers. Here's an edited version:

Duolingo turns 12 today. It's the best place I've ever worked. (It's also the only place I've ever worked.) Here's one of the earliest photos of me at the office. Happy Birthday Duolingo!

Here’s a trick for getting someone on board with your idea: distill it into one word. I’ve learned to use this method for many ideas I’d like to see implemented at Duolingo. I’ve told this story before: Early in our journey, I was convinced we needed to offer a subscription option. I would walk around the office and anytime I saw a colleague I’d simply say: “subscriptions.” I probably said it a thousand times per week. Pretty soon, people knew what I cared about, and wanted to learn more. How does one-word persuasion work? It’s similar to the domino effect that I’ve discussed before. Since I don’t have many direct reports, I have to influence people in other ways. People who don’t work with me often are curious about what I care about. Simplifying that into one word (“subscriptions”) gets the message across quickly. OK, there’s a little more to it. It has to be a sound idea—and I never pitch something that I haven’t researched thoroughly myself. And of course, I’d have to initially explain the concept to others (but keep it simple: one of the big data points I used for subscriptions was that 9 out of the 10 top-grossing education apps were subscription products). The key here: I only did this once. The idea was compelling enough that I could share my thoughts one time, and the rest, I’d just have to say the word. I was like a walking practice reminder. Not every idea or word works for this. For example, “AI” is a tough one-word pitch. It can mean so many things, and the action isn’t clear. It’s also fairly obvious—of course we should invest in the latest technology. What makes a good one-word pitch? 1. It should be surprising. At the time, we weren’t exploring subscriptions, so it wasn’t an obvious next step. A surprising word is the best way to spark interest and conversation. 2. The one word should be able to encapsulate the entire idea. “Subscriptions” = we should offer subscriptions. Enough said. 3. It should be “completable”, with an implicit goal attached. Once we launched a subscription product, I could move onto my next word. But if the word had been “AI,” would we ever really be done? (P.S.: If you’ve read “To Sell Is Human” by Daniel H. Pink, he also talks about the “one word pitch.” It forces clarity, and I strive for simplicity in my work.)

Since day one, “take the long view” has been one of our most important operating principles at Duolingo (and my personal favorite). It’s our secret to hiring great people. When we take the long view, we don’t just focus on tomorrow. We look ahead to what will be good for us ten years from now. We don’t make compromises for short-term gains that may hurt us over the long term. Because we take this mindset with hiring, we’ve managed a very low turnover rate across the company since our founding. Here’s what it looks like in practice: 1. When we hire, we try to fill the current role and we look ahead to future roles. We ask, is this candidate a good fit for both the role we have open and the role we envision for them in four years? Hiring for the long term means thinking about their career trajectory within our organization. 2. It also means we hire more generalists than specialists. If you hire someone to fulfill a very specific need, and that need goes away, there’s no role left for them. If you hire someone extremely smart whose skills apply to a range of different problems, they can grow along with the company. For example, even as we have expanded to new subjects, we haven’t made a single hire in engineering or design specifically for math or music. Our people were ready to shift to the new requirements for these new courses. 3. If you took the short view, you’d never hire new grads. It often takes them a year to ramp up, so you’re basically paying for their onboarding. But they make amazing and loyal employees. We often say a new grad with a year of experience at Duolingo is better than most senior software engineers at any other company. 4. We look for missionaries over mercenaries. The long view means hiring those who believe in your mission—even if it means passing over very talented people who want to sell their skills to whoever will pay the most. When you think long term, it means you’re investing in your people. And when you believe in them, they’ll believe in you. All that said, we’re hiring now. Check out our open roles: https://lnkd.in/ekeD8Jxt

It takes a while for people to figure out exactly how to work with me. As a result, the people I collaborate with best and most often are the ones I’ve worked with the longest. I wanted to change that… so I created an “owner’s manual,” AKA a guide for working with me that could act like onboarding for new team members. It’s nothing fancy, just a Google Doc divided into three sections: how to interact with me, how I think about things, and who I am as a person. Here’s a sample of what you’ll find in each: How to interact with me: - I am not a morning person, if you want the best of me, find me between 3pm - 9pm. - I am good at the first 20% of a project and I’m terrible at the last 20%. How I think about things: - At any point in time I try to have one high-priority project and spend 80-90% of my time on that. - I am motivated by purpose (mission) primarily and autonomy secondarily. - My time is my most valuable resource. Who I am as a person: - I never compromise with sleep and lunch is sacred. I’ve been slowly sending it out to people, and apparently it’s helpful (my newest direct report said so, anyway). I’ve been thinking Luis von Ahn should have one too. If you could ask a senior leader at your company to draft their own owner’s manual, what would you want them to include? *Edited to add: this is a picture from lunch with Tyler Murphy. That’s his burger.