Do you feel like you're the pioneers in
the AI-powered Entertainment industry? We had to jump in. We're really at the beginning of a wave
and a technological transformation that will have a lasting impact. What's the problem in terms of
synchronizing mouth and tongue movements when we do animation? An animator will do by hand, and this will
be extremely long. For example, when Asterix walks, he is throwing arms in the air, but... there's no actor who would do that, he would be a fool. And so it'
s rare that an Artificial
Intelligence can predict things it hasn't seen before. Or if she does that, it means it's going
to be a mix of things. Broad topic! As you know, on The Animashow, we deal with
all subjects. And for me, it's important to see all
aspects of our industry, whether it's artistic, business or technology. Our two guests today, Louis Abel and Sixte
de Vauplane, are founders of technological companies that take advantage of
Artificial Intelligence for... to work in the animation
industry and help
this industry flourish. They will share with the secrets of
fundraising, they will share with us the secrets of making products by using
artificial intelligence, and we will also discuss the future of animation in contact
with these new technologies. I hope you will enjoy! Feel free to
subscribe, to activate the little bell, and above all, to share this episode with
as many people as you can. Thank you so much for your support. I wish you a very good listening. Sixte, Louis, w
elcome to The Animashow! Thank you. Good evening. It's a pleasure to have you here. Today we're going to a special special
AI in a way, because you are both founders of companies doing
Artificial Intelligence, applied to the creation of animation. That's why we're
discussing it today. Do you have the impression or do you have
the feeling that you are the pioneers in the industry of entertainment powered by AI? I'd like to. Pioneer, we hope. It's clear that we're at the heart of a
transformation
in Animation. It's moving a with Artificial
Intelligence, we see it everywhere right now. Pioneer is a strong word. I would say more, we will be pioneered
maybe in five years, where we will become standards, but for the moment... I don't know. Interesting. Sixte, even... There's something going on in the
animation. You feel it, you feel sense effervescence,
you go to the markets, you go to the MIP, you go to the kids' screen, people talk
about the AI, it's on every mouth sometimes. We're afraid
sometimes we look at with
desire, but it's something that raises questions. And I think we're stoned. I think we should say it. There are a lot of actors who... I think we're really at very beginning of
a wave... and a technological transformation that
will have a lasting impact on it. And I think we are one of the first
companies that put the AI directly in the heart of the reactor. It's not just a feature we put to please
investors or whatever. The heart of the model is running with
this AI. F
or the listeners, thank you for this
answer. For the listeners and the listeners who
are to We're going to talk about your companies. Louis, you're the founder of Dynalips, a
company specialized in lip syncing, meaning how we coordinate mouth movements
with speech on 3D models, animation models, in video games, etc. You, Sixte, you're the founder of Animaj a
company that not only edits software that allows motion capture, but also motion
capture -free. So you will explain that in detail. And the
n who also edits YouTube channels
and other platforms for children. Not only on YouTube, we buy properties, we
buy IPs that were born or not on YouTube. So intellectual properties for those who
have intellectual properties. The english words. Sorry, I'll have to remind myself often
about that. But intellectual properties that we try to
transform into franchises. In fact, our goal is to create franchises
of tomorrow, the mega franchises of tomorrow, with an approach that is at the same time, base
d on the data. We use all the data from YouTube to better
understand audiences and build content that resonates locally, culturally with
audiences. A production assisted by AI, and it's
indeed a whole tool battery, including the tool you mentioned, but we have a lot of
others that we built to be able to considerably reduce the time to produce an
animation content. and then a digital first and multi
-platform distribution where we start distributing our content on YouTube
because that's where the
children are and then we distribute it everywhere they are,
so on streaming platforms, on TV, everywhere children are, our brand and our
content should be present. So the first two activities are in line
with what a TV channel or something that would look like a broadcaster and then
there's this library in the middle that uses artificial intelligence in the
manufacturing process. On Dynalips, how did you come to design
this company? With whom did you design it? Not alone, it's not the work of o
ne man. I'm off, but you are a researcher at
INRIA. At the University of Lorraine. I'm a PhD student. I work in artificial intelligence for my
thesis. And it's my thesis director, Slim who is
the origin of the works on articulation, on how to understand the mechanisms behind
human articulation, how someone speaks. visually. And with another doctorate he had a few
years ago, they created a technology that allows artificial intelligence to generate
the lip sync of a person. It was more his vocatio
n to understand how
it worked, but behind it they had the intuition that it could be used in the
world of animation. It's clear that talking to a character
automatically flows from the source. We're going put things in perspective. What is the problem with the
synchronization of mouth movements and language when do animation? As soon as a character speaks, we will
have to animate it. There are not 36 ,000 different solutions. Either an animator will do it by hand and
it will be extremely long, o
r he will use motion capture. But for lip animation, precision is
