Spatial Computing Catalyst #2: Autonomous Augmented Avatars with Artie

Ryan, you just gotta look at your first avatar now without telling us what you saw because they're not allowed to reveal the character or anything. What did you think? It was good. Was it fun to be able to talk to something virtual on your phone? Yeah. Yeah. So who are you guys?

My name is Ryan Horrigan, co founder and CEO of arty.

I'm Armando Kerwin, the other co founder.

Yeah. And so what is Arty for people who haven't heard of what you guys are doing or seen any of your work yet because your work is not yet out.

Not yet out.

Yes.

So we're a relatively new startup based in downtown LA and we're creating software for content creators to create. Create AI powered avatars. Yeah.

So think of something like a robot, like a R2D2 or something like that.

Yeah. It could be a movie character. It could be a doppelganger for a celebrity or an influencer. It could be a virtual celebrity or influencer. It could be a toy that's come to life. Really, anything. We were sort of inspired by things like Roger Rabbit, like a world of fictional characters kind of co mingling with real people throughout, you know, our everyday lives.

Now I can imagine a world, and I'm sure you can, where there's going to be thousands or hundreds of thousands or maybe even millions of virtual beings. Do you call them virtual beings?

We call them avatars. You call them virtual beings? Virtual humans? I don't know any other names because.

They could be robots or they could be inanimate.

Virtual robots? Yeah, they could. You know, I don't think anyone's really settled on the names yet, so there's probably a slew of names of mine.

It seems like avatar is a term that consumers understand. So we're kind of sticking with that for now. Even though when most.

Is that because of the movie Avatar.

Just people have them from games, you know, like. So it's like if you're. Most people think of avatars as human powered and what makes us different is that our avatars are powered by AI. So we think that can reach the widest possible audience that way.

You know, it's the most scalable solution anyways. I can imagine a world where we're going to have thousands of these. I'm going to, you know, tonight talk to my bank through one. I'm going to talk to software through it. I'm going to go shopping with one. I'm going to have an exercise coach with one. Right. That's going to run with me as I exercise and he's going to kick my ass like my real coach did in high school when I ran marathons. Do you agree with that? And then work backward to 2019, because 2019, we don't have that. That world is not possible quite yet.

We do think that we're going to move to a world very soon where apps don't need to be stored on your phone, where they're going to instant over 5G, so you can just conjure any app that you want and they don't have to, like, sit within the real estate of your hard drive on your phone. So if that's true, we also think that the user interface for apps might change and become more human and more personal. So if you think about going from, you know, the various forms of input we had with early computers to where we are today, we've progressed continuously to more kind of evolved forms of human input. So the ultimate form is talking to another human or human, like, being. Right. So we feel like, for example, rather than having to have the Geico app on your phone, if you get in a car accident, you should just be able to talk to the lizard instantly without actually having to go through an app and click through different pages. So we do believe that whether you go to Disneyland and you want to talk to Mickey and he's going to guide you through the park, or you're going to go to Walmart and they're going to have a special avatar there, or it's your favorite movie character, it's going to be a world of characters that are effectively different apps, whether it's for entertainment or for utility or for training, etc.

Now, I learned by just watching Ryan, who just played with one of your avatars, that kids are very willing to pull up a phone and see something virtual on the screen, talk to it, move around it. Right. And what else do you want to do with an avatar?

I would like to touch it and actually, like, feel it.

Yeah. Well, now you need haptic gloves. I know another company building those that.

Might be a little.

That's a little further out.

Not 2019. Right.

It will be possible and you're going to want to do it. Right?

Yeah. By the time he's out of high.

School, just like, ready player one. You know, he had to get his special suit and, you know, have the full. He could feel like other online players. That's right, yeah.

2019. Let's get back in the flow. 2019, what's possible right now, you can sort of see an avatar. You can talk with it, but It's. It's sort of following a script. Ish kind of thing. It's not really having a conversation with me where I could go completely nuts and ask it something. Yeah, well, it's not expected.

