

Fall (Features and Labels) discusses its $1.65B Series E & F, why generative media is a major AI product-market fit, and how compute/inference economics drive demand and pricing. They argue ads and marketing will be the biggest near-term use case, and that “infinite demand” is emerging as video quality improves.
Gerkem (co-founder/leader at Fall; Turkish; previously at Coinbase and later helped build Fall with a technical team) and Berkai (co-founder/leader at Fall; Turkish; previously at Amazon). They met in San Francisco in 2014 and began collaborating during a COVID Palm Springs house rental; company started ~1 year later. They emphasize hiring “AI/GenMedia-pilled” engineers with agency, humility, and hunger.
Generative media is one of three AI PMFs (search/chat, coding agents, generative media). Individuals can spend $10k–$20k/month generating video. Video model costs rise sharply with longer/higher-resolution outputs. Pricing spikes (e.g., VO2 from ~$0.30–$0.40 per 5s to ~$5 per 5s) didn’t suppress demand because professional ad/movie use cases expanded.
Sora (hosted on Fall; saw growth before it was taken down for compute/cost reasons). Ads: UGC-style, “throwaway” content; automated iteration and reinforcement-like testing (thumbnail/title testing analogy). Open vs closed models: open sourcing helps studios fine-tune and reuse unreleased styles, but trust/enterprise concerns remain.
Written by AI. May contain mistakes.
Discussion on the evolution and demand for generative media in AI.
“Now I'm just going to keep on saying it.”
Exploration of the transition from image to video models in AI and revenue impacts.
“I think you've also hit the point where you don't want to do any media on it anymore.”
Insights into the increasing demand for high-quality video generation in various industries.
“surprised by how much money there is behind this?”
The role of generative models in the advertising landscape and future predictions.
“So it's like the usage is going up and then the capabilities are going up and then everything's getting more expensive over time.”
Discussion on the evolution of Fall from a compute company to embracing AI growth.
“So let's talk about the engine behind Fall and then we can kind of get into how your culture is keeping up with the growth because it's crazy.”
Learn about the transition from traditional data infrastructure to focusing on AI and image models.
“You know, so in order to build the next generation of like the data infrastructure company, we also have to focus on compute.”
Discover how the founders built their team and the importance of team chemistry and motivation.
“like niche, which, you know, three years down the road, it's more well established and it's a whole like new industry.”
Examine the challenges of hiring in a rapidly evolving market and maintaining company culture.
“It needs power, land, and infrastructure.”
Learn about the emerging role of super individual contributors in tech teams.
“So we have forward deployed engineers and now we have super ICs.”
Understand the scaling challenges and communication strategies in growing teams.
“I think we're still keeping everything flat and see when things break.”
Discover the key traits needed for engineers to thrive in dynamic environments.
“It's like you can't just be a 10x engineer.”
Explore the dedication and hard work necessary to grow a successful startup.
“I think that helps maybe to identify hunger.”
Examine how compute resources influence the development and deployment of models.
“I asked him what he's most excited about.”
Learn about the diverse modalities in generative media and their evolving landscape.
“So that has become our responsibility in the ecosystem these couple months because compute is so hard to come by.”
Explore how creative naming can affect the recognition and popularity of models.
“So if you need to put more data into the model, there is abundance and it's seemingly free.”
Learn about the amusing and creative names used for AI models.
“I think when Nana Banana had its moment, everyone was talking about Nana Banana for like three days.”
Discover the origins and meanings behind the name 'Fall' in AI.
“So, you know, it's like your future prediction.”
Explore the ongoing debate surrounding open source and closed source AI models.
“Mercore's AI Productivity Index, or APEX, tracks it with every major Frontier model release.”
Discuss the perceived and actual impacts of AI on Hollywood and content creation.
“Talking macro, what do you think the biggest question is that people aren't asking right now?”
Discussion on how persistent worlds and AI will impact gaming.
“Before you used to like look this way you know see the flower look away come back it's gone.”
A conversation about influential figures and personal hot takes on the industry.
“So I'm curious, who are some figures that really inspire you and keep you motivated?”

Get the gist, explore an idea or ask about something said.
On Sourcery, what are the key points of this episode?
Automatic transcript. May contain errors.
0:00You guys have the coolest brand on earth, you guys are growing really fast, you keep on raising more and more money. AI really has three product market fits. One of them we believe is generative media. Three years ago no one really understood what that meant. Three years down the road and it's a whole new industry. You sit down in front of a computer and you generate videos all day and you spend$10 ,000 -$20 ,000 a month. And we were like okay who the heck is gonna pay all this money? Turns out there is essentially infinite demand. Every big company is gonna be a compute company. Just the storefront is gonna be different.
0:30Our storefront is unique and interesting. No one else has a storefront similar to us. Were you surprised by how much money there is behind this? 100%.
0:48Okay, cool. All right. Gurkem. Wow. You really don't need to. Gurkem. Gerkem. Gerkem. Yeah. Wow. That was the last one. That was the last one? That was very good. Gerkem. Yeah. Gerkem. Now I'm just going to keep on saying it. Yeah. That's perfect. Teaching AI. Where's my camera? Internet. You've been pronouncing Gorkum completely wrong. It's Gerkem. AI. Learn from the YouTube videos. Berkai? Yeah. Yeah. Perfect. You guys, this is great. You guys have the easiest names on earth. Easy. Welcome to Sorcery. Thank you so much. We're in a very special place. So do you want to maybe tell us where we are?
1:30Sure, yeah. So we're at Dalida. Shout out to Sayat and Laura. Dalida is an Eastern Mediterranean restaurant with a lot of Turkish influence. And fun fact, I got married here and Gerkem officiated it. Great wedding. Congratulations. Thank you. Amazing. Well, I'm excited. I can already smell the food. everybody is eating it behind the scenes. So as much as this conversation is going to be rich and delicious, I'm sure the food is as well. It's amazing. Okay. So I think I've reached out to Michelle like a billion times because your name kept on popping up. You guys have the coolest brand on earth.
2:11Like you guys are growing really fast every single time. And it's about around because we cover funding rounds. I'm like, I got to get them the next time. No, you guys already raised it. It's gone. It's over. The next one, gone, over. And I'm like, keep on trying to catch you guys, but you keep on raising more and more money. So congratulations on the recent round. Thank you. Thank you so much. You did like a couple back to back rounds, like any fast growing company these days, but this last one is the biggest one yet. So thank you so much. Congratulations. I think you've also hit the point where you don't want to do any media on it anymore.
