Factory Raises $50M from NEA, Sequoia Capital, NVIDIA, & JPMorgan

25:41Um, and the speed with which they're adopting is crazy to me. like they're adopting like a startup, which is wild because my understanding, and I bet if you asked anyone, it would be like, no, those, you know, those firms, maybe 10 years from now, they would do it. But seeing the speed that they're, that they're adopting is really exciting. We're also working with them and some of the big three consulting firms as well, going client facing, which I think is cool because it's a, it's another way that we can hit just more users faster. Do you have any partnerships? Like, how do you think about go-to-market was, how is that structured?

26:14A lot of the partnerships that we pursue on the go-to-market side are with, you know, the big cloud providers. But also I think there's a really something that factory has that not a lot of others do is the focus on the enterprise grade security. You know, I can't tell you how many CTOs or VPs of engineering have mentioned that these co-generation tools feel like they're giving toddlers machine guns. What? Which, you know, all that. And maybe it's not a very respectful way of saying that, you know, they're concerned about, you know, if you have an org of 100 ,000 engineers and you give them all this ability to start producing hundreds of PRs, it's concerning of like what's going to happen to our code quality.

26:51Are we going to have some massive incident in production because we were a little bit too eager to throw these things in without proper security considerations? and so we're working with some of the biggest like security dev tooling companies on having it's kind of you know two different things that we're doing so one offering that we have is basically ensuring with some static checks that every PR produced by a droid passes whatever existing security rules they have within the org but then also there are a lot of tools that will go in and measure within your code base here are all your vulnerabilities we can actually have a droid go in and solve each one of those as they emerge, which is kind of a nice partnership that we have because it makes sense for both, you know, the security companies as well as for us.

27:37Is there one key component that people are looking to optimize the most? Like what is the, how do you measure success? Yeah. So, and I think the org that we're working closest on this with is Snyk. And something that a lot of people are concerned about is basically like, there's kind of this duality between like what the CTO wants and the CISO wants. because the CTO is like, we want to adopt AI. We want to adopt agents. We want to do this agent native development. And then the CISO is like, what the hell is all this code that's going to be coming into our code base? How do we know it's going to be good?

28:07How do we know it's going to be like compliant with all these regulations that we have? And this is kind of a balance that right now there's been no answer. Like most of the people that have been doing it have kind of just been YOLOing and just like throwing the tools in. Some of the tools out there even have like YOLO mode, which I think to an enterprise customer is not exactly what they want to hear. And so figuring out how we can increase the leverage of all their developers while also keeping in mind the importance of the security and integrity of their systems is something that we care a lot about.

28:41I mean, you're particularly new to this as a founder. You were getting a PhD and you were pretty much like studying string theory. Now you're selling to customers some really big names. Was there any sort of emotional experience that you had? Maybe it was like your first customer or second, but like, was there any particular moment? Yeah. So yeah, I think there's one that comes to mind was, so I mentioned that 2024 was kind of a year where we really built out the product. Along the way, there was just like a lot of, you know, we had some ideas. We realized that we were too early in terms of people's behavior, because there are kind of two things you have to balance.

29:16Like what can the product do? And what are people, like how much are they willing to change their behavior and kind of lean in to adopt some new interaction pattern. And some of our earlier iterations just like, weren't quite hitting right. Um, which was obviously very frustrating. Um, and I think the big thing was, um, in January of this year, there was this one customer that we had initially done a 30 day pilot. We'd extended it another 30 days and another 30 days. Cause we just weren't quite hitting the mark. And then the first version of like the factory that exists now, um we finally had it and delivered it to them um it was way worse than it is you know barely functional but it was like a beta version that was actually working and um you know we had like one week left in the pilot and then uh I had a sync up call with our champion just to get a sense of like you know the 90 days is up they're not going to do any more extensions and we had to you know figure out if we're going to proceed or not and so he's like hey Matan um you know I just spoke with our CTO and he was, you know, trying to get a sense of like, Hey, like, what are we doing here?

