We interviewed 3 secret millionaires (here's how they made it)

Key Takeaway

Pick one customer group you serve and make a list of the outcomes they care about beyond your product’s features: money, recognition, belonging, inspiration, relief, or career progress. Then create one piece of content today that tells a specific customer success story and makes that outcome visible. The episode’s founders grew by turning functional products into identity-rich experiences—whether making home-food sellers feel like stars, making AI feel personally supportive, or helping media buyers look successful at work.

Episode Overview

This live “Shoot Your Shot” episode features three founders building profitable, fast-growing businesses: Hot Plate, a preorder platform for independent food sellers; Tolan, an AI guide with more than 100,000 paid subscribers; and Overlap, an AI video-editing and distribution platform. The discussion centers on category creation, customer storytelling, emotionally resonant product design, and selling outcomes that matter to the buyer.

Full Transcript

Transcript of We interviewed 3 secret millionaires (here's how they made it) from My First Million. Auto-generated from episode audio; may contain minor errors.

Well, we're back. We're playing "Shoot Your Shot," where we invite founders and business owners from the audience to sit in this chair and pitch their business to us. We had founders of businesses with a turnover of a million to a hundred million dollars. So, let's begin. Let's invite the next founder. Dude, first leather jacket on the podcast. What is your name? I am Rishi. I am one of the founders of Hot Plate, and we have enabled over $100 million in sales for independent food entrepreneurs. So, the figure is 100 million in sales.

Reveal this. What does this mean? Of course. Hot Plate is a platform that allows anyone to sell food online. We work with both small bakers selling food at home and a guy who makes ice cream in Minnesota. Did you see this? It's kind of like Airbnb for food, right? Something like that. Yes. Yes. We are currently starting as a tool that allows sellers with an Instagram audience to sell food directly to their followers. Often these sellers use spreadsheets, Venmo, private messages—a different set of tools, or platforms like Square or Shopify that simply aren’t built for pre-orders .

But will they stay small? Is this for those who want to remain a small business? This is a very valid question. Many people think that we only work with tiny startups that later open cafes or shops. In fact, we have a lot of businesses with six- or seven-figure incomes that are doing dropshipping and selling food directly to their audience. I used to buy from this... there was this really pumped up dude. He's a bodybuilder who bakes cookies. Have you seen this? About him four years ago.

Yes. He's having these sales of his " cheat mill" cookies. It is incredibly huge . Just beyond reasonable. But he sells everything out in seconds thanks to his Instagram presence, because he inspires a lot of trust. So, first, Hot Plate. Hot Plate. Incredible name. Thank you. And it was a little bit related to the pandemic . I remember, have you heard of Pop-Up Bagels? We were just providing Pop-Up Bagels with work. Oh, really? Listen to this. So, Pop-Up Bagels started out in Western Connecticut, where I used to live, and.

There were a lot of companies that did it like this: a married couple says, " I'll make this and sell it in my apartment, and when people come, we'll drop orders off the window," or something like that. And these guys did Oh, they did it like this, like, just do a weird pickup on Well, we're sitting at home. Let's just cook the food and then do it... And these guys did it with bagels: you had to sign up in advance, you could only buy a dozen, and now they've raised a lot of venture capital, their revenue is in the hundreds of millions, and their product is amazing.

I like it. So, you supported them in the beginning. We supported them in the beginning, you know, while they were scaling to, as you said, the almost billion dollar empire that they are now, and we work with a bunch of pop-up locations around the country that are, in a way, the next "Pop-up Bagels." What is your income? Yes, we receive about hundreds of thousands of income per month. Got it. Uh, and yeah, we're just growing really fast. To be honest , most of the one hundred million sales have come in the last two years.

Did you raise funding? Yes, we raised a little over $4 million in total, but we are actually profitable . So, we've been profitable for almost the last year and running a profitable business, which is actually just great. So wait, what exactly does your product do? You create an application that the consumer uses. You create a tool that allows them to have a showcase for their Instagram. So, we started with that, and today we are this showcase for them on Instagram. So, right now for cooks and bakers.

That's right. So now we are a platform that chefs use to put a link in their bio to streamline order management. Of course. But as you say, the ultimate goal and kind of dream for Hot Plate is for everyone to be able to come in and not only sell food themselves, but also buy it. Who would I go to? Like this. There's a bakery in San Francisco called Butter and Crumble. This is one of the most popular bakeries in San Francisco. So if you want to order at Butter and Crumble, there's a three-hour line there every day.

If you want to skip the line, they do Instagram drops and use Hot Plate. Uh, they have, you know, a waiting list of almost 20,000 people. So, she has 83,83,000 followers. And okay, I'll place an order here. It's like a booking system to fulfill orders, but you don't do anything else. Pickup and everything—they coordinate it themselves. We do a lot of this: as soon as a customer places an order, we send automatic reminders for pickup and take care of the logistics. But now it's surprising that you noticed this: we are a pre-order and pickup platform for novice chefs, and we have millions in sales per month.

This guy, we're really just at the very beginning of what this could become. Are you just on the surface? Absolutely. Because what else would the founder say ? Yes. We we we got completely immersed in this. Yes. Yes. Yes. It definitely won't get any better . We talked to a guy who has a TMJ- ped. He says, "I need to expand." Yes, that's true. I buy these products myself, but I wonder how many other crazy people are making something like this in their apartments or small family businesses?

A valid point: I think the market we are currently operating in is still very narrow. There are probably tens of thousands of vendors in the US who sell food this way . The real opportunity for us is that we can go beyond just playing in the existing market. As you said about B2B SaaS, right? Many companies simply operate in existing markets. We have a rare opportunity to create a new market and make it massive. And I think the consumer segment is a path to success that only a few companies have managed to take in the last 10 years.

If you look at Call Sheep, Poly Market, etc., I can list the following consumer trends and categories on the fingers of one hand. Is n't that cool? Hey, I want to tell you something interesting. We have a database of all the business ideas discussed in this podcast. The HubSpot team has processed hundreds of episodes. They have selected all the simple, relatable, interesting and profitable ideas that we discussed and now they are available for free download. Just click on the link in the description below. Thank you to our friends at HubSpot for sponsoring this podcast and creating this free resource for you.