2:49So you've passed the inflection point of like the hype cycle. And now you just have a humming business. Exactly. If you're rather focused on the business, the announcements, you know, it gives some, you know, push in the beginning. But as you do more rounds, I think it's irrelevant. It's the business that matters more. The adrenaline has worn off. Exactly. It's probably worked for hiring and that kind of stuff and customers, but you got to focus. Okay. So let's talk about the evolution of the business today. Generative media has taken off in AI. It's probably one of the coolest parts about AI, seeing all these videos being creative, all the creative stuff getting made.
3:31And so I don't know who wants to take it off, but feel free to just kind of like, where are we right now? And why is this just kind of exploded? The way we talk about it, AI really has like three product market fits. One of them is search and like initial chat GPT consumer product and now Google and Gemini. The other really clear one that's like growing really fast now is coding agents. And the third one is we believe is generative media. And to put things in perspective when we talk about it with coding agents an individual can go spend maybe 10 ,000 $20 ,000 a month on their cloud or open AI subscription.
4:11Generative media is the only other thing this exists. An individual, if this is your job, if you work at an AI native studio or a real studio these days, you sit down in front of a computer and you generate videos all day and you spend$10 ,000,$20 ,000 a month. This is the only other use case in AI that an individual can spend that much money. and search is so big that it gets a pass and there is no other use case or part of the industry that has this dynamic it's just coding and generative media and that's why you are seeing all this growth because people more people are adapting it and individually they are spending so much money on it i know this is kind of an elementary question but did you like were you surprised by how much money there is behind this?
4:58100 %! We were too. We were surprised all along the way. We were surprised how quickly we switched from image models to video models, where now we make, I would say, 50 % of our revenue from video models. And then after that transition, we thought video was gonna completely take over. It was zero and then it reached to around 70 % of our revenue. And then we thought But image models are like done and it's like it's not enough compute. It's a simple problem and from now on it's all going to be video. And then turns out image editing models brought life back to image models. And then it's also useful for video models as well because you can create different frames and video models does such a better job interpolating between those frames.
5:50So image editing all of a sudden made image models come back. and now like our revenues 50 50 even that was surprising right like when we thought like a portion of our business was dying but it like made a huge comeback uh but yeah video models keep getting better yeah currently what we can do like the best video model maybe creates 10 to 12 seconds of video like the new ones are going to reach maybe like 30 seconds but until you can generate like real-time streaming very high-definition video this is going to keep increasing and as as you increase the resolution as you increase the duration the cost increases exponentially because you have to have more attention in the in the inference during inference time and the resolution increases the amount of pixels exponentially so things are getting much more expensive so now we have a framework to think about okay there's a lot more money to be made going forward because the video output we get is still like very rudimentary.
6:50There was also like a very interesting point sometime last year when VO2 came out. I think at the time like Kling was probably one of the best video models and there was also like Minimax and the prices were like somewhat kind of going lower. I would say like they started at let's say 40 cents per five second video and then they sort of settled around there maybe like 30 cents for a five second like clip and then vo2 was very surprisingly it was like five dollars per five second video so it jumped like you know more than 10x and we were like okay like who the heck is gonna pay the pay all this money right like this is just another five second video Turns out like there is essentially infinite demand.
7:40So people started using that model and then even more expensive models came around and people started using those models. And what was really happening under the hood was we actually saw like more and more professional use cases start happening for like ads, like real, you know, people who are in marketing and like ads and actually like movie production. they started using these models because the quality went so much higher and there was just more demand that like other models weren't able to actually capture. So it's like the usage is going up and then the capabilities are going up and then everything's getting more expensive over time.
8:20So essentially, yes, like people ended up spending so much more money. And in terms of progress, it's still happening. Like we're nowhere near, you know, with LLMs, I think people talk a lot about hitting a wall or like diminishing returns. I think in this side of the market, we're like nowhere near actually hitting like diminishing returns. I think there's just so much new stuff happening with like CDes v2 and v2.5. It's just like the quality is going up, prices are going up and people still want to use all these models. For sure. I mean, do you guys remember Sora? Yeah. Feels like it was a lifetime ago.
9:01I know. I mean, that was a really cool moment, but clearly too expensive. Yeah, it was too expensive. Like, in fact, we saw, we also hosted Sora on our platform and like we saw huge growth in the like last three months before they decided to like take it down. we're like, what's happening? Like there's like infinite volume here, essentially. But yeah, it turns out, you know, the rumor is like there wasn't enough compute for everything else. And it's also expensive to run. So, so yeah, but I think it had a great future. Maybe, maybe it'll come back sometime. I mean, it's really interesting because earlier today we were down at Applovin and, you know, this will probably get released later than that.
9:42But I was talking to Adam, the CEO there and Gio their CTO and we were kind of talking about how they've created their algorithms for all the ads and what works best and like the best ads are like actually 60 seconds long and I was like okay well are you thinking about using like a Higgs field or any of these other kind of companies out there and offering that or acquiring any of these kinds of like generative ad platforms because some of them just really I mean most of these cases are ads and like it just it's just really not there yet it's like the best they can do is like five seconds and it's like not every brand can do that.
10:15In terms of where you guys came from and where it's going, advertising is going to be the biggest use case, you think? I like ads. As a business, internet runs on it. And also the content these models are able to make is perfect for these UGC style ads because the content is usually throwaway. If it doesn't work, make a new one. It works great. but also like the quality, like the lighting, nothing has to be perfect. It needs to be almost, it needs to look homemade. So it's like easier to do. And, you know, yes, the best ad may be 60 seconds, but you don't have to do a continuous 60 second generation.
10:57You can do fast, small cuts, which works maybe better in certain cases. And then what you can do is you get a lot of data back, right? I don't think anyone is particularly doing a good job at it, but because you can post these ads and see how people are interacting with it, you can have continuously improving or you can do small changes to what you are posting and just show more of the things that are performing well. And this can all happen automatically, right? So this is, I think it works great with video models. no one has done it perfectly but you would love to be part of this like automated ad story i think it's it's it's a it's gonna be a great business it was really funny i i was i was gonna mention this in the interview but then i realized it was probably kind of inappropriate but like the kind of the the example you just mentioned of like reinforced learning models on themselves and and just kind of optimizing over and over again there was and i don't know if you guys see this in your Instagram feed, but I do this because I make clips and I make like stuff for the internet and they're like, one of the best hacks that I had was showing my armpits.
