FYI - For Your Innovation: How South Park-Inspired Generative TV is Revolutionizing Content Creation with Edward Saatchi

Welcome to FYI, the four-year innovation podcast. This show offers an intellectual discussion on

technologically enabled disruption because investing in innovation starts with understanding it.

To learn more, visit arc-invest.com.

Arc Invest is a registered investment advisor focused on investing in disruptive innovation.

This podcast is for informational purposes only and should not be relied upon as a basis for

investment decisions. It does not constitute either explicitly or implicitly any provision of

services or products by arc. All statements may regarding companies or securities are strictly

beliefs and points of view held by arc or podcast guests and are not endorsements or

recommendations by arc to buy, sell, or hold any security. Clients of arc investment management

may maintain positions in the securities discussed in this podcast.

Hi, everyone. Welcome to another episode of For Your Innovation by Arc Invest, a podcast on

all things related to disruptive technologies. I'm Andrew Kim, research associate covering consumer

internet and fintech, and I'm joined by Nick Groose, associate portfolio manager and Angie Dalton,

advisor and theme developer for Arc, as well as CEO and founder of Signum Growth Capital.

Today, we have the great privilege of speaking with Edward Passacci, founder of Fable.

Hi, Edward. Thanks so much for your time today. We greatly appreciate it if you can

introduce yourself and let us know how you ended up where you are today.

I came to the US from England in 2007 to join the Obama campaign, created social and data tech

for the campaign. And then after we won, turned that into a startup and started working with

different campaigns, administration, nonprofits, Kaiser Permanente, the National Health Service,

all around visualizing big data for people in an org so that they could be more competitive

and accountable. And we worked on the reelection and then we're acquired by the main big data

provider, NGP Van, in the space, worked for them for a little bit, but I fell in love with

virtual reality. And I moved to San Francisco, found two friends from Pixar, a shorts director and a

technical lead, and we created Story Studio, which was then acquired into Oculus and became

Oculus Story Studio, a $70 million project to make virtual reality experiences and tools,

including Quill, the painting animation platform. We won two primetime Emmys for the

experiences we created in the Peabody. And our last piece was a Neil Gaiman adaptation,

starring a character called Lucy, who thought that you were her imaginary friend, so she'd

talk to you, guys would do things together. And I noticed that people were having, you know,

real considerable conversations with this character, even though she couldn't really

respond to you in any way. And I started to really think about, you know, what does it really mean

to bring a character to life? And, you know, what if we took that to the limit of AGI, where it's

a true being that exists beyond you and has its own independent life. So, spun out to create

Fable, whose goal is AGI, a truly intelligent person, and started to build the virtual beings

community around that. So, bringing together folks from Epic, a CTO Kim LaBrari, and Lukash

Kaiser, who wrote the one model to learn them all, paper, as well as folks from Pixar, to try to

create a community around AI people. 2018 to 2020, we focused on single agent chatbots,

so you're one-on-one chatting. And we unfortunately found that this is very boring. It's not much

fun talking to an AI. There's very platitudinous, a lot of small talk. They don't have any real life.

And so, we felt very panicked of, like, will we bet the whole company on this?

What the hell are we going to do? And we started to come up with the idea that these AIs, you know,

should be like us, and that they exist within a context of family, friends, enemies, rivals, lovers,

all the rest. And so, we created the simulation. So, this is an AI world, only for AIs,

where they can live and grow and have ongoing lives. And more recently, we had quite a lot of

attention for releasing a research paper on something that we've built for the simulation,

which is called a showrunner, and the show one model. And that's all around infinite story.

And the reason it got a fair amount of attention is because we use South Park as an example,

and it allowed you to, just through a text prompt of a sentence or two, generate scenes

or episodes of South Park based on your plot. You also could take a photo of yourself, upload it,

and speak for a few minutes, and be a character. So, you know, I created an episode of myself,

sort of a joke version of Fatal Attraction, where I'm sort of, you know, stalking Cartman.

And, you know, he's like, this guy's completely insane, and we have a big fight and all that.

But we also did one with Mark Andreessen, who discovers that Elon Musk has spirited away

all of the top actors in the 90s, and that since then, all the top actors have been deep faked.

So, you know, basically any idea you had, you could turn it into an episode.

And I think that captures the imagination around genres of TV that will come home,

and we'll say, I want this, I want a new episode of the show that already finished,

or I want to put myself in an episode, or I want to create a new show,

and it's going to mix together these elements, and this is the original thread of it.

