Tip:
Highlight text to annotate it
X
[MUSIC PLAYING]
JAREK WILKIEWICZ: Hello, everyone.
Welcome to YouTube Developers Live.
We have an exciting show for you today.
We'll be talking about Interesante.
Did I get that right?
ANTONIO ALTAMIRANO: Yep.
Perfect.
JAREK WILKIEWICZ: And let me introduce
our hosts and guests.
In the studio, I have Jeremy Walker.
How are you doing, Jeremy?
JEREMY WALKER: I'm doing pretty well.
I'm doing pretty well.
How are you doing?
JAREK WILKIEWICZ: I'm doing good.
I think I am quite excited that we have this show up and
running now that we had some technical
difficulty getting started.
JEREMY WALKER: A couple technical difficulties, yep.
JAREK WILKIEWICZ: All right.
And then on my left, Antonio Altamirano.
How are you doing, Antonio?
ANTONIO ALTAMIRANO: Good.
Good.
Thank you.
Thanks.
JAREK WILKIEWICZ: Great.
And then, in the studio, we also have Jochen Kumm.
JOCHEN KUMM: Hi.
Nice to meet you.
Great to be here.
JAREK WILKIEWICZ: Great.
And then, we will have Pablo Gamboa joining us live as well
on the Hangout from Argentina.
Can you hear us, Pablo?
PABLO GAMBOA: Yeah.
Perfect.
Hi Jarek.
[INAUDIBLE].
JAREK WILKIEWICZ: Great.
PABLO GAMBOA: Nice to be here.
JAREK WILKIEWICZ: All right.
So let's get started.
So Antonio, tell us a little bit about your company.
How did you come up with the idea and what is the problem
you're trying to solve?
ANTONIO ALTAMIRANO: Well first, thank you very much for
having us here.
It's a very, very exciting to be here.
The idea of Interesante came from the fact that we saw a
major, major need for the Latino market in the US and in
Latin America to be served by a tool that is culturally
sensitive and culturally aware, as well as
geographically aware.
So we all used the tools that are out there for social
networking.
And what we were lacking is that connection to our own
culture, to where we come from and to the things that matter
most to us Latinos and to the way that
we communicate online.
So the tools that exist right now work, but they don't take
into consideration the fact that we have different needs
in terms of communication and in terms of interest.
JAREK WILKIEWICZ: Hence the name, Interesante?
ANTONIO ALTAMIRANO: Exactly.
JAREK WILKIEWICZ: My Spanish isn't all that good but I
figured that one out.
ANTONIO ALTAMIRANO: Yeah.
Yes.
So Interesante comes into the picture.
It's an evolution of a prior product.
So what we did is we wanted to share our interests, our
favorite quotes online.
So we started just developing this and we realized that a
lot of people were coming in and connecting with what we
were doing because we were targeting this particular
population.
We were doing it for us, and a lot of people
connected with that.
So we saw this, but the brand name did not make a lot of
sense at that point.
It was called Proverbio and it was too narrow of a set.
So what we did is we were thinking about a word that
will unite interest across South America.
So what we wanted to do is we wanted a word that is easy to
understand, it's playful, and is understood by everyone.
Because people that are watching this show, they will
know that there are many differences between the
Spanish that is spoken in Argentina versus the Spanish
that is spoken in Mexico, but the word interesante brings
all of that together.
So that is what we stand for.
We want to be the place where Latinos can find their
interests online and on mobile.
JAREK WILKIEWICZ: Sounds like you performed a proverbial
pivot and became Interesante.
ANTONIO ALTAMIRANO: Yes.
JAREK WILKIEWICZ: OK.
Very interesting history.
So I was thinking, perhaps, rather than describing in
words of what the product does, maybe you could give us
a little demo?
ANTONIO ALTAMIRANO: Yes.
Completely.
Yeah.
JAREK WILKIEWICZ: Sure.
Maybe we'll--
ANTONIO ALTAMIRANO: I would love to.
JAREK WILKIEWICZ: --try the demo.
ANTONIO ALTAMIRANO: Great.
So this is the landing page and I want to show you the
landing page so that there are a few steps--
Interesante is very easy to use.
So basically what you have to do is you have to create your
account and you start to see this interests that are
relevant to you.
So as you roll over, you see this button that says mi
interessa and if you're interested in that interest,
then you click and we classify it automatically.
So and you can see that the collection is automatically
set and that we categorize that collection.
So you can select from a wide variety of interest.
And here you can see that Android is very popular in
Latin America so that's a very popular item in Interesante as
well as travel for places that are not found in sites like
Expedia or anything like that.