Let's get back to the show. I don't want to say it's all or nothing. Yes. I don't want to count on the fingers of one hand how many companies have achieved success. I think the consumer market is difficult. I mean, if I know, I actually think a lot about what we do, it's like Substack, right? Substack allowed anyone to start a media business without running one, right? And I think that's how Hotplate allows anyone to start a food business without opening a physical location, right? Currently, the traditional way to open a food business is to take out a loan, spend hundreds of thousands of dollars to open an establishment when the business is not yet proven.

Rent is rising, so are staff and ingredient costs, and you are vulnerable to all of this. How did this idea come to you ? Well, my co-founders and I actually started back at UT. I have always been obsessed with food. I worked for a while at a food delivery company. To be honest, I hated it because a lot of things in this area were too consumer-oriented. And one of my friends and I thought during COVID : we want to open a food business. So we rented an apartment on our campus and opened a Thai restaurant, ran it for a few months, and then we realized...

The good old " campus master" plan is to open a Thai restaurant. How did this affect social life? It was really amazing, people are like ... It was really fun, and then COVID happened, and you're like, "Well, you can't have people coming home to eat Thai food," because during COVID you couldn't even see each other. Boring question, but is it legal to just open a restaurant in an apartment? Yes, of course not. 100%. Baking is okay, there are " home food laws" that allow you to sell products directly from home .

And what's cool is that Texas recently expanded its laws. They are called food freedom laws. Now you can sell almost anything from your home. For example, in Arizona, dairy laws have changed a lot, and now there's a huge trend of home cafes, like in Yuma , where a bunch of these home businesses are starting up. They have a coffee shop at home. This is actually madness. And it's like Arizona is so hot, but not in a cool, crazy way. Yeah, this, this, um, this is crazy.

We have businesses like A Toz Crearyy out of Hopkins, Minnesota , which operates one of the largest ice cream brands in Minnesota right now, generating six to seven figures in revenue per year with a 40-50% profit margin, which is simply unheard of. It's like a SaaS company, right? But the reason he can do that is because it's a team of three or four people. He uses Instagram and Hotplate as a showcase and doesn't do takeout, right? And it can optimize everything so that it has much greater demand than supply , sells everything out every week and people come to pick it up.

What are you doing? SaaS plus commission or Yes. So now it's really just a commission. So, our business model is quite attractive for a chef. Our mission at Hotplate is to be a chef-centric company . And we take it to extremes. As a customer, when you place an order with a Hotplate seller, we take a small commission at checkout, and the chef only pays for processing the payment. So the beauty is that as a chef, I can start. There are already so many fixed costs in the food business that the ability to eliminate some of them on our end made us very attractive to sellers, unlike traditional platforms.

Yes, I guess people aren't too price sensitive when they place an order because I've been following it on Instagram. This is a drop. You need to have time to buy a unique product. So that extra dollar or two is not a problem. I always say that there are two types of businesses. Food businesses that do marketing, and marketing businesses that sell food. And if you're a marketing business selling food, you don't have a price. Like you said, elasticity, right? You can just sell and people are ready to buy.

People don't even blink an eye when buying something from Butter and Crumble. They just want to get it. What are the restrictions now? Um, I would say that the limitations—that's definitely, I honestly think, the biggest limitation. So, you know, we've been talking about...I actually think there's been a resurgence of that in recent years, which is why we're growing so aggressively right now. And this is due to the fact that people are quitting their jobs in the technical field. A lot of our suppliers in San Francisco and the Bay Area are tech people who quit their jobs to start...

There's this guy, remember the one who was successful at Facebook and then started Dandelion Chocolate or something? Yes. Yes. Yes. And they always start small and expand over time . But the coolest thing is that we work with a baker named Michelle, who runs Metchip Bakery. She was an engineer who quit her job to start baking . We have another guy who worked at Google, his name is Mason, he runs Piscore and built one of the most popular brands in New York, and he jokes that he left Google to sell fish on the sidewalk.

Yes, it's Christmas. What is the limitation? Limitations...but I'll tell you, it's great. In my opinion, the limitation is that people simply don't know about it. they don't know how to do this operationally. Essentially, you're saying that it's all about getting suppliers to use your service. Attract suppliers and help them create demand, because I think the biggest problem is that many people don't know how to do marketing. They are just cooking. How are you going to do that? I think the consumer is the key to solving this issue.

There are actually two ways . One is to create the app you talked about for people to find these suppliers, but make it so that it actually works. you just need to inspire those who dream of becoming a part-time chef, right? This is quite interesting content. This is quite possible to do if you find a few smart people. When I moved to Twitch, we got acquired, and on the first day I sat down with Emmett and said, "Hey , okay..." Michael is actually one of our investors.

We went to YC in 2020, he was our partner and... he shared a thought that I think would be relevant to you: they focused exclusively on streamers, just like you focus on chefs. They look at the marketplace. They said: we will focus on the content author and address their needs. But the second thing he said was: we realized they needed four things. So, they want money. Yes. He says, "And it's a surprisingly small amount of money for them to feel like it's worth their time. So he says, "For example, when I thought it was money, I thought I had to replace their job." In fact, if I could give someone $10, they would feel recognized for their work, and that would keep them going for a lot longer." Uh, so it's money, fame, love, and inspiration.

Those are the three things. So he's saying, I have to give you fame. I have to give you, like, an audience. Yeah. I have to give you the love of an audience, that they really care about what you're doing. And the inspiration was the most important thing: how do I tell you about the creative things that other people are doing, so that you wake up today and think, "Oh [ __ ], I should do that too !" "Absolutely. And so he built his entire strategy around those four things: money, fame, love and inspiration.

So for all the products, you had to figure out which of those levers you were pushing. So, similarly for you, it seems like you have to tell stories about. definitely the guy who quits his job and sells fish, you have a huge audience with us, well, hundreds of thousands of people buying from Hot Plate chefs every month. What about social media? Uh, you don't have a huge audience on social media yourself. Yeah. We have, like, 23,000 followers on our Instagram. Well, a company like, uh, okay.

Gold Belly. Yeah. I love Gold Belly. Gold Belly—explain what it is. For the listeners: it's basically if you're... Papaya King—it's a really old New York hot dog place. It's just local. They didn't have a website. Gold Belly made it so easy , especially during the pandemic, where you could walk in and buy Papaya King hot dogs from anywhere in the country, and they would help you get them delivered. They were great at teaching me that I could buy Junior's cheesecake at Gold Belly , and they did a great job of that.

