AI video director Billy Boman on making Super Bowl ads, directing AI commercials for Taylor Swift and Lewis Capaldi, and having his own Hollywood-scale sign

Speaker 1: on. What's the only thing faster than the AI market? Your business on MongoDB. Don't just build AI. Own the data platform that powers it. And our next

Speaker 2: And now.

Speaker 1: Is here with us in the TBPN Allergy Chat.

Speaker 2: Here he is.

Speaker 1: Billy Boman. Good to meet you. How are doing?

Speaker 4: I'm really good.

Speaker 1: What are

Speaker 2: you Yeah. How's your week?

Speaker 1: Well, yeah. What is your life like right now? Are you just getting blown up by everyone?

Speaker 4: Probably. Okay. Yes. I haven't even had time to check all the DMs yet. I'm just really doing my best just to take care of everything.

Speaker 2: Yeah. I was shocked. We were we were live on the show Yeah. Reading about this whole thing because we we had seen we had seen the ad on the one zero one.

Speaker 1: I saw it in person. Drove right by it.

Speaker 2: Hard to And then I we texted Nick, one of our producers, and he you seemingly responded within an hour or something like that. It was very fast. So Yeah.

Speaker 4: Yeah. Been perpetually online.

Speaker 1: Yeah. It's

Speaker 4: glued to the screen to not miss anything.

Speaker 2: Okay. Okay. So yeah. So anybody that missed it, there is a massive Hollywood sign scale. Yeah.

Speaker 4: 30 feet tall. 30 feet 30 feet long, wide.

Speaker 1: 230 feet wide.

Speaker 2: And on on a hillside that I haven't seen any out of home advertising on Yeah. Before so I wanna Yeah. I wanna get the whole story. Like, let's start I guess at the beginning. How how did this happen? Were you on Fiverr, very active, or you were friends with the Fiverr team?

Speaker 4: So actually, no. The creative director, Nir Rifua, so shout out to Nir for making this happen Cool. Reached out to me and basically, do you want to do this? And I thought, like, first of all, you're joking? He was not joking. Then secondly, right, it's just a print, like, it's a billboard. Right? No, no, we're building the sign. And I'm like, oh. Yeah. Right. And then, like, as they're sending mock ups and stuff, I could see, like, it's actually someone is this, like, the sheer scale of it. So it's Yeah. In the works since October. I don't have full insight into the full production. I've just heard the kind of the stories from the team, so full shout out to the Fiverr team and especially Eli Pier who who is the guy doing this. He pulled off the impossible together with the entire Fiverr team and

Speaker 1: the Yeah.

Speaker 4: Family who kind of just made this Yeah. What do say?

Speaker 2: I'm never gonna personally, I love ads. I'm never gonna look at any hillside the same way ever again Yeah. Because I'm surprised they were able to pull us off Yeah. Even with the city.

Speaker 1: It's remarkable. Take us back to the beginning. Like what did you did you study? What did you study in school? How did you get into your business? And then I wanna sort of go through how AI is affecting your work. But just start as far back as you think is relevant.

Speaker 4: Right. Sure. How the hell did I end up here? Yeah. I typically start with like it's nature and nurture. Right? So like creativity is the only thing I'm good at, like I'm terrible with everything else like bookkeeping or whatever.

Speaker 1: Yeah.

Speaker 4: And I grew up in a creative family. So my mom and dad ran a small advertising agency my entire life. My mom was the

Speaker 2: kind of exactly.

Speaker 4: Shout out to mom and dad.

Speaker 2: Dream upbringing. Yeah. Yeah.

Speaker 4: Grap designer and dad was the art director and I grew up, you know, with like putting little little micro machines in Xerox and just kind of making collages and all the all this stuff. So super creative. Yeah. My my whole family, my my sister works as a game designer. My my brother has done all kind of web design, runs a Dungeons and Dragons start

Speaker 2: up.

Speaker 4: So like we're all Oh, very creative. That's that's all I do. I'm classically trained. I studied at Beckman's College of Design in Stockholm. Probably haven't heard of it, but Stockholm, mean, it's Yeah. It was a pretty good year in Stockholm. Yeah. So classically trained. Ended up going into fashion for whatever reason. Worst decision of my life, but I did. I love clothes and just like aesthetics and Yeah. How to bring ideas into garments. I worked at All Saints, H and M, bunch of different brands designer doing clothes. No way.

Speaker 1: So you would actually decide like where the buttons go on

Speaker 4: the shirt. Yeah. Exactly. That level. Drawstring length of the hoodie and all Really? You know, all the all the stuff.

Speaker 2: That's Why was it the worst decision?

Speaker 4: Yeah. Well, so it's like, you know, you've seen Devil the Wears Prada. Right? Yeah. That's how it actually is.