12:09Like I got so many more views from showing my armpits. As the hook, right? Literally, literally, it's like I got 10x views just from like having my arms up and like, and I was like, what the hell?
12:24Imagine like you can have yourself in 100 different positions and let's go. Yeah, I mean, but that's like a thing. It's like the thumbnail is always testing. I mean, for like YouTube for us, it's like you got to test thumbnail, you got to test the title, you got to test like the copy, all of that kind of stuff because it works the algorithms differently. Exactly. I think there's like really sophisticated users out there that are doing it like at like smaller scale. Maybe they're like a creator, but also like large companies is doing it right now. I don't think it's like super spread out yet, but like it's out there for sure.
12:59And we work with a few companies that actually do it, but I think it's going to get even more available, I would say, because it's not very straightforward, I think, to do the whole reinforcement learning and all the technical stuff behind it. It's not fully there yet, but I think in a few years, this is going to become the norm. So let's talk about the engine behind Fall and then we can kind of get into how your culture is keeping up with the growth because it's crazy. But so you were saying you started off with compute. You started off as a compute company. You found your way into this area space.
13:35What was that evolution like for you and like how have you completely embraced this new category? Right. Yeah. So we started in compute, but like a pre-AI era, let's call it. Right. So we started in 2021. And the whole thesis was like, okay, there's, you know, this previous data infrastructure wave led to some amazing companies like Snowflake, Databricks. And what do they have in common? They own compute. You know, so in order to build the next generation of like the data infrastructure company, we also have to focus on compute. That was the initial thesis. And we thought the next big thing is going to be Python.
14:13So, SQL, basically like what Snowflake did and what Databricks did to a certain extent, was like enabling SQL to run on, you know, in the cloud. And they own the compute and they let you scale the compute elastically. So we thought Python is going to be the next thing because AI is the next big thing. So we ended up working with some really awesome teams, like trying to bring their data science and machine learning, old terms, like workloads into the cloud. So that was like our initial starting point. And then soon after we started, some insane things happened in the market. I think from a technology point of view, like we had a massive breakthrough with ChatGPT and in our space, StableDiffusion and DALI.
15:00So we kind of had a choice at that point. Like we can either stick to our like, you know, data infrastructure, your initial idea, or we could go into this like new world of AI. And there was language models. And at the time, I think people were very excited about a bunch of open source models. If you remember, like Lama came out shortly after and there was this like huge price race to the bottom. And we just thought, OK, we're like five, six people right now. we don't have that much money in the bank, you know, we're a pretty early company and, you know, we can't really like go and, you know, sell tokens at a negative margin.
15:45And let's do something more interesting. You know, the trend still holds, where the trend still holds is let's focus on image models. Because like there's something interesting happening there as well. There was no site of video models, like no one even talked about it, I think back then. But we thought that like we could take image models and make them run a lot faster in the cloud on the same infrastructure that we built. And so we did that. So we had a very technical team. We thought like we could do a really good job running these models faster. I don't think we had a very formed thesis on like where those models are going to go.
16:21So we kept like, we kept doing it. And that got a lot of attention from like very early customers. And we thought, okay, you know what, like with this other business. We ran it for a year, we didn't make a single dollar. With this one our revenues started to go like this. There's clearly something here so we decided to focus on the image model space there. And three to six months later roughly we actually realized that like this trend is probably much bigger and we thought that these models are like useful for very different things than what LLMs are helping with. We thought it'll help with ads and marketing.
17:05We thought it's going to be, the buyers are going to be completely different. So we decided to call it something else. So like, just be able to tell the story to people. So we actually came up with this word generative media and we slapped it on a website and we said, follow the generative media company. And we always have to explain what it is. Cause I think, you know, three years ago, no one really understood what that meant. So yeah, that's kind of like how we got into it and how we sort of found our like niche, which, you know, three years down the road, it's more well established and it's a whole like new industry.
17:42And people want to make feature films out of it. Exactly. Yes, yes. I don't think, I think like when you first play with the first Dali, like you would never think, okay, like we could make movies with this, but it took only like... Came here very, very fast. Yeah, we came here really fast. It only took like 18 months that people started thinking, oh, like this could actually make feature films. So we're in San Francisco. People here are super nerdy. Maybe you guys are too, but like, well, as you're growing super fast in this inflection period, how are you keeping ahead of research? Like, are you paying attention to research?
18:18Like, what are you, how are you learning as you go? In the beginning, we were extremely involved, obviously like we had a very small team and we had to be part of every single little edge we can make and then nowadays we have a really talented leadership team that most of them we worked with before in different companies so I would say that's that's the best part of our job that we can rely on these leaders and that's a superpower to ask their opinion what they think about like what's coming new what they pay attention to and a lot of information I get these days is from from that team I'm also very curious read everything but I get I learn more from the leadership team than than the internet these days oh that's pretty good what about you yeah I think just echo what Gerkem said it is it's very important like how you build this team I think I got a lot of inspiration from my days at Coinbase in terms of like recruiting the earlier team.
19:26I think one of the things we did there, so I joined Coinbase 2016. Team was like almost 100 people, maybe like very early days, similar to like our size now. We're like 150 people today. And one of the things they did was like they had this like crypto quiz that they posted on Reddit. right and and one of the things they assessed for was like you're literally like your knowledge about crypto like how much do you know about it you know bunch of technical terms also pre-ai like you know it's it was hard to find the answers to these questions but like there was literally a quiz and you had to pass this quiz to like actually get on a phone call with someone on the team how crypto pilled are you was there so this was like i think this was like a very key factor to like you know Coinbase's success in the long term because like you really want like these like fanatics about the space so that like they can keep things going because like you know most of the time like the whole job becomes about it becomes like staying up to date right so like how do you stay at the forefront one like you could collect really great people and you know get them on your team and hopefully they'll feel motivated or you can get people who are like already motivated this is all they can think about this is the thing that they're like most excited about today if you didn't hire them they would be doing this anyways so like that was literally like i still is this is kind of the criteria we have for for bringing in the early hires because like it helps with just naturally like we're up to date all the time um and i think it helps with people being just like constantly motivated about like what they're working on so like Like, you know, you don't have to try to do that.