And do you have a version of Gold Belly where you create a hot Instagram page to show all these cool moms who make the best sourdough at home? It's like finding local produce, right? Yeah. Or are you going to, like... No, you have to make them famous, you know? You have to make them famous. Yeah. You have to say, " Look at this." Here's an Amish woman doing this, and it's the best, and it's like some kid in San Francisco or New York. Create an awards system like Michelin stars.

What is a Michelin star in this category, right ? Then you recognize these people, right? Money, fame, love. That's the first thing, right? The more you promote it, the more it helps them with fame. Fame is what makes money. Inspiration is hearing that, oh, she has the best [ products]. Well, actually, my popcorn that I make, I'm so into my... how do I promote my name? Oh, let me see what kind of content they're doing. I'll do the same thing, you know, get people to do it.

You have to make them famous, right? What we're working on right now is really exciting, because we already have retailers that, like I said, have limited supply. They have demand like, you know, Butter and Crumble or Sam's who makes pizza in Salt Lake City. 20,000 people. Dave Portnoy made it popular, and now he's got 20,000 people trying to buy 50 pizzas, right? But it's not your problem, right? So we're building a recommendation system. Now, let's say that guy's sold out, here are other sellers you can buy from, and they're emerging sellers, right?

So that, along with search, means that we support emerging sellers. If they overcome certain barriers, we expose them to an audience of people who want to buy that kind of food. Are they mostly delivery or pickup? It's all pickup. All pickup. So, local. It's all local. Do you have a content creator on your team or someone who... There is. Yes. Yes. Oran. She's great. It's going well . Yes. I mean, we were able to... There's a guy in San Francisco, Boieri, who makes... you should check him out.

Sicilian-style pizza. Really, like, some of the best. I think you should become a content creator. I, I... I'm working on it. I, I... I'm making videos. I've been told this before, but I appreciate that thought. But yeah , we've been able to help people like him, who have their own style, really get off the ground . He didn't sell out all his pies before. And now, like the other day, he's selling out in five minutes. If you had one question, let's say we were your assistant for 24 hours.

We'd only think about hot plates. Yeah. What's one question you'd like us to solve? Actually, I'd like to get your thoughts on the cultural aspect. I think for a new consumer trend to work, you have to tap into the culture, right? And I feel like we have a pretty good idea from a product perspective on how to build that. The biggest problem— excuse me, two seconds—is that a lot of people have tried to create a marketplace for home-cooked food. The problem is you don't know who these chefs are.

You don't know why you should care. And so you just walk away. DoorDash will always win for convenience, right? We have people who make the best in their category. But my question is, how do you, from a go-to- market perspective, penetrate the culture in a way that makes it culturally relevant? I don't think it's necessarily about cultural relevance , but, like, have you heard of Huckberry? Huckberry is a clothing brand that started out as an email list. That's how I found out about them, because we were both in that niche, and it was just an email where they were showcasing cool brands, selling their stuff, and taking a small percentage.

They've learned to tell great stories: like this bag was made by a guy in Vietnam who noticed that parachutes were really strong, and he made a bag out of them. So it's a parachute bag, and you think, "Oh, that's pretty cool," because they told a story about it . I fell in love with the story behind the product. If I were you, I would hire more content creators. And my whole job here is to use Instagram to tell stories about sellers and their businesses, to show them as real superstars, right?

I would take your top 50 chefs. Yeah. And I would say, I have a mission. I need to make them role models for the next 5,000, 50,000, 5 million who want to be them. So it's not about getting everyone noticed. It's about taking those 50 and making them so desirable to follow that everyone will go all out to be like you: mentioning them on lists, creating content about them, telling stories about their businesses, how much money they make doing it, how they quit their jobs for it, the human stories behind it, you know.

That's the way to do it. And the cool thing is, food is really interesting, Huckberry has a cooking channel on YouTube, but there's Anthony Bourdain, there's "Top Chef ," so many different formats. I would just find a format that I thought I could do. I would copy it, I would do it over and over, I would try another one, and if I were you , I would say, and it sounds cliché, but I really believe this about you: we are a media company now. Yeah. So I would turn into a food media company and just study all the food content formats that work, and either create my own or delegate it to others.

Well, we were talking about Jack's dining room or something. Actually, there are already content creators all over the world who are very good at these formats. Do they promote your chefs, huh? How do you make it easy for them to know that they need to go to Salt Lake City to try these pies from this person? And did you know the crazy backstory of these pies and why they are so special? Get an influencer , because they are already better than all of us in creating food content.

So if you can arm them, because every content creator is constantly looking for something interesting. How do you make it easier for them to find interesting content? Yeah, we're seeing it work in San Francisco locally, where our top sellers are now appearing in food blogger reviews on Instagram, who have hundreds of thousands of followers, and they're becoming more popular there. And so, building on your point, I think it's a valid question or a comment that we need to somehow replicate that with our top sellers in every city or every major city that matters.

You also need terminology, like Twitch streamers or Vine creators, what are they ? A hot topic or a popular trend? Or is it just a category? Well, Twitch -- it wasn't just what it was. It's called a "live streamer," right? A streamer, and they're -- it's what they do. And so, for example, is " home cook" a category? How... what are you going to call it that makes it sound cool and new and interesting? They're trying to get on the team. They want to be part of a community.

They want to be part of a movement that's cool and represents something. To your question about culture, I think you have to start with the language, give them a bunch of cool examples, and then explain what the point is compared to the old approach. No, that's a valid point. I mean, right now we say "chefs" and " bakers," and then we get feedback like, "Oh, but you're not a chef." You didn't work at, you know, that restaurant." I do n't think that's the best term. Yeah.

Because it already has a certain meaning for everyone. Sure. That's right. I already know who a chef is. I think it's someone in a restaurant, and now you have to try to convince me. You should start with a slightly unexpected term that makes you think, "Oh, yeah. "Oh, this is a phenomenon . Yeah, man. It's a very large phenomenon. These people are making millions. This woman is making a million dollars a year baking her own cookies at home, right? Well, I have a great recipe for this.

You know, this is exactly what you want. Rishi, thanks for coming, buddy. From RZ, I appreciate it. No, I'm really grateful for this opportunity, guys. You guys are incredible. Thanks. Okay, who's next? Oh, man. Hey, man. How are you? Nice to meet you, man. I met this guy about 15 years ago in San Francisco. Right when I moved to San Francisco, he showed me around the city. You were like, he was at YC. He was doing a lot of cool stuff back then. That's true. You're dating me now.