This is just very localized type of travel that this is
growing exponentially in Latin America.
Then with that--
JAREK WILKIEWICZ: So if you want to find a destination
that isn't popular yet but it's cool, you go to
[INAUDIBLE]
Interesante.
ANTONIO ALTAMIRANO: Exactly.
Yes.
Exactly.
So then with that, we take that and we Johen here that
will speak more about that, but we have algorithms that,
as I said, they're culturally aware and we help you discover
new interests.
So for instance, if because I am into Interesante, then I
see this mug with the logo of Interesante on it.
And also, as an addition to that, you can collect them.
So you can collect this interest and you can select
your own collection, but the point here is that we don't
want people to be worrying too much about what they have to
do, where they have to put it, so we built the intelligence
in the back-end and we take your interests we auto
categorize them, and we use that so that we can serve
better interests to you.
And of course, we're in mobile as well.
And we'll be showing that later.
If you want, I'm going to go really quickly to one of the
actual screenshots of the product.
So for instance, this is an interest that many people--
this is in Spanish so basically it says that these
are not bags under my eyes, they're
just accumulated ideas.
That's perfect for you.
So you see that eight people have collected that.
You have access to those people, but most
interestingly, you have access to all the interests that
relate to that particular thing.
So and we're doing the same with video.
So again, this is something that is very popular.
This is Sofia Vergara is very popular in our site.
I hope we're live there now.
Let me talk about the integration with YouTube, in
particular.
Let me just get to the right URL so that I can share.
OK.
So in this case, we have this video that basically talks
about how difficult it is to understand Spanish.
So it goes through like 10 different words and explains
the different meanings in different countries.
So this also supports the fact that we need a place where we
can collect our common interest, not only our
differences.
So what we did is we integrated with YouTube and
you can see all of this recommendations that we get
directly from YouTube.
And the moment that you click on one of them, then it
creates another interest on Interesante so that you
continue to build your database of interest that is
personal to you and we can do a better job at suggesting
things to you proactively.
JAREK WILKIEWICZ: Great.
And that is using the Topic API that we'll talk about a
little later--
ANTONIO ALTAMIRANO: Exactly.
JAREK WILKIEWICZ: --with Pablo about.
ANTONIO ALTAMIRANO: Yes.
JAREK WILKIEWICZ: Great.
And then, you also have a mobile app, right?
ANTONIO ALTAMIRANO: Correct.
JAREK WILKIEWICZ: Actually we tried to look at the mobile
app as well.
ANTONIO ALTAMIRANO: Yeah.
So what we did in the mobile app is we wanted to have a
different utility because mobile is such a different
platform that we wanted to make it quick for our users to
be able to find their interests.
So for instance, right now we're in the category section
and what we do is we take into consideration the activity
that you have and we suggest interests.
For instance, what we're showing right now on the
screen, is their entertainment category and you can just
swipe through all of them and see what's interesting to you
and what is not interesting to you as well.
So in the top right hand corner, you'll see a counter.
So that counter just basically tell you, hey, you've done
this much today and we're going to tell you how you
stand versus other users.
So you can also discover other users as well.
But mainly what we focus on is what is interesting to people
and through that, it is a channel for discovering
interests more than people themselves.
So you remove a lot of the friction between who is the
person who is recommending me.
We're going directly into what is interesting to the people.
JAREK WILKIEWICZ: So the idea here is I may have many
different interests and someone may not be all that
interested in me, but they might be interested in a
particular area that I'm--
ANTONIO ALTAMIRANO: Exactly.
JAREK WILKIEWICZ: --collecting information about or may be
curating videos and so forth.
Right?
ANTONIO ALTAMIRANO: Exactly.
JAREK WILKIEWICZ: So then they would actually follow that.
ANTONIO ALTAMIRANO: Yeah.
JAREK WILKIEWICZ: And just subscribe to it.
ANTONIO ALTAMIRANO: And the cool thing is that I don't
have to actually go to your profile and
choose what I like.
The system will tell me, oh, because you like add
technology or because you like a certain topic and because of
your activity, then we're going to take one of your
interests and suggest it to [INAUDIBLE] for instance.
JAREK WILKIEWICZ: Great.
ANTONIO ALTAMIRANO: So then it's built in within this
discovery and the experience of the app.
JAREK WILKIEWICZ: Great.
So now I know Jochen works on figuring out exactly the
mathematics behind the interest accommodations and
then we also have Pablo on the Hangout so why don't we talk a
little bit more about the technical aspects of this
integration.
Let's see if we can bring in Pablo online here.
