Duolingo Raises $45M, Now Valued at $470M
techcrunch.com
techcrunch.com
All these social media sites that make their money off advertising or selling user data can all go bankrupt as far as I'm concerned. Their users are coming to their websites for the short-term dopamine rush brought by the novelty of social media interaction: these websites are not different in any meaningful way from drug dealers, and they're costing millions of people happiness and productivity. They sell themselves as communications platforms but in reality, they compromise on privacy and as a result some of the most important forms of communication are hampered by their influence. As developers I think we have an ethical imperative to work for organizations like Duolingo who are pushing humanity forward and not social media companies who are holding humanity back.
Youtube, however, is better off. Laborers are the people that upload videos and workers at Youtube, Capital is the servers and all the software and somewhat the content creators, Customers are people watching and businesses trying to advertize. Though it is not as clear cut as a sandwich shop, it makes more sense than Yo or the myriad other tech startups out there that confuse these 3 key groups.
Duolingo is also 'unconfused'. Capital is the language service, website, and software of repetition. Customers are, again, the people using the service and other companies and people trying to get translations. Labor is, again, the people translating and the employees. In lieu of being paid for the translating, the users are 'paid' in the free service provided. Overall, I think it works pretty well, and, according to the OP, so do many other people.
Why do you classify youtube watchers as customers and facebook users as capital?
The vast majority of youtube's revenue is from ads, so to me that says that youtube is in the same category as facebook.
You leverage your facebook profile to gain access to properties, be given permissions or access, and your value is measurable by the input output of your usage.
Travel around the world right now and show either your government identification or a printed copy of your facebook profile and wouldn't it be weird if a young person you meet takes more credibility from that printed page than the validity and authority of your government id?
>>>> “The reasoning for this was that, though we could make it a profitable business, we realized we’d quickly become a translations business as we’d have to hire people focused on quality control, sales people, etc. and eventually businesses tend to pay more attention to the parts of the company that are profitable,” a spokesperson told us.
“We are an education company and are here to help bring education to everyone in the world and to make education more effective.”
<<<<<
Good intentions but does not seem to support your key point or am I mistaken?
Are you referring to the $20 English certification? There isn't anything else that they currently ask their users to pay for.
Of course, user who "only want to learn" can "always learn for free"
Ankidroid: A flashcard app for adding vocabulary I need and phrases I want to come more fluidly.
Spanish Trainer: A spaced repetition verb conjugation app as Duolingo doesn't give me enough practice with this.
Google Translate
What would speed things up for me is the ability to easily import translations from Google Translate into my flashcard app.
Duolingo needs to provide better user control of the spaced repitition input and some setting for the skills practice sessions. I would love the ability to make a certain drill or word as Finished or Very Confident so I don't have to waste time practicing stuff I already know. I would also love the ability to mark certain words or drills as important or useful so that they get repeated more frequently when practicing skills.
(Edited for formatting and the friendly grammar red cross :))
It's my understanding that Duolingo (and similar tools) aren't quizzing you just to make you prove that you know a word. It's actually part of the process of ensuring retention, and the act of recalling the word is actually part of the process of storing the information more permanently in the brain. Therefore, the ability to mark a word as "learned" may be counterproductive long-term. Of course, the algorithm isn't perfect, but it's more advanced than simply checking off a box that you've learned the word and never need to practice it again.
That being said, I do agree that more customization would be helpful.
The two other spaced repetition tools I use (which I mentioned) allow me to specify how well I know the answer and use when I have a high level of confidence in my knowledge, the words will come less frequently.
(edit: added second paragraph)
Downvoters can read this: https://english.stackexchange.com/questions/114620/speeded-v...
Sped is a perfectly valid English word. If you don't believe so then you should question your own knowledge of the language.
In this context 'legal' can only mean 'having to do with the law' as there is no statutory definition of the English language (well not British English at least).
You might like to try my webapp for reading and learning a language:
It allows you to read shared texts or upload your own (.txt or .epub) or read any web page with a Chrome extension (https://chrome.google.com/webstore/detail/readlang-web-reade...)