So, I think that was very exciting to people.

It was certainly very exciting for us to see that it really captured our imagination, as you said.

So, I'm curious, what goes on behind the scenes? Maybe you can explain this to us

in more layman's terms, but you have these AI agents running around, you're prompting them.

How exactly is that working? I think most people are familiar with NPC

characters in video games, and they get this concept that you can have characters

that have a bit of context to them, they have a storyline.

But obviously, that's all pre-built in. What you're doing, it sounds like

it's much more advanced, it's more generative. So, we'd love just a little overview of exactly

how all this is working within the show one, and showrunner agents as well.

So, what we've seen so far at the intersection of AI and games is essentially

dumping an LLM into each character. There's this little backstory, and it's basically the

same as a single-agent chatbot. It's just lots of single-agent chatbots.

And that has the same problem, which is very boring. And I think that's why there are no games.

Even though we're five or six years now into GPT, there are no games that use AI NPCs,

which is kind of crazy when you think about it, right? The reason is not that the quality

isn't quite there, it's that it's not really a good idea. It messes up the pace of the game,

it's not appropriate, do you want to talk to the tab and keep it for 300 hours?

It doesn't really make sense. You want to play the game to complete the quests and do things,

and creating a massive system just to make secondary NPCs talk more is not a good use of

resources. So, our approach is quite different, which is instead of, let's say, the nature approach

of, we're going to give all these AI's backstories, and that's going to define their dialogue.

We're more on the nurture approach of, yes, these AI's are going to have backstories,

but really their personality and all the rest is going to be shaped through experience

and what they're doing, and their backstory is something that they will generate through

their actions, just like us. So, that's how we think of the simulation is that it's more about

behavior and less about a written prompt for an individual's personality. The show one model

is a little different, which is literally a model to generate TV shows of what's going on in the

simulation plus what a user wants. So, that is text to scene or episode. The stuff in the simulation

is operating off a different model called Gaia, which is more around behavior of AI's. So,

for Gaia, think a little bit about Westworld, where Anthony Hopkins is interrogating the AI's

saying, why did you do that? You know, you were speaking to this guest, but the way you responded

was very outside of your character. Explain that. So, players will be able to interrogate the AI's on

their decision making and rate them on the appropriateness of their behavior and continuously

train a model around appropriateness of behavior and simulations. So, you know, we're targeting two

models on modalities that people aren't really focused on. One is behavior and the other is

infinite story, whereas people seem to be much more on what we would think of as quite granular

things like generating three-second clips or videos or dumping LLMs into NPCs, I don't think.

I don't know that either of those are the right way to approach it.

I think in the paper that Fable published, you showed this giant value chain, right? That kind

of describes what's happening from the simulation to the end TV scenes, right? Output. Could we just

dive a little bit deeper into what's exactly happening after you sample certain events within

the simulation? How does it actually end up as a cohesive show? Yeah, so there was some confusion

about this. So, it's not recording. Think about an episode of Friends, right? The conceit is that

it might represent what happened in a week of the Friends life, but it's not recording 168 hours

of all the characters walking around and then editing it down. And I did see that some people

misinterpreted that. What we've created is not that. And that would be a bad idea, because that

would not lead to any kind of coherent narrative or anything particularly interesting. And I don't

know even also what the value would be of figuring out a system to edit from 168 hours to 22, like

wouldn't be the right approach. So, we take some things of what's been going on in the simulation,

but mostly what we're taking is your text prompt of like, what do you want to have happen this week?

So, you kind of say, you know, this week at a high level, let's say Springfield or something,

I want this to happen to Homer and to Bart. There's going to be tons of other stuff that

happens within the simulation of Springfield, let's say, during that time. But you as a user

are defining the kind of important bits. And the show one model understands story structure,

A plot, B plot, you know, the core narrative. So, that was a misunderstanding that I saw people

had is like, are you recording everything and then trying to find story within that? And that's

not the right approach. It's more about you take a prompt for the episode, it understands what an

episode is and understand story structure and all that. And it generates the episode. And then that

can take a little bit of what's happened within the simulation that week. But when we tried to do it

purely on what goes on in the simulation, like let's say you tried to make a reality TV of the

