Captions

“This reminds me I had Gaurav, he's the CEO of Mirage. He used to be called Captions, which is a really successful AI video company startup, and they have a policy of you ship a marketable feature every week.”

“I go single line because I like my captions, like appearing on screen, almost like one word at a time. I just think it's like a slot machine. It just sort of Psyche hacks of the brain. Kind of like how it works.”

“We want to build the best platform to make these end-to-end conversational AI agents, but we explicitly don't want to be building full digital avatar platforms, so that way we're able to be trusted partners to companies like Hey Jen, Synthesia, Captions, who a…”

“I invested in Captions... Invested when it was basically a Be Real competitor.”

“Canva for video as an idea has been sitting around for a while. And I think it's actually finally possible because of AI, right? Because the real value of Canva is you start with something, right? You don't have a blank screen.”

“It's just becoming a lot easier, and it's also bounded, right? Which means that there is a limit of realism, right? And then you've kind of solved it, essentially, right? Yeah. You can't get more real than real, and once you're there, you kind of have achieved…”

“Our goal Specifically for video is not to build professional tools, right? We're not building for professionals at all. We're building for the person who could not have created video before, right? They didn't have the tools, the skills, the means to be able t…”

“Our general framework for it is to look for user demand, right? And actually the easiest way to check for user demand is to just see what has virality, right? Usually what has virality. What people want to share and talk about. There's something at the core of…”

“Our engineering goal is every engineer should ship A marketable product every week.”

“If you did something unique that nobody else has done, you can go and show that to people and people will come to your app just for that. And even if your app doesn't have a lot of the obvious stuff, maybe it doesn't have justify alignment, right? People will …”

“Usually Q four for us is infrastructure quarter, right? We just like go and build all the infrastructure. Q four is generally, you know, we've already delivered a ton of products and stuff. We're feeling pretty good about the rest of the year. Things are windi…”

“We are using cursor. We've tried all the different tools. We were using Devon as well, which is another, you know, that's more advanced, I guess, you know, it's kind of solving bugs for you.”

“So it's like almost reversing the process a little bit and starting with design first, but it can often result in like finding unique ideas, basically.”

“And we actually divide our roadmap into two different roadmaps. So we have what we think of as the public roadmap. This is basically what people have asked us for, right?”

“But these are all features that every competitor knows about, right? Like these are public. Like if a user is asking us for it, they're asking everybody for it. And every team has essentially more or less the same list and everybody's prioritizing it. And yeah…”

“And a lot of times the biggest wins will come from the secret roadmap, right? That's the game-changing stuff, right? It's not gonna be the user requests usually that are gonna do that.”

“So we had this feature where it basically shifts your eyes to look at the camera, and we were actually the first company to build this.”

“I do think to fully get there to the point where it cannot be differentiated at all, it's still a couple of years away.”

“About a year ago, we would run AI video in ads and things like that, and we would get all these comments of people being like, oh my god, this is so fake, like, you know, don't show me this, and Around that time, the technology got just about good enough that …”

“Took two days to build. We put it out there and it immediately took off, right? It was absolutely shocking because I built it on a weekend. We put it out there. I call my close founder on Monday and like, it's at the top of the app store. We're getting like, 6…”

“And about a year and a half into the company, as we're working on other projects and stuff, I went back to a personal account, just opened it. And I saw that there was like, 500,000 dollars in there. And I looked at a chart and it was just like growing. The re…”

“The social network that we had at the time, we actually removed it from the App Store, so it's not available anymore. But till today, there are people, there are thousands of people that are using it, posting on it, and all the different things”

“I have to say here that I actually don't read books. It's actually something that I decided on purpose where I decided I don't want to build my skill in reading and I want to build it in listening and watching instead, because I think that's the future.”

“I think that's what makes it actually hard to innovate in these spaces is because people have so much built in muscle memory about how things have been done that if you come in and change a bunch of stuff and be like, here's a simpler video editor, people actu…”

“Like you actually can't templateize video because the nature of it is completely, it's unknown, right? It could be saying anything. It could be doing anything. Anything would be happening. The, template just can't fit with, it can't understand what's happening…”

“One thing that we did well sort of in the first couple of years with the company is we actually kept the application paid only. And that actually. It just filtered out anybody who's not. Right. Exactly. Right. Right. So it really left us with the people who ar…”

“The original app was made in a weekend, like literally in two days. And launched and it wasn't marketed at all. It was just put on the app store. And the next day I woke up and I just saw like all the numbers just blowing up completely. And we're like number, …”

“Microsoft OpenAI did much better in terms of reliability. It just worked perfectly for latency and never went down, which is what you want.”

“And then, but we ended up switching to Claude. So now we're on Claude.”

“So we are using Ideogram for that one because it generates really good text.”

“So we, the data that we're training on internally is all fully licensed, completed proprietary.”

“We are making the unique bet to say that we can actually build a better model than they can because of our unique data sets.”

“And very quickly for us, it became, wait, this actually performs better than if we were to, you know, hire someone to like make this video basically. Right. And the reason is because we can customize it to an unlimited ability, right? Like we can get all the v…”

“We were splitting the video into like 25 frame segments basically, and 25 is 25 FPS. So that's basically a second. And then we were generating overlaps on both sides, four frame overlaps with the next segment. And then we would check for differences, the overl…”

“I think hallucinations in text obviously can be a bad thing, potentially, because they're saying stuff that, you know, isn't true, but in a video, like, you are delegating a lot of decisions to the model, right?”

“I do think the point of, you know, complete perfection in terms of generation where there is absolutely no way to tell is still a couple of years out, especially in terms of this type of a role video because a lot of the core problems haven't even been fully s…”

“We run these things in ads like on scale. We also run these things on social media. People, big creators use, you know, these models and they get tens of millions, sometimes many, many tens of millions of views on their videos and nobody can tell that it was g…”

“Our goal actually isn't to disrupt professional video editing, right? It's actually to enable, and it actually probably creates more professional video editors because a lot more people can get into the profession, right?”

“I think for video editors, there's a small chance that it actually becomes the most important profession of them all. It actually is possibly the profession that rules them all because suddenly one video editor sitting in their basement can make A movie that g…”

“Everything from character consistency, which we, you know, everybody talks about quite a bit, but really what it is today is just face consistency, right?”

“So I think in our current biggest model, we're only using like less than one percent of our data that we have, our proprietary data.”

“So it started off being more expensive and it's only the price has gone down, right? It's literally only gone down and it's gone down more than 10 times every year that this You know that we continue down this journey.”

“Today, like, You know, for us, COGS is like the least of our concerns, basically. I mean, to give you an idea, like if we were to remove a lot of the sort of capital investment in you know, model training and things that are long-term value for us, like we wou…”

“We were a fully remote company. But as soon as COVID started exiting, we realized that to start a company, there's just nothing like being in the same room. And so we started doing the in-person thing. First start with one day, then two days, then three days. …”