30:21Like we need to, you know, and if it's, if it's not going anywhere. Um, but you know, since you gave us beta access last week, this is, I finally realized like, this is the factory that I was envisioning when you first pitched me three months ago. Like, this is what I've been waiting for this whole time. And so our CTO asked like, Hey, you know, would you rather I hire a junior developer for you? Or would you rather we procure this and, you know, give it org wide or we procure this and you are the only person that uses it. Like if we procure this and you're the only person that using that uses it, is it still worth it?

30:54Um, and he was like, and Matan, I wanted to let you know that I told him that we need to procure a factory. Um, and I remember just like, you know, ear to ear grin, like smiling immediately. Like when the call ends, I like run into the main room and I'm like cheering with the whole team and we're all like celebrating tears, like streaming down my eyes. It was, uh, that was truly one of the most satisfying, uh, moments ever. Cause it was like two years in, um, and we hadn't really gotten that developer obsession until that moment. Um, and yeah, I mean, that was, that was incredible. Wow. Yeah. You created like real, real value and affinity for someone that they stuck around that long to see the vision through.

31:36And then so quickly they started doing like org wide deployment. and they did a big hackathon with their whole team. They were sharing about it. We ended up like, the guy was based in Australia. Like we flew him out. You know, we had a great time. Yeah, we're still in touch. I think moments like that are just so special and I think so memorable. Things like that are what make all of the grueling nights and very, you know, opposite end of the spectrum of emotions. It makes it all worth it. Next time, go to Australia. Yeah, honestly. Yeah, I've never been. It'll be fun. It's a good excuse. Yeah, yeah.

32:07So I want to shift into the business of this, particularly the business model and pricing. Let's start with pricing. So how do you package this for your customers? You obviously have a smaller tier option and then enterprise. So how does this work? Yeah. So for the, for the team's plan it's$40 a month. You know, it's relatively self explanatory there. And we have a free trial just to, you know, make sure that, you know, new people can go in and get a sense of what this kind of new paradigm looks like. for enterprises, we do usage-based. And I think the important thing there is we're well past the like seat-based world.

32:42It also just makes no sense because if you do seat-based with agents, they're going to play this game of like, oh, well, I don't know if this is an active seat. So I don't know if I should take it. People are maybe more reluctant to adopt AI. We'll just be like, no, no, no. I'm going to stay in Vim and Emacs. Like, I don't want to use all that fancy stuff, which is kind of counter to what we want to do. Like we want to work with them. We want to align our incentives such that the more they use droids, the more value they get. And we both want that outcome. Um, and so, you know, because of that, we're usage-based and, um, I think it's also very helpful to drive that adoption, which is something that's kind of underrated in terms of how the companies that are building these tools, I think, talk about what they're doing because half the battle is having the best product.

33:28I think that's, that's really important. But the other half of the battle is how do you actually get developers to change how it is that they build to go and build in this new way? And for us, it's this agent native way, having their primitive kind of fundamental step of software development change from me writing a line of code to delegating a task to an agent that requires a kind of mentality shift. and if we can do that by saying, hey, you know, go try it once. It's usage-based. Just send one, you know, send one, delegate one task to an agency if you like it. We find that that's a lot more inviting.

34:04People can then record a video, share it with their team. Then another random person will try it once. And we also find that within the enterprise, when someone sends like one delegation, the retention at the 10th week ends up being 85%. So even though there is this big behavior change and we're not an IDE. Like we're kind of separated from that with this new interaction pattern. Despite that, it's like once they try it once and see what this delegation can provide them, they end up being pretty hooked. Sticky. Yes, indeed. Yeah, it was not like that in 2024. Times have changed. Yes, indeed. And the product's gotten better.