Sims, very boring, right, like it's just very primitive. And so that's why either the simulation

generates the prompt, right, and says, here's what's going to happen to the characters this week,

the major stuff that's going to happen automatically in South Park's case, a mix of like, you know,

what it thinks should happen next, plus topicality, because that's quite a topical show. Or the user

does it, and you know, it generates is the kind of major events that I want to have happen this

coming week. If you think about that, that human piece of it, there was a decision that you made,

Edward, and I love the timing of it. Because I know there's going to be a really interesting

answer here. But the timing was key in that we're in the middle of the biggest strike in Hollywood

in 60 years, as you said. And you're introducing this idea from a human perspective of this deep

big technology. And, you know, in the midst of, you know, actors and actresses saying,

even if they could, they doesn't mean they should. Can you just explain kind of your

thought process on the timing and then how you view IP in this context? Yeah, I think in any

negotiation, having the side that has more information is in a stronger position. And so

people negotiating without the context of what's possible, I think is very dangerous. So I think

it's great that, and I hope that this will provide more leverage to writers and actors,

although more writers in this case, to say, you know, AI can only be used if it's coming from

the artist, producers can't use it. And to fight for that, this is the moment of maximum leverage,

biggest strike in 60 years. Now is the time to negotiate the best terms. And I think a lot of

people are quite disappointed looking back on the deal struck around streaming 15 years ago and had

someone from streaming kind of, you know, whatever whistle blown out, just how kind of dire this

could all be and explained, because people at the streamers would have understood a lot better

than the negotiators on the other side, what the negotiators were giving up, I think that would have

been much better for the strikers back then. So the more information they have, the more aggressively

they can negotiate, that's our view. So hopefully it gives more leverage and clearer negotiating

paths to say, producers cannot use AI, only artists can use AI if they want to, if they don't, then

it's not going to get used. And that's what I would suggest is very strict bands on on AI,

if it's not coming from the artist. So the purpose is to arm humans, human artists, not to,

not to be not to basically create episodes with protected IP.

No, it's not just that I think it's it's also about it's not just about the the particular IP

of South Park or something like that. It's also about the entire structure of Hollywood,

which if a single person can create an entire TV show, that is massively disruptive. And

maybe there should be safeguards in place for for that to be done slowly and gradually and not done

suddenly. So I think that's sort of that's what I would go for. I think there is a whole other

question, which, you know, our sense around IP that currently exists is that if the people that

we're we're talking to, you know, do end up using the show runner, then I would see a world where,

you know, let's say you're a huge fan of a particular show, you would go behind a paywall.

You would sign a terms of service that say, I don't know anything. I'm excited to remix. I want

to make my own South Park episode, whatever, but they own, I don't know anything just because I

made it doesn't mean I own it. The IP holder owns it. I'm not going to share this outside of the

paywall site on pain of cease and desist and and fines. And but I can I do get to see everyone

else's episodes. And I do think that 1000 true fans of South Park, creating, let's say 10 episodes

each, which would take a couple days, maybe, I don't know, they're creating to 10,000 episodes,

I do think that the top 10 top 10 of that 10,000 would be as good as anything written professionally.

So I think that is a profound shift. And I think now is the time to,

to control it, because I think it is, it is very disruptive. And that's just for IP shows. I'm

personally more excited, obviously, about original. So people using the show runner, somebody has a

show Bible, they've been dreaming of getting the show made for years, but nobody says yes, it's

so expensive, or, or the people I'm really excited about are the very difficult people. Maybe, maybe

being good artists and being a good hustler, which has always gone together, maybe they shouldn't,

right? And maybe the production of TV shows and movies could be more like novels and paintings,

where one completely insane person can express their vision. So I'm very excited about that.

And so I don't want to, you know, I do hope this gives leverage to the strikers, but we're also

in a different industry, right? There are two, there's Northern California and the Southern

California, and we're never going to be friends. And we're in a different industry. But, you know,

that particular industry, I think should, that I love and admire,

but I'm more in the tech industry. I think they should be as aggressive as possible, because

any producer will just think, oh, God, we could get rid of everyone. Now, great, because they're,

you know, the enemy. So we should, we should limit them as much as possible. But it is not,

that's not our industry. And so we release this with our own interests and excitement. And personally,

I'm very excited for original things to be made with this, original stories and shows,

and for one person to be able to make things which has been impossible for the last hundred years.

