Start a Business With Just a Prompt (w/ Bolt CEO)⏐ Ep. #191

There's this guy C.J. on Axe. He posted a tweet and said, cost me nine bucks on Bolt to do this. And then someone asked the question, how much did you build a client? He was like $9,000. Most people still don't know that they can actually build software without being a developer. You can literally go from not even having a website to accepting payments in five or ten minutes.

Whatever software tool you use, replicate that in bulk. Absolutely. Worst case, you save money for yourself. Best case, you resell that to others. People will pay so much money for a dashboard. So go start a dang dashbo. What are some prompting best practices? This is the greatest time for entrepreneurs to be alive in the history of the world. Why? Because if you're marketing or business heavy, you probably don't know how to code. But now you don't need to know how to code. There are apps like Cursor, Replit, Lovable, Windsurf, Vercel. So many that are just vibe coding apps. Use natural language, Build a website, build an app. I am having the best time of my life and so I'm on a mission to interview the founders of all of these businesses because I want to know from all of them what makes theirs different or unique. And most importantly, how are their users using their app to make money and to save time. What are the common use cases? Building an agency on top of it, building apps on top of it for others. And what are the fringe weird use cases that people can learn from? So today I had the founder of Bolt Nuon. His name is Eric Simons, and he opened the curtain on how people are using Bolt to build big businesses. They went from zero to $20 million in annual occurring revenue in their first two months. Because people like me are building really cool things on top of it. They didn't pay me to do this interview. I just think it's a cool app and it's one I use on a daily basis. Hopefully you can learn a lot from it. Please share with a friend. Okay, well, Eric, why don't you just start by telling us who you are and what you do.

Hey, awesome. I'm Eric. I'm the co founder and CEO of Bolt New, which is the fastest way to take an idea and turn it into a real web or mobile application.

Okay. And what makes Bolt New different from, you know, other competitors out there?

Totally. Yeah. I think with Bolt, some things only focus on like web apps. And so we kind of do web and mobile. And generally it's insanely fast to get up and Running with something that, you know, looks, you know, not just like basic, but really stunning as far as your website or your mobile app. But you know, really importantly, like, we really focus on the end to end experience of like actually like building businesses online. I mean, that's not just like, you know, a basic storefront, but having like, if you want to build a SaaS application or like a new AI application, having payments built in and be able to actually use different APIs and that sort of thing. Yeah, all that is like integrated into this AI agent we've made.

Yeah, it feels like a lot of the other bolts of the world are like, they're really cool party tricks and they're really cool for getting an MVP out there. Something that looks flashy, but there's not a lot under the hood either when it comes to design or functionality or accepting payments, as you say.

Yeah, exactly. And you might be able to do it, but it takes a lot of like, of manual work to go and add those things in. Right, so you're kind of back at square one then, where it's like, well, you're kind of spending maybe as much time then as you would if you had started in an IDE on your computer. Whereas like with Bold, it's, you can literally go, we can, you know, we can do, you know, a demo of this, but you can literally go from not even having a website to accepting payments on your app in like five or 10 minutes and you know, like actually building a real application, like a course selling website or whatever have you, you know.

Okay.

And yeah, that's crazy. That's never existed before in history to be able to create custom software that you can actually sell like that.

What do you think that means for the startup ecosystem? Like that it's so easy, so frictionless, so affordable, like to zoom out a little bit. What does that mean for all these new startups out there?

It's an amazing thing, right? Because the thing that's always been really difficult about entrepreneurship is that, you know, the best entrepreneurs tend to be the ones that have spent a lot of years building companies. And it's largely because, you know, you can't really go and take classes on how to do business. I mean, you can, but like, but you don't really learn a lot of practical skills on how to actually go and build products and sell products and everything that goes into starting a startup. And by, you know, being able to like, go and take ideas and launch them into reality and get immediate feedback, that shortens the cycle. Right. Of you actually Being able to try and build a business and really accelerates the learning process for entrepreneurs. If you have to choose a single metric, like what's going to determine if an entrepreneur is going to be successful in starting a company? It's really just a, it's a numbers game. You know, if it's. How many hits at bat do you get on like building a product, getting feedback, iterating on it, et cetera. If in my view, like this is the tool that my co founder and I wish we'd had since he and I were 13, learning how to code together through all the things we build over the past two decades, because it's that, you know, you want to just be able to go and launch things, get feedback as fast as possible and iterate and adapt. So I think it's just an explosion of new things that would have never existed before. But I think we're going to see crops of entrepreneurs that, that you know, are able to be successful far faster and, and often in many cases where they may not have been able to, you know, afford to just keep taking swings on products.

On a scale of 1 to 10, how technical would you say you are?

I'm reasonably technical. Yeah. I don't know, like eight or nine or something like that.

So you have like a software engineering background?

I do, yeah. I never went to college actually back so my co founder and I, he and I grew up down the street from each other in a suburb of Chicago. And when we were 13, we. You learned how to web build web apps together. And we've been building, you know, startups and whatever have you ever since. And he and I have just been best friends for the past 20 years. When we went to graduate from high school, you know, we were looking at going to college and it was gonna be like 100 grand in state for four years. And we're like, this doesn't make any sense.

Like we're already making, we can build stuff. What are we doing?

Yeah, we were already making like 50 bucks an hour doing website contracting work or whatever. When you're like 18, that's like meaningful six figures. Yeah, exactly. And so, you know, just the math didn't. Math and this was very unpopular back in like 2010. But I think it's like the world's kind of come around to the idea that, okay, actually maybe forcing everyone to go to college and be in six figures, a debt isn't the best idea for Mo. For everyone, at least. Right.

