🎧🍌 Zero to $20m ARR in Two Months: Inside Bolt’s 7-Year Journey to Overnight Success | Eric Simons, Co-founder & CEO
Behind the scenes of hyper growth, almost shutting down months before launch, living on $1/day, building a community around your product, and how AI is changing software forever
Bolt is one of the fastest growing products ever. But they almost shut the company down last summer.
This latest podcast episode with Eric Simons, Co-founder and CEO of StackBlitz / Bolt, goes inside the journey that eventually led to Bolt, enabling anyone to build full stack apps from text, in the browser.
Bolt launched in October of 2024 and grew to $5m, then $20m, in the first two months. But Eric will be the first to tell you it wasn’t an overnight success - the product didn’t even work the first time they tried building it in February of 2024.
We go behind the scenes of the seven year journey building the tech that eventually led to Bolt, what changed between February and October, how to avoid distractions, being capital efficient, living in a frat house for $100/month, and squatting inside AOL’s headquarters for $1/day when he was 19.
Eric takes us inside the weeks after Bolt’s viral launch, figuring out a new business model on the fly, his strategy for fundraising and PR, why you should open source your code, Bolt’s playbook for building a community around the product that enabled their rapid growth, and how AI is changing software forever.
It's a great conversation around capital efficiency, grit, determination, and the future of software.
Let me know what you think after listening!
Timestamps to jump in:
0:00 Intro
2:24 Building full stack apps from text, in your browser
4:19 Running an operating system in the browser
11:48 Why Bolt failed the first time, almost shutting down last summer
20:18 How Bolt went viral from one tweet
28:33 Differences between ChatGPT, Claude, Cursor
39:38 Why AI code gen changes the software world order
42:01 What happened inside Bolt going from zero to $20m ARR in two months
47:32 Not sharing fundraises publicly + his PR strategy
58:57 Why the team never gave up for seven years
1:01:07 Living in a frat house for $100/month
1:04:07 How to be capital efficient
1:09:00 Living on $1/day in AOL’s HQ when he was 19
1:14:01 Inside Bolt’s Series B
1:21:03 Bolt’s hiring and product roadmap
1:32:58 Creating a new inference-based AI business model
1:38:07 Eric's playbook for building a community of users
1:44:05 Why you should open source your code
Referenced:
Follow Bolt on X
Check out Webcontainers
$0 to $4m ARR case study with Anthropic
Bolt’s open source code
Open roles at Bolt / StackBlitz
Eric’s favorite cafe, The Lighthouse SF
Lady Gaga’s “one person” montage
Living inside AOL’s HQ at 19 years old
Bloomberg coverage of StackBlitz / Bolt’s Series B
Joel Spolsky’s blog
Find Eric on X / Twitter and LinkedIn
👉 Find on Apple, Spotify, and YouTube
Transcript - (read on Rev)
Find transcripts of all prior episodes here.
Turner Novak:
Eric, welcome to the show.
Eric Simons:
Hey, thanks for having me. Excited to be here.
Turner Novak:
I'm really excited to have you. I feel like there's kind this lore around you and Bolt, went from zero to 20 million in revenue in two months after launch. I think that's probably the most insane trajectory of any startup ever. It's exciting to have you here.
Eric Simons:
Yeah, yeah. Thanks. Thanks for having me. Yeah, it's been a busy couple of months, that's for sure.
Turner Novak:
Really quick, for someone who is not familiar with you or with Bolt, or with StackBlitz, just what is it really quick?
Eric Simons:
Yeah, yeah. I think Bolt is the thing ... Is our main thing at this point. Bolt.new is the domain. You go to Bolt.new. It's very simply, it's just a text to app sort of tool.
Think of it like you go to ChatGPT, imagine if you could say, "Build me a full stack web application." Not just like a static landing page. You can certainly do that with Bolt, but you can say, "Hey, make me a CRM." It'll actually be able to build a full stack app inside your browser with this chatGPT sort of experience, all the way where you can just click deploy and it's live on the internet.
I actually need to make an asterisk because today we actually just rolled out the ability to do this for native iOS and Android apps. You can go to Bolt and say, "Hey, I want to make a dog walking application on iOS," and hit enter and it'll actually spin up this React Native-based tooling. You can preview it on your phone and deploy it to the app store.
Turner Novak:
Is this like it takes a day? Is it five seconds? How quickly do you get it?
Eric Simons:
Yeah. That's the magical thing is you're talking less than 60 seconds from when you hit enter. Whatever your idea is, you get a response, and it's building it and it's rendering entirely inside the browser.