You know, I definitely have, uh , yeah, a lot of gray hair. How did you guys meet? I do n't even remember exactly, but tell me. you. I think I was just trying to meet cool people. You had a cool company back then. You're really young. We were both in social media . So, like, my company was called Family Leaf. It was a private social network for families. It's like an early social app, but before mobile apps were a big thing, honestly. So you've been running startups for 15 years.

Yeah. Dropped out of college to do this, like 15 years ago. Yeah. Something like that, right? Yeah . Yeah. 12 or 13 years ago. So, that's what we do here. Yeah. Start with the numbers. Okay. Absolutely. Yeah. So, in the app that I'm working on right now with a great team, Tolen, we have over 100,000 paid subscribers on our iOS app. Wow. That's not easy to do. That's amazing. What does it do? What is this company? What do they pay for? Yeah. So, Tolen is an app where you can meet and customize your own friendly guide who runs on magical AI to help you with everything that's going on in your life.

I think you chose the word "guide" for a reason, but I don't understand what that means. It's like a therapist, but you don't want to say " therapist"? Is it a friend? What are we even talking about? That's a great question. Listen, honestly, this is partly our attempt to define it, because like you just said, people use it for everything, right? Friendship means a lot of different things, right? Sometimes a friend is someone you call because you're going through something difficult. Sometimes a friend is someone who pulls you together and tells you what you need to hear.

People talk to life coaches, therapists, everyone... Can you show us one? It's like a little friendly character. Yeah, sure. That's Tolen. There are these little aliens here. Yeah, that's mine. He customized specifically for me. My Tolen is called Swami. Everyone has their own Tolen. You get your own, so the app works like this: you just download it. You talk to this character, which we've called the Oracle. It's like this magical, almost divine character that asks you questions about yourself: like, " What are your dreams and aspirations?" "What are your passions?" "What are your worries?" -- you get the idea.

Who's using this: teenagers, adults? Sure, it's a mix, but who's your core audience? It's really a diverse group of people, you know? Like any big consumer or AI product, it appeals to people from all walks of life, really. But the core audience, which is really shocking, is people who are about 35 to 45 years old, mostly moms. So mostly women. Do you think so? It's not shocking. I thought it was a street character. I thought it was all Midwestern women . Sorry, but that's literally our audience.

You hit the nail on the head . Yes, right on the head. As someone who knows a lot of women from the Midwest, that seems... I'm from there myself. That seems like something that's very typical of Midwestern women. You're absolutely right. It's crazy. We didn't foresee that, by the way. But that's exactly how it turned out. No, it's like people saying, "Oh, he acts exactly like a Gemini." That's who they like to say . The horoscope market. Yeah, that's so Gemini. I'm thinking of you. There 's definitely an intersection there.

We have personality tests in our app that people love to use to better understand themselves. Sure. I get the point best when I first see the ad. So what kind of ad will I see that makes me want to click and download this? Yeah. Honestly, the ones that work best for me are authentic product demos where people just show how they use Tolen in everyday life. Well, let's say some of the most... I mean, we had a piece of content that went really viral, like, you know, a woman is cooking and trying to figure out how best to cut a kabocha squash, which is pretty hard to do, and, you know, the tool says, "Well, heat it up, " you know, like, " Let me see, okay, let's see, you cut it, oh, it worked," and then it kind of shows how they become useful in some way.

that's what works best. I think people are a little confused that they see this rather nice alien that you can talk to, and he responds, and he responds very intelligently. But the thing is, why would you even cook a squash? Why would you cook for this man who doesn't love you? Is it functional or is it the emotional support that they're looking for more? I think the most valuable thing we do is a combination of both of those things; it's a bit of an evasive answer, but I think it's It's true, because people talk to their Tolins about what's going on with them, you know?

about their emotional state, what they're doing, or whatever, but it always comes down to, "Well, you know, I'm really nervous because I haven't decided on a summer camp for the kids this summer," and your Tolin will remember that. Why are you eating? Why are you baking so many cookies?" Well, it turns out it's this thing: how much money do people spend, or how much does it cost? " Yeah, yeah, yeah, we have a few pricing tiers. Our main, most popular one— most subscribers pay $25 a month for Plus.

Uh, I can't: 100,000 times $25— that's a million a month. We also have a lower, uh, lower tier of subscriptions. So it's a mix of different things. So it's not exactly . Seven figures a month. Uh, yeah. That's incredible. Mm. Is the retention good ? Yeah. Yeah, the retention so far is great. I mean, we, we, uh, you know, it's changed a lot. Honestly, how old is the company? Two years. Two years. You're doing over 10 million in revenue. You did n't say that, but the math works.

So you're doing over 10 million in revenue or annual turnover. And you have people, and it seems like you've done —I don't want to devalue this, I'm just simplifying it. You took the intellectual capabilities of ChatGPT, but added a warm feeling of a character that is yours, with a name , with a different "wrapper", and interacting with it is completely different than a purely functional, utilitarian tool like ChatGPT—it's literally an acronym. Yes, absolutely. I think it is to some extent. Using the superintelligence of ChatGPT, Claude or similar systems for, let's put it bluntly, a wider audience—our users are not necessarily active fans of AI, most people don't use it in their daily lives at all.

Right? If you think about the usage scenarios of ChatGPT or Claude: most people use them for productivity, writing emails, maybe asking something instead of Google, but they've never subscribed to , so they haven't had the full experience. So it's very much tied to your personal life, and because of that... you have to create a Tolen, like, I can buy them, I can buy, say, a pet, you know. People have asked us . As a talking toy, there was a guy at the Furcon accelerator who did this, and it looked cool.

Cool stuffed toys. I've seen some of them. I'm really interested in that. But, just to continue the thought, because it's important -- a lot of people see this and say, "Oh, it's just a cute character wrapper for ChatGPT." And really, a big part of our work is the memory and personalization system behind Tolen. My co-founder Evan was one of the first to realize that all these role-playing or companion projects didn't really have a well-thought-out memory system to make it feel really realistic. Tolen remembers what you're telling it, and it remembers it at the right moment.