34:41Yeah. There is a controversial topic right now, especially here in San Francisco, the like AI capital of the world. Yeah. And it is margins. Yes. Margins in code gen companies and AI companies. How do you think about margins within this? Are they good? Do you have good margins? We have good margins, which is shocking, shocking to believe we have good margin. And when I say good, I don't mean not that negative. I mean, positive margins. And I think a lot of that has to do with kind of the approach that we are taking and the philosophy that we're approaching this with, which is it's not about like, you know, reselling LLM usage and like in a very familiar package, it's about this behavior change, this move towards this agent native development.

35:25And when you do that, and when you're focused on delegating these tasks, the ROI is that much more clear to these enterprise buyers. And when the ROI is clear, like, I don't need to do this migration anymore. I don't need to do this refactor. You know, you're kind of able to structure these things such that you're not kind of subsidizing their token usage, but instead you're providing leverage for their developers, which obviously, you know, as an, as a enterprise, you value a lot and they'll pay accordingly. The pushback on this is that sales and marketing spend is super heavy. It's a super competitive environment.

35:59Not going to lie. I saw some billboards around the city. Yes, that's right. And they had to do with droids. They did indeed. So how does this fit in? Yeah, well, so I guess, first of all, the initial, the first billboards that we put up were free courtesy of our friends at Brex because of the points that we had, which was a really good proving point of that as a way of demand generation because we got some ridiculous leads. And then the ROI calculus changed and it went from, wait, this is so cool that Brex allows us to do this. Like, that's fun. Like, you know, send it to my mom to wait. This is actually a serious channel for us to find new customers.

36:37Also to re-engage like, you know, pipeline that maybe slowed down. I cannot tell you how many texts I got from like CTOs and VPs of engineering, like one hand on the wheel, one hand like photo of the billboard. So the ROI on that makes a lot of sense for us now. I saw one touching grass. I saw one with fire in a coffee cup. Yeah. Droids trip software while you sip coffee. Exactly. But we do love Brex, by the way. Rex is the billboard capital of the world for AI companies. We're moving quickly from a world where developers wrote 100 % of their code to a world where developers will write 0 % of their code.

37:17But most of the existing tools that are kind of in the ether and in the mind space are like retrofitting AI to the IDE. And we see that as, you know, in that famous Henry Ford quote, you ask people what they want, they'll say faster horses. We see the AI within the IDE as like faster horses. And the reality is in this like paradigm shift that we're going through, there is going to be behavior change. That behavior change is going to be the automobile and that automobile is agent native development. You're not going to be writing lines of code yourself. You're going to be packaging up tasks for your agents.

37:52You're going to be delegating it to them. And the success will be judged by how well you scoped it and how well you set up the validation and verification criteria in a similar way to, you know, taking a horse on a, on a mud road works fine. If you take a Model T on there, it might not do so well. And so it's worth investing some resources in paving your roads, setting up some gas stations. It's kind of analogous with what we're doing with some of these enterprises. The competition is really hot. We talked about that throughout the conversation a little bit, but in different ways. So you went, you have a good branding style droids, you have billboards, but you have another differentiation and that's your agnostic.

38:32So could you explain the different buckets in which you're agnostic and why this is important? Yeah, absolutely. So something that's really important for developers, especially as they're learning this behavior change is understanding how the models work. And there are a lot of tools out there that'll kind of say, we're going to figure out what the best models are. Like you as a developer, you don't get to see it. You just use our tool. We're model agnostic because A, that teaches developers, you know, how the different models behave and it gets them kind of aware of the different, you know, flavors or personality traits that these models have, gets them familiar with tokens and what that means in terms of the fidelity of each, you know, model call and that sort of thing.