You can click or drag across any word or phrase to get an instant translation.
Every translation is converted into a flashcard, along with the context sentence.
You can practice with Readlang's own spaced repetition flashcards or export to Anki if you prefer.
Hope you find it useful!
(I've been bootstrapping this for the past 2.5 years)
If you aren't satisfied with the translation, you can refer to another dictionary in the reader sidebar (or a popup window in the Chrome extension). The default for Spanish-English is WordReference, but users can customize this if they have another favorite (http://blog.readlang.com/2013/11/07/custom-dictionaries.html).
On top of that, you can edit all the words you translate, and if in future you click the same word, you'll get your personally edited version instead of Google Translate. So you're effectively building up your own large translation dictionary. One idea for the future is to use all these edits to crowd source better translations, but there's plenty of lower hanging fruit to get through first!
Great job!
Your's is almost exactly what I was looking for for a long time.
Also memrise doesn't understand language the way duolingo does. The latter actually uses words in context, both in fragments of sentences and full sentences. These count as repetitions for the individual sentences, and there always seem to be some fresh sentences.
I think these differences make Duolingo a much superior tool as far as there courses go.
I disagree. I think what Duolingo needs is to do a better job of automatically detecting people's personal spacing and on which words (this isn't a trivial problem, BTW). Putting this burden on the user means that the user spends a lot of time tinkering with the platform instead of actually learning the language. This is a problem I ran into with Anki.
They courses were intended for international diplomats preparing for exams or relocation that require (at the least) conversational proficiency in their destination's local language. So, you learn the basics as quickly as possible without compromising understanding for rote memorization.
Do you know what you know, though? It's easy to spend an hour practising a bunch of words and feel as though you've learnt some of them, but come back a week or month later and actually not be able to recall them.
For German I used to try the exercises on my commute, but suddenly the app expected me to use "dem" instead of "der" without any explanation. I can't be the only one with this problem, right? Or is there a "new concepts" tab that I've never seen?
(Also, in case you haven't yet worked it out or looked it up somewhere else: you need to use "dem" with masculine and neuter indirect objects, often translatable into English using "to" or "for", and with certain prepositions. It is the dative masculine/neuter definite article.)
Duolingo has a full-featured web version: https://www.duolingo.com
I find it vastly easier to use than the mobile version for that reason.
I probably have a different perspective than most, but here's why I enjoy Duolingo's style:
The first foreign language I began learning was Spanish. I took it throughout high school and college. I was taught theory first, so I could understand most of what I'd read or hear following the new lesson. Running the input through the theory I'd established in my brain worked, but it was very slow and took a lot of practice to speed up my speaking and understanding.
When I was visiting my wife in Brazil while we were dating, I learned Portuguese very differently. Although they were aware that I didn't speak Portuguese, many of her friends and relatives did not let that reason hold them back from speaking it to me anyway. So I was bombarded with the language, and slowly began to understand more and more. First, I'd only understand one word in a phrase, then 25% of it, then 50% and so on until I reached a critical point where I understood enough of what was being said that I could use the context of what I did understand to infer the rest and learn that too.
Learning Portuguese was so much quicker for me. I still know more Spanish theory, but I still speak and understand it too slowly to consider myself fluent. On the other hand, Brazilians tell me all the time that they're impressed with how well I speak Portuguese.
I think the jumping right into the deep end style of learning Duolingo offers is fun, but may not be the best for everyone.
For the first German lesson, these are on capitalising nouns, grammatical gender, umlauts and verb conjugation.
I wish they'd add it to the app!
It took me ten years to be as proficient as a ten years old kid in my first language. So I don't think the way we learn our first language is necessary the best way to learn language in general.
And, I would say I was passably proficient by about 7. At least enough to read relatively simple non-picture books.