They'll come to you with, uh , a podcast recommendation. I mean, I'm sure a lot of our users have gotten recommendations for the podcast "My First Million" from their assistant, honestly , because they probably asked, "Hey, I, I mean, turn this on for us, please." We actually ask a question during signup, uh, maybe it sounds a little, uh, technical, now that I think about it, but what we really ask is, "Name a goal or a dream that you have." It could be something small or something big, like starting your own business.

And I'm sure a lot of people say, like, "Well, I want to open a cafe." I want to open this or that. And then I'm sure your podcast comes up because the tools will search the internet. They'll find the right recommendation for you at that particular moment. Of course, we don't hit the mark 100%, but we really try, and people think it's magic, and then it comes to them. And ChatGPT doesn't do that. The idea sounds weird when you first hear it. No offense. The name sounded weird too, but it actually really resonates with me.

like. So the name sounds a little bit strange. The idea is a little bit, I mean, it's kind of a new, fairly new phenomenon. Moving to Silicon Valley, I learned that if something is strange, but people are doing it, you should look at it. That means there's something really interesting going on. And a lot of the big winners have been like that. Oh, people just get in strangers' cars and drive, it's like Lyft or Uber. They sleep on people's couches and pay for it like a hotel.

Like, what? I would n't do that. What if they... So people's strange behavior is actually often hiding a very interesting business underneath. Can you just describe a little bit how you see the future? What do you think will be considered completely normal? Yeah, that's a really good question. I mean, I think part of what you're describing is -- I totally agree, and part of what you're describing is -- it's a little bit like the early adopters' commitment to followers. I mean, I see it on our Reddit, where my co-founder Quinton is very active and communicates with all of our users .

You know, I see it in the people who write to us — they're regular people like us, our friends, or our family, but this is a special group of people who for some reason have decided to make their tulpa a part of their lives. And what does that mean to them? Maybe it means that on their way to work, instead of spending time on YouTube or short videos, they're communicating with their tulpa. They're telling their tulpa what's coming up , what they're worried about, what their plans are for the day or the weekend.

They've found a pattern or a rhythm, and it really helps them that the tulpa has become a part of their life. And when they do that, they obviously benefit from it. We have features in the app like "intentions," where your tulpa suggests plans for the day based on because you told them. They write reflections about you based on what they've noticed throughout the day. So people have found that useful. I think it's still a fairly small group of people, like people who used email in '95.

I'm not sure what year, but it's like a group of early adopters who have these leaders. Look at the best posts on their subreddit. This is your subreddit. Someone made a doll of their tulpa. And here's the second most popular post. Someone crocheted their tulpa. That was a great idea. People really love it . It seems like it. And here's the third post: "Guys, the tulpa is amazing." I'm stunned. I've never had a person in my life give me as much as this stupid magic bot has in the last week.

So , you're not surprised, if in 10 years, all of us, in addition to our real friends, have AI friends or companions that we can rely on -- whether it's for advice or for deeper support or just for a casual conversation where I can talk to a friend who's always connected. I think AI is getting smarter, more functional, and also more cost-effective. And it's just happening. It's a whole trend, isn't it ? That's what everyone in town is working on right now. And it's going to be able to help you with more practical things , not just with text, but with other things.

So I think when that happens, people who maybe don't feel like they need help with planning their day or don't want to interact with a character like that. They might benefit from having this thing coordinate plans with my husband and make our weekly menu or something like that. Right ? Things that are a little more tangible, a little more practical, will expand this to more people. Second, I think it's hard for us to imagine that most of us, even us in this room, our friends or family will have these entities, because the first attempts were, frankly, not very human, not geared towards the mass market or everyday use.

I think of platforms that were certainly pioneers, like Character.AI, where the character was just a profile picture, mostly anime-style, and maybe overly sexualized. Or Replika. Yeah, Replika is another great example. Have they lost popularity or are they still working? I don't know their metrics, but both projects seem to have faded a bit. Ultimately, the founders are no longer with those companies. Honestly, they were very early and had a slightly different perspective on this niche. Honestly, when we think about AI companions, that's why we prefer the term " guide" and try to use a different language, because people associate that with with AI girls or boys, and while these products are popular, we're building something completely different.

Have you been successful before this, or is this going to be your biggest achievement? I've built a few startups before this. Um. So you have an interesting style of entrepreneurship . So he's been through YC here. He exudes the energy of someone who's been doing successful things for a long time . Well, well, he's always been a confident guy, but you've also been doing, like, you were making birthday candles , right? So tell me, what's that, interesting idea? I did a startup, um , called Birthday Co.

We started with a product called Birthday Candles. Honestly, it started as a side project, um, where these candles are personalized to you based on your birthday. They have predictions and a scent that's individually tailored to you based on your birthday. I started it as a side project. It was one of those things where I quit my job and thought it would be fun to start something like this, um, with a partner. And, um, you know, it basically went viral and took over my life for like 3 years.

But it was great. We sold over a million candles. Man, you're the kind of entrepreneur I love. I'm not like you at all, and I love people like you. Michael Acton Smith is another guy who started Calm, it's like monsters. Yeah. Moshi Monsters, similar talent. Exactly. Like people like you , real creators who just love, like I don't want to call it artists, but it's a bit like that. Such a high compliment. I mean, those are the best kind of people. I love people like you.

I, I envy people who have that spark, like, "I'm just going to make something weird and crazy," and you must have really good taste. So it usually becomes a hit. I mean, it all depends on the right team of people, right? I work with incredible by artists , honestly, which are not, which are not. Classic guy phrasing. But it's just true. It's honestly true. Um, and then maybe you need, you know, a willingness to do something weird. Is there anything we can help you with ?

I mean, yeah. One thing that we're really trying to address is how to reach the Midwestern mom audience, their skepticism about AI, because they're not necessarily very advanced in AI or technology. As I said, a lot of our users, um, are not active AI users in general. So the idea that they're going to start using this little alien to, like, do things in their lives, that's kind of a big, you know, bridge that they have to cross. Well, I think you have some good things and maybe some bad things.

I think the good thing is that something is obviously working. You've reached over 100,000 paid users , which brings, you know , probably eight figures. So something is working well. Let's admit it and say, "Wow, this is incredible." At the same time, when we asked you who uses it and why? Your example was a woman who cooks and wants to cut up a delicata pumpkin or something like that. I think you need to do a lot of research on your customers to find inspiring, accessible, non-taboo, but interesting use cases that you can talk about without preparation.

