The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
00:21:27 · Harry Stebbings, Alex Mashrabov
[00:21:27] S1: Can we actually just unpack the two different go to markets? Because you got consumer and you got enterprise. And I spoke to quite a few of your competitors in all honesty before this show. And I said, hey. You've got Alex coming on. What should we ask him? Everyone said the same thing, which was an admission of their respect for this particular kind of GTM. They said you've executed the most impressive influencer campaign in tech. And now what I wanted to understand was when you look at the consumer growth, what worked, what didn't work, and how [00:21:58] S0: do you reflect on that? That's a good question. So first and foremost, like, the goal is to make sure that the best commercial video content is generated on Hixfield, and we show all the workflows of how to make such professional looking videos. And we have an in house team of over a 150 creative professionals. 150. It's almost half of the whole work... Workforce, frankly. And those people, they make product launch videos. They make tutorials. For example, we made the first AI generated movie, which is also obviously a very sensitive topic. But what's important, we open sourced all of it. And what we learned is that for ninety minutes of, let's say, TV quality content, it was over hundred hours of yet generated content. So creative decisioning, picking the right piece is still very important. So that's really what we are focused on, and that's what's driving most of the most of the revenue. So you're saying [00:22:57] S1: the the growth in consumer subscription is through own content and distribution?
00:22:57 · Harry Stebbings, Alex Mashrabov
[00:22:57] S1: the the growth in consumer subscription is through own content and distribution? [00:23:02] S0: Yes. We don't do any page. [00:23:04] S1: Early on, you made an interesting architectural decision to have your own models, and then then you've since walked that back. Can you talk me through why did you choose own models and why the walk back? [00:23:16] S0: Oh, yeah. Obviously, this was obviously, this was my mistake. I'm gonna be... I'm gonna do my best to be transparent. What I need to admit, at some point of time, I really was thinking that chasing benchmarks is valuable, but I don't believe this is just sort of corporate psyops, frankly. I was part of the large organization, so I know what happens. What happens is that everyone just thinks, like, We need to show some progress, so we need to have some benchmark. But then when I talk to the top researchers from these labs, especially larger companies, what happens is that they start to put test data into the training. They start to kind of use leverage test data to use LLM as a judge for training of the models, use all the various tricks to basically gain benchmarks, get quarterly bonuses, so on. That's like, who cares, right? So if I make my couple million dollars a year in one of these labs, I can move to another lab easily. So that's unfortunately what's happening in larger organizations.
00:23:16 · Alex Mashrabov, Harry Stebbings
[00:23:16] S0: Oh, yeah. Obviously, this was obviously, this was my mistake. I'm gonna be... I'm gonna do my best to be transparent. What I need to admit, at some point of time, I really was thinking that chasing benchmarks is valuable, but I don't believe this is just sort of corporate psyops, frankly. I was part of the large organization, so I know what happens. What happens is that everyone just thinks, like, We need to show some progress, so we need to have some benchmark. But then when I talk to the top researchers from these labs, especially larger companies, what happens is that they start to put test data into the training. They start to kind of use leverage test data to use LLM as a judge for training of the models, use all the various tricks to basically gain benchmarks, get quarterly bonuses, so on. That's like, who cares, right? So if I make my couple million dollars a year in one of these labs, I can move to another lab easily. So that's unfortunately what's happening in larger organizations. [00:24:14] S1: That's- Can I just stay on that? Yeah. What do you mean? You're saying that they are incentivized by benchmarks. And so because of that, they are doing artificial things to improve their scoring and benchmarks, which actually don't increase output efficiently.
00:24:30 · Alex Mashrabov
[00:24:30] S0: Yeah, look, I think let's just look at the outcomes which we have today. Out of all the incumbents in The United States, when I look at Open Router Data, the only company which is relevant is Google out of all the incumbents. When I look in China, where probably obsession with benchmarks probably is less, We have Tencent, Xiaomi, Alibaba, three incumbents being completely relevant, and, obviously, like, Biden's obviously trying to catch up as well. What's your takeaway from that? I just do believe that there there is just... Obviously, in the tech bubble, there is a strong obsession over the benchmarks, which do not necessarily represent the reality. But I can talk specifics specifically in the... For video. A lot of benchmarks today for video is really text to video, which does not represent actual workflows at all. The way to think about video models today, it's just modern rendering engine. Think about this as like Unreal Engine or Unity, but just different types of inputs. And it's virtually impossible to really define a visual output and direct the execution just through text. If you just go into our open source projects like this movie, which I mentioned, average prompt length is over 3,000 words. That's the first thing. And look, all these benchmarks which we are talking about, they are not as comprehensive in terms of the details of prompts and people who are labeling, they obviously read 3,000 word long prompts. But also on average, there are at least 10 image references for every scene. The reason why it's important is because it's important to define how the characters look like, how the background looks like, how actually characters are located to each other in the scene and so on. And so that's why prompting and just the workflow is so complex. Benchmarks just don't represent that.
00:26:18 · Harry Stebbings, Alex Mashrabov
[00:26:18] S1: Going back to the model selection, why did we decide we're gonna do our own and then why walk it back? [00:26:24] S0: So it's true that with VFX and camera control, we got very, very quickly from maybe 1,000,000 to 20,000,000 in ARR within maybe the first three months. Then we released our own image model, which is really good at aesthetic photoshoots and product consistency. This is what allowed us to scale then from 20 to a 100,000,000. [00:26:50] S1: So help me understand, Alex, why did you decide that you were gonna do your own models and why did you abandon them? [00:26:57] S0: We still do them whenever we see specific use case like these photoshoots, but as soon as this is what our customers want. So it's all driven based on the customer feedback, not just by ambition to conquer the vaults and build the best model in the vaults. [00:27:13] S1: Do you think every company will have their own models? Like we're seeing Harvey, we're seeing Cognition, we're seeing McCore, Ramp build their own models. And we'll see every company have their own models with their own data, or we actually all use a series of providers?
00:27:13 · Harry Stebbings, Alex Mashrabov
[00:27:13] S1: Do you think every company will have their own models? Like we're seeing Harvey, we're seeing Cognition, we're seeing McCore, Ramp build their own models. And we'll see every company have their own models with their own data, or we actually all use a series of providers? [00:27:30] S0: So first of all, whenever, to be honest, whenever someone says we build our own models, very likely what they mean is something what's happened with Coursera. We we do remember. Right? A lot of companies, they actually take open weights model and just post training on own data. And post training can happen in two ways. Most important is whenever you have customer data around decisions they make, like sequence of decisions, and you can teach the model to actually learn how to compress these 10 steps into one step, this type of reinforcement learning is the most valuable. And I think increasingly more and more companies will have to do that. We see this in the market as well. So most of the companies in the world today, most of the businesses, they don't necessarily need Astra specifically. They don't necessarily need the newest Fable model. That's why OpenRouter reports that share of open source models went from below 30 to over 60 within this year. [00:28:28] S1: What do you think share of open models will be in two years' time? [00:28:32] S0: Look, I do believe that's just because the capitalism works. I mean, OpenAI Anthropics still are gonna have more than 50% of the markets. [00:28:39] S1: In terms of the dollars generally.