--:-- Host: She was like, "You keep saying 'scalable celebrity,' but like, what does that look like?" You've probably heard of Lil Miquela, also known as Miquela Sousa, a 19-year-old Brazilian-American influencer with millions of followers online. She's partnered with the likes of Samsung and Prada, but she's also not real. But here's the thing: Miquela is not alone. You might have also heard of the other digital influencers out there, whether it be Knox Frost, Ema, or even Olympia Ohanian's doll, Kwe Kwe.
--:-- Host: Well, in today's conversation, we get an inside look into the origin story behind Lil Miquela, a virtual trailblazer created in 2016 by Brud. And we do so with Trevor McFedries and Isaac Bratzel. Trevor was a co-founder and creative genius behind Brud, while Isaac was the Chief Design and Innovation Officer, so he was actually in charge of creating the avatar and running the design and tech teams. We get to hear firsthand what inspired this experiment, and also what early signs were showing that Miquela was not just a novel idea, but a character that people were really resonating with, even writing fan fiction about.
--:-- Host: And in an era where it's easier than ever to spin up a character, where does it all go? When an influencer can truly look however you want it to look, how do you decide? How do you create a narrative that resonates? How do you get past the uncanny valley? And what are the ethics of all this? We cover all this and more, and I'm truly so excited to have you listen to this conversation.
--:-- Host: As a reminder, the content here is for informational purposes only, should not be taken as legal, business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any a16z fund. For more details, please see a16z.com slash disclosures.
--:-- Host: It's 2016, and I just want to paint a picture for the listeners. Britain just voted to leave the EU. Rihanna just released "Work." OpenAI and TikTok were literally just founded, in their absolute infancy stages. We're in a very different world. And somehow, Trevor, you had this insight, this inkling that a virtual influencer like Lil Miquela should be created. So what was the key insight at that time?
--:-- Trevor McFedries: The reality was, much like you mentioned, we were living in this kind of heightened political moment, post-Trump, post-Brexit. And me just being terminally online had kind of spent time in these different evolutions from web one to web two into web three or the present, seeing these different media moments and the narratives they could create and how that could shape the ideologies and belief systems of young people, especially. Being in the States, I was hyper aware of what happened on 4chan and what was happening on social and meme magic, and what it meant to create narratives that could capture hearts and minds, and how effective those can be, and the double-edged sword that presented.
--:-- Trevor McFedries: In parallel, I had fallen in love with this data set around a show called *Will & Grace*, which I love. *Will & Grace* was largely responsible for gay marriage in the U.S. Public polling was tied to the ratings of that show. So it seemed like there was an emergent new media—social, visual platforms—that could be used for telling fictional stories. And that was really the dream. It was like, if you were going to build a modern Disney or Marvel now, you probably wouldn't want to start in comic books or in theaters. You'd want to start where the eyeballs are, which was really on social. So that was a vision. It was like, can we create Disney on social? Can we tell stories that are as engaging as a Kardashian or a Jake Paul, but kind of imbue these ideologies that make for a more tolerant, empathetic world? And that was kind of the impetus for all this madness.
--:-- Host: Yeah. I love that you mentioned some of these other characters or influencers that we see in TV shows. People have shown that they have an affinity to not just humans, but characters in TV shows and these illustrated characters. But was there some sort of insight that you saw where—to me, I don't know if we saw very many virtual influencers that looked like humans. What was the insight of like, well, Miquela is fake, but she also looks super real? She was almost like a new breed.
--:-- Trevor McFedries: Yeah. I mean, they were like zero. And that's largely why Isaac is in the picture. No one was doing photorealistic, bipedal human characters. There's this concept of the uncanny valley, this kind of middle ground between what an actual human looks like and what a cartoonish-looking Hello Kitty human looks like, where it looks too human but not human enough, and you get freaked out. The Polar Express girl is often cited. As a result, you didn't really see anyone doing any of that stuff. There were things happening in Asia—Hatsune Miku. We're standing on the shoulders of giants like Miku. But we were interested in doing something on social. And that was already going to be challenging a lot of norms. There was still this perception that social media was for nonfiction, and we'd be telling a fictional narrative inside that space and doing it with a character.
--:-- Trevor McFedries: If we could minimize as much of the novelty as possible and keep as much of it familiar as possible, I thought we had a better chance of succeeding. And I thought it would be really easy. I was like, "Yeah, we can just make a virtual human. No big deal." And six years later, we're still trying to figure it out. But Isaac really was the key in getting that figured out.
--:-- Host: So Isaac, let's hear from you. You get this idea from Trevor. You're like, "Okay, let's just make this virtual human." And you have to make this a reality. Talking about that design and actually making this real, what were your first steps? And what are all the little things that maybe sound simplistic but were actually really hard?