Speaker 1: Okay. So people have really big

Speaker 4: mean, strong you it's it's constant to this like, you should be happy to be here. Oh, okay. In the worst, it's like toxic. Sure. Sure. You can get into the sustainability part of it, blah blah blah. But like, there's no career prospects. You're basically, we say, licking ass for ten years and maybe maybe maybe maybe maybe you'll be the creative director.

Speaker 1: Yep. With a big Maybe. Yeah. Yeah.

Speaker 4: So not something for me. So eventually, I pivoted to tech Mhmm. Where it was like this is back in 2018, 2019. So then, that was like job security, like all the fun design stuff was kind of going on there, building like branded experiences on web and mobile. I worked at PayPal in Sweden, in Vodafone, a bunch of different brands Cool. Until I ended up in an AI startup in 2023.

Speaker 1: Okay.

Speaker 2: There you go.

Speaker 4: Stick up timing. Yes. Exactly. This was early. This was I'm tinkering with Midjourney since mid twenty two.

Speaker 1: That Yeah.

Speaker 4: Very first one came out and I'm like, I can actually get like a hovering sneaker in the air. Yeah. And it looks like something. Before then, was disco diffusion and Dolly and all these kind of weird, like a retro AI now, guess.

Speaker 2: Yeah. Vintage. Vintage.

Speaker 4: Vintage AI. Exactly. So from there, we did a bunch of stuff over 2023 with which Chapter has they're called? It's a German AI startup and eventually You were in German I was working remote from Sweden. So remote

Speaker 7: Oh, okay.

Speaker 4: That's, you know, happened as So I did that and eventually they pivoted. We parted ways and at that time, like 2023, I'd been posting some spec ads just like a little bit more art directed, maybe I don't want to toot my own horn too much. But maybe for the time, it was like just a couple of levels up what most people were seeing and all of a sudden, my DM is like people from like the first AI agencies, monks, the big agencies, like we have to rep you. That's You have to do AI with us. All organic on LinkedIn.

Speaker 1: Yeah. Explain explain what what what the workflow was for those first ads. Is that static images that are AI generated, then you're doing text layout and something like Figma? Or is it video from start to finish that you're cobbling together? Are you doing like mixed media where you're using some stock footage, some AI video, editing together after effects? Like what's the workflow?

Speaker 4: Well, that's where we are kind of now.

Speaker 1: Back Back then then?

Speaker 4: Back then it was like just getting anything that felt realistic

Speaker 1: Okay.

Speaker 4: Had a bit of grit and texture to it that was Sure. This like, you know, sheen plastic

Speaker 7: Yeah. Mid journey look.

Speaker 1: Yeah. Yeah.

Speaker 4: All in the industry.

Speaker 1: So you were able to get away from that with the Yes. With with the correct prompting?

Speaker 4: Correct prompting is having an eye for art direction, angles, lighting. Yeah. And then editing. Yeah. Editing. Yeah. Sure. Well, on on a where where to put the logo and you just get get like good pacing and everything into so it feels like an ad. Mhmm. That's I think was back then really kind

Speaker 1: of Interesting.

Speaker 4: A bit a bit lame

Speaker 1: Yeah.

Speaker 4: So to say. So so back 2024, did a bunch of work, did a music video for Hardy and Fred Durst, Valentine's Whiskey.

Speaker 1: Limp Bastard.

Speaker 4: Yeah. Oh. They really lean into AI so much.

Speaker 1: Ben over

Speaker 6: there Really?

Speaker 2: Know who Limp Bastard is.

Speaker 4: They've done a ton of AI videos. Yeah. Fred Durst and Limp.

Speaker 1: Yeah. I mean, I remember like the original the original cover art for their for their big album was I think it was airbrushed, but it looks like AI art now.

Speaker 2: Yeah. Were just so we were just asking the founder of Suno yesterday on the show what artist has leaned into AI the most. And he didn't he didn't want to name names Yeah. In part because he actually knows who's using the products, he doesn't and and a lot of them want to be quiet about it, but he didn't have, like, a really clear answer on, like, who's leaning in the most. But that seems like leaning in quite a bit if in 2024, you're like, we're gonna use this to make Yeah.

Speaker 4: Yeah. Yeah. A lot of crazy kind of glam rock video, all kinds

Speaker 1: Yeah. Of matters. And and there's so many times when an album comes out and they can they have the budget for one classic Yep. Music video or two or maybe three and it starts to get low budgets like, oh, let's use like a DV camera for this shoot or let's not bring the let's not rent a soundstage for this. We'll do something a little more organic. Yep. Maybe do some tour footage for this one and stretch the budget further. But, you know, you can with AI, you can probably just do a video for every single

Speaker 4: Oh, yeah. Mean, that's the real opportunity. Think people are still it's like staying in the live action box

Speaker 1: Yep.

Speaker 4: Which is a fine box Yeah. When you're doing live action. But like the truly and we can get into that later. Scale and the test to just do different specific stuff for different channels Totally.

