The Exchanges

Q Can you give me an example of that?

A Let's say on the coding space, there's like, you see Claude comes out with like Claude four and Sonnet four. And you're like, perfect. Let's use that. Um, and before that we were like doing a split between Sonnet and, and Gemini. And then, um, OpenAI comes up with GPT-V. And so, and you try to figure out for all of those folks, by the way, it's like, it's not necessarily that there's like a better model that you move your entire workload, but you see that there's like places where the model's doing better. So for the first prompt, Sonnet's like the best, like the best designer, and knows how to structure it, and is really good with like writing a, uh, like a project from scratch. But then what we found is that like GPT-V, and especially the, the thinking and the pro, Versions like really good at solving hard bugs. And so you move a lot of workload, especially on the complex applications and very large context, uh, API calls to, um, OpenAI. And so when doing that, like you're literally moving hundreds of thousands of spend, which is actually just like the simplest change. And if now, I don't know, but maybe Gemini three is going to be better than anything else, which is what people, um, is nowadays saying. Then we're going to move a lot of spend and a lot of workload overnight to a different vendor. And that's like an insane market dynamics to me.

AI assessment note: “Let's say on the coding space, there's like, you see Claude comes out”

Q Can you give me an example of that?

A Let's say on the coding space, there's like, you see Claude comes out with like Claude four and Sonnet four. And you're like, perfect. Let's use that. Um, and before that we were like doing a split between Sonnet and, and Gemini. And then, um, OpenAI comes up with GPT-V. And so, and you try to figure out for all of those folks, by the way, it's like, it's not necessarily that there's like a better model that you move your entire workload, but you see that there's like places where the model's doing better. So for the first prompt, Sonnet's like the best, like the best designer, and knows how to structure it, and is really good with like writing a, uh, like a project from scratch. But then what we found is that like GPT-V, and especially the, the thinking and the pro, Versions like really good at solving hard bugs. And so you move a lot of workload, especially on the complex applications and very large context, uh, API calls to, um, OpenAI. And so when doing that, like you're literally moving hundreds of thousands of spend, which is actually just like the simplest change. And if now, I don't know, but maybe Gemini three is going to be better than anything else, which is what people, um, is nowadays saying. Then we're going to move a lot of spend and a lot of workload overnight to a different vendor. And that's like an insane market dynamics to me.

AI assessment note: “Let's say on the coding space, there's like, you see Claude comes out”

Q With the greatest of respects, that's wrong. Dude, you're not gonna have people build their own Salesforce CRMs on the whole. Maybe a tiny, tiny portion, or build your own intercoms, or Sierras, or Decagons, and then do maintenance and updates. Are you serious?

A I honestly, I honestly think so. It might be not as like the current way that we, that we think about vibe coding, where you prompt like, hey, build a CRM for me, and then you get a fully fledged old features inside. It might be that software becomes more liquid, and so maybe you have a starting point, maybe you have templates, maybe you have like open source projects that people can take And then vibe code their own features and say, okay, this is great. Now I want this, uh, this version in, uh, Arabic from right to left, and I want, uh, on every lead to be able to add my own, uh, pictures of the lead or whatever, and then basically convert and transform the software to something that's more, uh, what they want. But I think the current model of one size fits all doesn't make a lot of sense. It's not that everyone will move there. It's not that enterprises, But it makes so much sense for the buyer. Think about it as, like, the software is going to be theirs. The code will be theirs. They will not need to share their data with the other, with other software vendors and be locked in and like, uh, in whatever. They could adapt it and create a linear version that's way more customized to their needs. Um, I think every I'll, I'll go even one step further. I think every software company, um, that needs implementers, right, there's like a Salesforce implementer or whatever, um, maybe …

AI assessment note: “I honestly, I honestly think so. It might be not as like the current way”

Q In the nicest way, dude, no one shares that. Like what, there's no way, again, educate me, does no one share it because it's so bad?

A No. It's also very decision dependent, so, like, you can change it from, like, a low margin to a very high margin, And just one decision is just deciding on how great of a service you want to give your users. I think what will happen eventually with all coding agents is that the cheaper open source models are getting way better. You always want to play with like the, the strongest models, but what will happen and what we're kind of like looking at it internally is like building this switch or proxy where when you get a prompt from a user, Maybe the prompt is so simple that it doesn't even matter, like, what model you're getting. It's like, change the color of the button from blue to red, right? I don't really need to use, like, the heavy sonnet maximum thinking, whatever, pay a lot for that, right? I can, I can, kind of, like, forward this request to a smaller, cheaper model, and this will also be, like, a way better experience to the user because it's going to be so much faster than, than bringing, kind of, like, the heavy guns of, like, the frontier models, and I think Doing that, and the more models, the more the smaller or open source models are able to take a larger percentage of their requests and prompts, obviously margins are going to improve drastically. I don't think it's bad. That's, that's my, uh, I don't think, yeah, I don't think current margins are bad, but I thi…

AI assessment note: “No. It's also very decision dependent”

Q What metric is user success for you then? If there is this nuance and complexity with regards to like success for user internally, what is your metric for user success?