I think right now what we're trying to enable is creating like a sandbox. So, you know, if you're talking to a movie character, should be able to talk to you about anything that happened in its films or anything that the content creator, say, movie studio wants that character to talk, talk about. Right. But at the same time, they may not want it to talk about politics, they may not want it to talk about something trivial like the weather today in Los Angeles. They're going to want to keep that character engaging you sort of with sort of a single author, sort of human author, you know, behind all this. Right. I don't think we're going to be into purely generative conversations that can go off into a number of different directions, infinitely for hours and hours, you know, tomorrow. But I do think that that's what we're going to get to. But first we need to reach a lot of, you know, consumers and a large audience, ultimately train the avatars to get better and smarter and learn more, to be able to do more. Ultimately.

That's cool. Today you're working with movie houses and celebrities and influencers. If I wanted to build my own arty, how would I pay for it? How would I get you guys to work with me?

We're still kind of sorting out sort of our model as we're working with our early customers, but it's going to be Probably a freemium SaaS model based on usage. So, you know, like any cloud services, the more you use, the more you'll pay. But ultimately for us, we want to make it free to publish and create, ideally so that any content creator, whether big or small, can use our software.

That's my other son who's autistic, jumping around there. If I wanted to create a robot avatar that would, you know, talk with me, sort of like what you showed me. What kind of assets do I need and how do I have to think about that before I get started with you? Do I need a script of what it might talk to me about or.

Yeah, today. So today, if you're think of it like the YouTube model, actually, like, if you want to be a YouTuber today, you have to, like, learn, premiere and learn how to, like, put together a basic little film and then do some work, you know. So our goal is to get to that level. We want it's going to require some amount of work. We want to make it as easy as possible to create, you know, an avatar experience essentially. So that's kind of where we're going for.

It's not quite today, like just, you know, shooting video for Instagram on your phone. Like, if you're going to be a YouTube influencer, there's some sort of storytelling, there's some production value that you need to have. So there's sort of a semi professional to professional still creating those videos, Especially if they want to grow a large audience. Right. It has to be fairly polished, but it's not, you know, Kathryn Bigelow or Steven Spielberg quality work.

Right.

There are different ways to tell stories, you know, with film or video. So I think the same for avatars today. You know, if you know, unity or Unreal, those are good tools to use in tandem with our software. However, you know, we're working on a solution so that you don't need to use either of those in tandem.

Today in 2019, what kinds of events can it react to? Like, if I move the camera away with my phone, you know, if I. If I'm looking at, at you with my phone and then I look away and then I look back, can I. Can my avatar react to that?

Yeah. So effectively the avatar can know if you're engaging with it, you know, through eye contact, it can know how close it is to you. We can know sort of what you're saying and try to make sense of that and understand that and then have an authentic specific reply to that. You know, we're working on sentiment analysis. We're trying to sort of understand the context of what you say, you know, and then also objects as well. So if you're holding a coffee cup, an avatar should be able to say to me, hey, put down that coffee cup and come talk to me. Or if he walks into your room and sees a tv, he should be able to say, hey, let's watch Netflix together. You know, so, you know, so you're.

Building a situational awareness engine.

Yeah.

So Effectively that's probably 2020. He's saying, sorry.

Today we recognize seven facial expressions and about 40 objects. So again, we're not trying to solve, for Google scale, millions and millions of different things. We're trying to solve for common things that we think will lead to a much more engaging and entertaining experience. Right now we're at this kind of toy stage. So if we can recognize 40 common objects really well and allow our customers to incorporate those into the store, objects.

Being like a Couch, Couch.

Coffee, Car keys, wallet. Yeah.

How many people are talking to the avatar at once? That kind of basic stuff. And then from there we'll grow. So it's about foundational tech today.

Very cool. What do you think, Ryan? What kind of skills is he gonna need in four years to build an avatar world on your platform? Where do you think his job is?

4 years, he's just gonna be able to choose what kind of avatar he wants to create. Is it a human? Is it a male or female? Is it an animal? Is it an alien? And then he'll go forward and sort of customize that character simply through a graphical interface. And then in terms of storytelling, he's going to sort of say, well, what's the story about? And then he's going to basically tell the software that, even potentially verbally. And then it's going to sort of chart some different courses for him and then be more generative so that it can sort of take a user down a story path. Almost like today I like to compare it to writing an email with Gmail. Right. Like, I still have to be the architect of what is the subject of my email, but today Google helps me compose my email and finish my sentences for me. Right?

Yeah.