Speaker 3: All this stuff.

Speaker 4: Yeah. We haven't really gotten there

Speaker 1: yet Cool.

Speaker 4: On the client side. Yeah. They don't know what to ask for. Yeah. They just know they should be experimenting with it.

Speaker 1: Sure. Sure.

Speaker 4: So but yeah. So then 2025 Okay. We did like So

Speaker 1: you're wrapped at this point by agencies, or they call you?

Speaker 4: A lot of agency rep at that point slowly built. Again, I was nobody on social media. I I mean, I'm not, like, 30 k on LinkedIn. I'm not some huge guy.

Speaker 2: But,

Speaker 4: like, I have a huge sign, but but not a huge guy. So it's building a following, building building attention. Yeah. And then eventually, that kind of formulated into a huge huge projects over 25. So Mhmm. We did activation campaign for Taylor Swift. We worked with Lewis Capaldi with Google and YouTube Yeah. Universal Music Group, bunch I mean, yes. A lot of a lot of real big brands together with Wonder Studios and a bunch of people I just got to know got to know over the I mean, I worked with Dave Clark back in the day 2024. Sure. If you hang out on LinkedIn, you'd seen

Speaker 1: Yeah.

Speaker 4: Seen Shadowkeep.

Speaker 2: And what is your what is your team look like to execute on all this?

Speaker 4: Great question. It's it's growing, but it's the the biggest bottleneck for us is actually finding people, like on the artist side. It's super difficult to find people who ask like the right circumstances. So I also teach at Sweden's like biggest creative school called Bergs School of Communication. They have, at least what we think, the world's only AI program at this One year in classroom, government funded Yeah. And I'm there for three weeks. So the ones with a little bit of a freelance appetite, they tend to come to me.

Speaker 2: Yeah. It's So like It's a great recruiting strategy.

Speaker 4: Part of the talent funnel is really good. I mean, like Promise, the AI studio, they bought the Curious Refuge, if you saw that.

Speaker 2: Yeah.

Speaker 4: AI online education. Cool.

Speaker 2: So Yeah. So so many questions. So like on the projects that you mentioned in 2025, what what does a team look like for an individual project? How quickly are you completing them? Like how would how would you compare those projects to the same type of production from, you know, a decade ago?

Speaker 4: Sure. Sure. Well, it's really different, I would say. I mean, everyone has a different take on this. I think having a lean and mean squad is better than it was like, we used 30 artists. Like, you can do that, but that creates a kind of a too many chefs problem,

Speaker 2: I think. Yeah.

Speaker 4: We have to have like a really strong lead and then delegate that. So structure is also everything in an AI Yeah.

Speaker 2: And there's a cost thing. You're running a small business, a creative agency.

Speaker 4: Right?

Speaker 2: Yeah. Like Yeah. Every incremental person that you bring on is just eating into that Absolutely. Overall project fee.

Speaker 4: Absolutely. And some are more efficient with prompting and others are not. I mean, the big cost is not necessarily on the credit side, it's hours, the human hours that's still Yeah. And that's the key on, you know, as you know, it's not what did they say, Kay Kay, the the LA news channel. I mean, he just pushed a button. I mean, that's the classic trope. Right?

Speaker 1: Yeah. Yeah.

Speaker 4: It's definitely that. Yeah. But smaller teams that can do more, I guess, is the is the short story Mhmm. Of it. And it's super different. Like, I wish there was a formula, always this. Always that, you

Speaker 1: know. Yeah, of course.

Speaker 4: But having a lean and mean squad of four or five. I think for this Lewis Capaldi, we were seven artists in total working under the Wonder Studios umbrella who And facilitated

Speaker 1: and for those folks who are joining you and working with you, do they have similar paths where maybe they were classically trained in art or design and then they narrowed down at some point to fashion or typesetting or typography or visual design or or video editing. And then this sort of allows them to zoom back out and and sort of almost retreat to the more classic train, the more aesthetic driven work?

Speaker 4: Something like it. I mean, I think the AI kind of family, the global one, we come from kind of all walks of life I mean, you have, you know, Gossip Goblin or Zach London? No. He's he's blowing up right now. Okay. He had a, I think, Hollywood Reporter article about him as well. Yeah. He's an ex anthropology student. Interesting. So like, the beauty of this stuff is that it's not only ex industry peeps. Of course, you have also live action directors. I have a weird non linear design career. Sure. People come from documentary

Speaker 1: Yeah.

Speaker 4: But all kinds but again, or just because what truly matters is what's what's in here Yeah. And how you can articulate

Speaker 2: And just an obsession with using the tools.

Speaker 1: Yeah. Yeah.

Speaker 4: Sure. Yeah.

Speaker 2: Yeah. We can

Speaker 4: that this never ending. The

Speaker 2: Yeah. Yeah. Yeah.