A There's a couple. Um, first we measure the sentiment of the messages they send in the chat. Um, which was also weird to me, but it turns out many people are like speaking with the chat as they would to a human. They either tell it, okay, fantastic job, this is exactly what I wanted when I was, like, asking for this feature, or they curse a lot, like, this is not what I wanted, oh, you deleted this button, and I wanted this, and I wanted that, and so we have this metric that we literally measure on a minute-to-minute basis of, like, how many of the prompts, because nowadays we get, like, tens of thousands of prompts, like, a minute or whatever in tandem, and so we measure the sentiment Um, and that's how we know our coding agent does what it's being asked to, and you see, kind of, like, how, uh, if we release an upgrade, you see, kind of, like, this, this, the negativity metric, kind of, like, goes down, or if, uh, when we jump from sun at four to sun at 4.5, you see a major improvement there, so this is one thing that we measure, and again, this Even the metrics change a lot with the business like base 44 because when we started, it was mostly around bugs, right? How many times like the model messes up something and the application doesn't work or there's like bugs and like hard things that like break your app. Nowadays, it's gonna, it's not easy to create bugs in your app. The…

AI assessment note: “first we measure the sentiment of the messages they send in the chat.”

Q With the greatest of respects, that's wrong. Dude, you're not gonna have people build their own Salesforce CRMs on the whole. Maybe a tiny, tiny portion, or build your own intercoms, or Sierras, or Decagons, and then do maintenance and updates. Are you serious?

A I honestly, I honestly think so. It might be not as like the current way that we, that we think about vibe coding, where you prompt like, hey, build a CRM for me, and then you get a fully fledged old features inside. It might be that software becomes more liquid, and so maybe you have a starting point, maybe you have templates, maybe you have like open source projects that people can take And then vibe code their own features and say, okay, this is great. Now I want this, uh, this version in, uh, Arabic from right to left, and I want, uh, on every lead to be able to add my own, uh, pictures of the lead or whatever, and then basically convert and transform the software to something that's more, uh, what they want. But I think the current model of one size fits all doesn't make a lot of sense. It's not that everyone will move there. It's not that enterprises, But it makes so much sense for the buyer. Think about it as, like, the software is going to be theirs. The code will be theirs. They will not need to share their data with the other, with other software vendors and be locked in and like, uh, in whatever. They could adapt it and create a linear version that's way more customized to their needs. Um, I think every I'll, I'll go even one step further. I think every software company, um, that needs implementers, right, there's like a Salesforce implementer or whatever, um, maybe …

AI assessment note: “I honestly, I honestly think so. It might be not as like the current way”

Q Penultimate one. You got married very recently. What's your biggest advice on partner selection?

A It shouldn't be, uh, like this person completes me. It should be someone that is, I'm enjoying this at least so much, that is identical to you. My partner is in their own, kind of like, uh, uh, field. Entrepreneur. She likes the same thing that I like. She's, uh, messy. She has ADHD like me. Uh, she likes to paint like me. This is so much fun, picking someone that's like you, because as an entrepreneur, it's like, you have to have sympathy for the person that's like missing day and night, um, and is constantly working, and you have to understand deeply why they're doing that, and, and that's, that, that, that, that's like a part of them as a person, same as it's a part of, of me as a person. So I think having the same mindset Is crucial. It's very, very hard to date a founder that's, like, uh, deep into, like, uh, building the company zero to one, and I think having a person and a partner that's, like, having the same mindset that they're building something, or they built something, is making it so much easier to empathize with.

AI assessment note: “It should be someone that is, I'm enjoying this at least so much, that is identical to you.”

Q In the nicest way, dude, no one shares that. Like what, there's no way, again, educate me, does no one share it because it's so bad?