Edward, I want to circle back to the more technical side, because you've been building

towards this release for a while. So what were some of the hurdles in getting this ready for

launch? And what do you think still remains in your ultimate goal of, you know, bringing this

and allowing people to create these shows on a one-off basis? Where, you know, where are you in

this, in this, in this journey? So it all has to be animated for a long time. It's not, you know,

it's not going to work with live action. And if you're using metahuman, I think it would be too

ugly and poorly, poorly animated. So it's going to have to be cartoony. Let's say that at the bottom

of the complexity pile of South Park, a little above that is anime, almost anime, that's quite

static and, you know, could be achievable. Above that would be something like The Simpsons, King of

the Hill, Futurama. And then all the way at the top would be something like Arcane, which is a

beautiful 2D Netflix show. I think that's many years away. But I do think in the next, over the next

kind of six to 12 months, we can go from South Park to an anime level of animation, whether it's

anime or not, whether it's just that level of staticness, and up to Futurama and King of the

Hill and the others. In terms of our present process, we are for our own benefit because we

just love it. So we're not going to release this one doing Star Trek the next generation,

mostly because I really want Captain Picard to say, Ensign Sachi, could you go to the

transporter room and join the away team? And that would be so cool in an episode of Star Trek.

And also, I want more episodes of Star Trek the next generation. So that's one thing we're doing

just to kind of test it out in a non comedic thing. And also to have a simulation that's a

spaceship is kind of cool, like there were 1200 people living on the Enterprise, civilians and

and military or I guess Starfleet. And I think that's so cool to see a simulation of the Enterprise

day by day and week by week in terms of what they're doing. After that, we're going to do three

originals. One is a satire of Silicon Valley with like Mark Andreessen in the simulation and kind

of going to whatever bully Peter Teal and all these little insane kind of founders talking about

saving the world while stuffing their pockets with money. So you know, we can do a little satire

of that. And a TV show around Silicon Valley, one that's around Starship and one that is a detective.

So those three are going to teach us about what happens when you don't have a show Bible,

how do you train it when you don't already have something that exists and what information is

necessary and what information is not necessary, probably something like a kind of pilot episode

plus a sense and introduction of all the characters, plus a pilot episode, not meaning

visuals, I just mean written out, plus like a rough kind of formula of what's supposed to happen

each, you know, in each episode works, works with formulaic shows, right, or like reset shows,

episodic, right, not, not these shows that are like 10 pieces of a movie, like a 10 hour movie,

which a lot of shows have now become. This is more like episodic content, it's more suited to that.

So yeah, that's the next, that's the next part is one more, instead of a comic, something dramatic,

and then three originals. And that'll teach us a little bit about, about, you know, how to get

original shows made and start working with creators to make their own generative TV and their own

simulations. I'm assuming you have tested out multiple large language models and, you know,

text image models in building the showrunner system. Could you discuss like how modular is it

today? Like can anyone plug and play use different LLMs and diffusion models? Or do you feel it's

quite specific to whatever architecture you, or whatever third party models you're using today?

Well, we're looking at Lama. We were, we already used stable diffusion for the fine tune model of

South Park, so that my photo could turn into me in South Park, or basically if there, if there are

characters or sets that don't exist in an episode, that it's in the script that's been generated,

it quickly can then generate those characters and stuff. But yeah, I think I suspect, if this is

what you're asking, I suspect that the particular advantage of each foundation model will reduce

over time and that open source will, will kind of keep up or even surpass. And so then, then the,

maybe the age of the applied AI company will become more important than the age of the foundation

model companies will become less important because what they've done is essentially commoditized and

open sourced. But we'll have to see. I say, you know, SDXL I think shows that you can have better

than mid-journey with, with open source, which isn't a new, a new release from stable, from

stability AI. So I think it's, I think it's a battle. It's a very exciting battle.

And on the language model side, please correct me if I'm wrong, but I think from the paper,

it stated that you largely depended on like the GPT-4 base model, since it already knows a lot

about South Park to begin with. Maybe can you talk about how you envision IP holders fine-tuning

on top of models going forward through the showrunner system, what that would look like,

and what challenges remain? Yeah, I mean, we sort of think of it as a sort of

traumatic fingerprint. So obviously, you have to figure out what are the core values of this IP?