Speaker 4: I'm sure you've seen

Speaker 2: How has your stack evolved? Like, what models what models were were you using in 2024, 2025? I'm assuming you're constantly trying all the new tools.

Speaker 4: Right. Yeah. I mean, up until '24, I guess, mid journey was king and queen and the it was everything. That was definitely a big model. Yeah. Then you kinda had Flux and a bunch of other more indie ones coming in and the scene becoming more competitive, which is great. I think that's I mean, I think that's also what's happening with Sora, know. Yeah. Not everyone's going to make it. It's just a testament to a healthy market. Yeah. That's how

Speaker 9: it So should

Speaker 4: then as Google kind of entered the chat like really bigly

Speaker 1: Yeah.

Speaker 4: To use

Speaker 1: a vo three and

Speaker 4: banana. Vo three, not a banana. The Google Suite has become kind of industry standard for for us. So definitely using that a ton. Okay. And then of course now, Cling three, C Dance two. They're kind of king of the hill Okay. Now in AI video.

Speaker 1: How is your access

Speaker 2: to Over over VO three.

Speaker 4: Oh, yeah. Yeah. That's it. Stands right now.

Speaker 1: Yes. How is your access to Sea Dance three or Sea Dance in general? We've heard that there's like long queues, there's rate limits. It's it's not something that

Speaker 2: Billy Boman does not get limited.

Speaker 4: Well, so I was fortunate enough to be in the Dreamina CPP, so creative partnership program. And I've had pre access for

Speaker 2: Told you. A month or so? Don't ever rate limit Billy Boman. So, yeah. The man with his own Hollywood sign.

Speaker 1: Yeah. Yeah. You don't want

Speaker 4: You gotta have C Dance. They're rolling it out now and a fascinating kind of play there, at least from the rumors I've heard, is that US might be kind of the last on the list. I agree. Because as Europeans, we're kind

Speaker 2: of So why is is their model better because they just downloaded every Hollywood film, every history and

Speaker 4: We can always speculate. Right?

Speaker 1: Yeah.

Speaker 4: I mean, that's the the dataset and the training that went into Yeah. Sure is, you know

Speaker 2: So it's basically higher value data than even YouTube. Right? Because YouTube is like Potentially. Yeah.

Speaker 1: Yeah. So that Although although it feels like at least trailers for every movie are on YouTube and I would Sure. Assume like there's that like every Marvel fight scene gets clipped and is on like the movie clips channel.

Speaker 2: Yeah. But that almost that that'd be like if LLMs were only trained on content above paywalls.

Speaker 1: I guess, you're right. Mhmm.

Speaker 4: Yeah. I mean, it's all kind of what you say, the inner workings of this So

Speaker 2: How how is the concept or the like how is prompt engineering evolved? Because there was a period

Speaker 4: Mhmm.

Speaker 2: With just regular LLMs where everyone was obsessed with prompt engineering and then the models got it much better and of course everyone still has their own prompts and now skills and things like that. But for the average the average user is not even really caring about prompt engineering at least with text now because the output is just good enough for so many different use cases.

Speaker 4: Yeah. Yeah. I mean, I I love that topic because I think prompt engineering was one and synthography and all these kind of titles around the stuff. Don't I don't care, man. It's like making images, whatever. So it's like people tend to want to make things more fancy than what they are Mhmm. Because they want to hold on to the old idea

Speaker 1: Yeah.

Speaker 4: That getting good output is technically difficult.

Speaker 1: And that's

Speaker 4: just gone.

Speaker 1: Yes. That is just And that was the that was the case with like you're a master of After Effects or Photoshop and like that you're a master of the tool. And so even if your ideas are mediocre, it's like, you're the only person that can actually instantiate anything. So if I just give you Yeah. These technical moats

Speaker 4: are around. Exactly. It used to be that way. Yeah. So Stephen Bartlett Yeah. Devourer of the CEO Yeah. This brilliant PDF article on LinkedIn last year that's called Your Mote Is Dead. Yeah. And that just summarized all of that. But but back to your question, prompt engineering has changed a lot in the sense that the models have gotten so much better understanding natural language. Mhmm. Like you don't need to go super fancy. Yeah. I mean, I do sometimes because the boring answer, it depends on what you're

Speaker 1: I remember old mid journey, you'd say like, not don't do six fingers.

Speaker 4: Right.

Speaker 1: Please please please do five fingers.

Speaker 4: Right. All the negative stuff. Yeah.

Speaker 1: You and you put in like the lens and the and the color grade that

Speaker 4: you Which you can still do.

Speaker 1: You can still do that. You can do some of that yeah. Some of that got, yeah, just abstracted. Yeah. Yeah.