A No. It's also very decision dependent, so, like, you can change it from, like, a low margin to a very high margin, And just one decision is just deciding on how great of a service you want to give your users. I think what will happen eventually with all coding agents is that the cheaper open source models are getting way better. You always want to play with like the, the strongest models, but what will happen and what we're kind of like looking at it internally is like building this switch or proxy where when you get a prompt from a user, Maybe the prompt is so simple that it doesn't even matter, like, what model you're getting. It's like, change the color of the button from blue to red, right? I don't really need to use, like, the heavy sonnet maximum thinking, whatever, pay a lot for that, right? I can, I can, kind of, like, forward this request to a smaller, cheaper model, and this will also be, like, a way better experience to the user because it's going to be so much faster than, than bringing, kind of, like, the heavy guns of, like, the frontier models, and I think Doing that, and the more models, the more the smaller or open source models are able to take a larger percentage of their requests and prompts, obviously margins are going to improve drastically. I don't think it's bad. That's, that's my, uh, I don't think, yeah, I don't think current margins are bad, but I thi…

AI assessment note: “No. It's also very decision dependent... I don't think current margins are bad”

Q You made eighty million bucks with the sale. Well done. Does money make you happy?

A I don't think money makes you happy. I think there's two things, um, that money becomes really valuable. Three things. Uh, one is your time. I don't like, like, fancy cars, so I didn't buy a car, but, like, nowadays I'm going with a taxi everywhere, and it saves tremendous amounts of time. Or other stuff that I can pay, you know, instead of, of, like, doing or wasting my time on. Um, and this is a lot of help in, like, personal stuff, especially if you're, like, busy people, you're entrepreneurs, and you want to be with your wife or family or whatever, so this is a huge benefit. Second is helping people you care about and helping causes that you care about, which is big, because at some points, like, there's not a big difference between doing X million dollars and 10 X million dollars. It's more about, like, okay, you can maybe with your family and stuff like that, you can, uh, Support. And the third, and I'm assuming, is like, um, you can use the money later on to build the bigger stuff. Or you have this Um, war chest that you can say, okay, next thing I'm going to do, or when I'm thinking about kind of like, ah, the, the, the, what I'm going to do in a decade from now, and it's like, okay, let's go to space, solve cancer, build robots, or something like that, because you can do that. You can take a riskier than that.

AI assessment note: “I don't think money makes you happy.”

Q Penultimate one. You got married very recently. What's your biggest advice on partner selection?

A It shouldn't be, uh, like this person completes me. It should be someone that is, I'm enjoying this at least so much, that is identical to you. My partner is in their own, kind of like, uh, uh, field. Entrepreneur. She likes the same thing that I like. She's, uh, messy. She has ADHD like me. Uh, she likes to paint like me. This is so much fun, picking someone that's like you, because as an entrepreneur, it's like, you have to have sympathy for the person that's like missing day and night, um, and is constantly working, and you have to understand deeply why they're doing that, and, and that's, that, that, that, that's like a part of them as a person, same as it's a part of, of me as a person. So I think having the same mindset Is crucial. It's very, very hard to date a founder that's, like, uh, deep into, like, uh, building the company zero to one, and I think having a person and a partner that's, like, having the same mindset that they're building something, or they built something, is making it so much easier to empathize with.

AI assessment note: “I think having the same mindset Is crucial.”

Q Where do you think no one is investing today that everyone should be investing today? And where do you think everyone is investing today where they should not be investing?

A Oh, wow. That's a good question. I think the not sexy industries Again, the finance, the, the restaurants, the whatever, the, the HMOs, the, like, actually having, like, entrepreneurs building that end-to-end is what people should be more, like, investing more into. I think a lot of people investing nowadays to very fast-growing on, like, a hockey stick, Um, revenue plotting, uh, startups, which is not what I think, unless you know a business or an industry very, very well, I think it's very hard to, for investors outside to understand if this is going to be with, with the moat or not. So a good example for that is I had the opportunity to invest in like this, um, agent companies, right? Like the Manus, the, the, those folks. I don't like, I felt like, I don't know the industry good enough to understand if at some point this would get commoditized by the model companies. And so this is like an area that I wouldn't invest in because I was like, okay, this is obviously dangerous. Like the, obviously the more that, that the capabilities that OpenAI builds into ChatGPT and then Google into Gemini, whatever they're like, they're going to be well, like they're going to have like serious and, and incredible agents. And I'm not sure this category, even if you get to, like, insane growth, like, you can see the Manus and the other players, like, they have this, this graph that I think ar…

AI assessment note: “I think the not sexy industries Again, the finance, the, the restaurants”

Q How does that change what you build or how you think?