1) You have the ability to translate. You already know the concepts and ideas - you don't have to abstract meaning from noises - you already know the meaning, you just have to correlate it
2) You understand grammar concepts. It may be quite different in your target language - I have been learning 日本語 for some time, and grammar is wildly different from English - but there's still a lot of things you only need to learn a few differences of, vs. having to learn the concepts from scratch
With our first language, we have to sit there and learn by what sounds right vs. what sounds wrong. This is so hilariously trial and error and completely inefficient that it makes the entire process far slower than it would be if we had any other way to do it.
How long does it take you to figure out verb conjugation the way you learn a first language? Quite some time. You probably don't get it fully right in all cases until you later enter classes that go over the grammar rules. In Japanese, I was conjugating verbs, including irregular and negative forms, at multiple politeness levels, in about two hours of learning conjugation. And I rarely made mistakes after that. Why? Because I already understand verb conjugation. It's a concept I have a solid grasp on in my native tongue, and only needed to learn a few new rules to graft onto the existing framework.
Duolingo is almost exclusively focused on translation, which is absolutely not how most of us learned our first languages.
I have a non-fluent understanding of what "subjunctive perfect" means in my home language, but I guess I deal with it OK.
As such, I don't find it valuable when I see that section on Duolingo and think "OK, it's subjunctive perfect time!"
Of course, we do study grammar in our own language. We do it at school.
http://www.amazon.com/English-Grammar-Students-German-Learni...
http://www.amazon.com/Hammers-German-Grammar-Usage-Edition/d...
But they definitively go well together. You see a few examples of case usage, then you read the theory, then after you use it again, it magically "clicks" (for me at least), and things make a lot more sense.
The 'explicit' reinforcement /'implicit' learning couple is a great way to learn I think -- you grasp it intuitively and yet if you are faced with a new or tricky example where your experience will fail you'll be able to make a principled choice by explicit reasoning ("Oh the case here is * such and such * because of * this * therefore the pronoun here must be * that one *").
As for what you're asking, I guess it's really hard for english-speakers. I'm bilingual, and coming from spanish and being fluid in english, it's a lot easier. The main thing is: nouns have a gender, and the article ("the") do as well. Do "die" is feminine, "das" neutral, and "der" masculine.
Verbs also conjugate differently according to gender.
That said, especially for German I would've liked more theory too. And some people just seem to have a preference for learning in a more formal/theoretical way.
Suppose they even managed to come up with that classroom solution for language learning. How does what they currently have help them? Teachers, schools & government education departments make the purchasing decision for classrooms, so what students use in their time outside class isn't especially helpful in driving more sales of a classroom solution.
If they came up with something good for classrooms, they might as well axe or install paywalls into their existing stuff.
It helps the user better understand the language, which I believe is their goal to begin with. For example, their current approach presents a conjugated word in the context of a provided sentence or phrase. When I'm refreshing a language this isn't a big deal, but if I'm starting from scratch I don't have the foreknowledge to understand exactly when and why the word appears as such. I don't really need to argue that learning the language structure is much easier than memorizing a large list of words and phrases.
More simply a better tool for learning, is a better product on their end. They don't have to sell it to schools, but if teachers see that it can work in tandem with their curriculum then it will translate to more users.
> If they came up with something good for classrooms, they might as well axe or install paywalls into their existing stuff.
You're only looking at monetizing and forgetting about growth.
Because of the valuation size, and because I think they've succeeded at growing quickly and becoming the most well-known language learning solution. Everything else about Duolingo looks really good, but the monetization side of things seems to be the elephant in the room.
What I'm also pointing out is that the growth among language learners cannot be easily translated into growth in institutions. To grow in institutions they will still need sales teams, and their consumer userbase doesn't necessarily gain them much of an advantage in that game.
All that said, in this article they hint at having plans beyond the $20 certification they've announced. To achieve the valuation they've now got, I think it's fair to say they do have a solid plan that they just haven't made public knowledge yet. It's probably something that will be a bit controversial among their userbase, like paywalls.
Evidently they're planning to make money from paid certification tests and they have "other plans to monetize".