--:-- Isaac Bratzel: Yeah, I mean, I think it's always really hard. But I think we got lucky in a lot of ways in terms of Miquela's design. One is we had a CEO at Forever who actually dove into some of the software—he found Daz 3D and these softwares—and kind of started making that design. So when I came in, it was really just taking that, looking at what's working already, and saying, "Okay, now let's turbocharge that. Let's make this able to look realistic. Let's make it able to be animated and all these other things." So I think for us, it was a little bit of having a really good sense from Trevor and from some of the cultural savants we had on the Brud team that really understand internet culture and how to connect with the audience they were targeting. Luckily, we found something that obviously resonated with a whole lot of people.
--:-- Host: Yeah. And it really did resonate. And I've heard you, Trevor, talk about all of these little decisions that you had to make about exactly what she looked like. When we're influencers online as ourselves, you only have the face that you're born with. But when you're starting new with a virtual influencer, you literally can make that influencer look like anything. And that's amazing because you have this clean slate. But it also is like, yeah, does she have freckles or not? How long is her hair? What race is she? How old is she? What are her beliefs? Given that you used the term "terminally online," how did you use your background? I mean, you also worked at Spotify. You've toured with Katy Perry. You have this cultural understanding of what people like. How did that feed into what you ultimately built into this character?
--:-- Trevor McFedries: It's an interesting question because when you're doing things like this, it's so open-ended. The other idea is that Miquela didn't need to be a fixed aesthetic or character. It could be this kind of shapeshifter, ever-evolving. In the narrative, Miquela's 19 forever. So there were decisions—I wanted to create constraints because as a creative person, I like having constraints. I think it breeds really interesting ideas. Some of the early things: I had worked in making music as an artist and as a producer for a very long time, and it was pretty clear to me that there were these underserved, really passionate audiences in Latin America. If you ever look at some pop stars' Instagram, 10% of the comments are "Come to Brazil, come to Brazil." So in the narrative, I definitely wanted to engage young people, especially people who spoke Portuguese. Miquela was programmed to be half Brazilian, half Spanish. The idea was that we could engage kids who I think were underserved by influencers who were kind of looking towards Europe and Western Europe especially.
--:-- Trevor McFedries: Beyond that, I think the age thing—I'm a big fan of pop culture, and I think generally have seen this motif where people engage with media that's five to seven years older than them. So Hannah Montana is running around high school, but it's kids that were in middle school, younger, watching that program. Miquela being 19 was quite intentional because I really felt that in a post-PG-13 America, movie ratings really neutered the middle ground of cinema and television where you used to have these John Hughes films, these coming-of-age tales that really talked about issues that people were facing and didn't talk down to them. One thing we always talked about early on was that we wanted to tell stories at eye level. We didn't want to talk down to kids. We wanted to address them like they were adults, and they had complicated lives. They're aware of the turmoils in the world, they're well aware of impending climate change, they're well aware of economic strife, pandemics. These issues affect everyone. It's not just cliché *Saved by the Bell* motifs. We can go deeper.
--:-- Trevor McFedries: Beyond that, we just made assumptions about what narratives would work and how we could kind of imbue these ideas of tolerance, of otherness. And then you just try things. I think people were probably like, "Wow, how did you know that she'd be so revered in fashion?" And I had no idea that—we want her to look cool. And all of a sudden, fashion raised their hands.
--:-- Host: I want to talk about the narrative that you built out because it seems really critical to the success of Lil Miquela. But also you mentioned you were surprised by things. What were some of those early signs where you're like, "Oh, wow, we've got something here"?
--:-- Trevor McFedries: For me, I get kind of theory brain and I'm talking about these concepts and ideas that I'm really excited about, and often they don't apply in the real world. But I talk a lot about this idea of parafiction, this idea of telling fictional stories in spaces reserved for nonfiction. The big one being professional wrestling, WWE. This idea that you're watching people in a ring where you watch sports take place, boxing, other things, and people are getting hit with chairs and they're really getting hit, but you're told that it's fake. In an infinite scroll world where nothing makes you take pause, these things that disrupt these patterns that you're familiar with make you pause and say, "What's going on?" So much of what we were doing was really trying to create these parafictional moments where you're like, "Wait a minute, Miquela's at Coachella? Is that really Coachella? Wait a minute, a brand is talking about her being at Coachella." And that led to credibility for the idea.
--:-- Trevor McFedries: The moments that jump out for me—obviously it was Shane Dawson when Shane Dawson made this YouTube video that included Miquela in these conspiracy theories about what she could be. That adds this other kind of meta-narrative layer to it, but then kind of compounds the reality of it all. That stuff was really special to me and quite fun and quite cool.
--:-- Host: And Isaac, do you have memories of things that really jump out?
--:-- Isaac Bratzel: Yeah, I 100% agree. The Shane Dawson one was really cool. And I think it was anytime there was a thing that kind of made it real to me. Because for me, the thing I was always worried about when I got there was, is this just a reactionary thing? People see Miquela on Instagram and they have a reaction, right? But when you could see fans really engage—somebody would respond to Miquela, and then there would be a whole bunch of fans that would come to her defense and explain the whole story. The Shane Dawson video, they tracked down all this stuff. You could see how much people really cared. There were fans that knew more about the backstory than I did at certain points. I was like, "Oh man, this is really deep."