Speaker 4: And and you can still bring it in. And I think that kind of goes back to like Mhmm. Everyone that comes into this industry from different walks of life will have a new perspective because they sit on a type of terminology and vocabulary and also taste and Yeah. And craft.

Speaker 1: And that goes to the anthropologist who's gonna come with a completely different set of language and get very different results.

Speaker 4: Yeah. Or or even me. Like Yeah. I'm not a live action director

Speaker 1: Yeah.

Speaker 4: At all. Like, again, I'm a weird designer guy Yeah. But I have a lifelong passion for movies. Mhmm. I've done my fifty thousand hours or whatever. Like, I could go mean I mean, the You

Speaker 2: should go movie for movie with John.

Speaker 1: Yeah. Have you seen Borat? Yeah. Oh god. Classic. It's a classic.

Speaker 2: It's actually Sorry. Sorry. It's a running running joke.

Speaker 1: Okay. It's the only movie that he's seen. I've seen a lot of movies. I I love movies. But we joke about it.

Speaker 2: But he's also Do you think people are how good do you think people are at spotting deepfakes today? Because people I think would like to think they're good at spotting it, but there's so many videos out there Yeah. In 2026 that people are like, oh, that's AI or that's not AI.

Speaker 1: I've gotten caught a few times. I've noticed. These funny vehicles, if it's a funny vehicle and it looks like it's filmed on a

Speaker 2: It's gotten to the point where people will be debating like physics.

Speaker 1: Yeah. Like

Speaker 2: that that is not impossible. And they're like, well, look at this study.

Speaker 4: Yeah. But I think Runway put out a pretty funny post a couple of ago about like basically an AB test of like Yeah. It mixed AI and and Oh, and I think they used real start frames, so like an actual photograph and then one was just the real Continue. And then the other one was animated with AI. Oh, interesting. Anyone can go and take that test right now, but like Yeah. We're we're we're there. Like on Bruce one

Speaker 1: The New York Times ran a similar thing with poetry, literature, science fiction Yeah. Sure. Yeah. All around AI generated content versus non. And a lot of the New York Times readers preferred the AI generated content.

Speaker 2: What industries so you're getting hit up by artists, you're getting hit up by companies that wanna make ads. What industries are or what is there anyone that's like not hitting you up and saying like, we're never gonna do this? Because even I I think people were pretty surprised when Gucci came out. Like Yeah. This was a

Speaker 4: was sending them a queen, all the big fresh

Speaker 2: Yeah. They came out and said like, yeah, we're making, we're doing and and there was it was they had an interesting approach where some of the images just looked exactly like normal Gucci photographs and then they had like some GTA esque Mhmm. Images that like clearly were were generated. Right. And and I thought that was an interesting decision because if they just came out and were like, we're using AI but it looks real. Yep. I think people would have had like less of a negative reaction.

Speaker 3: Mhmm.

Speaker 2: And then some of the stuff was just like obviously CGI. So kind of muddy did a little bit. But who is it seems like everyone is now starting to experiment. Yeah. But who is there anybody

Speaker 4: So so I would say overall for us, like being a Stockholm Sweden, maybe we don't have as much access to the big narrative stuff as you guys have here. Yeah. Will will that change? Maybe. But mostly for us it's been branded content and commercials. Just because like the the ins we have with agencies, just how we position ourselves,

Speaker 1: that's Yeah.

Speaker 4: Our bread and butter. I'd love to do more narrative. Yeah. If I ever had time, I would do more short films.

Speaker 1: I just

Speaker 4: don't have any time.

Speaker 1: Also, it just does feel like it like in terms of the actual workflow, doing a three hour, two hour, three hour movie from start to finish Yeah.

Speaker 4: Don't do

Speaker 1: it. It's going to be iffy to say, okay, does the character look exactly the same two hours later?

Speaker 2: What is what is the what what advantages are companies getting? Is it speed, cost? Flexibility. Like, even even not quality, but yeah, flexibility, like, can they do scenes that they never would have question and

Speaker 4: I think my kind of favorite sound bite that anyone can steal is, I'm not going to copyright it, but don't spend less on budget. Spend more on imagination. Don't spend less on budget because that's just like, you're just going to paint yourself into a corner if everything is about like Cost reduction. Cost reduction like Yeah. In this media landscape with Yeah. Everything that's going on on every platform and then get ever competitive and now ads are going to look better than ever. The rising tide lifts all boats, right? Yeah. So like, I just think that's a completely backward strategy. So rather

Speaker 2: couldn't there be an argument for keeping the budget fixed but having more variation You can the underlying creative?

Speaker 1: I think that's exactly what he's saying. Yeah. Yeah. He's saying that that if you're if you're a brand and you're spending a million dollars on creative assets and you say, great, AI is here. I'm only going to spend a 100 k. Well, your competitor is going to spend the same million, but they're to get 10 times, maybe a 100 times as much content, as many variations Yeah. As much creativity Yeah. And they're going to

Speaker 4: smoke you.