So I applaud you. 2010 is when I was in college And I've always been entrepreneurial. I didn't have to go to college, but I just had this nagging insecurity and maybe a nagging, like, future dad. I had my first kid in 2010 where it's like, this is like a form of insurance. It's a good thing to do. I've got two college degrees. I've never used either one of them. I feel like AI is taking away all of my excuses, like my macro excuses. Because when I was in high school to, you know, early 2000s, I used to say, oh, why wasn't I born five, 10 years earlier? I could have been in the dot com boom. I would have moved out to San Francisco. And then crypto comes around and it's like, I didn't buy Bitcoin until 2016, which seems early today, but seemed very late. Then it's like, oh, why wasn't I in the right message board? I could have bought Bitcoin at 5 cents, you know. And now with AI, it's like, okay, Chris, where's your excuse? You're on the cutting edge of this. You're on Twitter. You see, like when a new model or new app comes out, you're right there using it. You have no excuse. And the other excuse I used to tell myself is I'm not technical. I can't code. If I could code, I would be, you know, that's my Achilles heel. If I could do that and market, you know, if I could be the marketer and the coder, I could build whatever I want. And now it's like, I can't. For the last like six months, I've been in these vibe coding apps all day and it is so much more wonderful than I ever imagined. I'm just loving it. I have an idea and I just build it. Instead of like going to my brother in law who's really good at coding, like, hey, can you. Well, I'm kind of busy. Like I can just do it myself. It's amazing.

Yeah, it's huge. And like, and it's truly like, it's the tool that my co founder and I wish we'd had because like I learned how to code. It was somewhat intrinsically interesting to like learn about programming, but to me it was always a means to an end. Like I just wanted to build stuff. I wanted to build products, I wanted to build businesses. And you know, if you're going to, if you're going to build digital based products, like you have to, you know, building software is at least was a required part of that or knowing how to, you know. So I look at today and you know, some people have this opinion of, oh, well, you know, people should still learn how to code. Like it's, you know, I mean, and not, I'm not saying you shouldn't, by the way, but people, there's kind of this, oh, but we're losing something. Like, we also don't speak Latin anymore. Right. Like things, things change. Like the floor of complexity will kind of go away and meaningfully drop the bar on what it takes to go and build and do things. I think that's a good thing. And anyone that says software engineers are going away, period, like that's just, it's just not true. Like it's. What' to happen is you just have developers and this is happening already. Developers are just focusing on higher value tasks than just like going and changing basic syntax for styles or whatever. And so they're just being way more effective at their job and they are working on things that are very complicated that AI models are not well suited for. And so it's, you know, maybe a good analogy is like modern industrial man manufacturing. It's like there, there are still, maybe some people that say everything should be handmade, every car should be handmade. Like I guess Rolls Royce still exists, they still do that. But it's like this niche market. Right? Right. They're expensive, they break down a lot. Right. You know, in fact, it actually turns out having these great, you know, automated manufacturing processes overseen by humans with higher leverage tends to be better for everyone holistically.

Yeah. I mean, there's still people pulling, pulling around tourists in horse drawn carriages in Central Park. That doesn't mean that the right decision 100 years ago was to triple down on horse drawn carriages. You buy a dang car.

Exactly.

I think that big question of do I learn to code or not. It all depends on where you are. Like if I'm 20 something or you know, a teenager, like and I'm on the fence about like going hard on business or going hard on coding, like I'm going hard on business and I'm learning all of these vibe coding tools and that's my coding, like my shortcut. Right. Whereas if I'm already a coder, like I'm going to learn how to embrace this as much as I can. And like you said, delegate and elevate and just start doing more important. Like you'd be dumb as a coder to, to abandon coding because, oh, this is going to take my job. So let's start looking for like you are in the best position possible right now to further embrace coding, right?

Yeah, exactly, exactly. It just gives you so much like understanding how this, how software works, how to build software, gives you just so much more leverage on how to actually control and use AI tools and agents to, to go and do your betting to, you know, change the world or whatever. Right?

Yeah. It's like the difference between a college, like a business professor that like spent 30 years running and starting businesses, like his effectiveness in the classroom versus a business professor that's always only been a business professor. Right. Which one do you take? Like I would rather have a coder that knows these AI tools than someone like me, frankly. But I can build these myself and they know myself and they don't have to be amazing, you just have to be cool. So I, I love talking to founders about like 80, 20 of use cases of your tools. So I'd love to hear like the 80%, like what are people using it for? What's like the bread and butter? And then what are like those one off use cases that really just make you scratch your head but are really interesting to the listener.

Yeah, that's a great question. I mean this was like kind of crazy things. We launched Bolt like last October and before this is making a developer ID and so only developers were using our product and the scale of revenue. Just like we almost went out of business. So like a month after Bolt launch was when we were going to start spinning down the company. Because you've been around for seven years, we got into the company 700,000 of ARR. And so Bolt was the last experiment we ran like this in browser development, environment technology we made and Bolt, we paired this frontier AI agent with it that we'd made and overnight it just took off. Like in the first two months we went from 0 to 20 million of ARR, which gosh, it was crazy.

That's product market fit, right? Like if anyone's wondering, how do I know, like if you can't sleep because of all the inbound you have, that's product market fit.

Exactly. And it's the first time I've seen anything like it. I've been doing startups with my co founder for 20 years, I've never seen anything like this. And because every other launch we'd ever done, it's like you'll have this if you do great marketing and you'd spend a lot of time on the product. Most things you'll have this, you know, tick up on the day of the launch, next day it drops and then it just flattens out. You're back to the grind of, okay, we got to make this thing better.

And that's okay. That doesn't mean I should quit.

Absolutely not. Right? Like, and that is the normal thing. What happened here is from day one, it just, the revenue just was cranking. And so our team was super excited on the first day. And I was like, guys, temper your enthusiasm. Like, this is how it works. I've seen this movie for 20 years.

It's coming back down.

Yeah. And then the next day it didn't. And I was like, okay. So like maybe the word of mouth just kind of whatever. And then it just kept going day after day after day. And. And then the other kind of crazy thing was that most of the people that were emailing us as customers for support and stuff were not developers. They were designers or PMs or just people who are entrepreneurs and, you know, people that were salespeople making website for their daughter. It's like we were like, what is going on? Like, this is, this is not the market that we, you know, kind of thought that was going to pick this.

Tool up, but it's a much bigger market, right?

Massive. Right? Massive. And so after the first month, it became very clear to us, like, okay, this is something, you know, Bolt was really one of the first products that kind of hit that nerve where it's like, this is not just developers can build software now. And what does that mean? Right? I mean, it's like, dude, there's like a billion people on this planet that now actually can be empowered to go and take their ideas and put them into reality. That's going to change everything.