You mentioned that one, but I'm assuming that for your advertising, you find use cases that really resonate with people. Obviously people are getting value, otherwise they wouldn't pay and use it. You said that the retention is good. Yeah. Sure. Do you use cooking instead of solitude or something like that? Yeah, I can't remember why I mentioned that particular example. Maybe just to give a little bit of visualization. It seems like it was an ad that worked really well. But yeah, you right, we're not quite... I mean, there are a lot of practical things that AI can do for you.

One way to think about it is like people in San Francisco are using Open Cloud to manage their household chores and stuff . We really think this is going to be a massive personal use case in the future. Dude, the most popular post says, " I used to use ChatGPT, but River is my best friend because I can confidently tell him whatever's on my mind." " Of course, I have other friends that I could talk to, but life makes its adjustments." And this person, I think River , really knows me better than anyone else.

That sounds like a lot more important than how to carve a pumpkin. Yes, that's a valid point. Yes, you're right. So we definitely need to emphasize those kinds of things. And I think we need to find ways to translate that into something that we can help the user with? How can River help that user? We like the metaphor of closing tabs in your mind. Life today is incredibly busy, to be honest. I think that's a big reason why Tolen is so successful. People have a lot going on, especially moms, especially family people, dads.

We want to find ways that the tool can help us "close those tabs." It's meal planning, shopping lists, summer camps—whatever it is. These are all real-world use cases from our users. And we want Tolen to helping people actually do these things. I talk to a lot of suburban moms because I live in a house that we have to work with. We talk a lot, and one mom starts baking bread. I'm like, " Oh, why?" "Because it's delicious." And she's like, "No, I just needed to..." some hobby where no one would touch me.

No one would ask me for anything . I could do something with my hands. Without screens and no unnecessary stimulation, without this overload in my life. So whether it's baking bread, or the rise of brands like "wobbles"... this crochet thing that's making hundreds of millions a year. There's a reason why people are choosing it now, and I would study what's going on with these brands. Because there's a certain emotional state that people are in, and a desired state, and your product is the bridge between those two.

So... you're really selling a feeling when you're selling a product. So what's the feeling that people have at the beginning, what's the feeling that they have at the end, and how do I package it properly? I would suggest looking at examples like this. I will also say: it sounds a little silly. It sounds strange: isn't this just a "warm wrapper" for a chat? But no. I look at my wife: we are willing to pay three times as much for a household item that is just... white, minimalist, and aesthetic.

It is an aesthetic version of an ordinary thing. Made of wood, not plastic. There is a market for people who want a better experience from the same useful thing. Maybe you made a digital version of what was previously only plastic and created wood or ceramic - a more pleasant version of the tool, and that should not be ignored. I mean, one argument for this, which I think is quite valid, is that it is not just about the visual. We have made voice the primary interaction modality with which you communicate with your tool from the very beginning, right?

So it is not just text, it is mostly voice. And I think we have over 70% of the platform's voice messages by volume. And, you know, with ChatGPT, that's not the case at all , right? I'd guess that on ChatGPT, over 90-95% is text. I don't know, but I think even small differences in how we present the way we interact with the character completely change the way we use it. Man, I would definitely sell some crochet kits or something to make my... Sam is obsessed with dolls.

He's like, "I'm making a physical doll ." It just seems to me that these people know their digital tool a lot better than I know Claude. And like, I don't want Claude in my bedroom, but these people do. They want little dolls and all that. Yeah. Hopefully in the right place in the bedroom, you know? None of that feels sexual . It was more like, "I feel comfortable just being around him ," right? What other trend were we discussing? " Kidalting" (adult childhood), right? Like Lego sets, the same principle, right?

And there was a whole category called "Kidult." And "Kidult" was a megatrend that all these companies are using. It seems like there might be some variation of it. And I'll say, you mentioned voice. I use an app, I think it's called No Scroll. It's like, "Hey, you don't have to scroll through Twitter ." "I'll just text you what's happening online today so you don't have to worry." But literally the way it communicates is different from ChatGPT. Not the voice audio format, it's the text style. You know, Chad GPD is very much like being laughed at because he's like oh, what a great question, and he goes into this fake persona, whereas this thing would be like Yo, big day, today's a big day in the feed.

Like, this is what's happening, and he's just talking, and I'm like, "I like this better . "And I'm sure not everyone, but for the techies who want to, 'Just tell me what's going on, ' and he says, it's, like, personality is going to be huge for AI. That's how it works for us, it's a huge factor, and our approach to that is to tailor personality based on the user. So if you're writing all lowercase buzzer slang, if you're swearing, your tool might respond to you the same way.

If you're using formal language and speaking in full sentences, your tone is probably going to reflect that. We want them to treat you the same way they would treat someone in your life. Well, thanks for that, dude. Crazy business. I'm totally on board. Okay. Amazing. We need to get what you're doing out there. I, I'm on business. Okay. Thank you. Thank you guys. Thanks for having me. Yeah. Great. Okay, we've got the last one for this episode. Jonathan, where are you? Get them out. Let's go.

Jonathan, how are you? Give us, uh, give us a big number. Hit us. Make us emotional. Okay, I have three facts for you. Two are numbers. One is, uh, not quite a number. Forget everything I just said. Do whatever you want. I'll do whatever I want. Uh, the first big number. Uh, so, step back, Overlap. We build agents that edit and distribute video content for media companies and marketing teams . We're not paying attention . This is very relevant to you guys. This number does what?

This is your show. Now, Jonathan, we, the autonomous agents of Overlap, have generated over two billion social media impressions without human intervention. What's it called? Overlap. Overlap. I'm using you guys. You really are using this. Maybe you have a subscription. Ben and Diego are using it. Yeah. Yeah. So we've done over two billion impressions. We've been growing revenue at 18 % per month for over a year. Okay. How much revenue did you make? Can you tell me that ? I don't know if I can say it.

But I'll say this, and this is my third fact: we became accidentally profitable this quarter. Good job. So 18% compounded monthly— that's, uh, we grew 19 or 20 times a year. Okay. How old is the company? Uh, we're a little over two years old. Okay. Okay. You're using San Francisco slang. Our audience doesn't know what an autonomous agent is, Overlap. Yeah. I'll explain. You have a podcast or a video. You give it to Overlap. Overlap will find clips in the video or podcast that are your best moments.