The funny thing is Duolingo is now valued higher than Rosetta Stone: https://www.google.com/finance?q=RST
I don't have a business model but I'm trying to bootstrap an iOS language software company. Duolingo is really good but I think language learning is difficult enough that there are other good ideas worth pursuing.
Anyway, if anyone is learning another language, I'd appreciate any feedback on my apps. Here's my Spanish app:
https://itunes.apple.com/us/app/h4-spanish-lite/id388918463?...
> As a company spokesperson told me, the company actually tabled its ambitions around translations about a year and a half ago and hasn’t scaled the businesses or accepted new clients (it’s still working with CNN, though).
The DuoLingo Test Center is currently free but it won’t be for long. Give it a try if you’d like to see how it works. Once the DuoLingo Test is accepted at schools and employers, the company plans to charge $20 to take its test.
There’s an established incumbent (monopoly) in this market called TOEFL. If you’ve come to the US to study, you’ve probably taken this test. It’s a lot more expensive than $20 per test and DuoLingo is out to prove it can do this testing less expensively and better.
http://www.guidestar.org/organizations/21-0634479/educationa...
They are not wrong. I passed Duolingo Spanish course (it might have been extended since then) and barely remember anything now. I was also studying German in school, and later French by myself to the point where I could even speak with native speakers a little. Hardly remember anything either. English is the only foreign language that I haven't forgot because I use it constantly. But if you're not using a foreign language, after passing the test you will completely forget it in 2 years.
You usually take this test because you will have to speak this language, so the case where you take the test and then do nothing with it seems pretty unlikely. Also, I disagree with your assumption that you will completely forget another language in 2 years (even when not used). I cannot confirm this from my own experience.
I spent a lot of time barely using certain foreign languages (+ than 2 years). Of course you get 'rusty', but far from forgetting everything.
And it came up back quickly
This tell me Duolingo doesn't make you learn it, but rather memorize their game.
I'd love to see a independent study into its reliability/retention. But even if it is completely useless, it is still no more wasteful than any other mobile game people play daily.
I have seen the effects myself. After a month or so, even though I had learned some spanish before, native speakers suddenly slowed down (in my perception).
An important clue: Use the "Strengthen your skills" button frequently to keep the tree golden. Duolingo itself doesn't nag you enough to do that, but it's essential for mastery of the language.
Remember this: everyone is taught their perfect grammar and spelling and quizzed on it for years in school. How many people can say what a past participle is, or even what an adverb is? How many people who passed geometry can still do that level of math from the top of their heads?
Of course you can't permanently learn something. You have to keep it in practice.
It's hard to master and complete a Duolingo course without getting a grip on the language. You'd have to pretty much ignore the "Strengthen your skills" button and never repeat anything.
Even after finishing a course you should use the repetitions for weeks and months to keep refreshing your memory. That's the only way it may stay forever.
I can guarantee you that you didn't forget everything. If you completed the Spanish course, your brain has been rewired for Spanish...
I was thinking at the time (it was in 2005) that they were really out of touch with common sense, and right now I wonder if anybody outside academia gives some credibility to the TOEIC or the TOEFL.
Originally they were going to provide translation services, which hasn't worked out. Now they're trying to provide certifications for people who pay a fee. Will that work? Maybe it will. What feels odd is that the valuation is so high for a company that hasn't yet figured out a reliable way to make money from what it does.
If all else fails, they could install paywalls, but a substantial portion of their users might abandon them at that point.
https://news.ycombinator.com/item?id=7698130
So it's not free in the free beer sense.
He can monetize this in a similar way. Crowd sourced translation service where the more advanced language learners translate books or something. It will be higher quality than google translate for sure.
"As a company spokesperson told me, the company actually tabled its ambitions around translations about a year and a half ago and hasn’t scaled the businesses or accepted new clients"
However, despite not being profitable yet, Duolingo is doing a tremendous amount of good. I know quite a bit about language learning and learning in general, and I have never seen a better tool for learning the languages they offer. The combination of words, sentences, smart repetition and text-to-speech is pretty much unique, and the courses I've seen are quite well designed.
