Jordan Metzner: Make it rain mode.Hooman Radfar: I love that. I love that for us.Jordan Metzner: A custom make it rain disco song for Collective. It make rain game. You gotta collect all the money.Jingle: Built this week, breaking it down. Built this week, we show you how. A fresh idea, a clever tweak, you locked in. You built this week.Sam Nadler: Hey, everyone, and welcome to another episode of built this week, the podcast where we share what we're building, how we're building it, and what it means for the world of AI and startups. I'm Sam Nadler, cofounder, Ryz Labs, and I'm joined each and every week by my friend and business partner, Jordan. However, this week, we have a guest joining us. I'd love to introduce Hummin, the CEO of Collective. Hummin, if you don't mind, give us a a quick quick intro about you and Collective, and then we'll jump into the docket for today.Hooman Radfar: Fantastic. Thanks, guys, for having me. I'm Hooman Radfar, and I'm one of the founders at Collective. And just to give you a quick and dirty background on me, I've been doing startups for a very long time, actually, my entire career. So I've never held an honest job, and founded one company and sold it, helped join and was a founding partner at a VC firm, is how I met a lot of great folks like Jordan and and backed, think, gosh, over 70 companies there and 50 personally.And then I'm back in operating in the Collective, so I'm really excited to be here. And the quick and dirty on Collective is we're the first online back office that's specifically purpose built for solopreneurs. So think to yourself, there's a lot of things that you wanna do if you're the designer, the best developer, the best marketer in the world. I think all the things you wanna do don't include forming your company, running your books, running your payroll, doing your taxes. So our mission is to for you to do the things that you love, and we'll take care of those things as your partner.Sam Nadler: Amazing. So we'll jump into collective deeply, a little later on in the show, but just to give everyone the high level agenda. And before I do, don't forget to like and subscribe. We have new episodes of Built This Week every Friday. And listen on your favorite platforms, whether that's YouTube, Spotify, Apple Podcast, etcetera.But for the docket today, Jordan's gonna cover a tool he built with collective in mind. I think it has, some cash flow management views and some other, opportunities for these solopreneurs. Then we're gonna jump into the collective platform. And then finally, as always, cover some hot AI news. So, Jordan, why don't you give us the the quick demo and a little intro of what you were thinking when you built it, how it was built, and then we'll we'll jump into the tool.Jordan Metzner: Awesome. Thanks, Hooman. Great to see you. Thanks, Sam. You know, I've known Hooman for a long time and seen Collective grow from just small idea into a pretty big company it is today.And so it was awesome to get Houman, who's been a builder his whole career. You know, his first startup was really, really about builders and, you know, adding adding in the add this plug in. I wanted to do something fun today. So I asked him in a few ideas or future features that Collective was thinking about. And so while I'm not a Collective user myself, I decided I would build out some of these features.So I wanna give you a quick demo of what I build. I, in fact, I actually built two things today. So we have a little extra surprise.Hooman Radfar: Of course you did. Of course you did.Jordan Metzner: So, you know, this is just the basic dashboard. I'm sure the collective dashboard is much nicer, so, you know, we don't need to spend too much time on the dashboard. But let me tell you what I built today. I built cash flow, insurance, lending, and some AI insights. I'll talk to you a little bit about how I did that.But, essentially, since I don't know too much about how the collective system works, I allow you to simply update upload your profit and loss statement and your balance sheet. And with that, you'll be able to do cash flow analysis, the same for pretty much for your insurance, and then we'll talk about lending as well. Before I jump into that, the first thing I wanted to show you is just a little bit, you know, try to make, you know, the UI a little nicer. So I have light mode, I have dark mode, and just for Collective, I've made Make It Rain mode.Hooman Radfar: I love that. I love that for us.Jordan Metzner: Wow. I've made a custom Make It Rain disco song for Collective called Make It Rain.Hooman Radfar: I like it. So Jordan Jordan's gonna run marketing for us, by the way. That's what we Sam didn't tell you. There we go, guys. This is the feature we've been missing.That usually you could've just done this one feature.Jordan Metzner: It wouldHooman Radfar: be good.Jordan Metzner: I got more coming. Don't worry. Okay.Hooman Radfar: Alright. Let's go. I'm excited, man. I get to take this back to the product team and and get them jazzed up or maybe upset that they they need to speed up. You know?They gotta go Jordan's speed.Jordan Metzner: Well, you know, everyone likes to make some money, and I think Collective is a money management solution. So, you know, wanna wanna make it rain a little bit. So here's how the cash general forecast forecasting works. I kind of have, like, inflows, outflows, and I have it separated out here by month. Since I don't have access, like, to your data directly, what I built here