still... ...to be That's it, it's average. And here we intervene here, that is, we
take directly the speech, the audio, and we provide the trajectories of the lips. Almost instantly. The input file is an audio file that is
able to generate the movements of the mouth of the animated character depending
on what is said, with if I say an M it will come to touch the bass, if I ... Exactly. Does it take into account intonation? I mean,
if I smile while I'm talking... No, exactly. Yes and no. At the University of London, we did some
work on it, on the generation of lip sync, which we call expressive, to make the
contrast. But in Dynalips, we we call neutral But
it's a choice. It's actually an animator, he doesn't want
to do lip sync there, he doesn't want to do all the trajectories, where the M, it
must touch the lips, all that. But he wants to do the expressions. And an AI to guess the right expressions
with an artistic direc
tion, with maybe... ...with facial expressions, or things that
unique to the character. Something that's not human, not in the
voice. She can't do it. An AI is guided by the data she received,
so it means that she will predict the same smile every time she saw in her training,
but she won't be to predict a smile that is particular to a person. Very good. Thank you for your explanations. Sixte, on these AI bricks that you managed
to inflate in your teams to reduce the production time of these car
toons, what
did create? There are two things. There are bricks that on the shelves. The goal is not to develop everything
internally because there are initiatives that are excellent, like yours for
example. But we're talking about dubbing, we're
talking about lip sync, we're all these things, these are things that are coming
to maturity, which are not necessarily specific to children's they are more
generally usable in the gaming sector, in the FX sector, etc. And so as soon as there are on -sta
ge that
are available, we integrate them into our production pipeline, in our production
process. However, there are elements that specific
to children's In particular, and as you mentioned, the fact that there are two
major specificities. The characters we're don't have body
dispositions that are that of a human. They're not humanoids. You think of Paw Patrol Pocoyo, Obelix. Sometimes they look like humans, but
they're not... The proportions are not the same at A quick interlude to thank our 22
D Music
sponsors, the company that has been supporting the actors of individual
production in managing their musical rights and in the musical composition. As well as Toon Boom, the company that
edits the 2D animation, storyboards and production management Thanks to them. Back to the conversation. So, if you use motion capture, for
example, and then you to transfer it or retarget it, which is the output of motion
capture, which is overall... the skeleton of a human and transfer it to
Obelix or P
ocoyo, let's Obelix, it won't if I make the movement of touching my
belly because since Obelix doesn't the same body dispositions as me, it will
create interpenetrations. It's a bit technical, but that's what
makes motion capture often useless or unusable, or in any case, not very useful
in the process of making animation for children. The second specificity is that the
movements characters are not realistic. And that's also what makes the animation
beautiful. There's a graphic style, there's a
creative style. To get involved in a story. Exactly. And it's not just a human who does it. There's no actor who would be able to
perform and walk. For example, when Sterex walks, he walks
by throwing his arms in the But there's no actor who would that. We would be crazy. And so, the goal is to be able watch how,
one, we arrive. to take into the body dispositions that
are specific to the character and how to take into account the creative style and
the style of animation that is also specific to
the character. And so these are things that are specific,
one, to children's and two, to intellectual property and to the character
we're talking And so we come back to the subject of database. And all the models, we take the text -to
-motion models, so models where you can prompt a movement, the outputs, and
humanoid movements, so human characters. All the movements generated were trained
on databases, by people who made motion capture. To make it little more the AI trains on
huge which will b
ias the output result. Can you explain that to Can you make it
than me? I can I can take the example of my thesis
because it will be completely in that context. So I work on the generation of body
gestures from speech. So, we're the rest of Dynalips. And so we're going train a model from
data. In my case, it's data from a person who's
going talk with motion capture. And the artificial will try to understand
the link between speech and movement, for example. This is a precise case. And as it trai
ns, it will understand that
every time I make a word, I may make a gesture. on a sentence where I'm happy, I'll maybe
make more gestures, that kind of thing. And once trained, normally, if everything
goes well, if we give her a similar entry, she will make a gesture that she has
already seen. So we are on a of the reproduction of
behavior that we will see. And as you say, it will be a bit of
because it is rare that an artificial intelligence can predict things that it
has not seen. Or if she doe
s that, it means that it will
be a mixture of things that she has seen. But we will never be able to get out of a
frame that ... What Sam Altman called the hallucination. That's exactly it. We're at beginning of 2024. For those who would watch this show later,
2023 was the flourish of all generative AI, and especially the arrival of Chad
Dipiti, who really democratized the textual And not only, because with Dali,
we're starting to enter the image generation. And then there mid -days. Runways, Pi
caLab, etc. There are even Google initiatives where we
start to get into image animation on a whim.