So I'm very enthralled by what you're saying right now. When you made that pivot, so from what I understand, you were like a browser extension for developers when you made that pivot, was the reason for the pivot because you were just kind of looking around, seeing at these other guys, thinking, oh my gosh, like, we could do that. We're not far from that. They're crushing it. Or was it like some. Something deeper than that?

I was actually pretty skeptical, I would say, of the AI stuff from 2022 to beginning of 2024. Not that like ChatGPT was obviously, I think, for everyone in the world was like a mind blowing moment in my head. After the first month after that came out, I was like, what else is this beyond, like ChatGPT? What's kind of the killer use case here? How is this, you know, what are the next things that this thing's going to get incredibly good at doing, that's really valuable. And I mean, honestly, tangibly, there weren't, you know, there weren't a ton of things that, that were really progressing. And so kind of by the beginning of 2024, like our company, everyone else in the world was like, adding AI features into their app and they didn't make any sense. For a lot of these, they're just like AI, AI. Like a classic hype bubble thing.

It was like adding blockchain to your name five years ago.

Exactly. And so, so we were looking at it, we're like, we think there might, there might be another. Because back in 21, we, the exact same saw the same thing happen with the Web3 stuff. And people were like, you should have Web3. And it was. We were like, this seems like a bubble. People thought we were nuts. People were like, you should raise money, you should, you know, grow the company to 100. And we're like, no, this seems like a bad idea. It turns out would have been a bad idea, was a bad idea for a lot of people. So I was kind of had the same sort of lens where I was like, I haven't seen anything yet. That really strikes me as the breakthrough, like the next breakthrough thing for AI applications that will drive what they're saying is trillions of dollars of potential revenue per year. Right. Beginning of 2024, you know, I was using chat gbt to just generate code snippets and I would take it and I'd put into our product. Our product was an IDE for you to build web apps in the browser. And I was like, oh, you know, like, ChatGPT should basically have like our stuff built into it. So, like, when you tell it to like build a web app or whatever, it just like does it in the browser. Like, that's the magic of we do is like allowing you to run, you know, full, you know, dev environments in a browser. And our vision originally was basically to make building full Stack applications as easy as using canva or figma. Just you go to a browser URL just boots, etc.

Remove the friction.

Exactly. And so you know that it kind of clicked. I was like, okay, well, like, the code that ChatGPT is writing is like, I think kind of good. So, you know, let's make a demo of this. We spent a week doing that in February of 24, found out OpenAI's models just were really not great for code generation. Like, it just the experience didn't work. Because you would tell it to build something, it would just break, it was ugly and what would come out, which.

Leaves an opening for you.

Exactly. Well, and so. And we were like, this is a cool idea. Maybe it's not feasible right now with the current state of the frontier models. Maybe they'll improve at some point in the future. Then in May, a couple months after that, we got an early preview of Anthropic's Sonnet 3.5 model. It was immediately obvious. I was like, holy hell. Like, this is huge difference, step function difference in how good this thing is at writing software and the output quality of it. And so we green, you know, brought the project back off the shelf, greenlit it and launched it. But so that's kind of the long story short of it though is like the best things that I've ever done in my career have always been things where it's been like, intrinsically I'm like, this would be kind of cool. Like I am doing, I'm using ChatGPT in this way right now. I wish that I didn't have to copy and paste to an environment to run it. Maybe we should just kind of build them together. And I just want that. And our expectations when we launched Bolt is that we were like, this would be incredible if we added 100k of ARR within a year. We were at 700k at that point of ARR after 7 years. So we had no idea how explosive the thing would be. It really came from just my own actual curiosity and needs versus kind of looking at the market going, oh, we want to be part of that. Because if you look at like the 2022, 2023 era of AI coding stuff, there is tons of people just slapping AI. And for that exact reason, because they're all looking around like, oh, they're doing this, they're doing that. But like none of it added value.

How did you get the word out when you made this pivot?

Good question. We put together a tweet like that was.

I was hoping you'd say that.

Yeah, unconventional. But over the past seven years, it's always important when you're building a startup, you should always be building community around it because you never know which of your iterations is going to be the one that clicks. And ideally you want some amount of built in distribution to propagate great messages you're going to send out basically. So we built up some like a small following out of maybe 20,000 people were following our account or whatever at the time. And, and, but we put out one tweet, like, there wasn't even a blog post, There wasn't a press release. And it was a darn good tweet. We spent some time on it and that just basically explained, hey, this is what this thing is and why it exists, why it's different and better than the other stuff that's out there. And it went viral. That was important for seeding kind of the initial set of users that came in and tried the thing. But what kept that growth flywheel going was that the experience was amazing. Right. Like, people would. And this is still true today. When you go to Bolt New, there's no, like, marketing landing page you have to sift through. There's no, like, crap, you know, like, like most products, there's like all this crap you have to go through to even just try the thing out.

I'm not gonna lie, when I'm vibe coding stuff, I'm looking at Bolt New for my landing page example. It's like, remove the friction, remove the distraction anyway.

Totally. Yeah, exactly. It's like just kind of get people straight into the action. Because that if you have a great product experience, it will speak for itself, right?

Yeah.

And so, you know, the prompt box is there, you hit enter within a minute, you have a web app or whatever that's built that looks great and it's mind blowing. And so people would deploy that URL, they'd be like, wow, I have a website live on the Internet. And then they would share that, whether, you know, in imessage or on X or whatever, and be like, wow, look what I just made with Bolt New. And so that was like, we saw this crazy, you know, organic word of mouth thing happening, and that's like, still the primary driver of how people find out about Bolt. Yeah, there's no paid acquisition. Like, there's $00, you know, on the marketing for the launch of the thing.

I don't want to go too deep into this, but what was the anatomy of that tweet that made it so great, in your opinion?

Good question. By the way, have you ever heard of the Amazon PRFAQ approach?

No.

So this is a good one.

This is good.

Yeah. So Amazon PRFAQs, when Amazon wants to go and release a product, before they start making designs or writing code or anything, they write a press release and an FAQ for the product. What are you saying the day that it is launched? Because you effectively have to answer the questions of why is this unique? Where are we innovating on this? What problems is this actually Solving, Right.

You're proactively removing friction to people ever using your thing.