So it's really just about making that process easier so that I can go quicker. So I think for content creators, it's similar. How can we enable you to tell multifaceted stories that become generative and real time and unique for each individual? But how can we help you kind of organize those thoughts into a simple system that doesn't require a lot of work?

Very cool. Part of what I'm doing with my new video series is focusing on what I call seven voxels or seven areas that entrepreneurs need to worry about. How to get funding, how to get pr, how to get talent, stuff like that. Can we dig in on some of those? Like how did you fund your company?

Yeah, so we had some early investors like Founders Fund and Cyan Banister, who's a partner. There is also an angel, Chad Hurley, the co founder and former CEO of YouTube Venture Reality Fund, who you probably know. Tippitat and Marco WonderCo here in LA, which is Jeffrey Katzenberg and Meg Whitman's new venture, Macro Ventures, which is another Hollywood strategic, and then DCG in New York, which is a blockchain strategic.

So those are kind of amazing set of investors.

Yeah, those are our early.

How did you meet all these people? How many no's did you hear along the way?

You know, we definitely hear our share of no's like anyone does. But we were really looking for true believers. So, like, when we. When we, you know, met with Venture Reality Fund, we already knew that they understood our space, they understood this tech and what was possible. Or Armando can tell the story of him and Chad, but they worked together, and Chad obviously helped envision the future of entertainment, you know, over a decade ago. So he's always sort of looking to what's next as well. Cyan Bannister has invested in a lot of different companies at the nexus of AR and AI, So I think she's someone who really understood this space and then the blockchain folks as well. When you think about how are virtual goods going to be paid for? How do you create scarcity of virtual goods in a real time rendered space? These are things that some of the blockchain folks think about.

Yeah.

Yeah.

You want to talk about Chad?

I know Chad. You want to do the super quick backstory?

Sure, yeah.

Yeah. So Chad, when I was a very poor grad student, I was tweeting about something on, you know, the future of entertainment, essentially. And Chad, the story goes that Chad saw a tweet of mine that somehow had gotten retweeted and he followed me on Twitter, which is kind of a big deal when you're just a random kid. Anyway, long story shorter, he was coming to south by Southwest, and I couldn't afford a badge at the time, but I went to a startup crawl where they were raffling badges, and I won the badge. I put in like four bucks or something. I got the badge. I went and introduced myself to Chad. He was giving a keynote at south by that year, and he said, hey, let's have coffee. He didn't know that I didn't live in the Bay Area. So I flew out. I pretended to live in the Bay Area. I flew out there just to have coffee with him, and we ended up working together. And a few years later, he saw that this kind of avatar stuff that we were working on, and he said, hey, I want to invest in you guys. So he became our first investor. It all came from a tweet.

It's crazy stories. The stories behind business.

Yeah.

And then what was the other you asked about investors.

I have seven of them. Like, how are you set up legally? How did you guys set up your company?

We're a C corp. And, you know, we kind of got everything set up last fall. And, you know, we work with a really good attorney and startup ecosystem. So we made sure all of our paperwork was correct, all of our documents and, you know, we've been basically doing things, you know, as much as we can, sort of by the book and getting a lot of great advice.

And how big is your company?

Right.

We're almost 10 people right now. 10 people, and then we're probably going to be close to 15 very shortly. We'll kind of stay there for a minute and then see where we get this summer and fall after our first customers launch their first content, and maybe we'll raise a Series A and kind of go from there.

How are you thinking about culture internally? Well, it's a pretty early stage.

Yeah, We're a small team, but we already have a couple of women on the team. We have, you know, we care about diversity, obviously, so we've made, you know, I think, a diverse amount of hires. And, you know, we're in downtown la, so we're kind of at the nexus of a lot of different types of businesses in downtown la. We kind of chose not to be on the west side with all the other startups in la. But downtown is sort of becoming a vibrant place for artists and startups and other businesses. So we really like it.

Very cool.

Yeah.

And where are you finding the most talent? I guess around the world.

We've been really focusing on the LA ecosystem. And, you know, obviously when you're looking for machine learning, computer vision engineers, the talent pool is very depleted, especially in Silicon Valley. And I think it's no different here. But we're finding that if you look not just in your own space like say AR or AI, but you look for AI people potentially in other industries. Like we hired one person who's from the satellite space doing computer vision in the satellite industry and then sort of decided to switch industries into kind of what we're doing, more consumer facing. That's been a good approach for us, finding really talented people who are in other disciplines right now.