And it's starting to get serious. I always have this question, and I asked
to other guests, by Do you think that at some point we will see films only
generated by AI, whether it's in the story... in the image, in the animation and in
everything that is in the work. Is it possible one day? Technically, or do want to watch this kind
of film? That's a good question. I think there is an ethical, moral
point
Technically, if you compare mediocre films, or something that could be done in in 5 years, the answer is yes. We would be able to reach a quality level
that is good on a certain number of elements. We know that today, to make it you have
two types of models. You have two schools today. You have the 3D AI, so you the text to
motion. So when you want to stay in a 3D
environment, it's key for the world of gaming, and especially for the world of
animation today, if you want to have controllab
ility of your movements,
controllability of the creation you're going to do. And then you have other models. which are text to video models. Runway, Picalabs, where you prompt and
you've got a video. The problem today with text to video
models is, one, the consistency of the frames, so images, because it's a
succession of images that form a video. 24 frames per Exactly. If need 24 frames, it has to be
consistent. And that's what makes us have weird things
coming out at the end, and that the dura
tion is limited to a few seconds. We can't do longer things. And the second thing is... the controllability of your video output
compared to your product for the reasons you mentioned. Today, for example, we chose to stay in a
3D environment because what is key is to be to manage franchises and manage brands
when we have Pocoyo. We don't want to create something that
looks like a pokoyo or that becomes a weird animal afterwards. And so we want it to pokoyo from start to
finish. And that's why we
don't use text to video
models and we focused on text to motion models or on animatics to motion models,
animatics are the drawings, the rest of the storyboard, before the creation of
animation, to be able to answer these problems. of controllability and consistency of the
image. But honestly, in 5 years, will we be to
have something that lasts 30 seconds, for example, in text or video? 45 seconds? Possible? It actually all depends on the computing
power that we are able to do, because at the e
nd of it all, it's just a story of
computing power and data base size that you are able to absorb, that's we call the
memory of these models. You, as a researcher, are you... Do you see the emergence of all these
generative AI as a good thing and do you manage to stick to the page? Because it goes very fast. Two questions. Two questions in one. Yes, yes. The first one, first of staying on the
page, no. It's impossible. There are really a lot of labs that
release papers every You were talking abo
ut the Chat GPT ,
every week there was a new version, a competitor who was producing a new model,
saying that he is better, he is better, he is better, more consistent, and every week
we couldn't keep up. So no, it's... Impossible. It's super fast, and I don't think going
to stop. And the first part was about... About, do you see this emergence of
generative as a good thing? Yes and no, I'm shared. On the one hand, I'm in the research, I
love seeing things move. On the other hand, there are stil
l ethical
issues, as you said. Especially, the main issue is the right of
data, where it from, the use of data. You mentioned that people can be afraid to
lose their job, or that kind of thing. So, these are issues that... It's like when we went from silent cinema
to speaking cinema. It's something that is so disruptive that
yes, the law never thought we could do it. But it's not because the law never thought
that the law would not evolve. And I think it's precisely us as a pioneer
company and a
ll the companies that want to put a link to be able to look at how we
recreate a legal that allows for all the relevant parts to the authors, the
artists. the writers, how do we manage to recreate
the value? In each technological disruption, in each
profession, there has always been this question of saying, there is a problem