A You take big bets on things that are, if you can find those that are substantial to your users and will be really hard to copy. So I'll give you one example. Uh, the whole reasons for Base-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty- User management, authentication, integrations, analytics, and so on. And the thing is, is that we didn't use a third party provider like Superbase, which is what the rest of the industry is using, Superbase on Neon. And we've built that like built in, and we've taken a very courageous kind of like a step there, and there's still a lot more to build there. This is going to be very hard to replicate and very hard to migrate your, uh, Whatever, millions of users from one, from Superbase to a homegrown solution, if you, if one of the competitors tried to do that, and I think this is a bet that we're deepening further and further and further. For other stuff, it's a game of velocity, of delight, of like, taking the right UX decisions, um, and having the right taste, um, and just be, trying to be there before everyone else gets there. So it was like, uh, I think we started this, uh, so two weeks ago we came up with something called …

AI assessment note: “You take big bets on things that are... hard to copy. So I'll give you one example.”

Q How does that change what you build or how you think?

A You take big bets on things that are, if you can find those that are substantial to your users and will be really hard to copy. So I'll give you one example. Uh, the whole reasons for Base-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty- User management, authentication, integrations, analytics, and so on. And the thing is, is that we didn't use a third party provider like Superbase, which is what the rest of the industry is using, Superbase on Neon. And we've built that like built in, and we've taken a very courageous kind of like a step there, and there's still a lot more to build there. This is going to be very hard to replicate and very hard to migrate your, uh, Whatever, millions of users from one, from Superbase to a homegrown solution, if you, if one of the competitors tried to do that, and I think this is a bet that we're deepening further and further and further. For other stuff, it's a game of velocity, of delight, of like, taking the right UX decisions, um, and having the right taste, um, and just be, trying to be there before everyone else gets there. So it was like, uh, I think we started this, uh, so two weeks ago we came up with something called …

AI assessment note: “You take big bets on things that are... substantial to your users and will be really hard to copy.”

Q Does that not take away a lot of the low end of the market, though? If you think about that, essentially, if it's able to build, like, good front end, good landing pages, good simple projects, then it basically leaves the higher end champion users for the more, ah, complex service providers like you. Does that not eat away and cannibalize a lot of the market?

A I think it depends on like what you call the market. So like for base. We almost don't focus on websites, which also goes very well with Wix, right? Because Wix does a lot of like websites and marketing websites and e-commerce websites and stuff like that. For base. For our laser focus from our core focus from day one is the complex applications, the functionality. So if you were looking to build like a, A marketing platform where AI agents can generate content constantly and, and read your LinkedIn posts and see your, um, the traction that you're getting informed or try to digest, and then fill in your calendar, different, like, functional, complex stuff that you use day to day. And I think early on, the vibe coding category was focused a lot and still is, to some degree, on, um, Like front-end and landing pages, uh, where many use cases were on, like, uh, websites or e-commerce websites and stuff like that. I think this will get commoditized, um, for the category. There's still a lot to do when you're looking, like, there's SEO and domain management, and, like, if you're building an e-commerce, that's, like, a complex business. You want to integrate with, like, a full back-end, like, what Wix has or Shopify or whatever. Um, but for the functional stuff and the complex applications, the organizational applications, the personal applications, the full fully fledged SaaS platfor…

AI assessment note: “I think it depends on like what you call the market.”

Q sales forces in, like, the incumbent insertion point, and then you've got the canvas on the other, kind of more consumer end. I'd probably put figma in there with canva. What happens here? And then, sorry, you've also got the cognitions, and the cursors, and the Pro Tools there. Do the Pro Tools come down and eat your business? Do you go up and eat their business? What happens there?

A It's a very good question. I'll start with, like, the, what you categorize as incumbents, but I'll actually call it, like, the, the different software categories that exist. So, like, let's take a Salesforce or a Monday.com, for example. I think a lot of those tools will build built-in vibe coding tools inside their platform. You can see that, like, with Monday Vibe, which is a tool they've built on top of, like, the Monday.com infrastructure of, like, boards and, like, the database behind the scenes and the, the user management, roles, permissions, and so on. And this is what I, what I said when I meant software is going to become more liquid. So, I don't know if Salesforce is planning to do that at all, but I'm assuming that at some point, like, you'll be able to create your own user interface for Salesforce. And Salesforce would become like this system of record where like you can create the UI that you want and how you want to look at a lead or contact or account or whatever. And I think a lot of tools are going to do that because this is basically taking kind of like the, instead of like configuring different things and building different features, it would allow a lot of extensibility to existing products. Um, and so you start seeing that and you see also like, uh, Um, Microsoft building like some vibe coding tools on top of like the, the office and whatever. Like I, I ca…

AI assessment note: “I'll start with, like, the, what you categorize as incumbents”

Q How do you think about that analogy from LLM provider to cloud provider, given what you just said?

