[00:04] So, I'm from Guatemala.
[00:07] This is a public service announcement,
[00:09] that is where Guatemala is.
[00:10] (Laughter)
[00:13] Also, that is not
where they keep the prisoners.
[00:15] That is called Guantanamo.
[00:17] (Laughter)
[00:18] Not the same place.
[00:20] So Guatemala's right below Mexico.
[00:22] And for the Americans in the audience,
[00:24] and let this sink in,
because it really applies in most ways,
[00:27] for the Americans in the audience,
[00:29] you can think of it as Mexico's Mexico.
[00:31] (Laughter)
[00:33] Just like the US doesn't want
illegal immigration from Mexico,
[00:36] Mexico doesn't want
illegal immigration from Guatemala.
[00:39] It's a smaller country.
[00:41] It's a poor country.
[00:43] And well, what can I tell you,
[00:45] it has much better Mexican food.
[00:47] (Laughter)
[00:51] Guatemala is a very poor country.
[00:53] And a lot of people talk about education
[00:56] as something that brings equality
to different social classes.
[00:59] But I always saw it as the opposite,
as something that brings inequality.
[01:02] Because what happens in practice
[01:04] is that people who have a lot of money
can buy themselves a really good education
[01:07] and therefore continue
having a lot of money.
[01:10] Whereas people who don't have
very much money
[01:12] barely learn how to read and write
[01:13] and therefore never make a lot of money.
[01:15] And this is especially true
in poor countries.
[01:17] Now, I was fortunate that I received
a rich person's education
[01:21] even though I didn't grow up rich.
[01:23] And it’s because I’m an only child.
[01:25] And my mother, who was a single mother,
[01:27] spent all of her resources
on my education.
[01:30] And this allowed me to come
to college to the US
[01:33] and eventually get a PhD
in computer science.
[01:36] Now because of all of this
[01:39] about 10 years ago,
[01:41] I decided I wanted to do something
[01:42] that would give equal access
to education to everyone.
[01:46] Oh, by the way, this is what I want
to talk to you about today,
[01:49] giving equal access
to education to everyone.
[01:51] At the time, I was a professor
of computer science
[01:53] at Carnegie Mellon University,
[01:55] and I decided to work on this
with my PhD student, Severin.
[02:01] The way my brain works,
[02:02] all of education is just
too general of a problem.
[02:05] So I decided, let's start
by teaching one thing.
[02:08] And then I started thinking,
OK, well, what should we teach first?
[02:12] Should we teach math?
[02:13] I mean, we both love math.
[02:14] Me and my cofounder, we both love math.
[02:16] And, you know, we didn't
decide to do math.
[02:19] Then we started thinking, well, maybe
we should teach computer science.
[02:23] But eventually, and this is maybe
surprising to people in North America,
[02:26] eventually, we decided
that the best subject to start with
[02:29] was teaching foreign languages.
[02:31] And let me tell you why.
[02:33] There's a number of reasons.
[02:35] One of them is that there's
a huge audience for it.
[02:37] There’s about two billion people
in the world learning a foreign language,
[02:41] both in school and outside of schools.
[02:43] Most of these people, by the way,
are learning English.
[02:45] About 80 percent of them
are learning English.
[02:47] In this weird map right here,
[02:49] all the countries in green
[02:50] are countries in which people
are predominantly learning English.
[02:53] And the reason for that
[02:54] is because English can truly
transform your life.
[02:57] In most countries in the world,
[02:58] knowledge of English can significantly
increase your income potential.
[03:02] So this is another reason why we decided
to start with foreign languages.
[03:05] And it's because you can directly
make more money
[03:08] if you learn another language,
in particular English.
[03:10] And see, this is pretty different
than many other subjects.
[03:13] For example, take math.
[03:14] In the case of math,
[03:15] just knowledge of math does not
increase your income potential
[03:18] because usually you have to learn math
to then learn physics
[03:21] to then become a civil engineer,
that's how you make more money.
[03:24] Whereas with languages,
if you are a waiter and you learn English,
[03:27] now you can be a waiter at a hotel
and make more money.
[03:30] So we decided, well, let's get started
by teaching languages.
[03:34] Now, we also convinced ourselves
[03:36] that the only way
to really reach a lot of people
[03:40] was by using a mobile phone,
or a smartphone in particular.
[03:44] See, building schools all over the world
is simply too expensive.
[03:47] On the other hand,
[03:49] most of the world's population
already has access to a smartphone,
[03:52] and the trend is that that fraction
is only going to increase.
[03:55] So we decided at the time
[03:56] that we would make a way to learn
foreign languages on a mobile phone
[04:00] that was accessible to everyone.
[04:02] And then we called it Duolingo.
[04:04] Thank you.
[04:06] (Applause)
[04:09] Now in order to truly be accessible
to everyone, rich and poor,
[04:13] Duolingo uses a freemium model
to support itself.
[04:16] What that means is that you can learn
as much as you want
[04:20] without ever having to pay.