Bingo. And you're starting at the end, Right. You're starting it like on launch day and then working backwards from there. Right. So. And what that provides is then when everyone's built, when you say, okay, well, we're going to do this project, they're. The North Star is so clear on where you actually need to get to. Right. And so for us, the approach we've taken is we have like, our approach is a tweet thread faq. So what we do is because it's even more distilled format of like a blog post or a press release, because you have to. Every tweet has to have a purpose if it wants to be good. At least you don't have like a run on thread or whatever. In the first tweet, it was really the video demo that we had, you know, showcase, like just going to bolt on you typing a prompt and what you could actually build out of it. And one of the guys on our team did a great job with that. As I recall, the text of that tweet kind of started with the classic, like, you know, wouldn't it be great if, you know, ChatGPT could actually.

I think I'm looking at that tweet right now.

It took like maybe a dozen or two revisions just for the first tweet to get it to where we were, like within 280 characters and ET cetera.

Was this the one?

That is exactly it.

All right, why don't you go ahead and read it for those that are only listening.

Okay, so this tweet says, what if AI dev products, Claude V0, et cetera, let you install packages, run backends and edit code, Introducing Bolt New by Stack blitz prompt, edit, run and deploy full stack apps, full dev environment, NPM vite Next JS with Frontier AI free exclamation point. That's it. That's the tweet.

And then in the comments, then you get to the nitty gritty, right? You don't want to put that up front.

Exactly.

The most interested people will start scrolling. But if you put that up front, then you just turn people off.

And this is why I love that. Doing tweet threads is kind of like the focal point of how you make marketing messaging. Because the first tweet is the thing that needs to grab people's attention and be like, hey, this is a thing that sounds dope and is relevant and it's something you would want. Right? Like, we all agree on that, because that's really what you want to get on that. Because once you have them there, they will then read through all of the cool things that this thing can do. It's way easier to make all the subsequent, you know, kind of tweets, by the way, because you're really just speaking to the specific features that really stand out and, you know, et cetera. Not that we didn't spend time iterating on those, but that tweet specifically, we. What we really wanted to anchor to is like, you know, the way we find this out is we would just go talk to like our friends and prospective customers or whatever, and we'd say, hey, this is what we've built. Here's kind of, you know, the general pitch. What do you think they would ask that question? Well, how is this different from Claude or V0? Right. Why do I need this? Like, what are the actual things that has these? And that's. If you look at that tweet, that's exactly. It's like, hey, wouldn't it be great if these things that you know and love could do this? That's what Bolt is.

You lead with a question.

Exactly, exactly. So and you kind of have to, like, you have to talk to people to kind of get a sense of like, what is that? What's the first thing people are going to ask you? Like, yeah, but like, what about blah, blah, blah? You know, it's kind of like the first question on this pocket. What about kind of the other app builders, right? It's like, whatever. It's like, okay, well, how do you. If we deserve to exist, what exactly is that reason, right? And like, you know, how are we frame? That's super important to have. And anyway, so something that we've learned over the years when we've launched stuff is just really having, you know, a very well thought out hook and explanation of, you know, what this thing is, why it deserves to exist, why it's going to solve your problems. Way easier said than done to do that.

Well, I think one of the magical things about that tweet was you put the URL right there, not below the fold, not in a comment, but right up front. And I've looked at the Twitter code and there's a misconception out there says if you put a URL in your tweet, Twitter's gonna deboost it because they don't wanna take people off the platform, which makes logical sense. So most people, myself included for a while thought, okay, I'm never gonna put Links in tweets, I'll put em in the comment tweet, which I would still say 80% of the time is best practices. But Elon himself recently came out and tweeted and said, hey, we don't like automatically deprioritize or deboost tweets from the algorithm if they have a link. What we do do is, is we do whatever we can to keep people in the platform. And so it doesn't have to be a link. It might be, but if your tweet is not interesting enough, maybe because they're clicking on your link and leaving Twitter, maybe it's because it's just not compelling, there's not a good hook, there's not good payoff, we're not going to boost it. But if your tweet is compelling enough and it has a link and people stay on Twitter, like maybe they click the link and they just keep reading all of your comments, they're watching all of your videos. So Twitter's like, wow, I don't give a crap that he put a link in here because they're staying on Twitter even longer than they would have otherwise. Let's boost the heck out of this. Like, sometimes you can really find magic because you lose a ton of people if you put the link in, even the first comment. Like you lose 80% of people. But if you can find magic and go viral with the URL and the top tweet, oh man, 0 to 20 million ARR in two months.

Totally.

In this case, totally.

Yeah, absolutely, absolutely. And that was the thing too with our. The dot new domain, because we originally did launch with developers. And so one of the stipulations of having a dot new domain is that when you go to that URL, it has to immediately start some type of action. So if you go to like docs dot new that takes you to a new Google document or a Google like.

A. Oh yeah, that's what I used to open all my docs. I never even connected it to Bolt New, huh?

Yeah, exactly. And so, so and that's why we chose the dot NEW domain. And partially because the dot com was taken. But we were like, we really want people. The magic of this experience is zero friction between your idea into reality. And so the new domain actually requires you to, you know, remove the marketing crap, basically. Right.

I love it. Okay, so my ADHD is on full display. Let's get back to the original question that I brought us away from 8020 of how to use Bolt, how people are using it.

Most of the people that are coming to bolt, by and large are folks that are entrepreneurs actually like they're building businesses. And then there's the a large portion of folks that are, you know, assisting in building businesses, like PMs, designers, engineers, even like sales and marketing teams, et cetera. We have two kind of splits of where our revenue comes from, which is one is prosumer, which is, that's the entrepreneur starting businesses. And then we have B2B, which is existing businesses like Fortune 500s or Scale Ups. And that's where we see a lot of like, you know, designers, PMs, engineers, you know, adopting us to accelerate, you know, their prototyping and workflows and that sort of thing. Right, the prosumer side and just by sheer number of heads on the platform, entrepreneurs is like the number one thing. And they're coming here and starting, you know, businesses of like all types, which is insanely cool to see.

Do you have some examples? Like what, like selling info products or agency web, like they're selling websites or what are they doing?

Yeah, totally. So certainly a lot of agencies use bolt to go and build stuff for their clients, right? That's a pretty large one. And the ROI on that is insane. Like there's this guy, CJ on X, he has the most legendary ROI story we've seen where he posted a tweet and said, hey, this is the application I made for my client and it cost me nine bucks on Volt to do this. And then people got a good amount of pickup. And then someone asked the question I was going to ask, had they not, which is how much did you bill the client? He was like $9,000.