A Yeah. You do not switch a cloud provider. That's very hard to do, right? My previous company will try doing that. It was so painful. Uh, if I've known how tough it's going to be when we started that for like a year long process, probably wouldn't do it. For the LLMs, it's insane how easy it is. I think they're still like, so I think what we see for the model providers is that they're building multiple tools For multiple levels in the value chain. So you see all of them having this, uh, CLI coding tool and all of them having this very, um, high end or, or end user focused chat product like Gemini and chat GPT and so on. Um, you start seeing them bringing in. So Gemini is now everywhere in my Gmail and like a lot of like workflows and like, uh, building automation. And so you start seeing them go up. The market of software and trying to, and, and bringing their tools and their model basically to drive consumption all across the stack. Or if it looked like a year ago, it was mostly like you could consume it either in an API or you could consume it in like this, uh, um, end user type of chat. I don't know. It's going to be super interesting. It's just that to me, it's crazy. The fact that it's like, what, let's say one of them kind of like, uh, uh, storms ahead. For like six months, right? Something that haven't happened before. Uh, but it's like literally the best model for like s…

AI assessment note: “You do not switch a cloud provider... For the LLMs, it's insane how easy it is.”

Q What metric is user success for you then? If there is this nuance and complexity with regards to like success for user internally, what is your metric for user success?

A There's a couple. Um, first we measure the sentiment of the messages they send in the chat. Um, which was also weird to me, but it turns out many people are like speaking with the chat as they would to a human. They either tell it, okay, fantastic job, this is exactly what I wanted when I was, like, asking for this feature, or they curse a lot, like, this is not what I wanted, oh, you deleted this button, and I wanted this, and I wanted that, and so we have this metric that we literally measure on a minute-to-minute basis of, like, how many of the prompts, because nowadays we get, like, tens of thousands of prompts, like, a minute or whatever in tandem, and so we measure the sentiment Um, and that's how we know our coding agent does what it's being asked to, and you see, kind of, like, how, uh, if we release an upgrade, you see, kind of, like, this, this, the negativity metric, kind of, like, goes down, or if, uh, when we jump from sun at four to sun at 4.5, you see a major improvement there, so this is one thing that we measure, and again, this Even the metrics change a lot with the business like base 44 because when we started, it was mostly around bugs, right? How many times like the model messes up something and the application doesn't work or there's like bugs and like hard things that like break your app. Nowadays, it's gonna, it's not easy to create bugs in your app. The…

AI assessment note: “first we measure the sentiment of the messages they send in the chat.”

Q Why don't you think it's a winner-take-all? And if it's not a winner-take-all, does that still work for a venture model? Our business is predicated on winner-take-all.

A Not necessarily. Our business is predicated, or the venture capital business is predicated on huge returns. Right? If you look at, uh, all of my competitors, And on base 44, we're all growing insanely fast. And I think at least I can say about base 44, but maybe the other players like healthy businesses. Those, those are going to generate huge returns. They're all like fast growing businesses, and it's not a winner takes it all market, which is also, by the way, Why I got into this market in the first place. It was like, when I started, there was like Bolt and Lovable were like very viral on, on, on social media. And I was like, this is going to be such a big market. I feel like I have a very interesting take on this market with the backend built in, like focusing on functional and complex apps. Well, most of the other players were focusing back then on like frontends and maybe could connect the frontend to super base, but it was very hard. I think. We're at a different ballgame where AI is going to generate so much value that markets are going to grow very, very fast. Obviously some will get cannibalized, but markets are going to grow very, very fast. And the VC model is not necessarily a winner takes it all, but it's a huge returns type of market.

AI assessment note: “the venture capital business is predicated on huge returns”

Q How do you think about that analogy from LLM provider to cloud provider, given what you just said?

A Yeah. You do not switch a cloud provider. That's very hard to do, right? My previous company will try doing that. It was so painful. Uh, if I've known how tough it's going to be when we started that for like a year long process, probably wouldn't do it. For the LLMs, it's insane how easy it is. I think they're still like, so I think what we see for the model providers is that they're building multiple tools For multiple levels in the value chain. So you see all of them having this, uh, CLI coding tool and all of them having this very, um, high end or, or end user focused chat product like Gemini and chat GPT and so on. Um, you start seeing them bringing in. So Gemini is now everywhere in my Gmail and like a lot of like workflows and like, uh, building automation. And so you start seeing them go up. The market of software and trying to, and, and bringing their tools and their model basically to drive consumption all across the stack. Or if it looked like a year ago, it was mostly like you could consume it either in an API or you could consume it in like this, uh, um, end user type of chat. I don't know. It's going to be super interesting. It's just that to me, it's crazy. The fact that it's like, what, let's say one of them kind of like, uh, uh, storms ahead. For like six months, right? Something that haven't happened before. Uh, but it's like literally the best model for like s…

AI assessment note: “You do not switch a cloud provider... For the LLMs, it's insane how easy”

Q How do you determine if they're going to be eaten by models? Like, what is the question or criteria that will help you get comfortable with that?