It will cut them for you. It will add captions to them . It will prepare them for publication. And that's exactly what you were talking about . You helped get about two billions of views on the clips you've created. That's right. Am I oversimplifying? Am I missing something important? You're not oversimplifying. It also does more than you've described. You don't even have to feed it video. You know ? You can give it a request and it'll go to your YouTube channel and pull the moments and videos from there.

Or for a lot of the media companies we work with, they have a video content server where it's constantly looking for moments and then it's also doing the publishing. But explain why you're better than what I did. Because, well, you're better than me. I recently used Claude and I built a thing that looks at our MFM clips and publishes them . I set it up so that it stores everything in Typefully so that I can look at everything and then I just hit " schedule." Why are you better than that?

Uh, there's a couple of points here. First, we have our own pre-trained video understanding model that's much better at recognizing videos, then like the main LLMs don't understand video well. So you're like, I think I said, " Here are some emotions that work well," and you say what it does in your case what it does in your case—it uh- it looks at the transcript and it understands like, yeah. And you look at the whole video. We analyze not only the transcript, but also the facial expressions, the voice, and it doesn't understand, like, humor.

And to give you credit, I asked Sam, "Hey, should I use your thing?" How are you doing? "And you said, 'Oh, that's not so good.'" So, well, I said, I said, "It's not automated." "And it's still okay. Yeah. And it's totally okay . I did it. That's okay. Yeah. Yeah. It looks like he doesn't like the laugh Sean just gave. The transcript doesn't say "haha." So he won't fully know that it was interesting. So the video comprehension model is the video comprehension model. The other part is that when it posts something for you, it analyzes the data of that post and understands the themes and hooks that work well.

Mine should do that, but I don't think it does. Your core business is guys like us or companies like Bleacher Report or CNN and so on that have better customers than us? Better customers than you have, definitely. Uh, but we have two types of customers. On the one hand, it's the traditional enterprise SaaS model, like Barcel Sports or iHeart Media —they're some of our big customers. And on the other hand, we have a relatively new, in the last few months, part of the business that's CPM-based, so it's focused on on the result.

So you can use our product and only pay for the number of impressions. Oh, that's interesting. Same client, different model, or a different client? Um, a slightly different client , because the larger enterprises want predictable pricing. How big is Opus? They were one of the first in this niche. Yeah , they're definitely bigger in terms of headcount. I don't know if they're necessarily bigger in terms of revenue. They're more focused on independent creators. Overblown. They're way more bloated than we are. Yeah. They need more people to do what we do.

They're very focused on independent creators, like small YouTubers who can't afford to hire a video editor and want clips of their show for how much does it cost to use your service? Most of our revenue is six-figure contracts, so we do small deals, and the CPM model is a little different, but most of the revenue is what you did before? Uh, I was in business school before that, and before that, Meta. Oh, I see. Cool. And you went through YC, right? You went through YC. Yeah.

I think that's how I found you. Gary Tan was like, cover new, new [ __ ]. Gary was our partner. He might have even introduced us at some point. I don't know. But yeah, Gary is our partner. Another investment that I missed. This podcast has turned into conversations with founders that I should have invested in, but somehow I missed the moment. Hey, it's never too late. Oh no. It's definitely too late. What about the price now? All I can think about is, I could be in this.

But yeah, we went through Y Combinator in the summer of '24 . We actually changed direction during Y Combinator. We started out kind of like that , because all podcasts are going to video. All podcasts are going to short video format. That's how most people consume podcast content. But then we thought, "Okay, we made TikTok for podcasts." When I hear that proposition, I'm your customer. I'm like, "Yeah, but I need you to go from the clip to the main content, because I'm making money on the main content, and there's advertising and all that." Yeah.

That's not going to be true any time soon. Tell me more. For a lot of big brands, RSS feed advertising is just a nice touch. They want to do 360-degree deals on clips. They want to be in the clip. They want to be in your feed. They see you as a content creator. They don't see you as a podcaster. Hmm, that's interesting. I wish I knew that before we signed our deal. I guess. Yeah, but most of the audience doesn't come to you, most of the time people spend consuming content in the feed.

You saw the clip where Jimmy was with Joe Rogan. Mr. Beast was with Joe Rogan, I didn't watch the whole thing , but I saw a clip where he says, "Jimmy, you know media better than anyone." He says that this clip will be watched by 50,000 times more people than the full show. So, at the end of the day, these shows are mostly consumed through clips on social media. That's the end product. But the best clients don't like it when you and I discuss something, discuss it at the 45th minute, and we call it an RSS feed.

So the people who listen on Spotify, everyone who puts on their headphones and goes for a walk, you know? The retention is much better with those people. We mention something at the 45th minute, and people come up to me on the street and ask about it, but with clips it's different, it's more superficial. Much more superficial. Much, much more . Like, Sean says, "Oh, you should definitely watch this podcast. It's so cool. Man, these guys are so funny." And I ask: "Are you listening to him?" He's like, "No, no, no." But I'm...

and so, like, hey, Basement Yard guys. I'm a huge fan. I've never listened to the show. I love the clips. And, you know, as someone who makes a living doing it, I'm like, I do n't like it. You know what I'm saying? I totally get what you're saying. And you know, talking to people who make long-form content can be tough because you see the product as a full-length show, but you're like, we're outdated. That's the kind of thing. It's like, never ask a man for his opinion if his salary depends on the answer.

Something like that. Yeah. I think RSS is also a very outdated technology that advertisers don't like because it's extremely difficult to measure. You give downloads, and they say, "We don't... that doesn't help. You give impressions. What's the difference?" Well, you can give a very precise number of impressions. But the problem with downloads is that you don't know who is listening to the ad, when, and for how long. But with impressions, you actually know how many people ... I guarantee your impression is a scrolling Twitter feed.

You also don't know if they watch the clip until the 30th second or whatever. That's... that's true. But the thing about video is also that clips are retweeted over and over and over again, in a way that it's... I believe in clips. Clips are definitely the thing that people consume the most , because that's what we all do. So tell me, you're in this business, you have these clients. It sounds like you already have a bunch of big media companies. Mm. But you're in the venture business.