--:-- Isaac Bratzel: And then the Coachella moment was huge for me because it was a technological thing. I was like, "This is really real. We're not just doing still images in magazines and stuff. We're going to go in real time and try to interview musicians for Coachella." That was around the same time as the Calvin Klein ad. Rehashing the trauma of those moments where we're just working around the clock, trying to do things that take 100-person teams years in just a few weeks or months. It was totally mental.
--:-- Isaac Bratzel: I remember a couple of Halloweens where there were all these fans dressed as Miquela. One of the things we talked a lot about when we were designing the character was maintaining this Halloween costume. If you were going to be Miquela for Halloween, are there identifiers? People would know who you were. The space buns, the freckles, certain things like that were things we tried not to deviate from too much. I think that was one of the brilliant insights early on. It's recognizable instantly. Because if we change it too much, you lose that instant recognizability. So Miquela sticking with the space buns, having the freckles, these really recognizable features. I think that was really brilliant.
--:-- Host: Yeah, I think I've heard you talk about this, Trevor, but it was this right balance between somebody that almost anyone could resonate with—people aren't even sure exactly where she's from—but then at the same time, these iconic aspects where you're like, "Oh, I haven't seen that before." So it's a nice balance in her look. You mentioned fan fiction. It's always fascinating to hear that fans almost give the character a life of its own. It sounds like this was the case. Tell me more about this narrative that you guys ended up building over years, and also whether any of this fan involvement helped you curve that narrative or adjust that narrative. It seems like it's easier than ever for someone to create one of these virtual influencers. And given that that may be the case, that there's going to be this flood of them, how important is building a narrative in getting this traction, having people really resonate with a character like this?
--:-- Trevor McFedries: It's all extremely hard, but I think narrative is unthinkably hard. There's all kinds of books you can read about why you shouldn't build narrative into different media businesses because it's like building a fashion business. You have to keep things interesting forever. You can only keep *Lost* interesting for so long. *The Sopranos* runs out of steam—or maybe not that show, it's kind of perfect. But I think one of the ways we tried to solve for it, or I thought about it, was integrating Miquela into the real world.
--:-- Trevor McFedries: One of the things we had talked about a lot was inverting the traditional media pyramid. Traditionally, you kind of started with a longer-form asset like *Star Wars*, and then you iterate into a *Star Wars* video game, comic book, all the way up to a lightsaber you could buy at Target. You have this pyramid that moves from a big, expensive, long-form asset into this smaller thing. We wanted to invert that and start with just a character, a still image, and then work up that stack into longer-form television and film. The dream was to actually expand upon that television and film.
--:-- Trevor McFedries: One of the reasons I love the Kardashians and their narrative universe is that it's infinitely deep because they're human beings. You can work through all of Kim's father Robert Kardashian's OJ case into his family tree into Armenian conflict. You can go as far back as you'd like. So by integrating Miquela into our world—by having her work for Brud or fire Brud, or have these real-world boyfriends and other things—fans could go super deep. They can start to speculate, they can riff, and they can actually connect dots that we wouldn't see all the time.
--:-- Trevor McFedries: The other part of that was we really tried to build a technology organization disguised as a media organization. We had data science teams every week bringing back what worked and what didn't, building taxonomies on images and videos, trying to understand and figure out what motifs connect and why, and how things are kind of working backwards.
--:-- Host: Isaac, you were part of the team that actually helped this come to life. Like, "Okay, Miquela's apparently now dating this one person. How do we actually execute on that? And should we change it?" I think at one point Miquela was dating a real-life person too, right? So there's all these dynamics. Tell me more about operationally, how do you make that happen?
--:-- Isaac Bratzel: Yeah. Early days we were really small, and it was crazy fun trying to do things that were just way beyond what a team of that size and experience level should be trying to do. But I do think Trevor's earlier point about this really quick feedback loop—when you're starting off, you have the ability to do this quick thing and get that reaction immediately, and like you said, steer that character in our narrative to the fans that are engaging, which then informs what you're going to do when you start to do more medium and long-form stuff. I thought that was a really cool thing that I learned from Brud. It's this reverse pyramid that Trevor talks about. If you start with a movie, you're kind of like, "Here's a script. We're going to take a shot and see if it lands." If everything hits perfectly, then maybe you go from there. But this is almost like, let's introduce this character super early in the lowest form possible and see what resonates and what hits, and kind of build the story out of the fan engagement. That's just a radical idea—at least it was to me at the time. I had never really seen that. So using that to inform the longer-form content, and kind of having this pretty good idea that, "Okay, this is going to work now, so now we can go and try to create a five-minute music video out of a real-time character," which is a whole challenge.
--:-- Host: So I mean, seriously, how do you create that? I guess today the technology is getting better, but these music videos—I think she's on Spotify. What technologies are you actually using to accomplish all this?