Speaker 1: And so Or even just the competitive

Speaker 4: nature of the content because we can fucking put the camera on an arrow Yeah. Flying through a battlefield and over the monkey and then squirrel takes it up like anything. This guy's a director. Anything. Like, totally. So so, I mean, we're waiting for those briefs. Just to our earlier point, clients haven't gotten there yet. Sure. Don't know. It's just like, oh, here's we want to do a live action ad. Can we do it with AI instead? It's like, can we fit in this old box while the old box is gone? You can kind of do anything. So that's what's really exciting. I mean, did a protein bar ad with a Swedish Knicks. They did Kardashian has a big face a couple of years Yeah. Anyway, the brief was to like fly down through the clouds. Mhmm. A girl is glamping in like Okay. Argentina and she eats a bar and then feeds it to a capybara that starts twerking.

Speaker 1: No way.

Speaker 2: You capybaras can get violent, right?

Speaker 4: Yeah. But these were very friendly,

Speaker 2: the capybaras. This is a this is a good use case for AI because capybaras are some of the most docile creatures until they're not.

Speaker 1: Until they're not.

Speaker 2: We we played a video of a capybara attack earlier on

Speaker 4: the You might

Speaker 1: have saved a life by using AI to generate the capybara instead of using real animal handling.

Speaker 4: And there you have, know, the the sustainability issue as well. Like, okay, the the the carbon footprint of flying to Argentina Yeah. Doing all this the traditional way. Sure. Yeah. Good luck. And, you know, we could measure like, you'll never win Yeah. The sustainability argument against AI as well.

Speaker 1: Interesting take. Yeah.

Speaker 2: Any yeah. Any any I feel

Speaker 1: like the AI industry has very much been on its back foot playing defense on that saying, no, no, it's not that it's not that better.

Speaker 2: Yeah. The example the example that I've always used in the in the in the film entertainment context is the idea that, you know, somebody making a film would fly out 50 people to another continent to get scenes that would ultimately take maybe sixty seconds Yeah. Just to tear it

Speaker 1: Yeah.

Speaker 2: Yeah. It's it's yeah. It's certainly gonna be more sustainable to like do the right few hit the keyboard in the right way Yeah. For enough times to get an output.

Speaker 1: Yeah.

Speaker 2: What what advice are you giving to creatives? Because what you've done is exactly what we've talked about on the show Yeah. And what I think the general point of view at least from the tech industry is, yeah, if you wanna resist AI, you're gonna you might like feel better about yourself for some amount of time, but eventually you're gonna get steamrolled by somebody who is loves doing the same type of thing that you're doing is just gonna use a different tool set. So, yeah. What what advice are you giving to creatives across their kind of career on how to how to actually lean in? And I think there's a way to lean in and get just as much satisfaction, have just as much intensity of the creative process and just as much joy.

Speaker 4: And control now. Yeah. We're getting there. And truly, don't even have to choose anymore. I I typically say it's not about AI or it's AI and. Mhmm. Because we also did a recent like a Super Bowl commercial for a big airline. All the talent were shot green screen. Yeah. We background and relit them, put them in like a beautiful cloud environment and all this stuff. So you don't even have to choose anymore. Yeah. You can keep the human talent, the human performance and all that stuff. But the the choice the advice for creatives, well, just test it. Don't say blah until you tested it. Mhmm. Like, and if you have tested it and it's not for you, well, then wait or stuff but this is not gonna go away. Yeah. I I just fully accepted that that like this is not gonna go away. Yeah. I was also scared like I I'm not I'm not I've not always been this like happy go lucky optimist.

Speaker 2: So you started absolutely printing.

Speaker 4: Well, we'll see. The jury's we're doing fine. The jury's still out. But to that note, there's more opportunities than ever. Mhmm. The ad world is in kind of flux of, like, we need stuff in two weeks. You'll get good luck with CG and live action. You can kind of Okay. I don't always want these two week briefs, but you you can help. Yeah. You know, do this

Speaker 2: You'll take the the six month brief.

Speaker 4: Yes. Yes. Too long for AI, would say. Keep it about a month, a month and a half, I think, for for a couple of ads or like thirty seconds But

Speaker 2: still that is dramatically more than just like typing into the box what you want. Yeah.

Speaker 4: And Oh, yeah. Yeah. Because it's like just going into an AI project. You have to think about casting. Are we doing an AI phase or are we that we could where you can choose all the looks, demographics, everything? Or are we doing like What what

Speaker 2: feedback do you give the various labs on what they're doing badly today?

Speaker 4: Right. So I mean, we are in a bunch of these creative partnership programs and it's great because you do get the feedback and it can be anything from like the the if that's what you mean, like the AI providers. Yeah.

Speaker 2: Yeah. Like the five the the six finger thing

Speaker 1: Cheaper, faster Yeah. Better with a certain style.