39:09So we're model agnostic. That also helps for the enterprise because there are some enterprises that have special agreements with Azure or with AWS or with Anthropic directly or with OpenAI. And so allowing that to be kind of custom to whatever their existing setup is, is very helpful for them. We're also surface agnostic. And so, you know, there are some tools again that are like, you must use us in the IDE or you must use us in terminal or in a web app. Instead, our approach is we're asking a lot of developers to change their behavior. And so accordingly, we need to meet them where they are. And just how if you're working with a colleague, you wouldn't only ever interact with them on Slack or only ever interact with them on your browser.

39:48Similarly with droids, you should be able to delegate a task to a droid wherever you are, whether that's in your terminal, in your IDE, in a web app, in Slack, in linear. And so accordingly, you know, the droids are kind of ubiquitous. So wherever you're working, you can go and delegate to them. Cooperation compared to competition. That's what you're going for. Yes, exactly. We haven't talked about your team yet. I want to talk about your co-founder. You guys came together pretty quickly. Yeah. Could you break down that story a little bit more and tell us about him? Eno and I, we both went to Princeton together and weirdly we had like 150 mutual friends, but we were a year apart, 150 mutual friends, somehow never had a one-on-one conversation.

40:28Like if you had asked us, you know, way back when, like who the other was, we could probably give you a loose guidance, but we never like really spent one-on-one time together. And, you know, at that hackathon that I mentioned, we kind of saw each other from across the room and it was like, I know you. And, you know, we started talking and I joke that it was intellectual love at first sight because we started, we started talking that day and And legitimately, there's not been a single day since that we haven't been just like incessantly texting each other about, you know, coding and maybe some other things as well.

40:59But the next literally the next day we got we got together for coffee, talked about kind of doing this more formally together and then just got to work building, you know, that demo that we ended up presenting. Eno is so far and away the best engineer I've ever met in my life. It is crazy. Not only, you know, just in terms of technical ability or ability to, you know, go from working at Microsoft and Hugging Face to now like running 25 person engineering and product team at a factory. But he also is really good at behavior change, which is what we're providing to our customers. And the reason is when I first met him, it was only a couple months after ChatGPT came out.

41:38Most of the tools that we know now were not out. Most people didn't really change the way that they build software. but immediately I remember like working next to him he had you know chat gbt up and his ide up and he was doing all these interesting things of like how to make it more efficient to um you know get the relevant context in we ended up playing this game of like you know if you're not allowed to write a single line of code how accurately can you get chat gbt to uh you know get exactly what it is that you want and obviously that entailed you know pasting in the code running the code and pasting in the output.

42:10And obviously that is now what an agentic framework is. But he is just like any new AI tool, even outside of coding, like any new product, AI productivity tool, he is so good at adopting it the second it comes out, which I think is a nice compliment because I'm the most like routine and habitual person. Like I literally the exact same thing every day. I have the exact same like wake up time, all that. And so I think that balance is very fun between us. That's great. You have a true founder, co-founder love story. Yep. Yep. You're proud of it. Very much so. Okay. So how, how big is the team to date?

42:46How, how big have you scaled to? Yeah, we're, we're around 35 right now. Um, and, uh, you know, we started the year at like 15. So. Okay. Yeah. Doubling. Yeah. Should be ending the year at like 50 to 60. Nice. Good scale. Yeah. How are you thinking about composition? Like who, who's part of the team now? What are you looking to hire out for? Yeah. So big thing now is like, you know, obviously we're always in the business of hiring killer engineers. I think something that I've seen in some of the other players in the space is treating go to market as a little bit of second class, which I think there's a very fine line of like engineering is an art that deserves a lot of respect.

43:28But I think in some instances, I've seen some organizations kind of structured that, you know, these are tools built by engineers for engineers. And so it kind of puts them above. And then there's like everyone else. I think for us, something that's really core to the way that we work at factory, you know, we have an operating principle that's built together. And something that's important is anytime, you know, engineering and product ships a feature, salespeople say we, anytime sales close the deal, engineers say we. And I think it's really important because you can't drive behavior change, a 10 ,000 person org, just by having, the best product with the best features.