JEREMY WALKER: Can you hear us, Pablo?
Great.
Can you talk a little bit about how the integration with
the app came about?
PABLO GAMBOA: Well something Antonio said, our man use of
the YouTube API was to get related media on Interesante.
And the user has the opportunity to create
interests with YouTube videos.
So when another user opens the video on Interesante, what we
have done is to comb YouTube's API and also use the Freebase
topics that they provide to get related videos that we can
recommend to that user.
So this provides the users with a lot of more content
that maybe interesting for him.
And also, with just a couple of clicks, the user can add
those new YouTube related videos to his collection on
Interesante.
So it's like we are having two features.
One is recommendations and the other is letting the user add
this content to his interests [INAUDIBLE].
That feature is live on Interesante.com now, and we
have another feature that is still in beta, we are still
working on it which is basically the possibility to
import your playlist, uploaded videos, favorites [INAUDIBLE]
from your YouTube account or from another [INAUDIBLE]
account.
But this is still beta.
We don't have this feature on production right now.
JEREMY WALKER: So let's go back a little bit because you
said something that I'm really interested in.
Our new version of the data API, which launched recently
and Jarek posted it, uses one of the new features is the
Topic's API which uses Freebase a little bit.
So I know, and we want to get into this with you, if you
could talk a little bit about that.
So it sounds like you're importing and you're using
that portion of the data API to kind of get recommendations
from other people and kind of categorize them.
Is that correct?
PABLO GAMBOA: Yeah, that's correct.
Technically, what we are doing is once we tapped the video we
get the video in and we query the API to get topics related
from Freebase [INAUDIBLE]
release that [INAUDIBLE].
And once we get that topics, with our [INAUDIBLE]
we find which topics we are going to get to use and then
query again the API to get those related videos.
So it's pretty important to query that Freebase topic now.
JEREMY WALKER: And did you want to add anything to that
because I know you really jump into the
details of all that stuff?
JOCHEN KUMM: Yeah, I'll talk about the algorithms without
talking about the algorithms.
I think when I came to Interesante that the question
was, is there something unique about Latinos, in the most
general sense?
Having lived there, I think I know there is, but the
question was can we capture that and what's the point?
So the idea was you have two extremes here.
You can go and look at everybody in the world and try
to understand everybody in the world.
What happens is you lose your resolution.
On the other extreme today, you can probably track someone
as a person and get to know them very well but that wasn't
what we were after because we want to show them, we bring
them into the world.
So Interesante is right in the middle.
We're trying to find the community that you're part of
so you don't have to look at the whole world and just get
lost, but you're not limited to your own little world.
So we are about finding Latino communities that emerge
naturally from the data.
And so one more point here, we have two things going on.
We have algorithms that create communities algorithmically by
standard Markov-based sampling algorithms.
And there's an interesting phenomenon which I want to
think Pablo for and the whole Interesante team, the
interface is designed and it was working.
So in addition to algorithmic community, we have
self-organizing communities in Interesante.
And it is just a wonderful sort of in order out of chaos
emerging in Interesante.
From an algorithmic point of view, lots of people
interested in self-organizing community and Interesante is a
great example for it.
JAREK WILKIEWICZ: So are you modeling that in your basement
right now with Markov chains and such?
JOCHEN KUMM: Yeah, yeah.
JAREK WILKIEWICZ: This show amazes me because it's been a
long time since I had to read about Markov chains and this
was not my favorite topic in school.
JOCHEN KUMM: But it's so cool.
Well I'm really getting to know Latinos [INAUDIBLE]
Mexico and the United States and finding out what their
communities are and they're very diverse.
I mean, we've had Guatemalan writers, schoolteachers, the
Red Cross, everybody stepping in, self-organizing, and then
we can support them behind the scenes with our algorithms
which is sort of a socially aware way of using algorithms.
And hopefully we can do some statistical something that
makes a difference.
ANTONIO ALTAMIRANO: And just to add something, you
mentioned the Red Cross.
So the Red Cross right now is in Mexico.
They have these great videos about social
responsibility, basically.
But they didn't have a good way to get the word out.
So they're using us to do that.
And the advantages that by using the integration that we
have with YouTube and their own videos, they don't focus
only in one video, in one message, they have an audience
that comes in and then they see that there are other
relevant topics that they can learn about that is related to
the Red Cross and what they're doing in Mexico.
And there's a lot going on in Mexico--
JAREK WILKIEWICZ: Yeah.
ANTONIO ALTAMIRANO: --that is being communicated via video
and what is lacking is a good platform
that connects everyone--
JAREK WILKIEWICZ: Yeah.