maybe this will be easier to see in light mode.Hold on a second here. So what you do is you just upload your p and l and your CSV. You know, I made three different businesses under my Built This Week Holdings company. I have Sam's sewing business, I have Jordan's design business, and I have Humans Consulting business. You know, Humans Consulting business does the most revenue.Hooman Radfar: Thank you.Jordan Metzner: Yeah. So what you do is you you can generate a forecast. So first, I'm gonna just take my p and l, and, you know, I I just generated some p and l's here, but I have Humans' p and l here. And then I'll upload the balance sheet as well. So I have Humans' balance sheet from his consulting business.And then here, can say revenue outlook. If you think your revenue is gonna go up, if you think your expenses are gonna go up, or if you think your owner salary is gonna go up, you can also do stay flat or decrease, you can kinda just leave it as is. So if you have some idea of, like, you know, your major expenses, and this could be like, you know, it could also say like rent, taxes, other like kind of major expenses. I know for our business, we try to manage where we only have a few big major expenses every month, and so that's what kind of like moves the needle, like how big our credit card is or how big our payroll is and things like that. So I figured, you know, if you you know, or revenue also has a big impact.So, you know, if you think your revenue is gonna go up or your expenses are gonna go up because you're foreseeing something, then that might help generate the forecast. One quick thing here. I started with just using, like, rock one twenty b model just because I had access to it, but there was a lot of data inside these p and l's and balance sheets, and so it wasn't enough. I needed a bigger model, and so in fact, I moved to Gemini, and I'm using Gemini here because I was able to take the data in. So let me just click the generate forecast button.So Gemini is gonna analyze my financials, and this is actually happening. This is like really using AI. And here it is, and I'll zoom in a little bit here for you. But basically, here's next month, here's next quarter's projections, next year, and says Humans Consulting businesses healthy revenue growth, net income margins are fluctuating, and optimization expense management's business scales. So you get an idea, you know, of kind of one way of how you can generate some expenses and do some cash forecasting.So let me just run into these other ones. Love to get your feedback.Hooman Radfar: What'd you use for the front end work?Jordan Metzner: Oh, yeah. Good question. I think I built the first first version in Bolt just, like, really, really quickly just to, like, get something up. Yeah. But I quickly moved it into, like, my IDE, which is cursor, and I used OpenAI's codex to just do some coding for the front end and wire everything up and make sure, like, all my APIs work.So I like to use ViteReact as well as ShadeCN and Tailwind mostly because, like, they're pretty responsive and they look pretty good. And then I use this special library for filtering that I like that always, like, makes it look pretty sexy. Alright. Let's move into insurance. So I did something similar for insurance.So I had kind of, like, your general insurance policies here. I figured you would, like, input them or import them, and then you can run this thing called the compliance check. And then essentially, Gemini will review your insurance policies and and based on the data, it's gonna give us here some recommendations, and I'll zoom in for those. But it says, you know, moderate compliance needs, you know, some work in workman's comp and cyber liability. And then it says here, I need to renew my workman's compensation that it lapsed.So, you know, insurance is something that we don't think about top of mind every day, but, know, like, you know, I'm sure it's important to small business owners. It's critical to make sure you don't lose coverage. So, yeah, that's how the insurance insurance compliance check works. And then lending, something pretty similar. So we've got the holdings from our three companies.Remember, Huomount's company doing 27,000? So we've got, like, a little bit of financial readiness here with our holding co. And then what you can do is you can generate a risk report as well as run your, like, loan eligibility. So this is also using Gemini. I think down here, it kinda has, like, the idea of, like, whether we'd get approved or not and kind of what the status is, etcetera.Obviously, we'd use, like, out outside lending tools, but that's a little bit there. And then I have like an overall AI insight, so you can just kinda like get some overall insights of the business overall. I don't know if this one's working that well, but this is kind of the idea. So you get kind of these insights on your business and a little bit of a mix of all of these things in one single dashboard. So, alright.I got another thing to show you, but I'm not let's cover this one first, andHooman Radfar: Wait, there's more.Jordan Metzner: There's Yeah. Yeah. Okay. Let's go to the next one. Let's go.But wait, there's more. Alright. So, you know, I made a collective song called Make It Rain, which obviously was awesome. And so I figured what else would be, you know, a nice fitting built this week application, but a make rain game. So I made a make it rain game, and you can just start by making it rain.And I think you guys will see how