44:02You do need people to go in and drive that adoption with this go-to-market motion, getting the buyer and the champion and figuring out who the influencers in the organization are. And having a respect for that, I think is really important if you do want to scale to the degree that we're targeting. And so, yeah, for us, building our go-to-market is very top of mind. We're also across the board very much of the opinion that we would rather have one person that can do the job of two, but pay them like they're three. Right. So I think, you know, maintaining that density is really important for us.

44:39Are there any other principles that you've set within the culture? We have a lot of good operating principles, maybe one that I think is worth calling out. So obviously there's the famous Jeff Bezos, Amazon customer obsession. We have a little a bit of a slant on that, which our operating principle is create obsessed customers. It's not enough to be customer obsessed because that's a one way street. You could be obsessed with them and they want nothing to do with you, right? So our goal is really to build something so great and give them such a good experience that they become obsessed with us.

45:15And obviously the only way you can do it, it's table stakes to be customer obsessed. You also need to drive the output, which is them becoming obsessed with factory and what it is that we're delivering to them. Build a cult. Exactly. If you were to join any cult, what would it be? Honestly, I think the flat earthers seem like a, they're a pretty peaceful bunch. They like, there's so many funny videos of the flat earthers, like doing the experiments to test, to prove that it's a flat earth. And like the experiment obviously shows that it's round. And then they're kind of just like standing there and it's kind of like the office where it's just like, and it just cuts.

45:49I feel like it could be fun spending some time with them. flat earthers that's good yeah i haven't heard that one what about you i go between two um i think it would be fun to be like at a furry convention for a day i just think that's an interesting culture that would be that would be wild wild yeah and then the other one is like there's this one video of like this emo group underneath this bridge like doing this like rave dance thing to like mariah's carries like all i want for christ i think about this a lot clearly yeah all i want for christmas and i'm like thinking what would it be like to be like super emo for a day like what i feel like this is your halloween costume like you have to do it i think i should yeah but like what what do you think they talk about like it's just a normal group is it what's the difference between emo and goth i feel like in my head they're the same is it the same okay it's the same thing okay yeah i don't know i'm not sure we have to we have to find out well i don't know if there's an sf emo scene i feel like la would have it for sure i like SF definitely has it.

46:53I feel like SF definitely the furries. I don't know about that. I feel like emo is more LA, but. Maybe emo is like, I don't know. Maybe is that, is that, could that be more South Bay or is that more like East Bay? Like South Bay, like the kids of like the tech execs who are there. He's just like, it's not a phase mom. Not English, you know, really passionate. Yeah. Anyways, I don't know why that went that way. You're building a cult. This is exciting. Yes. Okay. As you build out your cult, I have to ask, are you at a stage? Okay. You're at a stage where you could probably choose between two things.

47:32If you had to pick, would you hire a mercenary or a missionary? Always has to be missionaries. I think it is like the most precious commodity in a startup is the people who are genuinely obsessed and passionate. I'd like to think everyone we have is a missionary. I think the thing that kind of keeps me up at night is like, how do I make sure as we go from 50 to 100 to 500? How do I make sure that we maintain that? Obviously, knowing the reality is like eventually you will have mercenaries. But I think the companies that do a really good job of maintaining that high missionary to mercenary ratio as they scale are the companies that end up doing the best.

48:14We already brought up Brex, but I have to ask the Brex performance question. Brex is all about spending smarter, moving faster and they give you billboards. But for you as a founder, like how do you keep your mind sharp? Like you mentioned you're pretty routine based, but like what specifically keeps you, keeps you in check as an executive? Yeah. I think the biggest thing for me is getting up early in the morning before work and working out. The delta of my productivity on days where I have time to do that versus don't is actually insane. Yeah. Like the morning workout is everything. Um, so that's the, do you track it?

