There's something weird happening to Dolingo. It's subtle. At first, you don't notice it.
The lessons are still there. Mistakes, sure, every once in a while, what's new? But after a while, it starts to feel hollow.
Stories don't feel like they used to. The audio lessons have become oddly empty, vacuous. And then you see it.
A big announcement from the CEO. Dualingo is officially an AI first company. Not AI assistant, not AI enhanced, just AI first.
Well, if it's AI first, who would be in second? Humans, assumably. Now, I use AI like all the time.
I know, especially online, you expect people to be in a camp of either all AI bad, evil, cannot use no matter what, no positives, or all AI good, it is the future. But I appreciate a bit of nuance here. But something about the way Dolingo is using AI, the features they're replacing.
It feels less like the bleeding edge of innovation to improve the app and the user experience and more like just bleeding of any originality left. My name is Evan Edar and this week I hit my 9-year streak on Dualingo. It's 3,285 days or something like that.
I've seen the app through ups and downs and for the most part I've very much enjoyed using it. I've updated this YouTube channel many times with different things about Duolingo. It helped me learn German.
It helped me learn Spanish. I'm still learning Spanish with it. But this recent update, which is more to the ethos than even to the app, though we're going more into that later, I think shows where technology and corporations are heading in the future.
And I think it's an important conversation to have. So, let's dive in. Now, I want to talk about the backlash, all the the weird things that I've noticed in the app recently.
But first, you need to know how this all started. Duelingo CEO Luis Vonan recently shared an email he'd sent to the team at Duolingo on his LinkedIn. He was uh clearly very proud of announcing Dolingo is now an AI first company.
What's that even mean? Well, he says AI isn't just a productivity boost. It helps us get closer to our mission to teach well.
We need to create a massive amount of content and doing that manually doesn't scale. Since when? What?
No, you don't. You can't teach well without more content. It can't scale.
What? No, you don't. You don't need that at all.
Actually, you could you could teach well with one book. This is a classic case of bull that's been uh pampered up a little bit for the CEO's LinkedIn. Look guys, I wrote this email.
We just need more content. Seems like he has no faith in his uh human workforce at all. He's like, "Guys, you are not good enough.
" Let's let him continue. He says, "We can't wait until the technology is 100% perfect. We'd rather move with urgency and take occasional small hits on quality than move slowly and miss the moment.
" What moment, Lewis? the Louie the moment of uh all the other companies I guess doing the same thing of gutting society. Yeah, what a moment to miss.
It's such a shame. In his words here, to teach well, they need to lower their quality. I mean, I I'm not taking him out of context.
He's literally just said that in his little LinkedIn post he's so proud of. We just need to lower the quality a little bit, at least for some time. I feel like saying there's been small hits to quality is an example of how no one at the team at Duolingo has been actually feeding any of the information about these small hits to anyone at the top.
Or rather, maybe they truly just don't give a because you can definitely tell there have been hits to quality. Dolingo has always been known for having little mistakes here or there. They were always very forgivable because it was like ah this is just a small team run by some people who really want people to get into language learning.
But the more it became a company based on taking all the money they can out of people, the less forgivable these type of mistakes are. Anyway, he does go on. He says, "One of the best decisions we made recently was replacing slow manual content creation process with one powered by AI.
" That sentence alone says so much. Reducing content creation to uh whenever humans have to make things, they're slow. They're so manual.
Uh imagining Lewis walking into a museum, he sees a painting on the wall. Oh, it took them 6 months to do this. I can whip that up in chat GBT.
One prompt. All right. Boom.
Artist. Let me finish up this dreadful email. Without AI, it would take us decades to scale our content to more learners.
We owe it to our learners to get this content ASAP. So, as I do have a 9-year streak, I can translate Dolingoes. What he's saying is you deserve to have lower quality AI generated slop if it means that Dualingo can create more content, thus keeping you on the app longer, thus making more money off of you.
You deserve it. Imagine if I'm uploaded a video here. You know, I uh I started the video by touching the camera a little bit, adding that flare, and I went, "Guys, you deserve more content.
It doesn't matter if the quality isn't really there. You're just here to consume. " As all of my content decisions are about retention and watch time.
I I figured the more I could just pump out, the more I could make money. And um that's why you're watching, right? You're just there to consume.
That's what you do. Anyway, I've uh decided to have a schedule of daily videos created with AI. Uh it's just a recreation of my face.
It just kind of looks at my previous videos and tries to do its best. Bruce versus American the Dualingo. Depressing.
Thank you as always for giving me money. I have no understanding of what art is. Anyway, he finishes with Dualingo will remain a company that cares deeply about its employees.
This isn't about replacing duos with AI. The changes are more about removing bottlenecks so employees can focus on creative work and real problems, not repetitive tasks. All right.