it works. Thanks, Molina. I got some music in the background. This is the OJ. This is So you gotta collect all the money as it makes it rain.I could a little bit. Can we get some cash money? And so, yeah, you gotta you gotta collect all the money.Hooman Radfar: What'd you use to do this?Jordan Metzner: This one I built in Bolt using ThreeJS. Yeah. And, yeah, just like a very simple game type. It's got a clock and a timer and like your best score, etcetera. And then I didn't have the Make It Rain song I made for you on YouTube yet, so I just used the OJ's as, like, a placeholder.I don't wanna get, like, locked out of YouTube or anything for copyright. But, yeah, it'd be great to put your Make It Rain song in here. I got eleven seconds left. You'll see it have like a little timer. Okay.My top score was oh, I I think I'm gonna get top score. Okay. Cool. And then I'll get my top score. Let's see.Alright. Top score, eight thousand nine fifty. Boom. Boom. And then, you know, the real way to master cash flow is with Collective, so that's that's the real That's right.Way to do it. So, alright, man. That's it. That's what I've built for this week.Hooman Radfar: Amazing. For everyone watching, Jordan sold me less than what he delivered. So he's like, I'm gonna pick one. And he's like, I'm just gonna do five. Just kidding.Plus, plus, plus. And I'm gonna make a song, and I'm gonna make a game, and I'm gonna perform the hallway. So, yeah, I'll I'll deliver. It's really cool. I mean, it's unbelievable what you can do quickly with these tools.I think people still are even on my team in the world, they're still, like, not fully grasping how fast you can get up to prototype. And I think, I mean, people like you might be able go from prototype to actual production. I think that's gonna be the big last mile because, you know, you look at tools like I don't know how much you guys are using it. Like, Lovable, for example, one of my favorite complaints from all my fellow founders is, yeah, it's great for setting up ideas, but then moving into production is hard. And so there's chains that you can fuse to your point.Like, you move over to cursor. You move over old stuff. And so the question is is, you know, can people do that or not? Are we just are these gonna be a lot of toy factories for for some time?Jordan Metzner: Yeah. Don't know if you've used Replit, but Replit has a great kind of like end to end getting like applications from idea and chat straight up, you know, into the browser that actually works. But, I think I think we'll see Bolt and Lovable probably build something similar, or is it just simply out of necessity?Hooman Radfar: And Figma is starting to get into, you know, the space too. So I think every everyone has an angle. Right? You wanna start with design first. You wanna start with a concept first.So everyone will have a role to play, and the raceJordan Metzner: is gone. For sure. And, I mean, of course, like, you know, Google is not gonna stop with Firebase, and Amazon's not gonna stop with Amplify, and, you know, so, I mean, it's gonna come from all angles, and everyone's competing for the same kind of thing. I I think we're gonna see, you know, multiple winners, you know, obviously, Versal's with o v dot zero. So, you know, we'll see, you know, all different players.But, yeah, I hope this was fun and, like, helpful for the collective community and the collective team on kind of, like, some ideas that could be cool coming in the future. And, I'm not a collective user, so, again, that's kind of my preface here.Sam Nadler: Yeah. I mean, just to transition, let's I would love to hear more about the collective platform and, you know, from you as a CEO, you know, how is your team thinking about AI, both, you know, just for their internal work and as potential features for the collective community?Hooman Radfar: Yeah. So, you know, I was, like, the ladder up from, like, our purpose. And as I mentioned, we there's 38% of the workforce that are solopreneurs in The United States, and that same phenomenon is occurring worldwide. There's $1,400,000,000,000 revenue generated by solopreneurs. I mean, it's huge.Right? And it's growing. And AI actually is a massive, massive tailwind on that. Right? Because individuals and we saw it with, like, base 44, but there's this whole, like, race to the first billion dollar business of one.Right? I do wanna point out for the record, and I was fond of doing this. We said it before Sam. So and it's documented. But, anyway, if if only you were paid for being first So in we wanna make sure we that we free that community to do what they they love to do.And where we saw the gap was a lot of solopreneurs were stuck between two different types of options. Either they had to chain together a bunch of DIY tools, right, so QuickBooks, Gusto, LegalZoom, so on and so forth, or, and this is the dominant solution, you know, in the market, go work with service providers, you know, accounting and tax firms. And so we said, let's focus very much on those 27,000,000 Americans. Let's build them one place for the first time, a single application for the first time, a subscription price. Right?You don't just get some bill, and you have no idea what's going on. And let's manage across all these tools. Because to the business owner, it's not bookkeeping. It's not tax. It's like, I need you to manage my finances.They're all connected. I don't even need to I don't wanna know how they're connected, actually. And the analogy I often use for