48:52What do you mean? Like my product? Do you have like an aura ring or a lot? Um, I used to use a whoop. I stopped it because I'm so obsessed with like numbers that it was like getting too much. It was just like too much. Yeah. Um, but, uh, also even like I have an eight sleep. And so like my sleep after, if I work out, like every call goes well, every meeting goes well. I feel like we're, you know, making so much progress. If I don't, I always feel like so much slower. I drink more coffee and then, yeah, sleep worse. So, yeah. I just threw out my Apple watch. I stopped wearing it like a couple months ago.

49:25Do you believe in like the RF? You know. I actually studied that in college. Oh, no way. Fun fact. RF. EMF and RF. Yeah. I studied that. But no. Although I did like notice something. Really? Yeah, I don't know. I'm sure I'm fine. Yeah. But do you do AirPods or wired? I switch between them. I like go on really long flights a lot. So like they'll die. I'm going into a rabbit hole I shouldn't go into. Yeah. Okay. Who's one founder that you really admire or look up to? Someone that I admire a ton. This is kind of cliche, but I fucking love Winston Churchill. Okay. I've read so many different Churchill biographies.

50:11It's kind of a, it depends on if I'm like on a walk or I'll typically get an audio book and the physical book and I'll kind of alternate. But Churchill was just such a badass figure in terms of decision making, you know, rallying people together. Also just even before he became prime minister, he had some crazy experiences just around the world. and the kind of relentlessness that he had and conviction that he had in what was right. And then also kind of ingenuity in terms of like getting things done around constraints of the real world with competing kind of incentives from these different parties.

50:52I think he's incredible. Any favorite book? It's tough because I actually, I think there is a current lack of a, of a comprehensive, like the comprehensive ones are like 3000 pages, which is kind of like, yeah, it's like not as feasible, um, to, to do that. But, um, I think probably the most interesting, there's some, I think there's some, uh, there's a 200 page one on his early life, which probably has all the info that you like don't learn in history class about like why he's a very interesting person. Just like from the kind of like individual characteristics trait. I really liked his integration in The Crown.

51:36Did you ever watch that? No, I didn't. It's great. You should watch it. Particularly that season. I don't know which one it was, but it was great. Do you have a founder that you look up to? Elon Musk. Yeah, fair. He can't lose. And he doesn't. Yeah. It's ridiculous. I actually, I remember our, I remember Sean, our board member from Sequoia. right around the acquisition, we were at this dinner and everyone was like, there's no way the Twitter acquisition is going to work well. It's like, it's definitely going down in flames, like all that stuff. And I remember he was like, mark my words, like two years from now, you all will still be using it.

52:09Everyone is going to be like, it's going to be more relevant than ever. And then obviously, you know, here we are. Elon never loses. I have the fastest growth on X. X is the best channel. I only care about X. It's great for Sean. He made a good bet. Is there anything you particularly learned from him? From Sean? I think the crazy thing about Sean is he is so deep on so many different topics. It's genuinely like we'll not talk for like a week and then he'll somehow still know exactly everything that's happened since we last spoke and like the likelihood of what this happened and that happened. And he does all of that.

52:48And he's like really relentless on all these kind of different things that he's working on. And also, you know, really values family, spends a lot of time with families. Kind of always like, you know, we always like talk about that side of thing. And I think it's just pretty impressive how he does it. Also, fun fact about our walk. That guy walks literally everywhere. Like he walks to the airport when he's leaving. What? Like from San Francisco to the airport. With his bag? Yes. He walks? Yes. Like with a rolly bag. Yes. Wait, like, is there a... London Heathrow Airport. Like any airport he's at.

53:21It's hilarious. You should have him on and talk to him about this. I haven't gotten the walking from him yet. I'll have to be the pod on a walk. Yeah, no, it's wild. It is wild. How do you do that? These are questions you'll need to ask him. But that is relentlessness. And I think that's, it's very contagious. That is. Didn't Jack Dorsey do that too? I don't know. I feel like he would walk from San Francisco to Palo Alto. That seems like his vibe. Yeah. There's got to be some paths. Maybe they cross paths. Sean and Jack, they'll see each other. There's some secret tunnel we don't know about. Yeah, yeah.