with sharing. We had the same thing at beginning of
Spotify. Exactly, I was going to about audio
streaming. Indeed, it shook the entire music
industry. And yet today we hav
e a model that, little
or no, But they still managed to redistribute the
value to the artists. Spotify is not profitable, so the value
for the 100 is not how it is really related. In any case, the artists make revenues
from their streams. After, there may be a problem of
distribution between the artists of the long tail, that is to say, who are little
listened to by those who are hits. And there may be another rebalancing to be
made. But to return to the legal framework. In the US, we're taking
the subject into
our own in Europe too, but not necessarily on things related to creation. In the US, I think it's the judge William
Orrick III you'll go and check it please, who tried to condemn Mid -Journey and
possibly another runway. Or they're thinking about how to include
the right to intellectual property. for all these artists on which the
training data of these AI are Because it's true that today I am able, with
Midjournée, to create an avatar of you two from a photo saying I want it in
a
Simpsons style. But I'm not the creator of Simpsons, I'm
not the creator of characters, I'm not the artist. And the artist himself, even if I use this
image, a commercial use or other, doesn't royalties. There's a problem with that. I agree. There's a major problem with that. Today, it's a black box. You don't know on what data it was really
trained. And I don't see why a runway or a Chat GPT
would open the box. Because if they open the box, the only
value of these boxes is the fine -tuning o
f existing models, the loss variables
they have, etc. So if they start to open it, I have
trouble understanding how they would have a competitive advantage over others. And then, you take the example of the
Simpsons, it's very simple. Obviously, if you do it the Simpsons
style, it's recognizable. But in fact, you want to generate, for
example, this cup. In the style of The Animashow No, but you
see, how can you find... The fact that it's the video of The
Animashow that was used to be able to reg
enerate that. That is to say that the only way to do it
is to make open source, that's why there is the question of... Well, it's to make completely open source
the data, the use of data and models. Today we are far from Vast subject. No, but in fact it's quite complex to
crystallize, finally, all thought and all... the moral and economic issues around the
AI. I feel like we are... It's too embryonic, in to decide today. Well, you're on that wave. So it's still... It attracts attention. I'm goin
g to to a subject that is also a
of company management, because you are both entrepreneurs. I would like us to come back a little bit
on that, on your careers, each one. Louis, on your side, you're an engineer,
well, a researcher engineer. Doctor. Doctor. Even more so. Doctor, excuse me, doctor. But... What has in your life since What have you
understood? No more free time. I don't anything. I'm a thesis. Did you free time before? Not really, but doing a thesis and a
business in parallel is spor
tive, but it's exciting, on the other hand. Especially me, who works at the level of
my research in the same environments as what I do in terms of entrepreneurship. It's super exciting. However, I had to train a little, I had no
skills in it. So that's where I was incubated, so to the
incubator Lorrain, which allows to accompany project bearers in the creation
of a company, which provides them with help in terms of understanding everything
that will be finance, research, subsidies, investments,
all that. It was fruitful to do that. Yeah, yeah, yeah. And I wasn't alone either, as I said, we
are three co -founders. The one who was originally Slim already
had some entrepreneurial elements for Dynalips. So we didn't leave anywhere, we didn't go
all alone, we knew pretty much what we were doing, but we had to go and we threw
ourselves into it. In any case, it's very encouraging to see
that there is still the will, even in an industry that has been installing people
who want to There are peo
ple in front of you.