--:-- Isaac Bratzel: Yeah. I mean, there's a lot of different ones. I think there's a huge—I would say an order of magnitude difference between the still images on Instagram and doing actual long-form animated content. That's stuff that's typically reserved for really high-budget films and AAA video games for a hyper-realistic digital human to be doing that stuff. So that was really challenging. I think we broke a lot of ground there, and the technology was coming along, so we were able to do it somewhat. But we're now in a phase where it's starting to become—I think even referring to this stuff—it's starting to become more and more realistic for smaller studios to be able to do that. But it was still a blocker.
--:-- Isaac Bratzel: I think what we were really hoping to do was somewhat blocked by the ability to create really expensive, long-form content of animated digital humans. That's really expensive. So we want to make more characters and create more content and do all these things, but you get stuck in a certain platform when you build out this technology piece. It's really hard. You can't just go to HBO and *Euphoria* and be like, "Yeah, sure. Here's Miquela, put her in your film." It's not easy like that. It's expensive and time-consuming. Anytime we had those kind of opportunities, we had to weigh the opportunity cost of diverting our whole team towards doing that thing and dropping Miquela's internal narrative. I think that was one of those big learnings: in order to scale this thing, we really would have to have a big investment to get over this bottleneck at this point in time.
--:-- Host: So did you end up getting over that bottleneck? Because I think you did end up launching a few other influencers—Bermuda, Blawko. How did you decide, "Okay, yes, we are going to expand this past Miquela. We are going to take on some of these really cool opportunities and invest in doing these longer-form video narratives." What were the next steps past that early traction?
--:-- Trevor McFedries: It's funny. I'm thinking back to the original pitch stuff. Initially, it wasn't clear to investors how this could be a venture-size bet. But the bet to me was always, okay, there's a coming computing shift. I called it spatial computing. I guess it's probably best now as the metaverse or whatever else. It seemed like all of the people that were building for the metaverse were building universes and expecting people just to show up—"If you build it, they will come." I always thought that a character would pull you into an emergent platform. Pikachu is the reason you get *Pokémon Go*. You don't want to just walk around your yard scanning stuff. You want to follow Pikachu somewhere. So if we kind of built a connected tissue between Web2 and this emergent spatial computing thing, we'd be way out in front.
--:-- Trevor McFedries: Looking back on it now, the things we kind of got backwards: I was like, "Okay, billions flowing into VR and autonomous vehicles, and we're going to have this huge shift. It's going to allow generative media to come really quickly. And then once we can do that really quickly, a lot of value is going to shift to this spatial computing world and be able to build digital economies." I think COVID accelerated digital economies before we actually had these tools to do generative media at the pace we'd like. So we had to make decisions about where we could be efficient and how we could build models that port really well.
--:-- Trevor McFedries: Anyone can kind of do a digital human that looks good in a still image, but speaking, building really compelling facial rigs that can be driven effectively—super tough. Isaac's really good at it. It's insane. But Blawko was like, "What if we just covered his mouth all the time?" Simple things like that, where he's always wearing a mask, he's always covering his face. It adds a little bit of mystery, some narrative, a constraint for the creative people, but it allowed us to create stuff more efficiently. Little things like that. Feet are hard—they often are floating. Let's not show feet. Hair is crazy difficult, so Blawko is bald. Miquela's hair is pretty much fixed except for a couple of things that move. Let's really limit things that could cause a lot of problems and try to figure it out from there.
--:-- Host: That is fascinating because with just the generative AI stuff today, the trope is the hands. Everyone's like, "Oh, don't look at the hands." I always look at the hands. So Isaac, are there other things where you're like, people who have not ventured down this rabbit hole, have not built this, don't know is actually surprisingly hard?
--:-- Isaac Bratzel: Yeah, there's so many things. I think one of the things—anytime we're outside of technology, we have this idea that computers can do one plus one equals two and computers are always going to do that thing. There's nothing farther from the truth when you get into this stuff. Everything breaks, everything crashes. This is why you have big studio models to get these high-end budgets that just go crazy expensive, because at the end of the day, you have to get pixel perfect to fit into this film, into this universe, exactly right. The only way to do that is throw time and money and people at the problem. What we were trying to do was do that at the lowest scale possible.
--:-- Isaac Bratzel: We're doing this on social media. We can get to 90% quality, right? People are not going to care if this one flyaway of her hair is slightly not quite right. So we're like, "How can we eliminate 90% of the cost and the time and get 90% of the quality?" Because it's kind of a power law in bursts. It's just so complex and so hard. Even today, I think one of the things we're going to run into a lot is this idea that, "Oh, anybody can do this now with AI tools and everything else." To me, it feels a little bit like suggesting that because Ableton is $5 now, anybody can go be a superstar musician because they have this technology. Technically speaking, and having the skill set to do it, it's not true. And then definitely narrative and culturally, and having the ability to connect with fans that way, it's definitely not true.
--:-- Isaac Bratzel: The technology piece is going to be so enabling. My hope is that it will really help people that have that ability and that desire to connect and tell these stories. I really resonate so strongly with Trevor's vision of storytelling being transformative. I think that's so true today more than ever. If we were five years later with Miquela, we would have been able to do things that would have been much bigger and even broader because I think a lot of the bottlenecks are going to be removed very soon.