21:17People are already like AI-pilled or Gen Media-pilled. This episode is brought to you by Brex, my favorite. You become what you spend on. And I refuse to spend my time on work that shouldn't exist. Expense reports, receipt chasing, and manual closes. The companies building what's next from Vercel, OpenAI, Anthropic, Granola, and Deepgram all made the same call. They all run on Brex. Brex is the intelligent finance platform that combines cards, expenses, and banking into a single stack with agentic finance built in. AI agents that handle expenses automatically, enforce policy before spend happens, and close your books in minutes.
22:00That's why Sorcery runs on Brex, so I can spend time on building and not busy work. It's time to get Brex AF. Learn more at brex.com slash sorcery. That's B-R-E-X dot com slash S-O-U-R-C-E-R-Y. Bye. From handshake to signature faster, Harvey, the AI platform built for legal work, is trusted by 80 % of the AMLA 100 and over 20 % of the Fortune 500. Learn more today at Harvey.ai. That's Harvey.ai. AI needs more than chips. It needs power, land, and infrastructure.
23:01So where are you hiring people from? What were their roles? Because this is all, again, this is all very new. that and this is a common question i ask in these conversations because it's like to remind people like none of this really existed like three or four years ago so how many people are trained even in the beginning it was an overlooked market because it happened at the same time with with the llm growth like maybe if it happened by itself it would have been the hottest thing but it was like the second hottest thing it's also hard because it's the creative industry it's kind of at odds with technology to some degree it's like there is a lot of technology involvement in creative industries right like even an imax camera is like you know size of that table over there like there is technology involved and in in different parts of like animation filmmaking bigger parts and then hit plateaus and maybe we are going through one of those times where in the next five to ten years it's going to have another outsized impact on on the creative industries But yeah, like AWS, AMD, they have massive like media businesses where they sell to studios and media companies.
24:10So this is just the next wave of it. Yeah. I think I was listening to something on Christopher Nolan and his infrastructure where he's keeping all that IMAX film. Yeah. 900 pounds. That's the... So heavy. Which is funny. Okay, so let's talk a little bit about the team and how you guys all gathered together. So you came from Coinbase, you came from Amazon? Amazon, yeah. Yeah. So how did you guys meet and what was kind of like the early culminations of gathering together a group? You go. So we are both Turkish. Really? But we didn't know each other back home. We met in San Francisco when I first moved here.
24:542014 is immediately I met him. we were both working at different companies and doing you know minding our own business and it was actually during covid time we this we've told this story a couple more times before but we rented the house in palm springs and that's when we first thought okay maybe there's like a working chemistry between us and then it took us maybe another year to start the company but that's that's how we met we were long time friends for over 10 years and never worked together until that COVID Palm Springs house. And here we are. We've been building the company for five, six years now.
25:35And you've been pulling people from Coinbase, from Amazon. Coinbase and Amazon. It's funny. It's like from both places. Exactly. And so what were you like looking for in the early days? What were the types of hires?
25:50I would say like just really good people. in in in many ways like i think one is obviously like they have to be excellent technically i think early on like we we we definitely had to get the technical part right um because like we were just doing a lot of like deep technical work we still are but that at that time that was our like differentiator. So people who are really good at, you know, very good technically, I would say also like people who are just like people that we get along with. I think like early team chemistry is really, really important. I think if you can, like figuring out like product market fit is probably the most challenging thing for a startup, right?
26:42So like you have to, in our mind, I think you have to keep a pretty small team and that can work really really well together that can like stick together and just have a really like great way of experimenting with ideas and then like throwing things that don't work and and just like keep moving and keep looking. I think this is one of the biggest advantages of starting a company in your mid-30s because you have your whole working life people you worked with you trusted you didn't trust that you can you can bring in together and And as soon as you reach out to people, if you have something worthwhile, they come and join you and they want to be part of the story.
27:25And if you are younger, you unfortunately don't have that. And you have to just go on your own and try to hire people and get trust. But here we build trust with so many people that early engineering team was almost all people like we knew before and worked with and trusted helped navigating the problems so much easier what are your views on scaling up or keeping flat so culture wise like our company I would say is like flatter than your run-of-the-mill startup I think like the deepest level we have is like two levels of like reporting and yeah I think that's that's by design I would say we we we really believe in this like super ICs I think it's becoming more popular now it has a name I don't think we have a name yeah we have forward deployed deployed agent not agents Is it agents?
28:28Any agents too. I should know this. I know Palantir quite well. So we have forward deployed engineers and now we have super ICs. Super ICs are, I think like now it's becoming a more common term. Basically, you want your like management layer to be super ICs. So people who are just like so epic at their job, they can like actually do hands-on work and then they can also like lead teams. I think we have a very similar, without like calling it this way, we always have the structure in a way that like essentially like we don't have pure managers, we don't plan on having pure managers. And we can gather the team around like these really magnetic like people within the team that are just like really freaking good.
29:14And that's just like a very, I don't know, natural way of like building leadership within the team. So that's kind of how we've been building it out. Like we just add really good people and then we see leaders emerge and we gather like the teams around them. But how long can you go this way? I think there's a limit, right? So you start hitting, I think like some communication bottlenecks, right? And I think that's part of the scaling challenge, right? Like you have to kind of think about, we think a lot about like how do we communicate with the team broadly? how do we keep everyone aligned? I think those are, as you grow, those start becoming the challenges.
29:59And so far, we just try to have many check-ins with the team. I think we're still keeping everything flat and see when things break. And then we would throw maybe more management layers. But essentially, we're really focused on this super IC strategy. And in a way, our growth was timed perfectly with how coding agents and NLMs became a huge part of software engineering. I think forever we can keep this as long as agents and coding models improve at this pace. You just need really talented leaders, super ICs to manage, maybe be the leader of, I don't know, five to ten engineers. and you rely on coding agents to do most of the work.
30:48And I think we can keep up like this to have like maybe just two levels of hierarchy and that's it. I mean, look, Applovin does multi-billion dollars in revenue. They have like 400 employees and 100 people on their engineering team. Yeah, yeah. Big fan of Adam. Big fan. Big fan. I mean, I was there today. It's crazy. I mean, like their market cap peaked at like 250 billion. Now they're at like 130, 150. But I mean, it's like he is such a proponent of what you're just saying. Like his whole thing is built on individual contributors and like people not just because now we have access to coding agents.