Speaker 4: Sure. Sure. I mean, what we're doing is what I call like live art directed realism. Yeah. That that that's if it doesn't look like a movie frame, it's not good enough

Speaker 1: for us.

Speaker 4: Okay. Like that's that's where kind

Speaker 1: of Yeah.

Speaker 4: My base level is. So I've we found a really neat I mean, you've seen some of the work maybe. Yeah. I have I sent it up if you wanna pull it up.

Speaker 1: Yeah.

Speaker 4: If it's not there, it's not good enough. So like realism, that's others focusing on animation, other focusing stylized claymation. There's you again, you can Yeah. Kind of do anything. What I want is better, more human focused interfaces. Mhmm. I think that's like making the tools more human and

Speaker 2: less More SaaS.

Speaker 1: Sorry? More software. Like software Like

Speaker 4: using the product. Like Yeah. Actually like I I wanna queue

Speaker 2: up You just want a chatbot that you can just No.

Speaker 4: God. No. No. We don't I mean, I don't generate in in OpenAI. Exactly. Yeah.

Speaker 1: Maybe Have you have you thought about getting into like the vibe coding world, building the tools that you want or having your team work on that, building a harness around all of the different models

Speaker 4: actually VibeCode in. Okay. To some extent, there's a really neat web flow plug in that you can use to as a bridge to Yeah. To get into the to bring Claude in into there. Sure. So there's just no time. Yeah. Of course. I I hear a lot of people feel the same like they're just the exhaustion of keeping up. It's insane. I mean, I do my best and I try to focus Mhmm. You know, branded content ad stuff that we're doing. But I do wanna

Speaker 2: How many Super Bowl ads do you think you'll do for the next Super Bowl?

Speaker 4: We'll see, man.

Speaker 2: Like like over five?

Speaker 4: There were AI Super Bowls. Yeah. Yeah. Yeah.

Speaker 2: So I but I figure you're gonna get at least a

Speaker 1: Do you think next year there'll be more than half? More than 25%?

Speaker 2: I would I would guess

Speaker 1: that I think it was like

Speaker 2: more than half 2%. Use AI. Year.

Speaker 4: Exactly. Exactly. Somewhere Background replacement. Yep. Relighting. Yep. You know, combining.

Speaker 1: Even even just like there is a new AI driven rotoscoping tool that just replaces the background and that's AI, but it's just a replacement for sitting there and cutting out something.

Speaker 4: Yeah. Yeah. Yeah. All these kind of menial tasks Exactly. That no one loves to do. It's just like a machine Yeah. Is better at that. Yeah. Can we let the machines do it?

Speaker 1: Little blemish replacement.

Speaker 4: Yeah. What I don't want AI to do is like writing the scripts, coming up with the ideas because that's the fun stuff for us. Will we get there? Maybe. Yeah. But right here, right now, it's trash.

Speaker 1: Like Yeah.

Speaker 4: That's not at all where it's at.

Speaker 1: Yeah. Yeah. That makes a lot of sense. What what does the integration with Fiverr actually look like? I it it doesn't sound like I can get you to make me a Super Bowl ad for $5. No. So so so so is Fiverr just a is Fiverr just a way to meet you? Is is this just more of like a a marketing partnership for Fiverr to open up people's minds about the type of people that they could hire?

Speaker 4: Sure. Great question. So besides the name Fiverr Yeah. You know, that's kind of stuff Yeah. Always there's a lot more high budget stuff on Fiverr

Speaker 1: these Okay.

Speaker 4: And what we're doing with the hub Yeah. Is that they've curated like a really solid selection of some of the best and with different styles also Yeah. That you can go and hire. Now because like our experience as well of talking to agencies like, yeah, we found this guy on Instagram and DM'd him and it was terrible. Like trying to find people who are vetted and really good industry pros is hard because there's like a handful of people in the world that does this at a level that is good enough for commercial production. Yeah. So I think that's like the opportunity they're presenting on on the hub, like come and work with these people.

Speaker 2: Here's here's an interesting question. So I've always felt that at any given point, at least when it comes to kind of startup style design and branding, that there's only maybe like a 100 people in the world that actually do great sort of like leading edge work. Mhmm. I've always thought that was funny because we for so long, you know, millions of people can make a website, can make, you know, product design, etcetera. And yet there's still like this small group of people that actually are pushing the frontier and kind of like leading from a style standpoint and kind of like setting setting the tone, setting the the actual taste makers. Do you think that that will stay the same with AI? My my expectation is that is that it will to some degree. It feels like an iron law almost that at any given point, regardless of the tool set, there will only be a small number of people that are true, true tastemakers.