Caveat: No, Duolingo alone is not enough to get fluent in a language, but if you used it as your only (or major) tool, you'll get very close.
For example there is also a course in the incubator teaching Catalonian for Spanish speakers, Catalonian being a language spoken in a part of southern Spain.
In the least, teaching Mandarin should be a higher priority than half the languages they support.
I think most people will admit that learning to write Japanese well is something not many foreigners will ever do, while speaking isn't that terribly difficult. I spent a little bit learning Japanese and I found that learning speaking + romaji only at first was not only easier, but helped when looking at symbols later (since I already knew how the words were said).
They do teach English to speakers of Chinese, Mandarin and Hindi.
[一] https://play.google.com/store/apps/details?id=com.chineseski...
[二] https://play.google.com/store/apps/details?id=com.pleco.chin...
But if you really want to learn the writing, get skritter. Absolutely worth the cost.
Their aim is to maximise their learning audience.
Lingvist is fairly new and French-only for now, yet I feel like I'm making more progress with that than I did with Duolingo. Not sure why that's the case.
I'm pairing both with French vocabulary decks on Brainscape, which is an Anki-like SRS app. Lingvist claims it will switch to a paid model soon, and Brainscape decks are in-app purchases. Lingvist is both mobile and web, though I always prefer the web interfaces when I have access to them. Wish it also had a speaking component.
I've started using Verbling for having conversations but so far they seem pretty unstructured and not worth the money. This will, of course, depend on who you choose. There don't seem to be a lot of French to English options there.
If anyone has suggestions for alternatives, I'd love to hear them. I'd like to rapidly improve my French before a trip at the end of August.
edit: ah sorry, missed the August bit on the last line. Probably not the value of soon you need.
And manage to run at loss, with it.
Now I use Duolingo multiple times per week on my phone and I wouldn't even think about paying for Rosetta Stone anymore.
I get that DuoLingo has a nice mobile interface, but do they have a larger marketshare than Rosetta Stone?
With significant effort I could type some Georgian or Armenian into Google Translate -- they are awful, awful scripts. Russian was easier, but not as widespread as I'd expected. I could also let other people type things, and they all know about switching between layouts, since they do it all the time to type email addresses etc.
If you're learning Russian, you have to learn a new alphabet anyway, a new layout isn't that much more trouble.
For iOS, just add the keyboard and it will become available (this is the same way that most people use emojis).
On Mac, you can add the keyboard as well. There is a phonetic keyboard for Russian that I've found to be much easier than the standard layout. I map keyboard switching to ctl+opt+cmd+space, which works pretty well for me. What I really like about Mac is that you can use dictation to read Russian words that you don't need to understand. You will need to download a Russian-specific voice for this, otherwise it will simply read the names of the letters.
Or just learn to touchtype off the bat.
Depending on your OS, you can also use a phonetic keyboard layout instead of the standard Russian one, but I really strongly recommend learning the standard Russian layout.
Their volunteer translation system may seem unreliable, but it's probably the only way smaller languages will ever get attention. At this point, I'm probably better off picking up some Polish books, learning a bit on my own, and contributing to their project.
Personally, I'd like to see language learning apps explore the parts of language learning that Duolingo isn't good at.
Also a high-quality, open source version of Duo would be awesome for super minority, low infrastructure, and/or conlang languages.
[1] http://utalk.com/app (App Store link)
Short, graded videos with transcripts and translations.
Now, can i have German->Japanese in Duolingo, please? :)
The code is on GitHub: https://github.com/euoia/node-reverb
Feedback appreciated here or via the site.
www.HelloTalk.com
You can translate the site to Vietnamese but they can't teach it to you, come on.
Edit: Found it, http://incubator.duolingo.com/courses/vi/en/status
Nice Viet-English pun there.
(Cám ơn == thank you.)