folks that are in the venture backed world, we're like we wanna be like an office as CFO. So when I go to my CFO and say, hey. Like, where are we on our taxes?I just assume they're going to get our books closed every month and worry about our cash flow, get our taxes filed. We wanna put that in the pocket of 27,000,000 Americans. And today, we've been very much focused on this concept of higher and income solopreneurs. Right? So 5,500,000 of them make over 100,000 a year.And all things being equal, if you do make that, there's this tax election that's called an s election. And a lot of people actually LA, Hollywood types, Ashton Kutcher had an s corp and one of our investors. All those types of people have that. It's quite it's not a new thing. It's just to do that, you know, you have to form an entity like an LLC, s select it, and then you have to do your books, your payroll, quarterly taxes.There's a whole set of things you have to do, but on average, we're saving our members $10,000 a year. And so when you go to someone and say, hey. Look. I'm gonna do all these things that you have to do anyway. I'm gonna do them for a fixed price.I'm gonna charge you $3.40 in a month, and I'm gonna save you $10,000 a year. To us, that's like a really fair, complete package, and we're excited. And there's all these things that we wanna do to extend it both in terms of adding new services, making it more accessible and democratizing it so that people who aren't necessarily high income can access some of the services, you know, in different ways. So, yeah, it's it I can work on this for the remainder of my career, and I won't be done.Jordan Metzner: That's awesome. And just one quick question before we run into the news and, you know, simply because, obviously, this is the core of our podcast. But how has AI had an impact both, you know, on on your company, on Collective, both internally, of, like, inside your organization, as well as kind of for your clients, for these for these solo entrepreneurs?Hooman Radfar: We look at AI for our customers, we call them members, in two different ways. The first is internally, and that was how we started. Okay? So people want to have efficient service. They wanna pay less.Okay? They want it to be bundled. In order for us to realize that member promise, can we use AI to deliver consistency in the service? Right? So instead of having a person know, doing bookkeeping, for example, and specifically account categorization or reconciliation, can we have an agentic approach to it?Can we have an automated approach to it so that every time you're getting a consistent result, but that also means we can do it more quickly. And so we will often, we are looking at each specific flow. We would benchmark it. We'd look at how much time, how much expense it has, and then try to we're trying to innovate in each of those flows kind of in rank order. I mean, just it's a very methodical process.And, you know, the hopeful outcome for the member is that they're seeing a more consistent and quicker experience. Right? And then then they get that price, you know, that lower price. And, look, who knows? With AI, there's could be deflationary pressure.I'd love to lower the prices, to be candid. Right? Amazon style. Now the next level is how do you build AI into the user experience, which we started to do. So we replaced our help center with an enabled tool for basic questions about application, about tax, and whatever, but we're extending it.And now we're starting to roll out features like you can ask about your account. Like, I'm a member of Collective, by the way, and that's real. I can ask what my payroll was. What did I spend last time? You know, things like, how much do I owe for my quarterly taxes coming up?Rather than finding the right place to interface, I can any anything that has an endpoint within the application, we're starting with that. But the natural goal is, again, how do I get a CFO on this? Right? And so that's what we're building towards. So internally, we're we've been moving to an agentic infrastructure for, I would say, more efficiency, consistency, low variance service.But that same infrastructure is gonna help power. So when you have your super agent that's accessible to the customer, they can query and say, okay. Go ask the accounting agent what the books are. Go ask the agent that we wrapped around Gusto. Gusto doesn't have MCP, by the way, for those that are out there.So we we have to build an internal gateway. And so that that's that's how it's gonna work. But imagine, fast forward, you know, in the future, what I see collective as is it is your collection or your collective of agents. Right? So, like, if I'm a solopreneur and I wanna spin up a business, I set up the core collective system that does my basics.But say, for example, I wanna go get insurance, like the example that you had built, or I wanna go get something I could go in to a chat and just say, hey. You know, like, what insurance is am I missing? It could tell me. I could say, okay. Go find me the best quotes, and then it could get it, and it would install the agent.And so Effectively Collective is a tool to discover, provision, and securely provide access to your data to these agents, just like you would an employee. Right? And then when you're done with it or you move to another solution, you can decommission that agent. Make sure you restrict that their access. So think just like Shopify and Salesforce and other companies created provision marketplaces