--:-- Host: Yeah, I love that you said that—that the technology is a democratizing force. But I can also see how people will look to Lil Miquela and be like, "Oh, I couldn't do this six years ago, but I can do it today." So on that topic, are there any—I guess any wisdom from your six years of doing this and building this and seeing what works, what doesn't work, how much harder it really is to get traction, to get people to care about something—that you would almost tell those people? Not so much to drive them away from this idea, but to understand what it really takes.
--:-- Trevor McFedries: There's kind of no easy way to say this, but it's a democratizing force, all this technology lowering barriers to participants. So effectively, if you're creating media on TikTok, you're not even competing for attention in your following or your social graph. You're competing against the globe. You need to have something more compelling than tens of millions of people that are generating stuff. So it's super hard. One of the things I tried to do is understand where you have advantages and where you have an edge. For us, it was music. Maybe 10,000 people are releasing music on Spotify this Friday. Can we be in the top 100? I think knowing where you're really gifted and where you can out-compete others is key, because it's hyper-competitive for sure.
--:-- Host: Isaac, on the technology side, I just want to ask you super quickly because you're building a company here that enables some of this, right? And from my understanding, your company is not taking the full generative AI approach, where you just go into something like Midjourney and spin up a new image. So how are you thinking about that? The different ways these new technologies are coming together?
--:-- Host: Sure. Yeah. I mean, I think right now, I think that there's understandably a lot of excitement about Midjourney and really tools that I call pixel pushing at the end of the day. If you're doing 2D morphing of images, then I feel like in a more immersive world—so use the dirty M word, metaverse, or spatial computing—like that 3D immersive world, when you create 3D avatars or 3D things and 3D worlds, they are fully immersive and they're able to now be interoperable within a more virtual world. And that's what I think is coming. Not necessarily the metaverse in terms of what we might think of it now, but just this world that's going to come more and more out of gaming, where things are going to be more three-dimensional.
--:-- Working in 3D for like a decade, you start to understand how much we underestimate what another dimension means when you're adding it to something, especially technological. We don't think exponentially, we think additively. So we're like, "Oh yeah, you add that third dimension," and it's like, "Yeah, just another layer." And it's not—it's completely a whole order of magnitude shift.
--:-- I also really like what Trevor said earlier about rather than building metaverse worlds and focusing on that, focusing on the character and the narrative and having them be that connected tissue. I really view that as one of the ways that some of this stuff can start to land. Which is if we can create avatars that are scalable—digital humans and digital people that can actually connect with fans in a real way—and have that be the way that ushers them through this kind of more virtual world and creates a more human connective layer. That's what I think would be really, really cool. Because a lot of times, if you look at what's going on now with a lot of the virtual world stuff, it just feels very impersonal.
--:-- I don't think that this is not including some of the generative AI tools. The generative AI tools and the stuff you see image-based right now is very two-dimensional. And I don't think that's actually going to be the solution. I think what it's going to be is when those tools are able to really generate and include into a three-dimensional world, which is something that we will be focused on.
--:-- Host: Right. So you're basically saying a lot of these tools right now are in 2D. You think that there's so much more fidelity in 3D, but then there's also things on top of that, right? Like voice and the ability to animate the 3D character. But I guess as we do see more of these virtual characters—these people who are not, as people say, not real people—how do you guys think this actually changes our relationship with creators or influencers, I should say?
--:-- Let me tee this up. Today, most people don't have the incentives to create a virtual influencer of themselves because, again, they didn't have access to the technology. It wasn't good enough. But as more of these characters come online, why would I want to put myself out there? Because I can be a more beautiful version. I can change my voice. I can basically become exactly what I want to be. And I'm not constrained by the physical nature of what I was born with. I also can't be canceled, right? There's these dynamics. But again, this clean slate. Do you guys see that being the case where with enough time, that is the reality where most things that we see online are not quote unquote real? And then also, how does that change our relationship with the things that we see online?
--:-- Trevor McFedries: I mean, I think we could argue about real and reality. I look at my Instagram feed and it's a lot of scripted pranks or things that are, you know, make people believe that are real. It's people that have been highly altered. It's situations and contexts that are presented as real—like sitting on a private jet when you don't actually own that jet. And so I think the sliding scale of reality will probably continue to move in the direction it's been heading in.
--:-- But beyond that, the thing that I'm most excited about is this kind of race to the bottom in a lot of media. A bottom's going to hit far quicker because of AI. It's going to be really easy to be good at things. Good at making music, good at making visual art, good at creating content. It's going to be really hard to be excellent. And because models effectively are backwards looking and data kind of begets more data, you're going to have some nasty feedback loops. Not entirely dissimilar from what happened to Netflix, where they kind of woke up one day and everyone realized that every show was a Chef's Table for a different industry. And they're like, "Wait, I like this format, and clearly the algorithm said it works, but I don't feel like I've gained any nutrients from this. I've just kind of been satiated and I'm looking for nutrients."