[04:22] But if you don't pay,
[04:23] you may have to see an ad
at the end of a lesson.
[04:25] Now, if you don't like ads,
[04:27] you can also pay to subscribe
to turn off the ads.
[04:30] And it turns out that the vast majority
of the revenue for Duolingo
[04:33] comes from people who pay
to subscribe to turn off the ads.
[04:38] Now, who are these people who pay
to subscribe to turn off the ads?
[04:41] Well, they're usually well-off people
in rich countries like the US and Canada.
[04:47] Who are the people
who don't pay to subscribe?
[04:49] They usually come from poorer countries
like Brazil or Vietnam or Guatemala.
[04:53] So what I like about this model
[04:55] is that it is a small form
of wealth redistribution
[04:57] because we're basically
getting the rich people
[05:00] to pay for the education of everyone.
[05:02] So I like that.
[05:03] (Applause)
[05:06] So with smartphones,
we can reach a lot of people
[05:09] and we can even get the rich people
to pay for the whole thing,
[05:12] which is great.
[05:13] However, if you're trying to deliver
education with a smartphone,
[05:16] you run into a humongous problem.
[05:18] And it is that smartphones come equipped
[05:20] with some of the most addictive drugs
that humanity has ever engineered.
[05:24] TikTok, Instagram, mobile games.
[05:27] See, delivering education
over a smartphone
[05:29] is like hoping that people
will eat their broccoli,
[05:32] but right next to it, you put
the most delicious dessert ever made.
[05:35] (Laughter)
[05:38] If you really want to deliver
education to everyone,
[05:40] not only do you have
to make it accessible,
[05:42] but also you have to make it
so that people want to actually learn.
[05:46] And with Duolingo,
we've been able to do this.
[05:48] And at the highest level,
the way we've done this
[05:51] is by making the broccoli
taste like dessert.
[05:55] I'll say it another way.
[05:56] What we've done is that we've used
the same psychological techniques
[05:59] that apps like Instagram,
TikTok or mobile games use
[06:04] to keep people engaged,
[06:05] but in this case, we use them
to keep people engaged
[06:07] but with education.
[06:09] Let me give you some examples
of these techniques.
[06:11] One of the most powerful ones
is the notion of a streak.
[06:14] What a streak is, is it’s just a counter
[06:16] that measures the number of days
[06:18] that you've used
the product consecutively.
[06:21] You just take that number,
[06:22] you put it very prominently
in your product
[06:24] and then people come back every day.
[06:27] And the reason people come back
every day is because, well,
[06:30] if they don't come back,
that number resets to zero
[06:33] and people don't want
to lose their streak.
[06:35] It works.
[06:37] Now, on the one side,
[06:38] streaks have been criticized
for, for example,
[06:42] getting teens addicted to Snapchat.
[06:44] But in the case of an educational app,
[06:47] streaks get people to come back
to study every day.
[06:50] Now, to give you an idea
of the power of streaks,
[06:52] in the case of Duolingo,
[06:53] we have over three million
daily active users
[06:55] that have a streak longer than 365.
[06:58] (Applause)
[07:02] That means they haven't missed a day
in the last year or longer.
[07:05] Now, fun fact about streaks.
[07:07] What country do you think
has the longest average streaks
[07:11] for an educational app?
[07:13] It's Japan.
[07:15] Of course.
[07:17] Shortest-ever streaks?
[07:19] Latin America, baby.
[07:21] (Laughter)
[07:22] But we're fun, hey, we're fun.
[07:24] (Laughter)
[07:26] Another important mechanism to get people
to come back to your product
[07:30] are notifications.
[07:31] On the one side, notifications
can be really spammy and annoying,
[07:34] but in the case of an educational product,
[07:36] people actually want
to be reminded to learn.
[07:38] In the case of Duolingo, we have
a very sophisticated AI system
[07:42] that basically chooses
when to send the notification
[07:46] and also what to say in each notification
[07:48] to maximize the probability
that people come back.
[07:51] Now, interestingly,
[07:53] even after all this sophistication,
[07:55] it turns out that the algorithm
for choosing what time
[07:58] to send you a notification
is pretty simple.
[08:00] Do you know what is the best time
to send people a notification?
[08:03] I'll tell you.
[08:04] It's 24 hours after
they used the product last.
[08:06] There's an easy explanation.
[08:08] If you were free yesterday at 3pm,
[08:10] you’re probably free today
at 3pm as well.
[08:13] So this is what a very sophisticated
millions of dollars of AI found.
[08:16] (Laughter)
[08:19] It's funny.
[08:20] Now with notifications
you shouldn't be spamming.
[08:23] And we're not spamming, with Duolingo,
[08:25] we actually stop sending notifications
after seven days of inactivity.
[08:28] So if you don't use Duolingo
for seven days,
[08:30] we stop sending you notifications.