Very cool.

Yeah.

And have you had success with PR yet or getting your name out there?

Yeah, we did some launch press back in December, early December. We kind of got and wanted to make sure we announced our company before the end of the year. And yeah, we felt pretty excited about that. We were in some publications in Hollywood and up in Silicon Valley.

Do you have any tips for other companies on Andes?

I think sometimes you could hire a publicist. They can be. Or a firm, they can be somewhat expensive, depending on who you talk to. Or you can hire people on a case by case basis just for a job, which might be a better approach when you're just starting out, trying to Keep costs down. Or you can really just try to build relationships with journalists. We've met a lot of journalists through just being in the VR space and AR space or, you know, you can do so on Twitter.

You know, you guys are fairly active on Twitter. It's really interesting.

Yeah, we try to be. And we don't have kind of the following that you might have.

I think you have a better following, might be smaller in number, but highly engaged, right?

Yeah, highly engaged. Well, it's really a following of our peers. You know, we follow other entrepreneurs, other startups and other people in the AI and kind of AR space specifically. And we follow some VCs and they follow us. Yes, I'd say it's highly curated and you have to engage there. What else would you say about that?

Yeah, I think, I mean, our team also does a fair bit of evangelizing. You know, we're lucky in that we're in like a really cool kind of up and coming space. So I think it's helping with PR and with recruiting and even with fundraising.

Yeah. And with customers. You guys have landed some amazing customers.

We're going to be revealing some pretty big customers.

Yeah.

And that's, you know, I think that's partly about the technology, but it's also partly about the fact that we're helping them solve a new problem. Like they didn't have access to the type of insights that we provide about their audience before, you know, so it's.

A pretty easy sell.

You know, it's not just like we go in there and we're so futuristic and cool. It's actually like, no, this is very tangible and helpful and that's, I think, helping us get these customers.

You're not the only ones trying to do this. I mean, even Magic Leap has this virtual human called Mika.

Right.

Have you seen Mika? Yeah, yeah. And what do you think of her and how does she compare to where you're going?

Well, I think they're taking a different approach. They're really trying to create a character. It's not an assistant character. They've been very clear about that. It's sort of a companion and someone to share experiences with.

So many lines over there.

But I think they're focused on building that character out to kind of do a lot of different things within the Magic Leap ecosystem, it seems. And most people have been taking an approach, kind of from an enterprise standpoint of like creating a singular character that can function to do many things, maybe not being an assistant, but to work with different content offerings or apps. We're really focused on creating software to create a number of avatars and a number of unique sort of digital agents, if you will. So we're trying to reach scale of, you know, hopefully thousands and then one day millions of these avatars through a wide net of creators. So we're not really focused on being the creators ourselves. We're focused on enabling creators who today are either artists or they're in AR VR, they're in gaming, they're in filmmaking, or they're influencers or celebrities who are trying to engage their fans in a new way. And really so engagement to us is the key. We feel like video is fantastic, it's not going anywhere. But on social engagement rates are dropping because there's a sort of plethora of video, There's a saturation of video. We believe that real time rendered content, specifically avatars you can have a conversation with, are just inherently more engaging. If they can be delivered via the web, then they can reach large audiences on pre existing fan bases through these media companies accounts. Or not only that, if you can.

Take a selfie with a spider man or something like that.

What if you could have a conversation with a character or your favorite toy if you're a kid and then ultimately post a clip of that conversation back to all your friends and say, hey, this is me talking to my favorite toy character. Or it's me and my favorite celebrities doppelganger talking about their next creative endeavor. Whether they're making an album or they're coming out with a new shoe, or they just want to kind of get closer to their fans in a new way that they haven't been able to do through video. We think that's really the ticket.

Very cool.

What do you think, Ryan?

You gotta start a whole new YouTube channel with a whole bunch of avatars. I mean, I don't know, it's hard.

We're gonna make it easier for you though, hopefully.

Well, thank you so much. And we learn about you guys where Arty.

Com would probably be the best place.

Where else, Twitter, Facebook, LinkedIn, all the usual places. Thank you very much.

Thanks.