--:-- So I think we're going to have this rapid ascent where things feel better on the surface. People are looking for substance. And that's what will separate the two Michelin star chefs from the very average. And that's when I think the really supposed to rise to the top that can augment this technology with unique skill sets.
--:-- Host: Do you think that maybe sometimes people misunderstand or misalign substance with what we just talked about, like reality? Because I could see how people over time will be like, "Okay, my response to this, everything feels fake. I want things that are truly in meat space." And that's what they look for. But to your point, just because it's in meat space doesn't mean it's real and has substance, right? And so it can be just as manufactured and less creative than actually something that is technically virtual, but has this dimension to it, this depth, this creativity.
--:-- Trevor McFedries: Yeah. It's a question that's been imposed for a very long time. I think even since the emergence of contemporary art, right? You have this constant dialogue where it's like mom and dad walk into the museum and they're like, "What are these scribbles? A three-year-old could do that." And then they walk in front of this photo-real landscape painting and they're like, "Now that, that's something I would pay for," because they associate the ability to take an image from someone's eye and translate it onto a page with craft and skill.
--:-- I think people are constantly doing the same. There aren't really good ways of understanding what is good and what is bad. It's so subjective that people are looking for concrete measures, and authenticity is one that's been tossed around quite a bit. And I think people look for things that resonate as authentic and often associate them with playing an instrument, having a craft. And I think the reality is some of the most authentic things in the world to me are like K-pop, where it's entirely scripted and fictional, but it's presented in a super authentic way. It's saying, "Hey, this is going to be so entertaining. We've plucked the most beautiful kids from Korean middle schools, give them the best choreographers and the best songs, and taught them to speak like Justin Timberlake. You're going to love it." And there's no deception there.
--:-- I think young people resonate with that. I think it's harder for adults who've been primed on these ways of understanding contemporary art. And as technology evolves and ways of changing those rules evolve, they feel very comfortable and fall back on the kind of frameworks they understand.
--:-- Host: Yeah. And actually, these virtual influencers, I think, are biggest in Korea and parts of Asia, which is interesting. Maybe it's a cultural thing. Maybe it's an age thing. So, yeah, I like that you use the word deception because we're all being deceived in ways in terms of what's being put in front of us, but we do fall back on these easy mechanics of "Is it real or is it not?"
--:-- Isaac Bratzel: Yeah, 100%. I think we talked about this earlier. Days abroad, nobody's going to see a movie of Mickey Mouse early days and be like, "Oh, this isn't real. How dare you guys try to present this as real?" It's too clear. And when those lines start to get blurred, at least for those of us who are used to something different, we have a lot of hang-ups or reactions to things or misunderstandings of what they are.
--:-- To me, it all comes back to storytelling. It's one of the most important things that we have as humans—the way we connect with others, the way that we evolve emotionally. I think it's just one of the most unique things about us as animals that differentiates us from anything else. And I think that enhancing our ability to tell those stories, especially with the world that I see coming—this more virtual world, this more inclusive, this three-dimensional spatial computing metaverse, whatever you call it—I view having digital humans be able to tell those stories as an essential piece of that. I want to help make that look a certain way and feel a certain way and allow people to tell stories in this new way. So when you think of it that way, all this other stuff feels like noise.
--:-- Host: Well, yeah. I'm actually curious from both of you, over the numerous years you've been doing this, what pushback did you see early on? And also, have you seen any of that pushback change? Like these ideas around ethics, what's real, what's not. Should you be doing this?
--:-- Trevor McFedries: I mean, yeah, so many. That was part of the premise though—that we were going to be first to the door and catch all the arrows. We were going to make all the mistakes quite publicly. But I think the really lovely part about that is that we get to set the tone for how these things are understood going forward. It was pretty clear that people were going to leverage generative media, virtual influencers, virtual characters to shape our reality. It's too tempting not to. And so how do we go in and set some boundaries and some ways of behaving so that people follow our path and make things that are maybe more noble and righteous than other things that could have been done?
--:-- That said, some of the obvious things—I think people were really concerned about virtual characters and people having accounts, right? It seems quite silly now, where they're like, "Wait a minute, how can this fake person have a verified account?" At some point you had to extend that into, "Wait a minute, what is Coca-Cola if not a fiction? And why do they have a blue checkmark?" And so there were interesting things in startup land, like Startup L. Jackson or all these Twitter accounts that were parodies. A lot of those things feel quite passé now.
--:-- There are interesting conversations that I think are happening around AI, NFTs, that feel very familiar to what we did, where it's like, "This is new and scary, and I don't like it, so I'm going to be upset about it, even though I'm uninformed about it." And so I think that's going to be persistent forever as the Overton window shifts. But a lot of things that we did feel quite passé right now, where it was like a blue checkmark next year. And so you're like, "Yeah, that's fine. That's totally okay."
--:-- Host: Trevor, what are some of those examples today? You mentioned like AI and NFTs where you see that parallel where back then it seemed strange, this idea of a blue checkmark. What do you think in five to ten years we look back on?