to applications.So we would provide, you know, provision access to these agentic applications and more and more. And and we'll see where the world goes. So that's like the big vision for us. So it's it's everyone you would wanna hire on your financial team, and we'll give the open market a shot to to do different things. Right?Like, because we can't cover everything, obviously. I'd love to, but that's not gonna happen.Jordan Metzner: Yeah. That's awesome. Yeah. I think about pricing and how difficult it is for small businesses to, like, you know, find the best price for pretty much anything. Legal.Hooman Radfar: Mhmm. Contracts. I mean, you pick an but, like, imagine you as a developer, and you guys are entrepreneurs, and you get this. I can go and you say, hey. Look.I will provide you a zero CAC distribution mechanism where you can go and acquire these users, which are incredibly expensive, by the way. This has been the challenge in the space. And you'll have access. You can with permission, obviously, you can get their tax information, their financial information, their PII, everything they can permission to you that you would have had to do flows that are very low conversion in one second. What application would you build?What agent? I could go to I bet you better.com, for example, would wanna do mortgages. Because what is the information you need to underwrite a mortgage? Go give me your w two. Go give me your k one.Go give me your pro I have it. Let me have access to your bank. I have it. What else do you need? And so we really can enable a whole level of, like, financial service distribution in a way that I don't think anyone's ever seen before because people are expensive to acquire, not just because of the CAC, but also because of the conversion because you're collecting all of this really tough to get information.Jordan Metzner: Okay. That's awesome. We're gonna move on towards the news. But before we do that, how do people find out about you or find out about Collective?Hooman Radfar: I am a very well SEO person because of the akis of my name. So early in my career, difficult, but now it's it's good in my life. So just Google Hooman Radfar. I'm at Hooman Radfar on most social media, and I'm Hooman at Collective. If you ever wanna email me, just go to Collective dot com if if you wanna learn moreJingle: about whatJordan Metzner: we're working on. Awesome. Okay. Cool. Alright.We're gonna transition into the a little bit of AI news, and then we'll wrap it up. You know, this story came out yesterday, but definitely, like, top story right now is just how much absolute amount of capital is going into AI. So, you know, NVIDIA stock jumped as they announced a $100,000,000,000 investment in OpenAI. Obviously, there's news in Oracle raising capital and them getting money and them investing. So it seems to be happening everywhere.But, you know, I guess this, to me, means this is just the early days, simply just of how long it takes to deploy this capital and build these data centers. You know? I saw Sam Altman on CNBC today, and he was standing in a dirt lot. So it seems like he's far away from getting those chips online. I don't know.What are your guys' thoughts as kinda where we are in the early innings of AI here?Hooman Radfar: Happy to to start. I think first question is where are we in the AI trajectory? I think we're early, but I think, as most people would hypothesize, the slope is increasing quickly. I think the other thing that's interesting here, though so there's, like, these two schools of thought, and I'll simplify. One group thinks this wave is the last wave, if you will.Like, we were gonna get we're gonna get to AGI superintelligence with this wave. The investment trajectory you see here, the willingness to invest in loss of the companies, all of that stuff, it's kind of with that hypothesis. Like, it doesn't matter because whoever wins this wins. Right? The second group thinks it might take longer, and definitionally, there'll be a cycle, which means this cycle will bust.It doesn't mean AI is broken. It doesn't you know, not, like, negative on all those stuff. It's just there will be a contraction because there's gonna be an overinvestment, and, you know, you're gonna hit that classic trough of disillusionment, all those things. Like, enterprises right now, they're trying to apply AI. They're spending a ton of money.Let's see if they can operationalize it. Let's see if it all works right right now. And that's that's an if statement. Okay? And so I'm personally in the latter camp.I think I don't think we're gonna get superintelligence in the next, like, two to five years. Open to being wrong, obviously. I I'll caveat it with, I don't know. This is the the weirdestJordan Metzner: Most difficultHooman Radfar: thing to call in in my career. So but that if you make me bet, I think that's what I'm betting. Part of it is around the structure of these deals, to be frank, and how the market's structured, and I'm happy to go into it. But I think this is it's like Jenga, and I can walk through that. But, Sam, I'm curious to see what you think.Sam Nadler: Totally agree with you. I'm I'm more in the latter camp. But, you know, I I do wanna hear your thoughts on the structure of this deal because I think it's a little bit unusual. And, you know, previously being, you know, a VC and an investor, what are your what's your take on the structure?Hooman Radfar: Well, let me take a step back, and I think a lot of people know this. Right now, the