You know, what is the structure of it? And, you know, what are they, what's, what's going to feel

right? And also what's going to feel like a proper episode. So we found that we had to go

further than what we were getting with, with GPT and start to figure out kind of A plot, B plot,

story structure, climax, as well as all the visuals of editing and animating. I guess the

objection that I've been starting to think about with Runway and with everything is it's not holistic

and that we should all maybe stop thinking about each modality and start thinking about how to

actually make a TV show, right, and all the different technologies and tools that you're

going to need to combine language just being a small, a small part of it. So that's, that's sort

of where my head's been going is what exactly is the vision over the next five years that like

the clip generation of text to video is going to go from three seconds to 10

or 20 seconds, but the, the mouth isn't going to match the dialogue. There isn't a dialogue,

right? Like, you know, there's no editing, there's no, these aren't coherent. And maybe we should

not be thinking about it from a single modality perspective of text to video, let's say,

but or image to video, but instead start to think about it from the perspective of

a show and work your way back from that. Because understanding how to make a great show

is probably more valuable than understanding how to make, I think anyway, images and, and,

and video, right? I mean, how much, how much value are we really getting from, from, let's say,

firing all these concept artists? What have you done? You've saved 5% of your budget, maybe three,

being able to actually generate an entire show and IP, that's a lot more valuable. So we sort of,

the language model, I think if each studio should, shouldn't think about it as just a fine-tuned

language model, I think they should think holistically, how do they, and they, you know,

not like how do we augment and help writers and that kind of stuff? Like, I don't know if studios

are ready to do that's why I point out that we're different industries. I think the valley may well

be the center of, of filmmaking and TV production in the future. So I don't think we should shy away

from the fact that these are two rival parts of the state and they're always trying to kill each

other and they don't like each other. So I don't know that studios are going to take advantage

of this, let's see. But at the same time, all the valley people are always so arrogant, right,

and they get onto their Hollywood and they're like, we're going to take you over. And somehow

Hollywood is much older than us, right? They've always been able to absorb our ideas, replicate

them and then, you know, remain in control. So yeah, it's an endless war and nobody will really,

will really win it. I was just thinking about the image that I had, thinking about the strike and

thinking about what you're doing. And it reminds me of my chocolate lab, Freddie, who stands in

her crate with the door wide open parking. Because there's a whole world out there of things that I

think can be embraced in terms of new innovations for, you know, adopted, for some of these, you know,

old Hollywood to adapt and really take, continue to have that lead. And I agree with you. I feel

like this shift is going to be felt pretty dramatically after seeing what you've done to

Northern Europe. Well, I mean, yeah, it's a community of artists outside of the producers.

And so I think quite rightly, they're only going to be enthusiastic once they see original,

exciting, dynamic art created. And there is none. All that we did was just remix South Park.

All that are older than any of us see with, with runway or mid journeys, you know, it's,

it's very generic. Often, it's all trailers or what if Wes Anderson directed Star Wars like that

is not art. And so I think that really the only thing that would convince an artist of the value

of this correctly is for another kind of argument would be for something original and beautiful

that makes them cry and is like, wow, that's a real masterpiece. And we haven't done that and

nobody else has. So I think that's, that's the humility is like, all we've done is create

stuff that's sort of just little remixes of someone else's artistic work. And,

but we are going to try and make original things. That's the only worthwhile stuff,

real works of art. So I don't, I don't think anyone, I don't think anyone's missing anything.

They are correctly saying that this is just boring theft and unoriginal. And they'll be

right until someone makes something original that's a work of art. And that's really hard.

So that's, that's, we'd like to be part of doing that, but maybe we'll fail.

Just back on the question topic of, you know, IP, given Hollywood, if we look at how video games

have evolved, right, we saw the rise of UGC, right, with Roblox, I guess now with Fortnite

Creative. And I'm just wondering if studios decide to kind of extend IP usage to end consumers,

like you or me, kind of iterating, injecting ourselves into South Park episodes, for example.

And do you see a future in which, you know, users are not perfectly content with the IP

that they've made and claim no ownership over it. You know, they demand a ref share on it as well.

Right. And what do you think happens to the bargaining power of the original artists if

consumers and consumers who are creating AI derivatives, successful ones, kind of want to,

want a slice of the pie as well? Yeah, well, there's kind of an interesting legal case

around clip shows. So I don't know if you guys remember this, but in order to meet their quota

of like 24 episodes a year, often shows would do clip shows, which mostly used previous clips.