A If most of the value Is on the way they prompt the application for a certain thing. Like they've built, let's say a photo editor. And so there's not a lot of features in a photo, like an AI based, like you feed a picture and say, okay, make me, I don't know, make me prettier or whatever. Um, and if a lot of that's going to be like the UX in the prompt, I don't think that's a, that's a, that's going to be a healthy business for the next day. I don't think there's like a lot of mode. If you're building a lot of infrastructure, or the company itself is vertically integrated, I think that's a very interesting take. So, like, you could nowadays build a competitor to Harvey and Spellbrook that would kind of, like, teach some middle limb better way on, like, how you, how to do law or whatever, or you could build a law firm. And I think building a law firm is going to be very, very, very interesting, because this is not something that a model provider would do, or any other kind of like sane software entrepreneur that wants to do just software. And I think there's like an opportunity today to build very healthy, new age, vertically integrated businesses. Um, and it could be again a law firm, it could be Buying a hospital, basically optimizing everything on every layer and making it the AI native and maybe in a few years being the first ones to introduce, uh, uh, uh, robots into, uh, as…

AI assessment note: “If you're building a lot of infrastructure, or the company itself is vertically integrated”

Q Where do you think no one is investing today that everyone should be investing today? And where do you think everyone is investing today where they should not be investing?

A Oh, wow. That's a good question. I think the not sexy industries Again, the finance, the, the restaurants, the whatever, the, the HMOs, the, like, actually having, like, entrepreneurs building that end-to-end is what people should be more, like, investing more into. I think a lot of people investing nowadays to very fast-growing on, like, a hockey stick, Um, revenue plotting, uh, startups, which is not what I think, unless you know a business or an industry very, very well, I think it's very hard to, for investors outside to understand if this is going to be with, with the moat or not. So a good example for that is I had the opportunity to invest in like this, um, agent companies, right? Like the Manus, the, the, those folks. I don't like, I felt like, I don't know the industry good enough to understand if at some point this would get commoditized by the model companies. And so this is like an area that I wouldn't invest in because I was like, okay, this is obviously dangerous. Like the, obviously the more that, that the capabilities that OpenAI builds into ChatGPT and then Google into Gemini, whatever they're like, they're going to be well, like they're going to have like serious and, and incredible agents. And I'm not sure this category, even if you get to, like, insane growth, like, you can see the Manus and the other players, like, they have this, this graph that I think ar…

AI assessment note: “I think the not sexy industries Again, the finance, the, the restaurants”

Q Michael Eisenberg said I had to ask you, make the case For how you beat Cursor?

A When I started Base-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty-Forty And so I said, okay, let's build for the very non-technical folks. And then when doing so, obviously like, and, and base for the four started getting a lot of traction. And obviously like we also got some developers and some technical people that said, give me the code. Um, and, and like, you should invest in like a GitHub integration, allow me to edit files and whatever. And, and at some point we also implemented that, but the more time passed, And the more the models got better and the more like faster and whatever, there were less bugs. Nobody wanted to see the code. Uh, and it was less important for, for people to actually be able to manually edit and change and whatever. And I think the direction, and again, that we're going to is like more and more and more percentages of like more and more types of software will, you'll be able to build inside base. 44 that. Nobody wants to see code. I don't think it's like if you'll get to like this flow state where you're pointing with the AI and it's going to be super fast and it's going to be almost immediate. Every point, like you'll see that and it's going to change. I don't think you want to see the code and I don't think you want to start saying, okay…

AI assessment note: “even the most technical users Would prefer building a site base... than to do that in cursor”

Q Do you think we will hit a speed bump? Everyone is very concerned about the AI bubble and not delivering on the revenue that we promise. Do you think that we will have a speed bump where there is a sign that we don't deliver the revenue that we promise?

A I think the question is whether it's going to turn out to be a platform play, which means when building better models, the entire ecosystem and the market benefits from, or whether what will happen is like model companies are going to fight, their margins will get worse and worse because they will need to lower prices. Um, maybe at some point, um, we'll need to figure out or correct something better in the LLMs because maybe we'll hit a wall. Um, so this will be the bigger, bigger question. Like, is the value that nowadays, uh, GPT five to GPT 5.1 adding is only to open AI for the rest of the industry, uh, or the rest of like the market. If it's for the rest of the market, I don't think we'll hit a speed bump. Me personally, I don't believe we'll hit a speed bump. I think even with existing models, I don't think also we'll hit like a wall with like the, the improvements of LLMs, but I think even with existing models, we're only scratching the surface of economic value that you can create. If, uh, if a one person team can, uh, can get sold for eighty million dollars and building a viable business and healthy business and profitable business without people. It means that you'll be able to do a lot more stuff with very like less people and that companies will become more efficient. And, uh, and, and as a result of that, every product that we buy. It's going to be better. It's goin…

AI assessment note: “Me personally, I don't believe we'll hit a speed bump.”