Yeah, that's it. And you, on the Sixth side, you used to
the founder of Nestor, the co -founder of Nestor, we met at this time, a long time
Today, there is an image of how you got into this industry, what happened, with
whom did you come this? A lot of questions, indeed, I launched my
first company at 21 years old, Nestor, so a very different we were delivering meals
at the office, it was exciting, it was incredible, it was an adventure. I started with the Food Tech in 2015,
the
company grew, I sold the company to the group Elior in 2021, and in of 2021, then
I managed the transition for 12 months. And in fact, when you have an entrepreneur
like that, who in the skin, you don't want to stop at your first company. So I had ants in my legs, I wanted to
start my second company. In the tech ecosystem, most of the
companies today are B2B, SaaS. Honestly, I don't shit I need a company
that speaks to It's completely stupid, but I need a company that speaks to me, a
compan
y that has an impact on people's That's why Nestor was perfect. I saw the impact every day at lunchtime. And I still have three kids in their early
ages, who are between one and five years old. And I thought, if I can make a box that
makes me a hero for them, that's great. And I found myself that... When I showed the cartoons on YouTube,
with my phone, I thought, there's something wrong. It's like... Every time I try to watch an anime, I
never had the reflex to watch TV, I never had the reflex,
I had my first reflex, I
take YouTube and I go. And in fact, from there, it's where the
vision and conviction of an image was which is to say, fact, the next franchises
of tomorrow, they will be born on YouTube, they will be born in digital, they make us
have a radically different and that we create a new model. And I met Greg like that, Greg who was the
former Global Managing Director of YouTube Kids. and who had built the whole Kids ecosystem
in the world outside the US. He was at the forefron
t of these issues. He experienced this shift in the media
consumption between TV and YouTube. That's how we created an image. It was created in May 2022. There is also Gilles Gaillard, who was the
former CEO of Micros, who gives him all the expertise in the production industry. In the animation industry. Exactly, in the animation industry. And so we launched this in May 2022. We needed enough capital to deploy, to go
and buy the right properties. And so we raised after a few months 100
million t
o go both... buy the best properties in the world and
transform them into franchises and invest massively in the the AI on the objects
that we have mentioned to be able to create a tool that allows to solve what is
today a major problem for animation producers currently which is how I finance
my show in a world where TV channels no longer have enough funding capacity since
they have fewer audiences. And less advertising revenue. And less capacity to finance. There's another business, the SVOD, a
nd
all the streaming. Netflix, HBO Max, and many others that
share market share. Yes, there many. There was a bit madness that was made,
where everyone thought it was the new golden age. and that we were going keep the TV, and
that the market was exploding. What happened in 2022 -2023, where
streaming platforms have invested massively tens of billions, there, for a
year, it's not the same story anymore. Warner has frozen costs, Netflix has
considerably reduced its content spending, and so there
was a kind of euphoria on the
animation market, where people, where everyone thought that the financing
problem was not the same, and that if France Télé no longer fully financed the
show, It was enough to go back to Netflix, today
it's more more the case. And so what's the problem today is that,
one, if we don't have digital audiences and we're not present on YouTube, by
definition we cut off part of the audiences and therefore part of the
advertising revenue. YouTube is 30 % of the audiences t
oday in
the US, for children from 2 to 11 years old. And two, if we don't have a manufacturing
tool that responds and that works in the digital economy and that allows us to
exist on YouTube and exist on other platforms, We are stuck in a model that worked when
TV was predominant in the audience's case, but that no longer works in a world where
YouTube and the mainstream platforms are predominant. It's economic equation that doesn't hold
up in terms of production and distribution. There is anoth
er industry that we often
talk about on Zia Lema Show, it's video games. And on video games, Dynalips, you already
have customers who use your... It's in progress. Nothing definitive. Not yet a game released that is on our
technology, unfortunately. Maybe next year. And it's this video game industry that
also has IPs, it also has big audiences more and more. And I'm even going to further. I recently a discussion with the two
people who are in charge of sound and video here, who told me that they
could
spend evenings connecting to video games. streaming live and meeting up with friends
who are not necessarily in the same geography and that's how they're going to
spend their evening. I may be a big boomer, but for me it's