--:-- Trevor McFedries: Well, I think the idea that digital goods have value, right? I can totally understand why it seems silly. The "right click, save" argument makes a ton of sense until you think about it. If I have a wallet that's following me wherever I go on the Internet and effectively becomes the way I play status games, it may be really important for me to have a digital asset that shows that I was there or I'm a part of this thing because it leads to IRL love or jobs or wealth or whatever it is. So I think it's probably going to be quite silly that we thought for a moment that digital goods wouldn't have value. Much the same way, I'm sure it was really silly to see luxury watches sold. You're like, "What do you mean? It's the same thing my watch does. Why would you ever pay a markup for it?" And it's like, "Well, because I get a handshake at the country club that I wouldn't get otherwise."
--:-- Beyond that, with the AI stuff, people try to dunk on the fingers thing or the eyes or the hair looking wonky. And I don't think they recognize how quickly those things will get resolved. Much the same way for us early on, they'd be like, "Wait a minute, this shadow is not real." Oftentimes it was totally real and we matched it entirely. But one thing we had to do early on was add additional shadow to things, or do things to make people believe what they were seeing was real, because the reality of the image wasn't real enough. As bizarre as that sounds, you have to kind of embellish.
--:-- Host: That's interesting. What are some examples of that, Isaac? Like once you look at an image and you're trying to find what's fake, you think everything's fake.
--:-- Isaac Bratzel: Right. And how many times the things that people called out were actually a photographic element—they were real—and people were convinced, "Oh, this is something." And then the things that were CG, they were totally fine with. I would say it was like 50/50. Half the time, the things they complained about were an actual element that was from a photograph.
--:-- Host: Have you guys started doing that? Like whenever I see a picture on Twitter, someone will literally be just taking a picture of their lunch or their house, and I'm trying to figure out if it's generated by AI. I'm like, "This isn't real," and it's totally a real picture.
--:-- Isaac Bratzel: Yeah, I mean, I've been doing that for years. I've been in VFX, right? CG movies. People always ask me, and I can always tell because the first thing you're actually doing is asking, "Is there any reason for them to use CGI?" So the first test that I can do that nobody else can do is, "Of course that's not CGI. That would cost like 8 million extra dollars and there's no reason to do it. So there's no chance that it's CGI." But then there's a lot of other little tricks once you get really familiar with the technology that you can see that maybe isn't.
--:-- So one of the things that's actually hard for me is trying to step out of that and understand, "What is the regular person seeing here? What is scaring them because it looks this way? What feels off?" That's kind of hard to step in and out of. I think the same thing for AI now. I think artists especially have a really strong visceral reaction to it. There's a whole art community, 3D art community, that's really anti-AI because they have this fearful reaction that's happened every time any new technology has ever happened. That doesn't mean that I'm dismissing any of their concerns. It's this fearful reaction that my job is going to be taken away or this other thing that I don't like is going to happen. So I think we're seeing that very strongly with AI right now.
--:-- Host: Yeah. And are you basically saying that you think in years this will not be the case? Like, what do you think happens to, let's just use those 3D animators, for example? What do you think either changes in their perspective or changes about the world?
--:-- Isaac Bratzel: Well, I think right now what I'm thinking is that the idea that you're going to reject this and then it's just going to stop in its tracks is just the most wishful thinking I can ever imagine. So the thing to do is try to identify the parts of it that you don't agree with and try to figure out how can they make that better. So if it's rights to artists' images—that they created the artwork and they should get fractional compensation, or you should be able to opt out of having images—these are all very valid things. So again, I'm not dismissing any of them. It's just that there are a lot of people who are going to react so negatively. It's like, "No to AI, cancel all this stuff." And it's like, that's not going to happen. Even if that was the right move, if you were right about that, it's not a realistic option. So what's the positive thing we can do here to try to make this better?
--:-- Host: Well, on that note of positivity, the technology is advancing in many different ways. So I'd love to hear both of your perspectives on how not just AI, but technology advancing as a whole, may actually change our ability to create these virtual influencers. For example, might these influencers actually be able to engage with their audiences more, like use AI to actually be the one responding? You train an AI model based on this character and they actually become like—not a sentient being—but someone that's actually able to engage. That's one example. You could also imagine a decentralized influencer where basically not only is their worth distributed amongst a bunch of people, but actually what they do, their narrative, might be influenced by the people who have some sort of ownership over them. So Trevor, why don't we start with you? Any ideas on how technology actually advances, empowers our ability to create these characters?
--:-- Trevor McFedries: I mean, I could do this for years. We could just riff. But obviously, the decentralized one is something that we're very excited about. And we're trying to build it at Dapper. One thing I think about a lot is why culture is not viewed as intellectual property. You could almost see influencers as these rent-seeking middlemen that have the aesthetics that allow them to do a dance that was popularized by some six-year-old girl in the hood, and capture a lot of the value that should have flowed back to that person.