stock market is dominated by seven companies, and those seven companies are very much tray tracking along one company, specifically like NVIDIA. And so but NVIDIA's market, they're very concentrated. I think on the magnitude of 37%, and I may be incorrect on this exact amount, but I think it's some somewhere close to that, is, like, two companies, which they can't disclose. But they're probably in the max seven, because who has that much cash?Right? And so what people are doing is they're they're in this chip war, and they're buying and they're buying, again, with this assumption that this might be the last time. Okay? Now what happens if Google builds their own chip? And what happens if they're the customer and they pull back on NVIDIA, and NVIDIA misses by 20%?What does that look like? And then you compound that with some of these they're like round tripping deals. By the way, for people who are old Internet people, this didn't end well. AOL was the king of round tripping. So what what what do I mean by that?There were these deals where AOL would go, and I don't know if people know this, AOL owned part of Google. AOL owned part of all of these different companies, Yahoo, Amazon. And they'd come in and say, okay, you're gonna give us equity. We can invest, but you're just gonna give us the money back because we're gonna go like, a lot of it because we're gonna we're gonna drive traffic. And then everyone wins because you go public and your stock price goes up because you're growing, and we're growing, and everyone wins.Right? Apparently, know, DOJ and others didn't look too fondly on that. It's not dissimilar to what you're seeing here. Nvidia, for example, great company, and I'm sure they diligence this, so I wanna disclaimer, disclaimer, disclaimer. But they're going out and investing in the startup ecosystem.So they put 50,000,000, 100,000,000 into a company, but then they okay. And this company's buying a bunch of NVIDIA chips. And I'm sure, like, know, you look at Horweave. You're seeing permutations of that deal. Right?Where Sam goes out and says, I'm gonna do get 300 how are you gonna pay $300,000,000,000 to Oracle, you know, over the next n years? Like, it was a five or or something like that?Jordan Metzner: The numbers are crazy. I mean, Facebook's promise at the White House last week. I mean, it's everywhere. Right? It's everywhere.It's crazy.Hooman Radfar: Okay. Great. So you're gonna do that deal. Right? And then it happens that NVIDIA is gonna invest.So NVIDIA is gonna put money in for data centers. Who's who's building those? Who's gonna be it's Oracle. Right? Like, who's gonna be part of that?Not just the infra, but, like, you know, some of those. So my point is there's a lot of shuffling of capital, equity on one person's balance sheet, revenue on another balance sheet, and it's highly concentrated. And that dynamic, it's very susceptible to change. So if there's anything that slows down in this market, anything that stops it, I think you're gonna see a, like, a big it's not gonna be like a a slow stop. It's gonna be like hitting a wall.That's what I meant to him by kind of the the challenge. And, I mean, the startup ecosystem is, I mean, gonna be even worse. Like, you can take any company. Some of these companies have negative gross margins or AI companies have 50% gross margins. Can you imagine a SaaS company raising at the valuations they're raising at 40% gross margin?It's a joke. I mean, Jordan, you know how much how much pressure were the margins for you in in the company that we worked on. I'm just like, it's not gonna last. At some point, they're gonna be like, I want cash flow.Jordan Metzner: Okay. Well, let me let me just present an alternative here, though, just on this one argument here. And, you know, and I and I don't know if I'm right or wrong here, but, you know, if we just take collective as an example and you try to extrapolate the amount of AI you've applied into your business, and then you say, okay. Well, like, how many, you know, how many NVIDIA servers am I using right now to support these things. Right?You've got your customer service bot. You've got your, you know, a few other small things. You've got your your your your categorization bot that's going. Right? So let's say you're using like one or two NVIDIA servers.Right? As you start to implement some of these additional servers, we're gonna see that your consumption of NVIDIA servers, especially as your business grows on top of that, is gonna be maybe five a month, or 10 a month, or even 20, or 30 a month. Right? And now extrapolate that to all these businesses who are currently playing maybe zero AI inside their organization and will start to apply some AI. And even if they're slow, I mean, like, the largest of companies.Right? Even if they start to apply and they go from, you know, zero NVIDIA, you know, chips concurrently running to one or two to three or four. Right? That times the scale of all of these businesses could be massive. And I think to your point about how many small businesses there are, if each of those is using, you know, a half a chip a month or a chip a month.Right? How how big does this become? And, you know, my final point on this is I think NVIDIA just announced some new, like, video chip, and, you know, we've seen how good, like, you know, VEO three is and things like that. But, obviously, none of it is good enough to make, like, production quality television at scale. You know, it it's getting close.It's getting close. But