53:51That's wild. Okay, Matan, as we close out, I like to do some Kalshi markets, some future predictions. Great. Number one, with all of the recent releases, what do you think the best AI will be this month? I'm afraid to admit, I think I use Kalshi a lot. You do? I do. I think - Wow, Kalshi Brex. This is amazing. I swear this wasn't sponsored. I'm tired of Turing. You know, they have 4 million engineers. That's not a bad lead gen. Yeah, we should work with them. You should definitely get integrated with Turing. Yeah. I know them. Yeah. You should put us in touch after. Okay. Yeah. My understanding is that I'm pretty sure Google is at the top there.

54:32The thing is, the one thing that's tough is I'm pretty sure the market is not necessarily for coding in particular, but like a kind of general performance. I think for coding, it's so tough because I think it's so task specific. For my personal coding use cases, I'll probably be using GPT-5 still by the end of the month. GPT-5 is pretty good. Yeah. It's all based on LM Arena. So read the fine print. Yes. Don't forget that. Given that, I was going to ask this after the next one, but do you think open AI will raise their prices this year? Yeah. You do? Yeah. How much? I don't know. I think model by model or it might be, they might do like plant, like certain plans or, you know, they, there are a lot of things you can do to make it not clear if you're raising the price or not by like change, just changing some of the structure.

55:24And, um, I think the unit, unit economics will certainly change. Yeah. So at the time of this, the market price for pro is 20 a month and the price of plus is 200 a month. So if it's any increase on time. Oh, I see. Okay. But what if, what if, and I guess read the fine print. Um, but if they like change what you get for 20 or 200, does that not count? Cause if that doesn't count, then I'll say no, they won't change it. Cause I think people like 20 and 200, but I think what you get with the 20 and 200. Yeah. Yeah. They're just nice numbers. They're nice numbers. What you do is you add another tier.

55:58Yeah, exactly. Like 2000, the$2 ,000 too. Just skip that. Just go with 10x and you'll be good. Yep. Okay. All right. We figured that out. Okay. So we're going to do some of these like IPO predictions. Okay. So CalShe also just added this new market. It's for IPO predictions. Nice. There are some fun ones. I'll bucket them out. Okay. So first, who do you think is going to go public first? OpenAI, Anthropic, or XAI? That's a good question. If you believe in the conspiracy theory that like Elon's using like the X, XAI thing to like be acquired by Tesla and then have more ownership of Tesla. I don't know if you've seen this, but there's a, I think there's a pocket of Twitter that's like really into that.

56:41I'm going to say OpenAI, but I don't think it'll be particularly soon. I wonder what, like how that like restructures them. Cause you know, they're. Yeah. I think, yeah, it's also subject to a lot of. But like that could make it attractive to you. Yeah. See if it just flattens everything out. bit help. We'll just go into broader AI companies. There's Glean, Databricks, and Cerebris. Which one do you think is going to IPO first? I would say Cerebris, Databricks, Glean, in that order. That's the order? Yeah. Why? I've just heard, I don't know, vibe-based. You've heard? What have you heard? No, just vibes.

57:14Just vibes in the ether. These are just vibes? Yeah. What's your take? I'd probably say that's the same. Glean is growing really fast. You don't hear about Databricks too often anymore. But it's like the one that's like, everyone's always like, oh my God, when is it going to... It's like one of those things. Like when is Databricks finally IPO-ing? Maybe they're quiet because they're getting ready. That's a strategy. Yeah. Okay. Latan, it was great to have you on. Congratulations on the big round again. Thank you so much. Thanks for coming on Sorcery. Thanks for having me. Hey, it's Molly. If you enjoy our interviews, check out our newsletter, sorcery.vc 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