And originally they didn't pay anyone who worked on this. And there was a case and it was agreed

that actually when there's a clip show, you got to pay the director of that episode or that clip,

plus the actor, plus all the rest. So, you know, I could see something where eventually maybe a

creator gets a little rev share, but most of the money should go to the, to the original people

involved. They're the ones who did it. If you're really just remixing and putting yourself in it

and for your own pleasure, I don't think that's, I don't think that's original work and it shouldn't

be compensated as such. In terms of its artistic value, all that I could claim for

generating new episodes of shows that exist or personalization is that it's better

artistically for the world than rewatching old shows that you've already seen for the 20th time.

But I don't think it's as good as even watching a failed work of original art. And I think,

you know, the creators should be compensated in a way that is, that's commensurate with that,

which is that they're only remixing. But some creators won't do it, but a lot of fans,

the ability to create a new episode of something you love, any fan would find that amazing and

wouldn't want to dishonor the creators of it, I don't think. But they could become famous,

right? They could be the best creator. I'm going to stop creating unless you IP older,

hire me to do it, or I learned how to do it, and I got famous and attention from that.

Now I'm going to create my original show, like many YouTubers, O'Burnham,

went from being a YouTuber to directing eighth grade, which is a terrific

movie from 2015, like, you know, these are, these are well understood trajectories

where you start off just doing stuff for free to show off, to build up pure reputation

so that then you can create an original work of art.

I have a question just how this could extend to the video gaming space. You mentioned, you know,

you don't find it that interesting to just, you know, produce NPCs with LLMs behind them.

But is there, because when you, when I hear you talk about generative TV, I feel like that starts

to blur the lines a bit with kind of interactive video, and that's, you know, then leads into

video gaming. So where do you see this type of AGI application extending itself into video gaming?

Yeah, I mean, what we're creating with the simulation is basically a game.

And it's a, it's a 3D world that you, all these AIs live in and they interact in and they do stuff.

The South Park one is top down 2D because it's South box, it's a bit simpler, but we're big

leaders in 3D. So I do think gaming is huge for this, but I guess I think of it a little

differently, which is Goethe said, do you must umber or do you must umber sign, meaning you must

be hammer or anvil. And right now, the games industry is being told that they are anvil.

And that the AI people that said, great, they're, they're hammer, they should be in charge.

They should have the power. And we should just be their client, right? We should just give them

money. But I think the games industry could be a lot more powerful than that. And that a games

company might be the next open AI by gamifying training, right? Like RLHF that is actually

gamified, that's fun, that is part of the story, part of the mission. And so I just think it's

very unambitious and unimaginative to just, uh, you know, have these second, it is cool,

like the secondary characters instead of just saying, Hey, buddy, get away from me. They would say,

like, Hey, buddy, get away from me. I'm a, you know, I grew up out on the fields there. And,

you know, let me tell you my story and what's, you know, I just don't think it's that interesting.

We should be more ambitious and a games company shouldn't just think of themselves as using

AI tech. We should be the leaders and in charge of the whole thing by harnessing the competitive

will of gamers who are convinced that they will get to the quest of AGI faster than companies

or academics by gamifying RLHF. Yeah, that's super interesting. I really never thought about

it from that lens. I would have said, you know, it would be interesting to have a

generative storyline. So I have, if I'm a user and I'm playing in Fortnite, you know,

a lot of focus has been on generative assets and, you know, 3d models. And I think that is

interesting in and of itself, but it's not really what we're talking about, but prompting, you know,

the map to generate a storyline for you. And then you play through that. I feel like that kind of

aspect of it could be really interesting. But what you've said, I would never even have thought

about it from that lens. No, I think we'll have all succeeded as the games industry when

why the hell is Mustafa going to the White House and Congress and testifying like,

should we stop your model? It's going to be when, when, you know, we're all testifying in front of

Congress and they're saying, God, could we, should we turn off your video game? It could, you know,

it could lead to the end of the world. I think that's very exciting. We may as well try for it.

Why should all the others be laughing as, as Oppenheimer or something? I think it's ridiculous.

We've, we know how to do incentive design, gameplay design, and encourage players to do

RLHF and they're good at scraping petabytes of data from the internet. Let's, let's see who wins.

I think it's going to be very exciting. So the games industry should be much more confident and not

just think, oh, we could maybe make some 3D assets. I think they should, they should be the leaders.

It could be pretty interesting to see, given this kind of merger of Hollywood and gaming,

you just made me think of another thing, which is some of these smaller studios might actually get

supercharged by a game company who comes in and says, hey, you don't have any games expertise.