you can imagine that a chip that comes out in two years and how much content creation that could generate, and then how much content consumption that can generate. And so, like, everyone becomes a content producer. And so, you know, maybe you see downstream that, like, actually, NVIDIA is, like, actually undervalued. And that maybe, you know, Google and these other guys can never catch up, and they can never build a chip that's that will be as good.And these guys will just own and dominate the market almost the same way that Apple dominated the phone market in The United States. Right?Hooman Radfar: Well, I mean, the the I'll give you two two counterpoints to think about. Everyone said the same thing about Intel, and then what happened? Right? It does get commoditized. It does fall if they miss a couple bumps.So I don't think that they have a completely career path. Now you have an amazing leader in Jensen, very forward thinking. I I I don't wanna say that and then I don't know if the space is the same. Right? So that's one one area.My argument is not that I think you're right. There will be more consumption of AI. NVIDIA can be more important. Right? So I'm not I'm not betting against NVIDIA.My point is is the way that the valuations are set up and the game is being set up, it's assuming perfect model. So, like, I'll give you an example. And and I don't mean any like, you know, to pick on, like, Cursor, for example. Like, there are there's any of those com there's some of these companies that have their either a negative gross margin or very low gross margin. Right?And I I had a conversation with an investor this week, and they said the AI companies just fundamentally can grow faster. There's a new way of growth. Right? And I I disagree. I think some of the products are so unique.Like, ChatGPT. Phenomenal. Like, brand new step function change. There's a couple of those. Okay?But, like, some of these other products, at the end of the day, and I mean, you sell this with the on demand ecosystem. If I offer you Sam a dollar for 50¢, you're gonna be like, alright. Sure. I'll take that deal. Right?That's what they're doing. They're literally they're the reason I'm saying a dollar for 50¢ is they're just saying, well, at some point, this compute is gonna be cheaper, and I'm gonna keep getting venture dollars until that happens. So there's a certain slope of costs that they're assuming. Okay? And it's very aggressive.Secondly, they're all built on the same couple foundation model of companies who also will have the same pressure. So it's not that actually they're giving you, you know, just the the dollar for 50¢. They're giving you an even cheaper because they're giving for 25¢ because they're not paying OpenAI enough. They're not paying Anthropic enough. Because Anthropic and those guys are getting money from The Middle East that, like, in bucket loads and other other companies.So they're like, yeah. It's okay if we're at whatever gross margin, 50%. Right? Now what happens when OpenAI goes public and the and, like, they have margin pressure? And then now they're win they're won.They've won. What would they usually do? They will start charging more. They will extend their dominance and say, okay. I wanna have 80% margin.I wanna be a cash machine.Jordan Metzner: But the switching costs here are zero. All of these have zero switching costs. That's why that's why Cursor is such a dangerous position. That's how they grew so fast, and that's how they they can fall so fast. And same with OpenAI.Right? Same with all of them.Hooman Radfar: You're kinda jumping on my argument here. So my point is actually now let's just do this this this is why I'm saying if a crunch happens, OpenAI goes back to Cursor and says, Cursor, I'm charging more. Okay. Is Cursor going to make the product more expensive, or they gonna eat the margins? Well, I imagine the pressure that goes to OpenAI that year is gonna go to Cursor's board, and their board's gonna say we need to get to 80% gross margins.Okay. Now if I told you, Jordan and Sam, who you if you use Cursor, a great tool. I have no problem with it, obviously, that it's two times more expensive. Would you do you think that it's gonna grow at the same rate as if it was two times as expensive? My argument is no.Jordan Metzner: Oh, of course not.Hooman Radfar: Oh, okay. So this is my argument to VCs. I'm like, you're not experiencing a phenomenon that's like this brand new thing with AI for most products. Some of them are step function, different pro I'm I really distinguish between those. I'm just saying you're literally selling a dollar for 25¢.You're asking and you're acting like it's a miracle that people are coming and taking you from you. No shit. Like, I I would do that deal all day. We've seen it in all the on demand ways. HomeDrew was a class example and all these other things.So I believe in the AI stuff. I think it's we're on track. I think, Jordan, what you said about the NVIDIA chips and all that is is totally relevant. If you believe, though, we're you're in the second bucket, not in the first, where this is the last wave, and but there might be an like, it's not it's gonna take longer, there is going to be a dislocation because of the these dynamics that have been set up where there's so many companies that are these low margins, and, basically, they're all depending on this stack of seven, if you will, to keep these round tripping deals working.Jordan Metzner: I I don't disagree with you, and that's where the