31:26Like what about the other side of the brain? What about the like, you know, the human personal things and also holding the business weight of it and making business decisions. But like having, you know, the team effort around it. So I think that is like kind of a structural component that's just fundamentally changing. It's like you can't just be a 10x engineer. You have to be a lot more. I think other CEOs are also talking about it, right? Like Brian Armstrong from Coinbase, he recently talked about it. Literally describes a blueprint. This is how we're building the company going forward. I think Elon is a very big proponent of this, for better or for worse.
32:02He talks about this a lot. Yeah, I think it's really picking up, especially among... I think it's harder to do at non-tech companies because there's so much hierarchy. But I think tech companies particularly are very... They still need to, I think, go out of their comfort zone. But at least they're better suited to be able to pull this off. But it is something I think a lot of people are talking about. So what are the key traits that you guys look for? You mentioned good people. Agency is one thing that we talk about in the office. I think everyone would want people with agency, but you just mentioned software engineers even making business decisions.
32:45I think it's unheard of in a big company like Amazon, Coinbase, whatever. Software engineers do software engineering. But because the market is moving so fast, the customers are emerging, changing between different models so fast that a lot of the times we want our software engineers to make business decisions or work with customers, therefore have to make business decisions around them. So that's the hardest type of engineer to find. And I think you look for agency. If you think this person can act independently, take responsibility and own the decisions, then the other things we want just comes naturally with it.
33:27Hard to test for it. I think the best way to know it is if you work with him before or her, that's the best way to know it. but that's something we care about a lot. Yeah, I think we also look for, maybe correlates a little bit, but like hunger, I think is something that's like very, very important because like, I think you need to be, you need to be able to like really push in this market like to actually win. So you have to like just really want to, you know, win and build a big company. So we try to test for that, which is very hard to test, but I think like you can kind of get a sense when you like sit down and talk to a candidate.
34:08We do a little bit of like reverse selling. I think that helps maybe to identify hunger. Like, hey, like it's a lot of work. We don't like believe in, I don't think we believe in like, I don't know, toxic work hours or things like that. We go home during weekends, for example. Really? Yeah. What? Imagine that. This is the first I've heard of that. So nothing to, I think, brag about. like, but like, you know, like nothing to brag about, about not going home. You go home during weekends and you're still successful. Yes, imagine. Crazy. But, but I think like there is a certain baseline. I think that we, we try to be realistic about like what to expect.
34:49And so that's part of the reverse sell. I think, hey, like there can be a lot of like time crutch, you know, like, you know, where we really push and we have to spend a lot of time working and 60 hours and 80 hours and i think that like filters out maybe like certain you know certain people so i think looking for hunger is very very important and then i think like culturally i would say anyone like who interacted with our team um like on the customer side they would probably attest to this is like our team is like pretty low ego like pretty easy to work with i think we look for like humility in general like they got to be really talented but at the same time like not be a jerk about it I think that's like pretty unique to find and I think like anyone who's interacted with like our team probably knows that and that's that's I think why they really like us like people are really easy to work with and like really friendly I think I think you can you can have both at the same time so so we try to I think filter for humility as well.
35:58What have been the biggest surprise challenges that you guys have come across? And not just like the standard stuff, but it could just be like something you had no idea you were going to expect. I'd say like being at the forefront, like as a company is pretty hard. I think one, you have to really want it, right? Like you have to want to build a big, big company and two you have to like keep up with it right which requires a lot of effort so that i think in order to you know build up build a big startup you really have to just be very committed and be very hungry and and i think like that's you know that that definitely that's a lot of work you know that that takes that takes a toll on you i think adam would say this probably uh you know uh but But yeah, it just requires a lot of hard work.
36:53I don't know. I asked him what he's most excited about. He says he doesn't feel excitement. So he's beyond that point. You're going to get there one day. No emotions. Our industry is incredibly dynamic, right? It's when there's no good model for a month, everyone starts panicking. Oh my God, it's been one month. Nothing new has happened. but the answer is just wait another month and something incredible is going to happen. It's been true since we started and that keeps excitement going and that keeps people very encouraged that, okay, there's going to be the next model that's going to unlock these other use cases, which is going to unlock these other types of customers.
37:36And that has brought so much growth to our business. And so far, it just keeps happening for us. Where have you been seeing some of the best models come from? Do they just come out sporadically? Is there any consistent theme between all of them? So right now we play a bigger role in the ecosystem because we are an infrastructure partner for the model providers, but we are also the API partner for the application layer or studios, people who are consuming the models. So we got a good sense of supply on demand on both sides. And we are in such a compute constraint moment right now that if you are coming up with a new model, you really have to plan ahead of time.
38:18You have to make sure you have enough compute to run inference. If you are successful, you are likely not going to be able to procure it on your own. So it's usually a good idea to partner up with an inference provider like us to do your launch. So we work very intimate with the model providers. And for customers, they usually use multiple models at the same time, but they really care about model releases because they usually tie their product releases with model releases as well. So we keep in very close communication with them as well. And then we have to sometimes anticipate how much growth is going to come on the model side to procure our own compute, to maybe do some investment on our side to enable the whole space.
39:04And someone really has to do that because otherwise all the compute is going to go to coding agents or whoever is doing the bigger investment. By planning ahead and investing on compute a little bit, we can enable both the model providers and the applications being built on it. So that has become our responsibility in the ecosystem these couple months because compute is so hard to come by. I'll add a few more points. I think generative media is very unique in two ways. One, I think there's just more modalities. I think when people think of gen media, maybe they only think of video. Yeah. But like you would have to think about text to image, image editing, video, 3D, text to speech, you know, speech to text.
39:53I don't know if I said 3D, like for gaming use cases, whatever. 3D, 4D, 5D. Exactly. All of these. So there's actually like many modalities within generative media. So like, and there's like, I don't know, five to 10 companies in each of these sort of sub categories within GenMedia. So that's like one interesting aspect. So there's just like a lot more happening. There's more fragmentation. And then the other is going back to what I said about like diminishing returns. I think it's still like so early in the space that there's actually a lot more leapfrogging happening between more players. because we're not at the 98 point whatever percent.