I mean, if we were to take this bit at face value, I would say that made sense. If all of their employees were proficient in AI, hypothetically, they could let the AI do the repetitive tasks, the menial tasks, and they can focus on more of the creative work, the stuff that, you know, fuels them as a business and also as humans that work for the business. But humans don't work at Dualingo anymore.
That's not actually what they're doing in the slightest. That's just how they're selling it. They're like, "Guys, we can we can keep doing the creative stuff, but the repetitive stuff, we'll we'll have AI do that.
" from my experience with the app that no, they're they're letting AI do the creative stuff. You could very very easily tell from these small hits to quality. But there are two possible reasons for this LinkedIn post.
Either Lewis is a delusional person with his head so far up the AI bubble that somehow neither he nor any of his higher-ups at the company predicted how tonedeaf and out of touch this post would be seen as. Or they ran the numbers and realized even with it being so tonedeaf, bad press is press and they only make decisions based on numbers, which that clocks. Hm.
Here I am contributing. I filmed a stupid sketch in my last video about the worst Dualingo update. It was like in 2023 where I referred to the Dualingo update as akin to a burger joint increasing the proportion of sawdust in their burgers as a way to increase profits.
And I actually think I was pretty bang on about that money. Dualingo stories used to be really great. They were at their best when they were just real stories that you could get at a bookstore, albeit short.
We used to have things like The Telltale Heart. It was it was actually quite nice. Eventually, uh, with no rhyme or reason, Dolingo removed all of those from the app and then started having their own original stories.
This was very frustrating as a user at the time, but I understood what they were doing it for, like they just wanted everything to be in-house. And you know what? They used to be really fun.
You could tell that they'd paid a team of people to write funny little stories up using the characters. It was a fun little way to learn a language. But I think the whole small hits to quality is nobody at the company having the heart to tell anyone up top, these are not small hits.
The stories, especially the little podcast audio stories recently, have been so horribly bland. Like someone at Dualingo, I don't know why I'm saying this like it's a hypothetical. Someone at Duelingo just went, "Hey, Chad GBT, I mean our version of Chad GBT, look at all of our stories.
" Yes, write some more. And then they didn't have a single goddamn human review the stories. They just went, "We need more content to scale.
We can't teach well without small hits to quality. " The stories make no sense. The stories have no beginning, no middle, no end.
They just exist. Now, I will say, yeah, the audio lessons use the words from the chapters. So, it's useful to hear the words used in context, but Jesus Christ, it's it's the most bland blah nothingness where it used to be so joyous.
It's frustrating. They're like, "Guys, we we have to all over your food. " In exchange, at one point, the food might be worth more money.
That's how it feels. We just need more content to scale. Our users deserve to be handfed AI slop.
Just pump it out. Thank you, large language model. Here's the thing.
Like I said, this is a weird one because I am not anti-AII. There's just one thing that I feel like people are missing that anti-AI people are quite right about. Writing stories is an art form.
It's creative at its core. There are important elements intrinsic to mankind within stories. Stories are not repetitive, menial tasks.
They are the creative work the company is supposedly able to focus more on now. But that's just the BS they're selling to people on an effort to make more money for their shareholders. You can't say you want to teach well and produce lower quality slop without a severe amount of cognitive dissonance.
Stories written by AI aren't bad because they're technically wrong. They're bad because they're nothing. They're average words on a page based on patterns.
Empty of meaning. They're gray. When it comes to language of all things, language shouldn't be a gray.
It should be great. I don't think an AI could have come up with that stupid pun yet. If you look on my YouTube channel, you'd see some of the most successful videos that I've made are actually about Duallingingo, ups and downs.
But it's helped me so much with learning German and Spanish. And I've had many comments through the years of people like, "Oh, when's your next Dolingo video? Why don't you make more Dolingo videos?
They clearly work. " Well, because I've already said what I had to say. I don't just pump out the same thing over and over because it's a formula for more views and more money.
I have to have something new to say to to add to the conversation to be creatively fulfilled. If I were solely a money-driven AI first business, I would probably have cut all the varied content that I've been making over the years and just give you a formulaic cookie cutter video with no new info. Here's Dualingo stuff, constant, constant, constant, until you get bored of it.
A lot of YouTubers go through this process. They find something that works. They ram it into the ground until people stop liking it.
They try and change. It's now too late. People don't expect it.
That is a recipe for disaster. I know the best way of being a online brand these days is to be on the Tik Toks every single second and to be on the YouTubes and then to make an Instagrams and make Twitter. Life is worth so much more when you touch grass, says the guy who's making an entire video about a language learning app.
I understand. I've got the self-awareness, but I can't touch it. I don't do anything about it.