That's like buying Bitcoin in 2010.

It's crazy, right? If you like search like Bolt Arbitrage, there's kind of a growing number of people that like are kind of connecting the dots on this. So there's a lot of agencies that use this for just accelerating the development workflows. And this is becoming a much more, you know, very common thing at this point that folks are doing. But then you have folks that are starting like SaaS businesses. Right? There's a lot of different types of these that we see. So you have people that are, that will make course building websites for themselves, for example. So instead of going to, you know, one of these other platforms, you can actually own all of your revenue and have your own site to do it, et cetera, instead of giving a 10 or 30 or 50% cut or whatever it is to go and you know, Use some other hosted platform and that sort of thing. So that's like, you know, just, you know, kind of one example of something we've seen other ones too is like people are making AI applications. One example is this guy named Paul who's you've been part of. He was like one of the first couple of weeks of folks that were on our platform, but built an entire CRM with Bolt that had AI integrated into it. And so you can literally industry specific. I think he was just kind of going after a general, just general industry, like anyone who's like a CRM to.

Fit his needs, hoping that maybe other people would like it as well.

Yeah, absolutely. And you can go sign up for it, you know, et cetera. We've seen a lot of different, you know, cool types of businesses and folks, you know, making like real money now going and actually selling monthly subscriptions and.

Yeah, etc.

For this stuff. Right.

What kind of app did CJ build? The $9,000 one. What was that for?

That was for one of his clients. It was a dashboard as I recall, for some type of reporting and their product, et cetera. And he was, you know, being. Acting as like an agency consulting with them and.

Yeah.

You know, building out web apps and whatever have you.

That just reminds me of just a framework that I have. First principle, like, people love dashboards. I don't care if you own an auto repair shop, you're CEO of a Fortune 500 company. And I think the first principle behind it is like, it just makes us feel good that we have everything under control. Like my whole business is on one screen. I know what's going on. Really, you don't. But my point is people will pay so much money for a dashboard that is simply just aggregating Together data via APIs to. Through an interface that you build with something like Bolt. So go start a dang dashboard agency. That's the thing. You won't be the first person to do it. Like, AI is totally, totally.

And so we're actually running this hackathon right now for the entire month of June, where. I don't know if you saw this, but like, so we're.

Yeah. Million bucks.

The world's largest hackathons. We have over a hundred thousand people that are participating in this at this point. We crossed the world record, which was like 66,000. Guinness is like officially going to market. You know, like, I've never done it before. I've never been in the world. World record. But we're all going to be there. It's been so, so cool to see all these folks going and, like, and building ideas that they would not. And the reason we did this, by the way, was that most people still don't know on planet Earth that they can actually build software without being a developer. And so that's, like. That was why we kind of created this lightning rod of attention for this hackathon of, like, hey, over a million dollars in prizes. You do not need to be technical. In fact, it's better if you're not technical and you come to do this and, you know, come on and, you know, try Bolt for free something and, you know, see what you can build. And there's been amazing stuff for, like, folks. There's an app that someone's building in South Africa for giving access to, like, emergency services down there, because apparently there's not really great infrastructure for. If you needed to go to the hospital, like, where do you actually go? Like, which kind of blows my mind. You would expect that this would be kind of a solid problem in 2025. I think it was, like, a month or two ago, one of his friends had an accident in South Africa, and they didn't have Internet access and had no idea where to go. And that's, like, a common thing. So there's stuff like that where people are building applications that are actually meaningful, kind of even. Even beyond just a business sense, like, actually, you know, obviously important for. For humans.

I think it's very key that you said, like, it's actually better if you're not a developer. Like, I. I think of this study where they took a bunch of executives and a bunch of kids, and they gave all of them a bunch of marshmallows with toothpicks. And they said, all right, build the biggest tower you can. And they put them in groups of, like, three. And all the executives, they got together, and they had, like, two hours to do it. They spent the first hour, hour and a half planning, game planning, engineering. Let's get the load bearing right on this. Let's have a big enough foundation. And the kids just started fricking sticking marshmallows with toothpicks. And they started building and falling over building. And the kids won, right? Because they weren't biased. And they just started trying and failing, trying and failing. And when I think of a developer going into something like Bolt, they probably have all these preconceived biases, biases and confirmation biases. And they go into like, well, I don't want to build this, because it's probably going to. And they might be wrong in a lot of cases because they have these assumptions that just aren't true. Whereas someone that's totally ignorant like myself can go like, yeah, build me an iPhone app. And then it's like, okay, and you're going to hit problems, you're going to hit friction. It's like, oh, okay, let me take this over to ChatGPT, let me export my repository to GitHub, okay, huh. And like you just start working through these problems, which that's all entrepreneurship is, is problem solving. And entrepreneurs know how to do that very well. And I could actually see a scenario where the non technical vibe coders, if you will, are, are crushing it for sure.

Right? I mean like that's 100% because like the big problem with like, you know, the building like apps and software being gated to date on having development experience is that once you're a really proficient developer, you lose touch, you know, naturally with what the experience of using software is as a non developer. Right. And, and this is something for us as a company, like we were building tools for developers before, now we were building tools for non developers. And so for us we are having to completely transform how we approach building, building the experience of our product because a lot of our engineers own instincts no longer actually match with the end users using the product to that. And so who are we hiring to bring in is like folks that are like great PMs and designers that have really good perspective of okay, what really represent the end users experiences. And then for our engineers there's lots of ways to kind of get that context on the engineers. Like some companies, every engineer has, you know, knows a customer personally that they talk to every week right. To like they have that built in like empathy and you know, you know, thread with them. Right. But it's a huge advantage for folks that are, are building apps is if they don't, if they're not kind of actually bogged down in the bajillion details that go into building a software product. And that's like the stuff that's come out of this hackathon is some of the coolest ideas that I have seen from a hackathon. Largely because it's not just engineers making stuff for engineers, it's stuff that's like, wow, this is like we brought a girl in our live stream that is a new mom and built an AI app for new moms that helps you like, you know, you can ask it questions about, you know, what do you do, like how much you know generally should you be feeding the baby or like, how do you get them on the right sleep schedule? Blah, blah, blah. Which, you know, I had, you know, my first kiddo a year and a half ago, and I was like, yeah, that's a real problem. Like, that's like, you know, a real thing. And, you know, but as an engineer, my first thought wasn't, hey, I should go build an app for this. And, you know, like, it just. Yeah, it didn't click. That's what's really cool about this, is that it connects people that are directly exposed and feeling the problem to. With the tools to actually solve it, you know?