40:34We're only at, I don't know, 50 percent of where we're ultimately going to go. So there's actually more people leapfrogging. There's more diversity in the number of labs. So it's not just like, open AI, Anthropic, open AI, Anthropic. It's more like there's eight players and they're all leapfrogging each other. One big reason is you're not running out of data. There is like so much video data out there and so much new video data is being produced every day that no one has, no one claims to train all of that data, right? They have to pick, the hard problem with training a video model is actually you have to filter out the interesting video data so that you make a good model.
41:20So if you need to put more data into the model, there is abundance and it's seemingly free. There's enough free video out there that it works very well. But also the problem space is bigger, right? You are trying to reason around the whole world, physics, everything. So it's not as specific as coding or as specific as certain LLMs are trying to optimize around. So there's a lot more angles people can take to create a more interesting and appealing model. And it's been happening every other month almost. I think that's where the naming comes in too. There you go. You have to differentiate with your naming.
42:00I mean, they've got some funny names. Yeah. Happy Horse, Nana Banana. What do you like? There's Happy Oyster. Happy Oyster. It's like a game model. There's Minimax. Sea Dance and Sea Dream. Sea Dance, Sea Dream. I think those are... I think when Nana Banana had its moment, everyone was talking about Nana Banana for like three days. Even my dad was talking about Nanobanana. I was like, what the hell is going on here? I think the backstory is that it was an anonymous model. So people do these cloaked releases where they will release a model. They won't actually disclose who it is. It's a codename.
42:43Get people excited about it because it ranks highly. And I think somehow they use that Nanobanana as a placeholder name, codename. Yeah. And then it just stuck. So it became the original name. If you guys were a model, what would your model name be? Sorbe. Sorbe. This is Gerkan's favorite word. I don't know why. I wanted to rename the company at once. Sorbe. Sorbe? He likes Sorbe a lot. I also have this. I think about names. I also want these exotic colors. Like burgundy, turquoise. You know, like, I think that would be a good series for model names. That is a good name, I guess. What about you?
43:30I don't know. Sorbet could be a color, apparently, too. It's like the peach color associated with sorbet as a color. I like... you can do birds. There's so many birds. Birds! Do you have a favorite bird? I'm asking for a name. I don't know. Like, you can even do owls. Like, there's so many owls. You can literally just do, like, different owls. I like owls. You can do insects. I don't know. Nature, I think, would be my theme. Well, how did you come up with the name Fall? All right. Let's talk about this. Yeah, yeah. So Fall is... So we have a longer name, Features and Labels. It's a double entrande.
44:09Wow. It's the initials of Features and Labels. And it's not Fall. It's Fall. It's Fall. Yeah. It's Fall. Yeah, it's Features and Labels, which if you're a real OG, you would know those two terms. It's like from the machine learning world. It's like very old, like pre-AI. And so it's the initials for that. But also it's, I don't know, have you had Turkish coffee? Yeah, I have. Okay. So have you been like told your fortune from like the coffee grounds? No. All right. So that's foul. It's like your fortune told by the coffee grounds. You should get it done. We'll do it sometime. Okay, okay. Let's do it.
44:50So, you know, it's like your future prediction. Someone looks at, like, when you finish the coffee, looks at what's remaining. Are you guys trained in that? Can you tell portions? You know, everyone's trained in Turkey. Everyone has done it before. Yeah, yeah, yeah. Should we do it right now? We could try. Yeah, yeah, we'll get some coffee. I think they have Turkish coffee here. Oh my gosh. That's fun. Do you think the open source, closed source debate is still going to be going on in two to four months? six months how long do you think that's going to keep on going i think so because like anyone who is falling behind a little bit uh open sourcing their model is an incredible marketing tactic and the whole world starts talking about you if you open source your model and there are real advantages like if you talk to anyone any of these studios in hollywood they they want to be able to fine-tune the model they want to be able to apply different lauras filters to change the lighting to be able to use their unreleased characters their unreleased and styles on the models and the only way to do that is is open source so there are real technical advantages to it also it's an incredible marketing tactic and so people are just going to keep doing it some someone's going to create a frontier model people who are falling behind is going to open source it and the pendulum is going to keep keep swinging we thought this was a bigger threat to our business in the beginning because we were running mostly open source and then we started partnering up with closed source models and running inference for them as well so it's a it's a win-win situation for us but open source makes it a more interesting ecosystem and i think it's it's good for the industry in general So we always welcome more open source models and we support as much as we can the community.
46:42I think the fact that like people want to protect their IP and like that that's gonna continue right not just for media but also like you want to protect your source code you want to protect your customer list whatever right like all of that that's always going to be a thing. so I think yeah if you have something local you know you don't have to worry about it so I think I think it's it's always gonna be there this this debate like when we talk to our customers if they know the model is run on our GPUs they trust the whole process so much more open or closed this this works the same but that that has been a big differentiation for our business when and obviously when it's open like they know for sure that they can even self but they trust us because we can do it better than them.
47:32But trust is really important with video models especially because people don't want their unreleased content to leak. And it's usually an input to the video model when these creatives work with it. So we've been really pushing on that with the enterprise customers and it resonates. How good is AI at real professional work? Mercore's AI Productivity Index, or APEX, tracks it with every major Frontier model release. Built with partners like Harvey and Cognition, see where the Frontier actually stands at mercore.com. Today's episode is sponsored by VCX by Fundrise, the public ticker for private tech, allowing investors of all sizes to invest in venture capital.
48:16Learn more at getvcx.com. Public is a multi-asset investing platform where you can automate your investing with AI agents. Create your own custom agent with a simple prompt. Agents can monitor markets, execute trading strategies, and manage your cash across stocks, bonds, options, crypto, and more. Want to buy at market close and sell the next day's open? Just tell your agent. Automate your investing at public.com. Founders scale faster on deal. Set up payroll for any country in minutes, hire anyone anywhere, get visas handled fast and get back to building. Visit deal.com slash sorcery. That's D-E-E-L dot com slash sorcery.
49:02Talking macro, what do you think the biggest question is that people aren't asking right now? I would not exactly that. But if you if you listen to like any Hollywood podcaster or look at the news about Hollywood they always have this angle of AI is gonna kill Hollywood and we are like really in this business like we talk to these people all the time so far I would say I had zero impact on anything in Hollywood it's just like people are projecting to the future a little bit and the business we get from Hollywood has been increasing in the last six months I think it reached to a meaningful percentage of our revenue.