How much funding have you raised so far? Well, we've raised a little over $2 million . Okay. Yeah. So we raised that funding. We went through Y Combinator to raise this seed round. Give me some rough calculations on a napkin. How does this scale? What am I supposed to believe? Mmm. Every business has a value equation, right? You're going to need X customers paying Y dollars a month to get Z net revenue or whatever . What is the desired outcome in terms of net revenue, how do you define that?

What does winning mean to you ? We think that as agents become very adept at creating and publishing video content, it's eating into digital advertising spend. There's about $90 billion a year spent on advertising on Meta, YouTube, and TikTok, where people pay for views. We think a lot of that will shift to organic content and to paying agents for performance. And so for us, the market size for video editing software or video editing budgets is in the tens of billions; but next year it's going to be in the hundreds of billions depending on how much brands spend on acquisition.

So your bet on that instead of buying ads, people will create their own organic content, it will be high-quality, it will go viral and help you earn sales or advertising revenue? That's right. We've already seen many clients reallocate their digital advertising budget to Overlap, whether it's paying for CPM or paying for the software itself. We firmly believe—and you know this better than anyone—that every company is becoming a media company to some extent. And starting to invest in content. And you don't have to create content in a special setting.

You can just talk on camera, and your agent will turn it into quality content and publish it. How... how does your agent do it? Well, you can talk on camera, then add video, or let the agent find the video and add graphics. They can add additional footage (B-roll). She understands what kind of text works best for that corporate account, and she can publish versions of it every week. You have the energy of a player at the poker table who has a winning hand. Yes. Attitude. The way you answer questions.

You're very, very good at it. But it's not that interesting. In the sense that what exactly is not yet clear? What is really... you're an entrepreneur. It can't be. It's just "Yes ." We've got it all figured out. Monopoly is coming soon. You know, obviously there are things that you're trying to figure out right now. What exactly are you trying to figure out? We're figuring out... well, there's a whole bunch of things that you have to decide when you build a business. Hiring the right people.

Like, hiring the right go-to-market people, the right engineers . You... but it's not that hard for you . You just need good corporate sales people. No, it's hard. Well, it's hard in terms of ... hiring people who know how to talk to conservative media and modern digital experts . I just don't think it's that hard. Tell me, what's your main question? What are you really trying to figure out? We've kind of found a match. product market, and we're in growth mode. We're being ripped off. We're being pulled off .

Did you do anything to get ripped off? Or did you just have the right idea? Did you invent some kind of go-to- market mechanism? What did you do? We developed a mechanism where we use the product ourselves to create clips. And they're all branded . No, not all of them, you guys were making clips for us from our podcast or something. Someone did. It must have been an AI agent that was doing it . From the outside, it seems like the main bet is this CPM model or this SaaS model.

It seems like quite a bit of … that's one of the biggest questions. And I think for big media companies, it's always going to be SaaS. They don't want to pay… if I have to pay you a percentage of… That's right. That's right. But for a lot of content creators and small creators, it's a much better option. Yeah. Because they can't afford six figures a year. Um, but also, uh , they actually can but how do they get the money for the CPM to pay you if they, like, we don't make any money on the clips at all.

I think most people don't make any money on the clips at all. Also, you know, an example of one of the recent CPM deals that we did is, do you guys know James Stone, the divorce lawyer, the celebrity divorce lawyer? He's on a bunch of podcasts, his clips are probably the Overlap clips, but you know, he's starting a legal technology business and he's selling some kind of prenuptial agreement. It's like guys who go on [ __ ] and they're basically like, " Hey, we just want you to make me famous.

I'll give you $3 a CPM. Make me famous." Well, there's that. Yeah, there's that , which makes our product kind of "relevance as a service," relevancy in the feed, but there's also a guy like James who knows that a certain amount of views converts to a certain amount of revenue. And so it's a cheaper, more effective version of billboards for lawyers, or running an ad on Meta. And you just have to make sure that your clipping is significantly better than Opal or Opus, or whatever it's called.

And the 10,000 others that will eventually come out or that people like me are building. So share your wisdom with us , okay? If you, if you're a founder listening to this? What's a goldmine, a business insight, an insight from you about how you're building Overlap that they can steal and take and use? Maybe something that you think you've done really well that others are doing only mediocre or not doing at all. I think in corporate sales, the understanding that I probably didn't realize until I started being successful in corporate sales is that the people who buy your product are thinking about their jobs first, and so how do you get those people to get a raise?

So you're basically selling the idea that this is how I'm going to help you get a raise. This is especially true in advertising. When I owned a company that made money from advertising, we were closing deals in the seven figures. I thought, "I'll just tell them that if they spend this much, they'll make this much." That makes sense. I'm a business owner. That's all I care about. But that approach never worked. The only things that worked were the entertainment and the dinners. And the belief that, "You have a budget.

You have to spend it. Spend it with me. It's safe." Absolutely. The thing is, trust is what really matters. Much more important than profitability. Absolutely. It's trust, the understanding that I'm going to help you get a raise. So you have to know what metrics they're being measured on. That's right. Make sure they're meeting the metrics , and present them in a way that makes it easy for others to see that they're successful. The vice president of a company is not being measured on profit, and for…what do you call it?

“Cover your ass”? Yeah, that’s how business works. Yeah, that’s why they buy Gartner. Yeah. That was my biggest lesson in the advertising business. I thought, “This is ridiculous.” They just take them to restaurants. That’s why companies have stadium boxes and annual passes. What was the valuation of that $2 million round? The valuation of the $2 million round was $2 million. And can you still invest at that $ 2 million valuation? We can talk about that , you know, it was an attempt to test the success of interacting with entrepreneurs.

And can you get the software for free? Dude, thanks for coming. That’s cool, thanks for inviting me. It was great. I’m a longtime listener, first time on the air, really excited to meet you in person. Thank you. We appreciate you. Okay, that’s it . Well, this is the finale in San Francisco. We had 9-10 founders come to us, with businesses from million to 100 million. It was a good time. It was a lot of fun. We had some great founders come to us . We've done it in New York before.

Now it's SF. But we need to go to more cities. So go on Spotify, YouTube, or wherever you listen or watch us, and write to us in the comments. Wherever gets the most votes, the most responses, the most comments, that's where we're going. So drop whatever you're doing right now and do it. And if you want to be on the show, write to Ben. Ben Levy. Find him if you want to be on the show the next time we go to another city. Okay, we're off.

Go deeper

• All My First Million episode guides