Yeah. What are some weird, unexpected ways that people have used Bolt to make a ton of money? Maybe they built an agency on top of it. Maybe it was a random website, Anything.

That's a good question. I think the agency's one has been so interesting, but it makes a ton of sense why that's the case. The other really interesting ones that I've seen have really revolved around folks that are going and building, you know, platforms where they are owning the content that they're making and selling it directly, in effect, kind of cutting out the middleman of some of these other tools. Right? So whether they're selling courses, whether they're selling, you know, materials or access to community or, you know, et cetera, like, that's. That to me is like some of the most interesting stuff because some of these tools will take a substantial cut out of your earnings. Right? And like, yeah, before you could have gone and signed up for Stripe and done yourself, but like, you know, the previous company I ran before this was, you know, we were selling courses online, kind of like pluralsight. And for my co founder, and I was just him and I was, you know, like a. Basically a lifestyle business. And it took him and I, like, a month or two to get just the infrastructure set up with payments and login and da, da, da, before we could even start selling courses. And in retrospect, it probably wasn't worth our time, but we just didn't, you know, we were like, we're bootstrapping. We need a 30 or 40% cut is brutal. Like, we gotta pay rent. That to me is, you know, some of the most interesting things from like a monetization standpoint that. That I've seen, you know, for Bolt.

But that could be a good framework of just like whatever software tool you use that is more expensive than you think it needs to be, or takes a bigger percentage cut than you think it needs, try to replicate that in bulk.

Absolutely.

At worst case, you save money for yourself. Best case, you resell that to others.

Absolutely. One of my friends, actually a tangible example of this and this. They're not making money on this, but they have. There's like a teleprompter iPad app, right. So they're using it to like, you know, they're recording videos and it's like 30 bucks.

Know of any good teleprompter apps? So I would love to hear about this.

Yeah, so this blew my mind because he sent me a message. He's like, yeah, like, you've been using this app. It's like 30 bucks a month and it sucks on the iPad. And so I went to Bolt and just one shot said, make me a teleprompter app. And it did, and it's better. And it, you know, cost basically, you know, 2 cents or something to me.

Yeah.

And then they own it forever. Right? So there's a lot of things like that where, you know, tools that you're paying for, you actually may not need to pay for them. And in fact, you can customize them for your own use cases for your business, which is like a very common use case, actually, across the spectrum of people we see, whether it's entrepreneurs or it's actually Fortune 500s or whatever. Like, they're also using Bolt to just build internal tools for themselves that are kind of bespoke, you know.

Do you have any examples of big businesses that have already been built with Bolt, like from an ARR perspective or number of customers perspective?

Yeah, good question. I think I'd have to pull. I know that there's startups that have launched using us for parts or full parts of their web apps, whatever, have you. There is a YC startup called denormalize that used us to build their application and they're doing stuff around AI data pipelines, et cetera, and they've raised a couple million dollars at this point. There's a lot of these though, right. So the one thing is, because, you know, folks own their entire stripe accounts, we don't have observability per se into how big some of these folks are getting. But we hear this a lot, though, where they like, hey, we use this to build out this entire thing. And we certainly see the companies that are growing by headcount and adding seats on us. You know, we see that as well. But yeah, I mean, there's a lot of cases where, you know, certainly startups are going to raising. Raising money and are making substantial amounts of ARR, but also existing scale ups and Fortune 500s that are like, you know, launching new products out of this thing that are, you know, doing responsible for tens of millions of dollars in revenue per year sort of thing.

Yeah. I could see someone saying like, well, like, yeah, I want to get in Y Combinator, but I can't just build something with Bolt to get in. And it's like, you, you can, you can, you can.

Yeah, you absolutely can.

I'd love for you to show us, like, show us how the average user with no technical knowledge can go from nothing to something on Bolt.

Cool. Yeah, absolutely. Let me go share my screen here.

Okay. And for those listening on audio, I'll be describing what's going on here.

All right, so we get the Bolt homepage here. Simplest thing in the world. What do you want to build? And there's just a text box we can tell it we want to build. Maybe let's just have like, you know, create a course selling website, perhaps something like that. So I'm Eric Simons. I want to build a course selling website for my new course on how to use Bolt New to make production web applications. Anything else you want to add here?

No. I'd love to ask you questions as we wait for this to prompt. What are some prompting best practices?

Like the first prompt, start with like, something, you know, relatively simple. Like you start with like the homepage and start with maybe some of the, you know, just static parts of the application and like, figure out like how the. The colors, the UI UX you want. That's what this thing's going to go do right now you want to like. Because if you look at. From a typical product development process, that's usually where you want to start is just with, okay, let's make sure the general flows, the layout, the look and feel, et cetera. Let's make sure all of that stuff works properly. And then we start like layering on functionality.

Yeah.

And so in the next steps here, depending on what comes back, we'll either make some changes to the colors or whatever have you, or to the layout. We're like, okay, cool, this looks great. And let's start adding in some functionality. And at that point, you kind of want to break these things out, you know, into specific tasks, like, you know, prompt by prompt, just so that it's easy for you to. If you have to roll back, you're like, okay, this, that, that didn't really work. You're. There's like only one thing that I was working on there. I can easily roll that back. But generally you don't want to, like, completely jam a ton of stuff, you know, into the chat just because you might end up with results that are hard to roll back and kind of have fine grain control.

Okay, Right. So me, I'm, like, super impatient, and when I'm building something, I'm like, do this and then it starts. I'm like, also do this and then do this, and then do this and then do this. Would you not recommend that?

It depends. So especially if you're doing, like, if you're changing kind of parts of the UI and that sort of thing, totally fine.

Yes.