49:45But I would say it had zero impact on the labor. It barely made any content that's worthwhile in Hollywood. It's going to happen. There are signs of it. But we've been talking about it for four years. And it's always been fear-mongering and nothing real. I think that's the thing that people don't realize from a distance. Yeah, most of LA, like, that's where this is happening. it's really regulatory and what they're doing there with different laws exactly it really has nothing and that's been the reason the adoption has been slow because it's it's not only uh the the quality of the models the the lawyers of the companies need to be okay with it as well but finally slowly we are seeing that everyone is is going full force at it what do you think people should be asking what are what's what's going on out there i think people are asking everything.
50:38They're asking too much. They're asking too much. Yeah, I don't know. There's a lot of skepticism, I think, about use cases. Again, just speaking more broadly about AI, right? There's certain skepticism about, oh, all of this CapEx investment, where is that going to go? And is AI actually useful but I think like I'm more on the optimistic side I don't think a lot of the people who are making these claims I don't think they they really actually understand like what's going on how like AI is actually being used at companies and and you know I think gen media is still like still very early if anything like it's just starting it's just like gonna I think ads we're starting to see it Hollywood social whatever we're starting to see it I think there's there's real value like we were just estimating like how much faster we're going like our software engineering team is moving and I estimated somewhere between like three to four x like we would we would need pre-ai we would literally need like four times the people to go at the same speed so and we're definitely paying less than that yeah right so if there's clear ROI for us uh so so yeah I think I think if you if you really like go deep into the ROI for the for the companies who are using AI like if you really go deep and actually try to measure it uh not read it like from a bunch of reports or like I don't know non-technical people uh I think I think you'll see a different different picture so so yeah I'm very in favor like very long compute as they say like we talked about it every big company is gonna be a compute company just the storefront is gonna be different our storefront is unique and interesting like no one else has a storefront similar to us some has coding agents if you go to a neo cloud there is no storefront if you go to Databricks, whatever, it's about like data.
52:45Everyone has a different story and different way to talk about it, but behind the scenes it's who's running the most compute. That brings the revenue. And yeah, we have a unique interesting storefront and it keeps growing. And there's not enough of it. Like all the numbers you hear, you know, like 200 million in you know capex and whatever like a lot of that stuff is like not built yet so yeah like it's very slow right like building these things is it takes a lot of time and there's not enough compute right now so i have two final questions and then i'm going to ask you what your hot take is i'm letting you know that now so you can like get it think about it this was a hot take You guys have been crushing it.
53:34You have great answers and they're quite opinionated, so it's awesome. But before we get there, what are you most looking forward to in the next 12 months? I'm looking forward to the first timeless AI content. And the way I define this is you and I, we are going to watch this. And it's going to be a year later. We're going to watch it again. And we are still going to like it the same way we like it today. So far, AI content has been about capabilities. It's been about, oh, look, this is possible to do with a video model, or that's possible to do with an image model. But we need to reach that point where things look fresh all the time.
54:14And I'm looking forward to that. I think we are, in the next year, there's going to be a widely accepted, maybe a short movie, maybe an animation that people are going to look back at it in a couple of years, and they're still going to enjoy it the same way they're enjoying it today. It's pretty good. Evergreen content. Yeah. It's very important. I try to do that, but... There you go. Like, people say, Toy Story is that for computer graphics. Like, that was the first time. It came out in 1996, something like that. Isn't it crazy how long ago that was created? You know, we can just go watch it and it's going to be good.
54:50Yeah. I'm going to enjoy it. Did you like The Odyssey in that way? I haven't seen it yet. You haven't seen it? All you've been doing is talking about Christopher Nolan. You haven't seen it. I'm waiting for the IMAX, the big IMAX, and I couldn't get a ticket yet. Go to the Spear. Did they show it there? I don't know. It's too big. Not as Nolan intended. No. For me, I would say, like, world models, we haven't spent too much time talking about them. I think they're super exciting. I was, like, I was very skeptical. like even three months ago. So world models are these models like you can actually like play around and like move around.
55:33It kind of like looks like a game and you can actually like put in actions like with your keyboard and you know, you can navigate the world. But like seeing what I've seen recently, it's pretty incredible. I actually started feeling very bullish about it. It's a matter of I think the quality going up and then the speed getting so much better. and I think there's like new architectures now that like people are able to actually have like more persistent worlds. Before you used to like look this way you know see the flower look away come back it's gone. So they're able to like persist things now and you can actually you'll be able to like actually build like a game with a prompt but I think that space is like just gonna blow up.
56:18So again the big challenge there is compute but I actually think that because you need like a gp running at all times to actually like play that game but i think the quality of the games itself is gonna go a lot better and i think it's gonna be probably yet another product market fit like gaming with ai that's not like a thing yet i i actually think there's a big chance that may happen and it might be like this fourth category maybe very expensive yeah damn okay so um this is we have one more question until the hot take okay i'm just letting you know again i'm warning you okay so sorcery is very proud to be sponsored by brex i don't know if you know what brex is but of course brex knows where you are so uh they're all about spending smarter moving faster performance a performance corporate card i like to take this question in the personal bend things.
57:15I mean, you guys have worked for legends, whether it's Brian Armstrong, Jeff Bezos. I mean, you like Chip, so you might like Jensen Wong. I don't know. And then you also like Christopher Nolan. So I'm curious, who are some figures that really inspire you and keep you motivated? Who keeps you, like, running at it? Great question.
57:38To be your dog. I don't know. of of the you know previous generation of founders i think jack dorsey is very underrated and i think he has more range than than anyone else and and then like now he's he's like maybe maybe coming back i think watching his his journey with twitter and then square i i never got to work close to with him But that's an inspiring figure. When I was younger in Silicon Valley, I always look up to what he was doing. Yeah. I wouldn't say it keeps me going level, but I really like Ali Gozi. I think he's transformed his company many times over time. I think I like to think...
58:36Keeps reinventing. keeps reinventing and I think that's very important. I think that's like a secret to building a long-lasting company and navigating through like all these changes that are happening. So yeah, really great leader I think. The time has come. What is your hottest take right now? Do you have it? No. There's so much pressure.