The thing is, Lewis says that Dualingo owes it to its users to get them this content ASAP, and they've already cut their contracted workforce twice in the last 2 years. But that AI productivity boost, that decrease in company overhead, that hasn't gotten translated to a better app for us. It hasn't translated into a price decrease in spite all the money that they've now made from it.
No, they actually just locked it away behind another tier. So, they expect you to pay even more money to access features that used to be part of the app that they took away and replaced with AI. So, we've got price increases.
We've got tanking quality in the story department. And what's left feels like a bland corporate fast food meal of an app with an incredibly fine-tuned lesson structure. Used to have a human element, now it's just a robot.
And that's supposed to be something we're like, "Oh, yeah. I wanted no humanity in my communication app. I just wanted to give you my money.
Jesus Christ. I see all these articles like why aren't there any like big social media companies or big tech companies coming out of Europe. I just feel like Europeans not only have more laws to protect consumers but just have such a different understanding of what matters most in life.
And there's something sick about a lot of the people that are the top seuite CEO type people at most American companies. Something truly dystopian and disgusting. Tell us how you really feel.
I'm disappointed. That's what it is. It's it's a sheer disappointment.
You know what I mean? In order to be authentic and to share my opinions on this topic, it's going to come off as loud and hurt because I am hurt and I am American so I am loud at my best. So being a datadriven company has allowed Duolingo to make some difficult decisions that have resulted in people spending more time learning within their app.
That is a positive, but that is very transparently a bonus to the real reason of increasing shareholder value. Morally, it's such an interesting question. They're doing all this data mining and they found a way to trick people to using their app longer.
It's to learn a language so that they can make more money. But what used to be language first, money second became so overwhelmingly money, shareholder value, profits, and you get to learn a language along the way while you're giving us money. Why would you want to run a company like that?
I wonder if it is that famous quote, power corrupts everything. Money, the the greed, seeing big number get bigger must really genuinely break some people's brains and then they must not see anything else. They lose a bit of humanity.
Being a datadriven company is is a good thing. I'm not saying that's necessarily bad, but the data should be a tool, something that helps you make decisions, not something that you myopically go straight towards without any other consideration. Similar to Dualingo with learning a language, most people recommend you do Duallingingo alongside some other methods.
And it's a really good method. You know, it's just a tool in your arsenal. You're not supposed to go 100% AI cuz then this type of thing happens.
You become some sort of bland corporate entity without a soul, without a meaning, some dystopian sort of Batman villainesque person who posts emails to your LinkedIn being like, "Look how good I am. I'm It's so embarrassing to everybody else, but I guess people within his psychopantic circle. " I feel like Lewis is the type of person that sees the trolley problem and goes, "How is this a problem?
" Doesn't matter. Just pull the lever. Don't know these people.
kill them. I'm the guy that owns the train track anyway. I'm making that money.
What used to make social media such a fun and exciting place was that everyone was on it. It was social. People were there.
Your friends were there. You'd log on and see what Carol was up to or or your mate John. But then over time, in an effort to increase time spent on their apps to generate more advertiser profits, these same sites started algorithmically removing posts from Carol and John in favor of glittery influencer posts or posts designed to get you angry.
Once they discovered that, hey, angry people stand our apps longer. We make more money when you're upset. These apps became less about the people and more about the content.
The content to keep you there. Do I miss the vlogging era of YouTube when it was still social? Absolutely.
I do a lot actually. There was such a stronger sense of community to the place. Dulingo had a similar community within the comments of their exercises which they swiftly destroyed and replaced it with the ability to talk to AI for a fee.
This is the problem with having 100% of your business decisions dictated dogmatically towards profits and data as opposed to using this information to make informed decisions with other variables involved. This way ensures you create a soulless feeling entity. And like I said, I'm not against AI.
It can be incredibly useful as a tool in your arsenal. But at the end of the day, AI is an average of patterns that it currently knows. It can recognize patterns and predict from them better than any human.
But there exists an element to human creativity that exists far outside this data set that elevates the human more than the average. And no, it's not just Duolingo. Weird tangent.
Why is every Disney Marvel film, or rather, why have most all Disney films been pretty terrible in the last 5 to 10 years? Well, because executives base all their creative decisions off patterns. This exact pattern made us a lot of money.
Let's do it again. Let's do it again. Let's do it again.
Oh, shoot. Stop working. Let's go for this pattern.
Let's do it again. Let's do it again. As an audience member, you feel that this one wasn't as good as the last one, was it?
Something's off. Something feels missing. In my opinion, what's missing is that intrinsic human creative element that is necessary for a good piece of art.