something that surprises me. Before we went to the movies with friends,
etc. Now we connect on video games. You know, 10 years ago, I was surprised
that my generation would spend their evening on YouTube watching people filming
themselves with their webcam. Yeah, wherea
s today there's a whole
YouTube industry, absolutely. So, in the video game have seen other
artificial intelligences that have marked There are many, especially the text to
motion or the speech to motion. The generation of animations, it's
starting to become artificial intelligence, in same way as in the
analysis, but on movement. GPT is also starting to integrate. The generation of NPC dialogues, instead
of having three predefined answers, we can naturally speak to a NPC. A NPC that is... Non P
layable Characters. So basically, you're playing an
exploration in an open world, typically, we're to take GTA, maybe next year, GTA,
and you meet a character, and instead of having a dialogue that's pre -designed,
like it was before in a video game scenario, there he is able to answer a
little bit of and anything depending on what you send him. That's starting to be a bit crazy. It's ultra, it's hot. It's ultra crazy. Imagine, it's so much, it becomes hyper
-personalization. But it will happen
in animation too, in
general. For me, it's the dream of every kid and
everyone to say that that's it. It was the Bandersnatch you had on
Netflix. You imagine, you make a Bandersnatch, but
not with these predefined stories where you are the makes the story. It's crazy. No, but that's why it's going to enter a
dimension. So there is that. The generation of dialogues. Dialogues, the generation of voices too. We had seen prototypes where you could
copy the voice of the player. and it was also in the
game. So it's not for the immediate. It's the ultra -personalization. And there is also the generation, at the
moment when they create the games, the generation of environment. It's like prompting, where you say I want
to make a Jurassic Park or at least inspired by Jurassic Park to make a battle
game space. It's possible to do imagine the time you
spend on the production. It's extremely... Especially because it's very expensive to
make a game. So, video game, animation, VFX too. We're getting
more DIAs will generate
things. And we talked about it last time with
other guests about the fact that potentially tomorrow we'll be able to use
avatars, exact avatars. to the same level as all the superstars
we've known in our lives. So, I know, De Niro, who in a few years
will maybe disappear, unfortunately. But we could keep making movies with him,
and maybe the 8th with De Niro, and his game, his voice, and all that. And is that... I don't know if we want that, but it's
totally possible to d
o if he himself gives his agreement. So, that's All I mean is... It's the prospect, but it brings us back
to the fact that you are at the beginning of this revolution and I wanted to know
what were your biggest obstacles? Obstacles that could be of technical order
or adoption or things like that. You have technical every day. It's the heart of the thing, we were so
happy at beginning that you get a So you arrive and you try to go on a path, you
see that it doesn't you go around, you go on the ot
her path and you test other
things. So that's a lot of obstacles we have, but
that's what makes the beauty of these models. I think that today in animation there is
indeed an obstacle in being able to adopt these techniques. because I think the biggest challenge is
not so much technical as cultural, but the fear that is now in sector and all that it
implies. Because in fact, it will permanently
transform the industry and will permanently transform a certain number of
jobs. And so that's what cre
ates a lot of
frustration, and we saw it in the US with the strikes of the authors on the subject
of AI. and that's what makes the adoption of LIA
in the sector much slower than the technological capacity that LIA has today
imposed in animation films. Okay, so the first obstacle is the
adoption of people. Is that something you feel? I agree with you, as soon we talk about
our technology... People will say, you want to steal
animators' jobs? And we'll say, no. On the contrary, we did some tests w
ith
animators, they're very happy to have our technology, it takes away everything they
don't to It's true, there are lots of mini -tasks that are very chronophagous, or
debilitating, that can be reduced thanks to grasping. When we put technology in their hands,
they said, I won a week of work, I'll be able to focus on the AI, it's super cool. We also met video game or independent game
developers who told us, with your technology, I can incorporate words into
my games, because otherwise I wouldn
't it's too complicated, too expensive, I
don't the skills, now I can. So it also opens new ways on that. But you have to go in contact with people,
you have explain to what technology is. And that's where we, for Dynalips, lack
visibility, because people don't necessarily know When they meet us at the
first time, they say, they want to us. And then when they learn to know each they
say, no, it's great, it's perfect. It's funny, there's always this
psychological barrier to overcome. Yeah, yeah.