59:07Well, we can take our time. Yeah, let's take our time. You can cut it. And then it's like dark outside. Should we prompt you with some? Like, Kyle? No, no. Let's do it. I have a good one. It could be on like IPOs. It could be on like companies disappearing. It could be on fake GPUs that people sell from. Oh, wow. I don't know where they get them. Could be. Yeah. I had a hot take like a year ago, the OpenAI meta merger. Yes. Oh, could still happen. It could still happen. They're not doing too great. I think it makes sense. My hot take would be bad for the whole industry. Sure, sure. Okay. So, not, I guess like not too spicy or anything, but I think it makes a lot of sense.
59:57Like OpenAI was like known for this like, you know, consumer like edge, right? Right? Like that was their whole thing, right? Oh, like OpenAI is winning in consumer, therefore they're going to win because like consumers so much harder to win. And then Anthropic sort of came with, you know, their enterprise thing and, you know, and then Google started adding, made their own consumer move. And so I see like OpenAI, not necessarily like winning very strongly in either enterprise or consumer. and then meta on the other hand i think is in a similar situation where like they they need something and i thought what if they like combine their forces that would be really cool so like they can really just like really start winning consumer so i i thought it made sense but again i have no information about any of this so i'm completely completely speculating but i i thought it would be a good move.
1:00:57That is very interesting. Another one. This is a common hot take. I'll do it for both of us. Sure. I have more now. Okay, good, good, good. It's coming. The consensus view was there's going to be more margin in the application layer. Yeah. And I think it's completely reversed. Margins are close to zero in the application layer. There's so much competition. People need to spend so much money for marketing. You can only create margin if you are in the compute path and you can innovate in that path. That's the only way you can create margin, which wasn't the case when we started the company four or five years ago.
1:01:35Everyone thought the infra layer is going to end up with price going to the bottom and application is going to have margins. Looking around, application layer doesn't have any margin today as well. So, wake up with this. It's a good one. We talked to Tony Kim about this, that raise, and he was like, the chips have much better margins. They're the only ones with margins right now. I have another one. I think this is a very recent one. I've been kind of like toying with this. I think coding is part of Gen Media. Oh. Really? Yes. Well, it's also part of Gen Media, right? Like it's its own thing, but at the same time, like we're seeing so many developers building like richer applications that are not just like a website.
1:02:27Like they can be, you know, programmable video, right? Like content, like maybe you've seen these websites that like look more like a video. Like there's, I think coding is becoming more and more like a very key part of like what the developers on file are using. so I you know I think we're gonna like have to invest in it because there's just so much demand from our our own customers everything is a computer everything's a computer everything is computer that's the hot thing well that's a great place to end thank you guys so much I'm very excited for dinner and the coffee and the coffee I can't wait to know what my fortune is there you go cool thank you thank you Molly hey it's Molly if you enjoy our interviews check out our newsletter sorcery.bc where we deliver a once a week top deals and tech headlines email and also go deeper on our podcast interviews subscribe to Sorcery today and don't forget to subscribe to the podcast on YouTube Spotify Apple or wherever you listen link in description to sign up
fal has raised $1.65B in a combined Series E and Series F, with Sequoia Capital leading the E and Bessemer Venture Partners leading the F. Its founders see generative media as 1 of only 3 product market fits in AI, alongside search and coding agents.
Burkay Gür (Co-Founder and CEO) and Gorkem Yurtseven (Co-Founder and CTO) started fal in 2021 after building machine learning infrastructure at Coinbase and Amazon. fal runs the infrastructure behind image, video, audio, and 3D models, serving developers, studios, and the model labs themselves.
Recorded at Dalida in San Francisco, Burkay and Gorkem explain how fal moved from Python data infrastructure into generative media, why video and image now split its revenue 50/50, and how demand kept rising as video prices jumped more than 10X. They also cover fal’s 150+ person org with only 2 reporting levels, the compute shortage shaping model launches, and why they believe the margin in AI sits with compute providers over applications.
“AI really has 3 product market fits. One of them is search. The other really clear one that’s growing really fast now is coding agents. And the third one is, we believe, generative media.”
We cover: › The $1.65B Series E and F › Why individuals spend $10K to $20K a month on generative media › Veo 2 pricing and the shift to professional ads and film work › Super ICs and hiring for agency, hunger, and humility › World models, open source, and Hollywood adoption
Burkay Gur: https://x.com/burkaygur
Gorkem Yurtseven: https://x.com/gorkem
Molly O’Shea: https://x.com/MollySOShea
Sourcery: https://x.com/sourceryy
𝐄𝐏𝐈𝐒𝐎𝐃𝐄 𝐋𝐈𝐍𝐊
YouTube: https://youtu.be/woVqLHMiwfk
𝐒𝐏𝐎𝐍𝐒𝐎𝐑𝐒
• Brex—The modern finance platform, combining the world’s smartest corporate card with integrated expense management, banking, bill pay, & travel. https://brex.com/sourcery
• Harvey—The AI platform built for legal work, is helping the best firms and companies close deals faster. More than 200,000 lawyers, 80% of the Am Law 100, and over 20% of the Fortune 500 build with Harvey. Visit: http://harvey.ai/
• Zone—develops next-generation data center campuses, partnering with AI companies, site developers and technology leaders to bring compute online faster and at scale. Visit: https://zonefrontier.com
• Mercor—The human intelligence behind frontier AI. Frontier labs and leading enterprises use Mercor to turn the knowledge of 5M+ experts into the data, evals and benchmarks that make AI better at real work. Visit: http://mercor.com/sourcery
• VCX—VCX is the public ticker for private tech, allowing investors of all sizes to invest in venture capital. View The Portfolio at http://GetVCX.com
• Deel—Deel is the global people platform that helps startups hire, manage, pay, and equip anyone, anywhere. Trusted by more than 35,000 fast-growing companies, Deel is the people platform that just works, so teams can scale without the chaos. Visit: https://www.deel.com/sourcery
• Public–Investing platform Public just launched Generated Assets, which lets you turn any idea into an investable index with AI. With Generated Assets, you can build, backtest, refine, and invest in any thesis with AI. Gone are the days of one-size-fits-all ETFs. https://public.com/sourcery
Follow Sourcery for the latest updates!
Disclosure:
Paid Ad. Agents are an experimental AI tool. Its output may be inaccurate and is not investment advice. Public Holding’s, Inc. makes no guarantees about its accuracy or reliability—verify independently before use. Brokerage accounts and services provided by Open to the Public Investing, Inc. See additional disclosures at https://public.com/disclosures/agenticdisclosures.