All the big guys do work. We have the games expertise. You have the creation

expertise. Let's get together.

Totally. That's probably why it's okay to live in Northern California because there's not as much

loyalty to another industry in city. You know, it would be cooler if this stuff was all

decentralized and disrupted because maybe it isn't so good that it takes a committee of people

to make a movie and 30 writers. And you know, that used to be an insult that your movie had all

these writers. It's not necessarily better to have more writers on a script. So what's it going to

be like when a smaller number of people who are much weirder and aren't as good at raising money

and making friends? What happens? What's in their minds? I want to see their vision for gaming and

for the rest and so much smaller studios making really weird and interesting stuff. I think that

could be fun. It plays into the decentralization kind of everything theme and just creators to

the front of the line, right? I mean, the whole UGC platform thing, I was so excited to see Gossamer,

the movie being first launched on UEF and players having such a big role in influencing the movie

line. But it's pretty exciting. I think a lot of people who I'm friendly with and

in the film world, you know, do correctly say that this is a collaborative art.

And the great things come from there's this actor who disagrees with the director,

who disagrees with the writer, and they all have to work together to create a masterpiece. And that

probably is always going to be the way to get to the most amazing outcomes. It's also the way in

which sometimes things completely derail. But it reminds me of a line from Larkin, which is,

why did he think addition meant increase? To me, it was dilution. And in some cases,

more people make it worse and dilute a singular vision in some cases. So I just want to see

what that's like. What is it like when one person, you know, who's talented and understands the tools

wants to make a TV show that has not physically been possible so far? You have a history of

of of experiment in that space. If I'm not mistaken, I remember us talking a while ago about

collaborative content generation. And you have some experience there. Did you find the content

was watered down and not as interesting? Or were there some gems in there where you thought,

wow, all these humans in a decentralized way can actually make something amazing?

Well, I don't like that stuff. No, I don't think that works.

No, but that's different to Eliah Kazan working with Marlon Brando and Tennessee Williams.

You know, these are fierce people at the top of their game versus a kind of

here are just people all around the place who kind of vaguely are interested in this medium

and will do some stuff. Yeah, I don't I don't I haven't seen that working yet.

It comes up a lot in, you know, because I spend a lot of time thinking about decentralization

and media and entertainment gaming. And there's so many people who say, oh, it isn't going to be

great when we can have people all over the world writing a movie. And my response is always, no,

that sounds really awful. Yeah, no, I'm more excited. I have more faith in something very narrow,

where it being one person able to bring their vision to life, like that hasn't been possible.

It hasn't really been tried because it isn't. It just isn't viable. And I just have to believe

there are people out there with personalities that are so annoying, that they just couldn't

manage running a set, right, or convincing all their team to work with them. That shouldn't be

the only person personalities I've who's making movies. Orson Welles said, I have to spend 98%

of my time hustling and trying to be nice to people and 2% doing what I love. And that's no

way to live a life. And he's the greatest filmmaker of all time. So I'd like to have seen more

movies by Wilson Welles, personally. Well, Edward, thank you so much for your time and

insights today. It's so cool to dive deeper on what you and Fable are building. And before

we stop bothering you, we'd love to ask one more question, given that we were on the topic of

content dilution. You said you are excited about original content with already existing,

already successful, pre-existing IP and infinite derivatives created by AI now possible.

How does new IP break through the noise, in your opinion?

I mean, I think people always want to connect with another human being through art. And derivative

content is just that, it's derivative. And so that's why we don't only have why we have

Oppenheimer and Barbie original movies by original visionaries crushing

Marvel 32, literally making much more money than them. So I think people,

they want, what's that thing? It's like that madmen thing where he says the most powerful idea

in advertising is the new. But there is a second one, which is nostalgia. And so you're talking

about that with derivatives, nostalgia, but the new is a more powerful concept than

nostalgia. So I think the new will always be very exciting to people.

All right. Well, thank you so much.

ARC believes that the information presented is accurate and was obtained from sources that

ARC believes to be reliable. However, ARC does not guarantee the accuracy or completeness of

any information. And such information may be subject to change without notice from ARC.

Historical results are not indications of future results.

Certain of the statements contained in this podcast may be statements of future expectations

and other forward looking statements that are based on ARC's current views and assumptions

and involve known and unknown risks and uncertainties that could cause actual results,

performance or events that differ materially from those expressed or implied in such statements.

Machine-generated transcript that may contain inaccuracies.