ANTONIO ALTAMIRANO: --culturally.
JAREK WILKIEWICZ: Yeah.
I think we talked about that about pivoting earlier on this
show and I think that one of the ways in which I see you
using the API is that by user sharing the video, you can
actually find what's related and have them actually explore
things that maybe they didn't know existed--
ANTONIO ALTAMIRANO: Exactly.
JAREK WILKIEWICZ: --and this uses the Topic's API and then
the Freebase on the back-end to figure out all the terms
that might be related.
JOCHEN KUMM: Yeah.
Freebase is pretty amazing.
It gives us a look into the whole world and it made our
algorithmic development so much faster because that's our
starting point and refining is much easier than trying to
build Freebase from scratch so we love Freebase.
JAREK WILKIEWICZ: Do you find that Freebase's coverage of
the Latin America Spanish terms this is adequate for
your use case?
JOCHEN KUMM: In a way I have to say the Latino community is
under-served a little bit.
JAREK WILKIEWICZ: Yeah.
JOCHEN KUMM: So what we're finding is in addition to
switching existing content, Interesante is continuously
creating content that wasn't on the web before.
So that is not in Freebase yet and we'll have to figure out
ways to ultimately bring back to the
community what we are gathering.
JAREK WILKIEWICZ: I'm sure that the Freebase folks will
be happy to have it.
JEREMY WALKER: Yeah.
They'd love to hear about that.
But just quick, to jump back in the technical end, Pablo,
can you tell us a little bit, if you can, about--
I'm always curious what you used on the back-end, what
language you actually used to do your implementation.
PABLO GAMBOA: Well I can brief you something about our
technical aspect.
In the back-end, we are using like three different
[INAUDIBLE].
We are using PHP and Python for the main core of the web
application and our API that uses the application and also
the iPhone application that you've seen.
And we are also using Ruby on Rails for our
recommendation engine.
And everything is connected to MySQL in database.
And that's for the back-end now.
The front end is pretty simple, a lot of Javascript,
jQuery precisely.
And we are using Twitter [INAUDIBLE]
framework also up front.
So that's the main thing which is that we are using now.
The decision to use these languages are related to the
way that we work.
It's a lean start-up that's how we define us.
We always try to move quickly on implementation and try to
deliver fast to the production server so that's what we are
using right now.
JEREMY WALKER: Great.
I'm always curious, did you have any surprises when you
were integrating with the YouTube API?
Anything that you weren't expecting?
I'm always just curious when you were
integrating with our APIs.
PABLO GAMBOA: I don't if surprise is the word, but we
were really happy because it was really simple for us to
understand and incubate [INAUDIBLE].
I mean, as I told you, we try to work fast and try to
deliver fast what we are doing.
And I think the main feature for recommendations went very
well, went pretty quick, and that's because the API is
really intuitive.
And also the [INAUDIBLE]
of the new API is also excellent.
You have examples there.
You have all you need to know so it was really simple for us
to do that and in a really short time.
So that was surprising for us.
I mean, something that you think that may be will be
difficult to do was very simple.
JEREMY WALKER: That's great to hear.
That's great to hear.
Do you have any features that you wish were part of the API?
Like anything on your wish list?
PABLO GAMBOA: I think that it's pretty complete now.
I was checking all the other features that we didn't use
yet and it's pretty complete.
I would like to see how it's going to be more integrated
with Freebase.
That is really important for us.
That's more.
JEREMY WALKER: So deeper integration with Freebase.
PABLO GAMBOA: Yeah.
Deeper integration with that.
JEREMY WALKER: Great.
JAREK WILKIEWICZ: Great.
All right.
So let's check to see we have any
questions in the moderator.
I see one question that I think might be addressed at
you, Jeremy.
And the question is, is YouTube
Developer operations hiring?
JEREMY WALKER: Yes.
We're always hiring.
Specifically for the YouTube team too so if you're
interested in video, you know coding, check out our site.
It's on developers.google.com/jobs.
So check that out and even if you're not interested in
YouTube, we have a bunch of positions even outside of that
that are available.
So please check that out.
JAREK WILKIEWICZ: All right.
I think this is it.
Thank you very much for joining us today live in the
studio and thank you, Pablo for joining
us live on the Hangout.
And we'll see you live next time, same time 10:00 AM
Pacific, Wednesdays on YouTube Developers Live.
JEREMY WALKER: Yes.
Thank you.
ANTONIO ALTAMIRANO: Thank you.
JAREK WILKIEWICZ: Bye.
JOCHEN KUMM: Later.
[MUSIC PLAYING]