In fact, the subject is completely
passionate. You use the word AI, in fact. It's very very clever. Today, you have two big schools,
especially in animation. You a school of people who don't want to
about AI for all the reasons we mentioned. And some reasons are good, you see, on
legal subjects, etc. And then you have another school that
says, in fact, AI will not steal your jobs. It will transform jobs, but AI by
definition... In fact, I think that... and a tool that is well used here can do a
lot of good to the industry because it allows you to put resources back on
creative subjects, on naval subjects rather than on technical subjects or
subjects that were much more chronophagous, which did not necessarily
bring a little less value. And that's our thesis. For example, Guillermo, who is the
creative director of Pocoyo, or even Billy, who is the creative director of R
.A. Kids. Well, in fact, they are always in need of
tools because they have so many tools that allow them to give life
to their
imagination and their creation much faster. There nothing more frustrating when they
imagine or when you draw on your storyboard than waiting six months to see
it on the screen, while it took you a few minutes to do it and you just want to see
it and you want to iterate. Because in fact, maybe you realize that it
doesn't And the fact that it takes time is what makes your iterations. in creative are much weaker and that
suddenly, especially when you have budget problems, what is the fir
st thing you
remove? The re -take of animation, so the creative
iterations. And so the AI allows you, at the same
budget, to be to do more re -take, so more creative iterations and so in the end, a
better quality of your show. They sell us things well at Sixte. There is a stake rather on the
entrepreneurial path. The fundraising, so this fundraising is
pretty impressive, 100 million is nothing, there is debt, there is equity, you have
in your investors, Kima, Xange, Daphne, we have an American f
und, Leftlane, Resonance
too. Today, the discourse that resonates with
them in terms of thesis, is it the side, we are here to create the IP2 tomorrow,
and in fact that's what they buy in fashion. Maybe we have a future Disney or
Illumination and why not give him the means to fly away or is it the side I'm
here to reduce the costs of production and therefore the manufacturing time in the
animation? No, it's really the first one. I do not believe in tech as such, just
alone and we are not a 100 %
tech company. Tech is there. to bring value to the IP, and therefore to
intellectual property, sorry. It is there at the service of the
property, it is there at service of history. And above all, we are a company of
intellectual property, a company whose job is to tell stories, to tell the best
stories and to create the best brands. And so what interested our investors is
actually this whole thing, this model and this new model that we are creating. We are trying to create what Disney 100
years
ago, in their capacity to be able to build, develop, produce, distribute franchises and stories
that have made generations laugh. But we try to apply or try to rebuild on a
world that has become completely digital, on a world that has become completely
fragmented in terms of audiences, where you have to be on TV, but you all the
tools, including gaming, which is a major part on which you have to be present, and
in which the AI is the only way today. which is given to the owners of franchises
to
be able to run this multi -platform strategy by being able to produce more
content, more often, and optimized for each platform and personalized for each
audience. And so that's the whole thing. Which takes, which makes the right
cocktail. You saw the news like what, I don't know
if it's done or not, but Disney that takes... a participation in Epic Games, the editor
of Fortnite, and they also put a part of their IPs in the game. I saw things coming, we heard about that,
but it was quite amazing
to see the trailer. I don't know if you're a gamer? Totally. That's why I recognize myself in the story
you told. When you see worlds coming from our
childhood, or... or Star Wars, things like that, that
appear in a game, what does it do to What does it evoke? It's great. Yeah. I mean, it's the convergence of the two. And what does it evoke for you in the
future? What do you... Well, I think it's great because it's
clear that sometimes we're going watch movies in the movies and then we say to I
would like to go into this universe, do things a little bit more different. And video games, I think it's a good way
to do that. Where, precisely, we can project ourselves
into a universe that we like. And that's where artificial intelligence
will be necessary to help the production of these contents, which are still
extremely long. A game on Star Wars... I thought Star Wars The Old Republic. It's magic, these things. If we had that today, with content
produced by an AI we would never finish th
is game. If I told you about dialogues generated
automatically, you could have that for all the characters in the universe, and you
would be at the heart of the movie of your childhood. In fact, it would be that Apple buys Epic
Games and buy Disney, and there we are with Apple Vision, with a world of Star
Wars and in fact we never leave this world again. But no, but in addition there are papers
of prospects that indicate that tomorrow there will be a whole economy in this
virtual world. And that
's kind of case today, that is to
that you have online GTA that are played on Twitch, or there are transactions that
are operated. Roblox is a parallel You have games, I don't know how much it
generates. I think the first ones generate several
tens of millions per year, like 50 million first. The first game on Roblox? Yeah. And on top of it's editors who benefit
from So there's a whole economy with positive external effects to take economic
but basically that's what's going on and it's pretty am
azing. In any case, I'm delighted to have spent a
little time discussing AI I think it's important that we talk about
it on and not put under the table. On the contrary, opening the discussion,
I'm very happy to have you. Thank you very much. Thank you. I wish you the best for your companies. Thank you. Goodbye.
Comments