question is. It's like, where in the stack is there permanence? Whether that's, like, at the NVIDIA level, even maybe a level higher of Or isHooman Radfar: it By the way, they're gonna be fine. Yeah. So They're gonna be I'm not suggesting look. This thing's gonna like, what I'm telling you is, those big company like, OpenAI, maybe they will experience some stock compression at some point in the value. That's fine.That's what happened to Amazon. That's what happened to all these companies, and they became huge. What it's really gonna do is crunch the startup ecosystem, crunch it's like musical chairs. When the music stops, it's gonna crunch the weakest. These guys who are really powerful, for the most part, they probably have hit the escape velocity, they'll be okay.Like, NVIDIA will be fine. So because the AI trends that you're saying are fine. The fundamentals are fine. It's just it's gonna take long.Jordan Metzner: Yeah. I I would just say, like, just to wrap up that point, you know, in the last, like, twelve weeks, my engineering kings have moved from using cursor exclusively with, you know, anthropic models to using cloud code exclusively to using now OpenAI exclusively. And, you know, what I mentioned to them is like, look, OpenAI is obviously subsidizing compute here and offering us, you know, this codex tool at a subsidized price. And we should take advantage of that. It's free dollars.And so what are we doing? We're maxing out the codex usage because it's cheap as hell and it's really good. And, you know, again, you know, to your point, Hooman, like, I don't know if that will create enough dominance to force all the others out of the market, but if there's a new sexier, cheaper model out tomorrow, we'll switch to that one. You know? We got no loyalty.SoHooman Radfar: And and look at all it's the same thing for some of these, you know, I don't know, like, cogen platforms that are, I would say, more user facing, like, Lovable, Replit. I'm not they could be fine. Like, the Replicios, Gangster, and all these guys are good. So they have to hit this escape velocity where they move from everyone's trying it, everyone's using it because there's four or five that are working, and they're prototyping to, okay, this is actually part of my stack. Right?I'm stuck. It's in there. And I don't know a single one of them yet that is an irreplaceable part of any of my company stacks that I've seen. Everyone's using them, so they had the right they started there. And so that's that's to the point there.For for what we have, and the good news, bad news for Collective style companies, we're in an area that's not in the crosshairs. Like, OpenAI, Google, and what have you, they're they're they're fighting a different battle. Right? They're finding fighting the foundational battles. And code, for example, they have to fight.Because whoever wins the code battle kind of gets you have that, like, looping, if you will, where I have better coding tools, which means I can make my, you know, better, faster. Eventually, it's coding itself, Devon style, and, like, you can't lose that battle. That is, an existential battle. But on our case, we have a huge TAM. You know?$100,000,000,000 plus, but Intuit would be the incumbent. Now you know what I mean? Like, okay. That feels fair. I I can deal with that.You know? Not to not to put down into it. It's it's a strong company, but that's not like, I'm going up against Sam Altman in the thing he said he's gonna do. He said he's gonna own coding and all these guys, like, that's that's a tough battle.Jordan Metzner: Well, I think that's your I mean, honestly, just wrapping up on Collective, that's a huge competitive advantage for you guys is being able to integrate AI while big companies like Intuit will be super slow to integrate. And that, you know, that will just make your platform so much better, and I I think allow you to bring, you know, more more customers over.Hooman Radfar: We are like to use a political analogy, I think if you look at one of the reasons The United States succeeded, part of it, obviously, is I think it's an innovative country, and it's great, but we were surrounded by the two oceans. And when World War one happened, when World War two happened, it didn't hit us, and so we were the only supplier left standing. And so everyone that we watch is going to war, and there's this massive battle, and we're just sitting there quietly building up something for a group of people that has been classically very overlooked. Even now, as much as it's sexy and people talk about billion dollar business one, how many venture backed companies do you see in my space? They're not.They're all like, I'm gonna build an agent to do this. Like, I've seen, like, 50,000 secretary agents and 50,000 video editing agents and all this stuff. Those are tough battles. They're worth it. I'm not suggesting I wouldn't fight as an entrepreneur, but I'm not fighting that battle.Jordan Metzner: Awesome, man. Well, this was an awesome, awesome episode. Love to show you off Collective and just for you to talk about your opinion here on on the thoughts on AI and the market. Sam, maybe you wanna wrap it up?Sam Nadler: Yeah. Great episode. Thanks for joining. And as always, we have new episodes every Friday covering the latest in AI tech and something we've built. And thanks for joining, Hooman.Hooman Radfar: Thanks, guys, for having me. Was fun. And Jordan, way to go, man. You you never cease to amaze me.Jordan Metzner: Thanks, man.
.jpg)
.jpg)