If you're kind of getting into more complex functionality, it's good to, like, kind of break it down and take it more step by step. I'll show you one feature we have, actually, that allows you to kind of consolidate changes like that that are maybe large and etc. It's called discussion mode. And this down here in our chat box is this little, like, discussion icon. If you click on that, then you can actually just have a conversation where the agent isn't actually building anything. It's just helping effectively answer questions and build a plan on what to go and build with you. Right.

Okay.

So this is the app we got back. So it's called Bolt Master, gave us some branding. Yeah, like, pretty good for shot.

Wow, that looks really good. Jeez. I'm not like a design master, but I think that this looks legitimately good. Let's say, for instance, it's not good enough for me. I want to look something. I want to see something much more sleek. Maybe I have some examples. I can go screenshot it. What prompting tips do you have for asking it to change the user interface to look a certain way?

For sure, yeah. I mean, if there's, like, certain, like, colors or, you know, like, hey, I want more animations. You can just tell it to do that. You can just say, hey, like, I. I want it to like a purple theme, or I want to like. And if you have specific examples, you can actually just take screenshots, drop them into the chat, and it'll use that as reference then and go, okay, I will make something that looks more like this. Right.

How well does it work with screenshots? Let's say I went and took Airbnb's homepage, which is known for design. I just screenshotted it and said, I know this isn't vacation rentals, but make it look like this.

Sure.

How well would it do with that?

Let's give it a shot. I'll go ahead and drag in the screenshot of Airbnb, make the theme look more like Airbnb.

Like, design is something that is very easy to ignore, but I feel moves the needle much more than people would assume.

Absolutely, Absolutely right. And so that, to me, is what's cool, is I'm building something from scratch. And Bolt, what I'm often doing is I will go and find, you know, great design inspiration to pull from and, you know, tell Bolt to go and look at that and then, you know, spin up design based on that. Right. And hopefully, hopefully the demo gods play well with me here. But it's. It's really looking at the screenshots, like, I don't know. There we go. Cool.

What happens? Because this has happened to me in other apps where I just hit brick wall. For instance, I'm trying to build this website right now, and it will not let me log out. I click log out, it just refreshes, and I put the code in GitHub and ChatGPT. And they all say that, oh, I see what's happening.

It's just.

It's actually logging you out, but it's not refreshing on the UI over and over and over, and I cannot fix it.

Yeah.

And it's a very simple thing. What are your tips for getting around that brick wall?

Absolutely.

Because it's going to happen.

Yeah. Yep. And this stuff does happen. Right. And so, and that's like, one of the reasons we rolled out discussion mode is it's very useful for solving these types of issues because it. It's, you know, instead of it, you know, typically when you go to other, like, even with the. The normal build mode agent and bull. But. And this is how, you know, ChatGPT and a lot of the other things by default will work is they'll. Instead of being like, hey, let's like, maybe talk this through and like, have you do a couple of things to test stuff, it'll just be like, oh, I definitely know what's going wrong. It's like. And has no idea. Right. Whereas discussion mode will say, okay, let's try this. Okay, can you give me this information? No. Okay. And it'll actually start instrumenting the code base. If things are still going wrong, it'll start adding logs and et cetera, to then pull in and go, okay, well, what is it? What's happening now? Right. That typically goes a very long ways. Right. And the cool thing about discussion mode is that every message of discussion mode takes like 1/10 of the price of a normal agent message. Right. And so you can send Way more discussion mode messages, which is important for debugging and not just burning through your entire, you know, monthly allotment. Right. So this thing, you know, shamelessly.

Shamelessly, that's pretty close. I mean, there's no shame on this podcast. So we're all about copy paste.

Yeah, yeah. You know, obviously when pulled the exact color scheme and this is like, you know, took exact inspiration, you know, from those listening.

It looks like Airbnb.

Yeah.

Like it took about 90 seconds and it looks like Airbnb.

Airbnb for courses. Yeah, yeah, exactly, exactly.

I mean, how much time and effort and care and money does Airbnb pour into design? Oh, I mean, so like, in all seriousness, why try to become a good designer when you can just do stuff like this?

I think for, for a lot of applications. Right. I think that you don't need to for Bolt. Like we, we spent time on the UI ux, but as I mentioned earlier, like we started Bolt's development in July and we launched in October. So there's 90 days. Right. And we poured log into the UI UX of it to the degree you could. But we didn't even have a mobile responsive view for Bolt and we didn't have it until the beginning of this year. Like we went to 0 to 20 million ARR without a mobile responsive view. Right. And I'm not.

There's a lesson in there. There is just friggin launch.

Absolutely right. If the value you're providing is actually insanely great, which it should be like, that's the, that's what you have to do. Right. There's all these other things that you should do that people were yelling. It's like, how do you not have a mobile responsive, you know. Well, we made cut corners somewhere to get it out online.

Yeah.

And for a lot of applications, folks build. It's like the design of this, what comes out of these models is excellent. And it's. Why would you go and spend cycles on something that is usually kind of, you know, somewhat ancillary to the core value you actually have to go and provide. Not that you should ignore design at all or something, but there's a certain bar you have to hit and unless you're doing something that is design specific, and that's one of the key reasons, you know, that your app is going to win and is competitive or something. Right. And those are few. Like there's not a lot of those, you know, it's. You should be just be focusing on solving the end user's problems, you know.

Do you have any other, like, prompting hacks, like shortcuts, like, talk to it this way or say this or this phrase or. I'm very curious about that.

I encourage people to just, like, you know, throw vibes at it, especially for design. And, like, you say, hey, like, go and you make it beautiful. Like, make this thing pop. Like, you know, add a page that, you know, really focuses on X, Y, Z. Like, sometimes it's just better to let the AI take a crack at whatever you have to do versus you trying to, like, really spec out exactly how it should be. Because I think that some people will just get really, really, really detailed and then what comes out is, is, you know, will. Will, like, work, but it's, you know, they're spending all this time on things where, you know, in these models that they're trained on, you know, the best, you know, apps in the world. And so you get to kind of leverage a lot of the best practices just by saying, hey, we need a pricing page. Can you just kind of go whip something up with these three prices? And it's like, what comes back?

Like, yeah.

Oh, okay. Yeah, that's great.

Yeah.

Like, I went to thought made out that way, you know, so in a.