Q not, depending on where you shift spend. Can you advise me, honestly, dude, I'm an early stage investor, I invest in AI apps today, and everyone is posting insane fucking numbers. Like, you know, you look at yours, and it's like, a hundred million, you know, your, your lovables and your replets, like, two hundred million. How much weight Should I, as an early stage investor, place on revenue today?

A That's a very good question, which also me as a myself as an investor, like I'm asking myself. So obviously like the metrics today is like, uh, we are like different, right? When I started my previous company, we got to like, two million dollars in ARR. I think it was like a year and a half and everybody were cheering. It was saying like, this is insane and like, whatever. Nowadays it's like, if you, if you can show the, the hockey stick of how fast you got to Uh, ten million dollars or twenty million dollars. Everything changes. I think it depends on like the, the model of the company you're investing in. Uh, for me is like smaller checks into very healthy businesses. So the thing that I look at the most, okay, fast revenue growth, obviously that's what you want to see. This is like the most exciting, but there are two things. One healthy business is like, how likely is this business to get eaten by other players, the model providers or whatever. And the second, which is very much related. Is a vertically integrated business. Um, so here's something really cool that, that, that we're looking at right now. So Kerser came out, I think it was two weeks ago, or maybe a few weeks ago with their own model, right? Composer. Um, and this is like an, an, a nice way to like think about strategy of those types of businesses. So think about Kelsa, right? I don't know how the margins were.…

AI assessment note: “fast revenue growth, obviously that's what you want to see... but there are two things”

Q They don't have what they think they have?

A I don't know if, how they're thinking about this, but It's like, what's the, what's the end game for the race? Right? It's like, either it's not going to be one winner takes it all. And there's going to be like more and more companies getting into this market. It's going to be like the Chinese players that are now like changing the entire market dynamics. I think eventually it's like, if like it takes one better model or lower price, To drastically reduce your revenue, then I think, or at least they used to be delusional about this. I don't know if they used to be delusional about this. Maybe they always like, they were always planning to build kind of like up the stack. I think nowadays they're doing smart things. For example, on topic building cloud code SDK. Uh, or CloudCode. Like, even if now Gemini three comes out and it's a better model than Cloud, uh, and so on, then still there's gonna be a huge number of people staying with CloudCode because they're used to it and they built some things up the stack.

AI assessment note: “or at least they used to be delusional about this”

Q Final one for you, my friend. You can go back to the night before you start Base-Forty-Four, and you can say, ah, Mayo, you should know this. Future self has learned this today. What do you go back and tell yourself that you wish you'd known when you started?

A When I was, when I was by myself, it was scary as hell. I used to wake up every two hours in at night just to check that the platform is up. And at some point, um, Baseball started growing really fast, and I was still by myself, and people started building, like, businesses and applications on top, and There was, like, I remember this week when it started, like, really going up, and the infrastructure and the thing that I built didn't, didn't hold up, and so we had, like, many, there was, like, this week or two or three that they had, like, four incidents where the platform would go down, and people were, like, screaming and shouting at me as, like, uh, I trusted you and whatever, and I took it very, very hard, and it was, like, it was, like, emotional, very, very hard, and I wish I could have told myself, like, This is just going to be a bump in the road. Community is going to recover from that, and eventually it'll be like, like, this is going to be sustainable. This is going to be great. Don't take, like, don't take the bumps in the road. Also consumer, or the consumer market is so much different. Like the, the emotional spectrum of the feedback that you get is so much wider than building a B to B SaaS, that you get like some, ah, very, ah, not, not, not too positive, not too negative, ah, comments. Uh, I think handling it better emotionally is important. I didn't know that.

AI assessment note: “I wish I could have told myself, like, This is just going to be a bump”

Q How do you determine if they're going to be eaten by models? Like, what is the question or criteria that will help you get comfortable with that?