Speaker 4: Absolutely. I mean, this is true already today. Like, if you are on socials, lot of people are asking for work. Yeah. Like, kind of mid I mean, people are new. People are kind of happy amateurs playing this stuff. Most people might not have a

Speaker 2: And there's still there's still demand for for kind of the middle of the pack. Sure. Yes. It will always It's just that you're not gonna get the, you know, the the like, quickly people identify the 100 or and and realistically, it can get down to, like, 10 or something

Speaker 4: like Something like that. Yeah. I mean, I think what's happened is that this the the floor is raised, but so is the ceiling. Yeah. So, like, really standing out is also more difficult because Yeah. It's like people make awesome stuff. Like there's so many talented people in the in the industry. So Cream will rise to the top but it's also more competitive because as the baseline, anyone can make a cool fight scene now. Okay. What's next? Like what's what's actually makes us tick, so to say?

Speaker 2: Yeah. Yeah. Where how are you building the company? You're you're here in LA because you have your own Hollywood Sign, but you're based in Stockholm. Stockholm. You're building out the team there or do you Yes. Remote?

Speaker 4: We we really really kind of what do say, step by step, brick by brick, really really Mhmm. Carefully selecting who we bring into the team. So we have two assistants at home now holding the fort. Shout out to you guys for Mhmm. So we could be here. But like to our earlier point, it's really difficult to find people. Yeah. So what's been a good funnel for me is working with teaching the stuff as well and Sure. Identifying some of the leading talent that are kind of AI native that grow up kind of going to school with this stuff. Yeah. They might need more traditional schooling in terms of tools, art direction, and stuff that they don't have because How to work with clients? That too. Yeah. That too. Being able to take a brief, all this stuff. So really really slowly and like human at the at the front and center of everything. Mhmm. But we have like a temporary office. I mean, it's literally just been me and my co founder for the last years doing this.

Speaker 2: That's great.

Speaker 4: So we're some type of yeah.

Speaker 1: It's Yeah.

Speaker 4: I don't know.

Speaker 2: Amazing. Amazing. I mean, it's so so fun to talk with you and because like, we haven't had we we've had plenty of kind of creatives come on the show that are using using the the different tools quite effectively. But nobody that's just like made it like Yeah. I am gonna recreate the modern ad agency.

Speaker 1: I I I think people sort of wrote off how like the models are so amazing and so it's so magical when you fire off one prompt and you get something that looks real that people just completely wrote off the middle, which is like actually making a product of Yeah. Like a a story or something. We had this idea for an an AI video that would be me and Jordy on the microphone throughout history different like big innovations. So the invention of the steam engine, the invention of the telephone, the invention of the, you know, Wilbur Wright launching the first airplane and we're there like sort of talking about it and we and we generated a still frame, tried to animate them between it was just like it was quickly turning into like a multi week process that should have been like a month that we budgeted. But we were like, let's spend thirty minutes on this and like hopefully we'll make it work. We never quite got it there, so we just never published it.

Speaker 4: You know what you should do? Should you should film it.

Speaker 2: You have

Speaker 4: a green screen here. Yeah. Yeah. Mean, just your background places. Oh. It The voices are real.

Speaker 1: You don't have

Speaker 4: to clone your again, you don't have to choose. It's AI

Speaker 1: and That's really good.

Speaker 7: So we like

Speaker 4: You have everything here, so we could do it.

Speaker 1: Okay. Last question. Aura recently partnered with Chrome Hearts for a ring. Yes or no? Was this something you'd

Speaker 2: pick up? Is this something you'd reckon?

Speaker 1: Is this fake news?

Speaker 4: Yeah. Departed already. Well, for the record, I don't I can't afford Chrome Hearts just yet. You know,

Speaker 2: it's Hey. Let's next Super Bowl season, you might be crummed up.

Speaker 1: Might be. Might be. Okay. Well, it's been fantastic.

Speaker 4: Likewise. Likewise.

Speaker 1: Thank you so

Speaker 2: much for coming down. I'd to meet you.

Speaker 1: We should definitely work together on something. Let's figure out something to work on. Yeah. That'd be amazing. Yeah. Yeah. Follow-up. I'd love it.

Speaker 2: Awesome.

Speaker 1: Have a great rest of your day. We'll talk to you soon. Let me tell you about Plaid. Plaid powers the apps you use to spend, save, borrow, and invest, securely connecting bank accounts to move money, fight fraud, and improve lending now with AI. And let me also tell you about Lambda. Lambda is the superintelligence cloud building AI supercomputers for training and inference that scale from one GPU to hundreds of thousands. And that brings us into the Lambda lightning round, and we are kicking it off with Benjamin Miller from Fundrise. But first, will pivot that camera around, show you the entire TBPN UltraDome, and we will show you the Lambda cloud because we are in the Lambda lightning round right now. And I believe we have Benjamin Miller from Fundrise in the restream waiting room, so we will bring him in to the TBPN UltraDome. We'll try that again to bring him into the TBPN UltraDome.