Previous podcast, I suggested people do this, and I'm curious what your thoughts are on this. Let's say you have an idea for a website and you pull up Bolt alongside four other direct Bolt competitors and you just start prompting all five at the same time. Do this, make this website same prompt in all five browser windows. What result are you going to get from Bolt? Either good, better, you know, different, worse, whatever. It doesn't matter. How is Bolt going to be unique when you're prompting five different, you know, vibe coding apps the same way?

Totally. Yeah. I think we've got pretty. The best, like, design output at this point. Of all the things out there, we've spent a lot of time on that because, like, again, it's like we're trying to get you from idea to business, you know, as. As fast as possible. Right. And in less than five minutes. And so if you're having to spend a ton of time on just making the website, like, look like something that could actually sell a product, that's a problem. And so that's the first prompt. That is what you're going to see with bold on the second prompt.

Save time.

Absolutely.

Just on the design port.

Yeah. And what comes out is stunning. Right. And. And on the second prompt, you're actually, you have a real backend and payments, etcetera, Integrated with The thing, right. That's what you're going to really notice and the speed of it too. Like that's like we really have. And this, a lot of it comes from the seven years we spent building this technology. Like so this is like how we're running this in browser here by the way. This is the seven years of technology we spent before Bolt on our previous product. And this is no one else has the. A pure in browser based dev environment. All the other providers have to use Cloud VMs for every time you open up one of their things. So you will notice there's they takes a long time to boot those things often sometimes they won't boot at all because you know, the server is borked or whatever, blah blah, blah. With Bolt, every single time like when I refresh this chat window, this is booting a fresh operating system on my device inside of my browser, booting up, installing it, getting it running, et cetera, you know, within seconds. And that's. And you know on the other guys this can take minutes to do that. So that's like, I think those are the things that you'll tangibly notice when you kind of go compare these things against each other.

How viable is it to build like a production ready iPhone app with Bolt?

You can do it. Absolutely. Yeah. We have an integration with Expo which is like a react native thing. And so on the home screen if you just say like, you know, build me a mobile app, mobile app that looks like Spotify or whatever, this will actually, you know, get a real native application spun up that using just a. You can take a picture of a QR code, you know, that'll pop up here. But you can actually run that app on your device and as you are prompting it'll up like hot reload it, you know, in your hand in real time. Right. And so, so that's like we're one of the few providers that can actually do this again. It's because we're running all this stuff locally on device that we can let you build mobile apps and not just kind of like specific types of web apps or something.

Yeah. What type of functionalities are still difficult for Bolt? Like let's say I wanted to build a WhatsApp clone and I wanted people to have encrypted chat between, you know, anywhere they are on the earth. Is there any like sort of feature or product or database or something that like Bolt is still kind of young at but improving?

That's a really good question. I think for all of these things there's kind of A complexity curve of when you start adding in functionality that's reasonably complicated and not super common, you.

May start like an Uber clone.

Like an Uber clone.

We're using like GPS data and stuff.

Yeah. And especially with Uber, because it's like a lot of what makes Uber work is the dispatching algorithms and software behind. And obviously there's tons of humans that have to support that operation. Right. And, and that's where things get a lot more complicated. Because now you're not just talking about, you know, really a like what, you know, software is like a crud based application. Create, read, update, destroy, where you're just, you're putting, you're uploading things to a database, you're downloading. Now you're dealing with real time systems where you have to have, you know, servers that are doing sophisticated algorithms that you have to invent to, you know, be matchmaking drivers to, to passengers, et cetera, et cetera. It's like things of that nature would be very difficult to do with Bolt or any tool for that matter, without having knowledge of it. But it also depends on the use case too. If you're building an app where you only ever have 20 drivers to be assigned to different orders.

Okay.

You know that, that actually would work like, you don't need some crazy sophisticated, you know, mapping algorithm or something for that. Right?

Yeah.

And you can just be doing a more direct matchmaking thing. And again, this is like where you're spending a, like as people build their projects up, they're spending a lot more time in discussion mode in Bolt because they're going, okay, I need to go add this functionality, blah, blah, and then, and discussion mode will come back and go, okay, we're going to have to go do this. That here's kind of the trade offs, like, what are you actually trying to do here? And you know, really honing in on the plan. Which is exactly.

Yeah.

You know, when you, when engineers write software that, you know, they don't just kind of go, oh, we're building Uber great. And then just start coding. It's like there's, you have to do a lot of planning to kind of really think through, okay, how is this going to work? What's the actual, you know, intended, you know, experience we're delivering here, et cetera. Right. And so the same thing is true with this and coming back to earlier in the conversation where it's like knowledge of building software being important. This is why. Right. It's if you want to build more complicated things, but it's not even like Even knowing how to code, it's also just understanding at a high level. Okay, what are like the tools and services that, that you can use to solve different types of problems? How do you glue these things together? Those sorts of things. Right. And yeah, what we see a lot is people will come here, they will, they'll build their first website, which is something that's maybe, you know, maybe a personal website or it's like a course selling website, something like that. And then they're like, oh, I want to go build something more complicated and. And then they have to go learn some stuff along the way, which again, discussion mode in our discord communities help with, et cetera. You know, folks, like, there's something really powerful about being intrinsically motivated to go and learn about this stuff because you're like, oh, I know I can build this.

Yeah.

And therefore I'm going to go dig in and like, you know, understand how I would do it versus if you don't even know how to code, like, it's really hard to kind of get past that point. Right. And so here's a mobile app. So this is like previewing it. You saw this QR code. So if you scan this on your device, you can actually like run it.

Oh, wow.

But yes, this is Spotify, our Spotify clone. And you know, wow. Et cetera. But that's zero shot, you know.

Yeah. I've never even attempted a mobile app with any of these vibe coding apps, but I want to.

Yeah, it's cool, it's fun. I mean, it's. Yeah, it's so different to be holding this thing in your hand and then seeing it update live as a native app. It's really, it's something else.

This was amazing. I know people are going to love it. Where can they find you? We know where to find Bolt. Bolt New. What about Eric?

Yeah, you can find me on X. I'm Eric Simons over on X. And then our X account for Bolt is Bolt New, like spelled out D O T B O L D O T N E W. Yeah, that's where you can find us.

Okay, thanks for watching. I hope you enjoyed it. Share with a friend and if you want to use Bolt, I'd love it if you use my affiliate link in the show notes.