A If most of the value Is on the way they prompt the application for a certain thing. Like they've built, let's say a photo editor. And so there's not a lot of features in a photo, like an AI based, like you feed a picture and say, okay, make me, I don't know, make me prettier or whatever. Um, and if a lot of that's going to be like the UX in the prompt, I don't think that's a, that's a, that's going to be a healthy business for the next day. I don't think there's like a lot of mode. If you're building a lot of infrastructure, or the company itself is vertically integrated, I think that's a very interesting take. So, like, you could nowadays build a competitor to Harvey and Spellbrook that would kind of, like, teach some middle limb better way on, like, how you, how to do law or whatever, or you could build a law firm. And I think building a law firm is going to be very, very, very interesting, because this is not something that a model provider would do, or any other kind of like sane software entrepreneur that wants to do just software. And I think there's like an opportunity today to build very healthy, new age, vertically integrated businesses. Um, and it could be again a law firm, it could be Buying a hospital, basically optimizing everything on every layer and making it the AI native and maybe in a few years being the first ones to introduce, uh, uh, uh, robots into, uh, as…

AI assessment note: “If most of the value Is on the way they prompt the application”

Q You made eighty million bucks with the sale. Well done. Does money make you happy?

A I don't think money makes you happy. I think there's two things, um, that money becomes really valuable. Three things. Uh, one is your time. I don't like, like, fancy cars, so I didn't buy a car, but, like, nowadays I'm going with a taxi everywhere, and it saves tremendous amounts of time. Or other stuff that I can pay, you know, instead of, of, like, doing or wasting my time on. Um, and this is a lot of help in, like, personal stuff, especially if you're, like, busy people, you're entrepreneurs, and you want to be with your wife or family or whatever, so this is a huge benefit. Second is helping people you care about and helping causes that you care about, which is big, because at some points, like, there's not a big difference between doing X million dollars and 10 X million dollars. It's more about, like, okay, you can maybe with your family and stuff like that, you can, uh, Support. And the third, and I'm assuming, is like, um, you can use the money later on to build the bigger stuff. Or you have this Um, war chest that you can say, okay, next thing I'm going to do, or when I'm thinking about kind of like, ah, the, the, the, what I'm going to do in a decade from now, and it's like, okay, let's go to space, solve cancer, build robots, or something like that, because you can do that. You can take a riskier than that.

AI assessment note: “I don't think money makes you happy.”

Q Does that not take away a lot of the low end of the market, though? If you think about that, essentially, if it's able to build, like, good front end, good landing pages, good simple projects, then it basically leaves the higher end champion users for the more, ah, complex service providers like you. Does that not eat away and cannibalize a lot of the market?

A I think it depends on like what you call the market. So like for base. We almost don't focus on websites, which also goes very well with Wix, right? Because Wix does a lot of like websites and marketing websites and e-commerce websites and stuff like that. For base. For our laser focus from our core focus from day one is the complex applications, the functionality. So if you were looking to build like a, A marketing platform where AI agents can generate content constantly and, and read your LinkedIn posts and see your, um, the traction that you're getting informed or try to digest, and then fill in your calendar, different, like, functional, complex stuff that you use day to day. And I think early on, the vibe coding category was focused a lot and still is, to some degree, on, um, Like front-end and landing pages, uh, where many use cases were on, like, uh, websites or e-commerce websites and stuff like that. I think this will get commoditized, um, for the category. There's still a lot to do when you're looking, like, there's SEO and domain management, and, like, if you're building an e-commerce, that's, like, a complex business. You want to integrate with, like, a full back-end, like, what Wix has or Shopify or whatever. Um, but for the functional stuff and the complex applications, the organizational applications, the personal applications, the full fully fledged SaaS platfor…

AI assessment note: “I think this will get commoditized, um, for the category.”

Q Why don't you think it's a winner-take-all? And if it's not a winner-take-all, does that still work for a venture model? Our business is predicated on winner-take-all.

A Not necessarily. Our business is predicated, or the venture capital business is predicated on huge returns. Right? If you look at, uh, all of my competitors, And on base 44, we're all growing insanely fast. And I think at least I can say about base 44, but maybe the other players like healthy businesses. Those, those are going to generate huge returns. They're all like fast growing businesses, and it's not a winner takes it all market, which is also, by the way, Why I got into this market in the first place. It was like, when I started, there was like Bolt and Lovable were like very viral on, on, on social media. And I was like, this is going to be such a big market. I feel like I have a very interesting take on this market with the backend built in, like focusing on functional and complex apps. Well, most of the other players were focusing back then on like frontends and maybe could connect the frontend to super base, but it was very hard. I think. We're at a different ballgame where AI is going to generate so much value that markets are going to grow very, very fast. Obviously some will get cannibalized, but markets are going to grow very, very fast. And the VC model is not necessarily a winner takes it all, but it's a huge returns type of market.

AI assessment note: “venture capital business is predicated on huge returns”