As someone who makes YouTube videos for a living, I always appreciate when I watch a video from someone and I can tell just how much effort they went to to get a certain shot or do a particular edit. This shows what they, as an artist, decided was worth their time to add value to the video. I don't care how interesting a topic or how gripping a thumbnail, but if I click onto a YouTube video and I hear some sort of AI voice telling me something, I immediately click away.
This is content with nothing in it and no one behind it created solely to capture your time. And that's just not what I'm interested in. I want that human creative element.
It's a reason why I avoid YouTube videos run by big companies. My favorite ones are just run by people like me, people who spend way too long on a niche topic that they care about. That's interesting to me.
The frustrating part with Dualingo is that for me, I am still making great progress with my knowledge in Spanish with the app, and I can appreciate just how fine-tuned the lesson lengths and challenges are as a way to keep me engaged in the app. thus learning more Spanish. In that regard, I really do enjoy it.
It is working and I am learning more Spanish, but I'd be lying if I said all the joy behind it was still there. That's the thing. Just because something works doesn't mean it's good.
And just because something is efficient and scalable doesn't mean it's the only variable that matters. AI isn't going away anytime soon. It's like a genie that's been let out of the bottle at this point.
I don't think people fully grasp just how much of a paradigm shift we're living through at the moment. There will be a time in the very near future where you won't be able to find a company in any field not using AI to some degree. Similar to how rare it is to find a company without a website or social media these days.
It's only a matter of time. I hate being doomy and depressing, but AI is a thing that is here to stay. You just have to find the best way you can to continue living in society.
Find things that you really appreciate about life. For me, it's cycling, hanging out with friends, making these YouTube videos on whatever topic that I find interesting. for you.
Hopefully, it's consuming my content all the time. I've got more videos from you. I've got an AI farm generating videos every single day.
You deserve more content, and so I'm bringing that to you. And I hate to say this, but we'll probably be seeing more and more of this sort of thing as time goes on, companies becoming AI first and whatnot. But likewise, I feel like there will always be a stronger appreciation for anything made with the human element that's a bit rough around the edges.
As much as it upset me to notice that in a video I uploaded last month, I had an editing error with a missing graphic for 5 seconds of the intro, I didn't really beat myself up over it as much as I thought I would. Yes, it was a mistake, and yes, I wish I hadn't made it, but to error is human, and in these times, maybe that's not so bad a thing. So, yeah, I am still using Dolingo.
I strongly dislike its cartoonishly myopic focus of quantity over quality, but I am still learning using it. I don't know what to say. I'm a messy, flawed person myself, but I just missed the version of the app and the internet in general that felt like it was built by people who cared, or rather just people.
I'd like to hear from you if you have any thoughts on AI, Duolingo's implementation, and the overall path society's on in general. I'd love to hear from you in the comment section below. But speaking of AI, let's see if you can tell which part of this sponsored brand segment was written by AI, and which part was written solely by me.
I'll label each part with A or B and reveal the answer at the end. And speaking of websites that weren't made by an AI hallucinating its way through a language lesson, this video was sponsored by Squarespace. Now, I know what you're thinking, Evan.
I'm not starting a website. That's what I thought, too. And then I needed a portfolio and a newsletter and a place to sell things.
And suddenly, I was like, "Oh, right. That's what websites are for. " Squarespace makes it incredibly easy to hit the ground running.
You choose a template you like from their large selection of professionallook styles, and then you get to editing it up and make it your own. I started off with a simple homepage, then added a portfolio, a blog with info on my kit, and then I added a shop for good measure. Now I've got a sleek looking site to call my own, and I did it all myself.
Plus, everything is mobile optimized out the box. I didn't have to fiddle with weird formatting or CSS. I just clicked preview on phone, and it worked.
They've also got proper 24/7 customer support, so if something doesn't work, you're not stuck Googling your way through it. And for anyone running a brand, the marketing features like built-in email campaigns and basic SEO actually help get your stuff seen without needing a marketing degree. If you'd like to build a website of your own for your business, your blog, you name it, sign up to squarespace.
com/evanediger for a free trial today. And when you realize you love it, you can get 10% off. So, thanks Squarespace for sponsoring.
Now, let me wrap this up. Hello. Welcome to the outro.
You see, we've got the master of pun bit. It's the end of the vid. Did you guess right?
A was the AI. Obviously, I don't have a newsletter. It made up that I have a newsletter.
That's weird. I do feel like these days it's getting significantly easier to identify AI speak, but who knows what the future will bring. Either way, I would love to hear your thoughts below.
And hopefully I'll see you back here next Sunday for uh a light-hearted vid. I've got some uh really nice, happy little ones coming up about words and a big one that I've been spending a lot of time on that features lots of beautiful compositions around East London. So, hopefully I'll see you here for that one.
Thanks for watching. Goodbye.