--:-- So I think the really challenging part with culture and why it's not intellectual property, and why it's often caught in this crosshairs as being either cultural appropriation or unlicensable, is that attribution is really hard. And remitting payments to the people that you identified as being a part of a culture or being owed for using that representation is hard. And I think digital payments, like blockchains, and tools that allow you to quickly identify someone where they exist culturally, can help.
--:-- So to me, the idea that as a black person, you could identify a community and say, "Look, all of this rad shit happening here has created value that we can repatriate to your community." I think that's really compelling for generations of people that watched Elvis walk off the shit or the Rolling Stones or whatever else. It's sort of like TikTok dances and emotes. I would love to see technology enable stuff like that. So there's all kinds of stuff I could riff about, but things like that to me are kind of exciting and maybe on the frontier.
--:-- Host: Totally. Isaac, what about you?
--:-- Isaac Bratzel: Yeah, I mean, I think to me, there's three elements that I'm really excited about. And these are all elements of the metaverse. To me, what the metaverse really is, is persistence. So a character, a digital character—which to me is a brand, it's a living story—it's persistent. It doesn't exist in one single world anymore. So you don't have Nathan Drake from *The Last of Us* existing just in a video game that people went after in that one situation and then that's it. It starts to be able to go into any other medium instantly.
--:-- So this persistence—that character actually has a history, a story, a life, and that's there. It's synchronous, meaning you can actually interact with it. So that's kind of something you just mentioned there, where maybe you can go and talk to Nathan Drake. That's the simple thing. But I think there's a lot more interesting implementations of that. And finally, it's interoperable. And that means these walls that create barriers between different mediums and different things, they kind of go away. Technology's an enabler and a democratizing force. And I think what it's going to allow is for really compelling characters and brands and stories to be told in really unique and new ways that in five years, we're going to look back today and be like, "It seems obvious," but it doesn't feel obvious to a lot of people right now.
--:-- Host: Yeah, something that's coming to mind is you mentioned synchronicity, but the reason that celebrities can't interface with every one of their fans today is because they are in meat space. They only have so much time. They need to sleep. They can only engage with so many of their fans. But actually technology can change that, right? And that could be for the human celebrity who now just has leverage through technology, but it also especially can be used for these virtual influencers. The idea of being parasocial maybe doesn't exist at some point because actually you're not just listening to people talk. You can actually, after a podcast, go have this virtual conversation with the podcast host. I don't know if I'm extrapolating too far, but that just dawned on me where I'm like, "Oh, actually, the limitation existed before because we didn't have technology."
--:-- Trevor McFedries: The first person I heard say that—trying to make this celebrity scalable—was from Trevor originally, I think. We ran into problems with Miquela just because we were so early. The expensive cost of creating animated content at this time was like, that wasn't really scalable. But I think we're getting to a point pretty quick where it is. As we kind of thought about scalability and what makes that experience exciting, creating some type of scarcity is important for making that thing valuable. It's yet to be defined. I think the variable rewards component of social—that brain hack, the slot machine brain hack—is what makes a lot of social experiences sticky. And so the idea that you could just message Emma Chamberlain and she responds right away or pops up on your screen might devalue what Emma Chamberlain means to you. This is all stuff that people have to work through.
--:-- Host: So it's going to be really interesting. Maybe an interesting question that I didn't have planned, but since we went down this rabbit hole: Do you guys think that at some point we will have a virtual influencer that a religion is built around?
--:-- Isaac Bratzel: Yeah, I mean, I never thought about it, but now that you mentioned it, I'm like, is it possible that with enough time goes by that it's not going to happen? I don't know. It seems like at some point, yeah.
--:-- Trevor McFedries: I will say, problematic part of the pod potentially, but our first employee at Brud was a woman named Savannah Pecanian. I remember we were talking and she was like, "You keep saying scalable celebrity, but what does that look like?" And I was like, "Christ. That's what it looks like to me." These narratives that have shaped our reality because this figure has been able to touch people all over the world and shape the way they behave. Some would argue that there are people in our current lives—Kardashians, whomever—that have near demigod power. And I have to imagine that at some point you'll have someone or something that matches some of those traditionals, some of those trad gods.
--:-- Isaac Bratzel: I mean, it's kind of mind bending. We talked about the sliding spectrums. It's human to computer generated, but then it's also what is the sliding scale between listening to everything someone says to religion?
--:-- Host: On that note, sometimes I come away from podcasts with, "Oh, I know more about this subject." And I do feel like I know more about this subject, but I also was like, "Oh gosh, my brain is spinning. I don't know where any of this is going." But yeah, Trevor, Isaac, thank you so much for going through all this with us, for being early movers in this space and setting the tone, setting boundaries, experimenting. And yeah, thanks for spending this time with me.
--:-- Trevor McFedries: It's a pleasure. Thank you.
--:-- Isaac Bratzel: Thanks.
--:-- Host: Thanks for listening to the a16z podcast. If you liked this episode, don't forget to subscribe, leave a review. We also recently launched on YouTube at youtube.com/a16z underscore video, where you'll find exclusive video content. We'll see you next time.