[08:32] Now, at some point it occurred to us,
[08:34] if we're stopping
to send people notifications,
[08:36] we should let them know.
[08:37] So we started sending
this notification to people saying,
[08:40] "Hey, these reminders
don't seem to be working.
[08:42] We'll stop sending them for now."
[08:44] You know what people do
when they get this notification?
[08:47] They come back.
[08:48] (Laughter)
[08:50] Passive aggressive.
[08:51] (Laughter)
[08:54] Works for my mother, works for Duolingo.
[08:56] (Laughter and applause)
[09:02] These passive aggressive notifications
[09:04] are really good at getting
people to come back
[09:07] back because they feel like our green
owl mascot has given up on them,
[09:10] so they come back.
[09:11] And speaking of our
green owl mascot, by the way,
[09:13] because all our notifications come
from our green owl mascot
[09:16] and well, he's passive aggressive
and also pretty pushy,
[09:19] this has given rise to a lot
of memes on the internet
[09:21] that make fun of the great lengths
that he will go through
[09:24] to get you back to learn a language.
[09:26] Here's one of my favorite ones.
[09:28] This is it's a meme,
it’s one of my favorite ones.
[09:30] (Laughter)
[09:32] Basically looks like you forgot
your Spanish lessons.
[09:35] And then there's an intruder alert,
[09:36] presumably the owl broke into your house
to get you to learn language.
[09:40] Now, Duolingo has entered the zeitgeist.
[09:43] And there's thousands of memes,
there's SNL skits about it.
[09:46] And it's because we've managed
to get people to want to learn a language
[09:50] by using the same techniques
[09:52] that mobile games and social media use
to get people engaged.
[09:57] And this is a really important point,
[10:00] let me say this.
[10:02] I don't actually believe that there's
a way to make an educational app
[10:05] be as engaging as something like TikTok
or Instagram or mobile games.
[10:11] But the good news is that --
[10:12] And by the way, the reason
I don't believe that
[10:14] is because ultimately you have
to teach people something.
[10:17] And it's hard to compete with,
like, cats and celebrities.
[10:20] But the good news is
that I don't think you have to.
[10:22] See, here's the thing.
[10:23] When you're learning something,
you get meaning out of it.
[10:26] Whereas when you're scrolling
for two hours on Instagram,
[10:29] a lot of times afterwards, you feel
like you just wasted your time.
[10:32] So I think it's actually OK
if your educational product
[10:36] is only 80 or 90 percent as engaging
as something like TikTok,
[10:39] because the other 10 or 20 percent
[10:41] will be provided
by people's internal motivation,
[10:44] though of course, not much more than that.
[10:46] This is really a key point.
[10:47] If we want to get people
to do something meaningful,
[10:51] you can use the same techniques
that apps like social media use
[10:54] to get people to do it.
[10:56] And even if you're not as engaging
as those apps are,
[10:59] you can still get hundreds of millions
of people to use your product.
[11:02] In the case of Duolingo, for example,
[11:04] there are more people learning languages
on Duolingo in the United States
[11:08] than there are people learning languages
across all US high schools combined.
[11:12] (Applause)
[11:13] And this is true
in most countries in the world.
[11:16] My hope is that -- I know
we can do this, but, you know,
[11:20] my hope is that as humanity,
[11:21] we can do what Duolingo has done
for learning languages
[11:24] but for all other subjects.
[11:26] Where we can get people
to learn math with mobile phones,
[11:29] like millions of people to learn
math with mobile phones
[11:32] or physics or whatever.
[11:34] I hope for a future in which
screen time is not a bad thing,
[11:38] in which we can deliver high-quality
education to everyone, rich or poor,
[11:42] using a mobile phone.
[11:43] But the single most important thing
that I can end this talk with,
[11:47] is a reminder to please,
pretty please, I beg you,
[11:52] do your language lessons today.
[11:53] Thank you very much.
[11:55] (Laughter)
[11:56] Thank you.
[11:57] (Applause)
[12:05] Helen Walters: Luis, thank you so much.
[12:07] I wonder if you can just say
a little bit more about that last point,
[12:10] how do you think you can apply
this type of thinking
[12:14] to other subjects?
[12:16] So you mentioned math,
and things like that.
[12:18] How do you do that?
[12:20] Luis von Ahn: I think in particular,
[12:22] subjects that are learned
through repetition
[12:24] and it turns out most things
that are kind of really meaningful
[12:27] are learned through
thousands of repetitions.
[12:29] You learn to read through repetition,
[12:31] you learn elementary school math
through repetition.
[12:33] Most things that you can learn
through repetition,
[12:35] you can actually gamify
and turn into something like Duolingo,
[12:38] where people just
do it a lot and do it fun.
[12:40] It's a little harder for things
like explanations.
[12:43] That probably is going to require
some really good videos.
[12:46] Sal Khan is doing a really
good job with that.
[12:48] But for things that require
a lot of repetition,
[12:50] I think we can use the same methods.
[12:52] (Applause)