A lot of this, for us, it's the fastest experience that's ever existed for this because we've actually been working. Before Bolt, we were building in browser dev environments that are built with WebAssembly.
We designed this, the operating system we made, from scratch. WebAssembly to boot in hundred milliseconds. When you go to Bolt and you say, "Hey, I want a full stack web app that's like a CRM," we boot an operating system in your browser tab in 200 milliseconds, like instantly.
Turner Novak:
What does that mean if I'm not super technical, not an engineer, which by the way that's me. I'm kind of like ... I maybe think I get it, but I'm also like, "I don't know if I" ... What does that mean? You're creating an operating system in the browser?
Eric Simons:
Yeah. By the way, for everything I'm saying here, doesn't matter if you understand it, because that's the beauty of this thing, right? Is that you can come ... I think 60, 70% of the people coming to Bolt are not developers. They're PMs, they're entrepreneurs, they're VCs. My mom used this thing to build her first website at 73, right?
That's the power of this. It's the simplest interface in the world. It's a text box. You just tell it what you want it to build, but when you break under the hood, okay, well, if you want an AI to actually build a full stack web app or a mobile app, how do you do that?
If you look at how do developers do that, they have to set up developer environments on their computer, which dev environments are basically just tools you install onto an operating system and run.
The problem, if you wanted to have an in browser experience for this stuff, is you need an operating system to run it. Browsers don't give you access to your local operating system.
Turner Novak:
It's kind of lik you would have to download a program that just runs on your computer and that's what you use, but you can actually create that in the browser or recreate that experience in a browser.
Eric Simons:
Bingo. The example ... The closest technical analogy to this is Figma. Back in 2012, my co-founder and I actually had the good fortune of bumping into Dylan Field and Evan Wallace when they were starting Figma. We got to see that that story play out for them.
Their first pitch for Figma when they went and raised their seed or their A, or whatever it was, they didn't even have a design tool. They just had this 3D demo of a ball dropping into water in a browser, and that was it.
The pitch was like, browsers have gotten powerful enough with web GL, and what was the predecessor of WebAssembly, to actually do complex graphics rendering entirely on device within the browser sandbox.
Because that's possible, we think we can build a design tool on this. We're going to have to write an entire rendering engine from scratch that's built around just Web GL, because that ... None exists. You can't take what just works on local, like what Photoshop or whatever uses. You can't just port that to the web. It'll take forever. It'll be slow, whatever. You got to write it from scratch. It'll take years to do. If we do it, we can make this design tool. Obviously we've seen how that story's kind of played out.
That was the key insight we had, my co-founder and I, back 2016, 2017. There were a whole bunch of APIs that landed in browsers. We were like, "Holy heck. It actually seems like theoretically possible for the first time, where you could actually write an operating system from scratch that could run entirely inside of a browser, that if you could do that, what that means that you can actually run full development environments for building web applications, and et cetera."
That would enable the same sort of ubiquitous access that Figma has allowed for design for actually building software. We call that technology web container. That's really the crown jewel of how our stuff works.
Turner Novak:
You sort of quote unquote invented that sort of like, what is this? Nine years ago at this point? Eight years ago almost?
Eric Simons:
It took us four or five years to really pave that technology out, so it was kind of a classic deep technology play, right? Where a lot of the risk was front-loaded to can it be done, because there's no prior art. You go and talk to subject matter experts that know stuff about browser engines and they're like, "This isn't going to be possible." Right?
Turner Novak:
Did you get a lot of that back, like, eight years ago?
Eric Simons:
Absolutely. It was back when we raised our seed. Sarah Guo was the one who ended up leading it, but it was really tough to get people to believe, because they would hand this to the SME at the firm or whatever. I imagine those guys look at it and go, "Eh. We don't think this is possible. If it is possible, how big of a deal could this actually" ...
This is classic startups, right? At the earliest stages, it's about finding those one or two people that are like, "Okay, I'm going to take a flyer on these guys and believe here."
It took us years to do. Fortunately we got to work and it just works incredibly well. Yeah, this is the same. If you look at all the productivity apps on the web that have ever grown to be used by hundreds of millions or billions of people, it's always the same technological compute approach where ... Like Google Docs. If you look at Cloud IDEs ...
Turner Novak:
What's an IDE, for somebody who doesn't know?
Eric Simons:
Interactive development environment. It's like the thing that developers are writing code in, right? Back in 2009, I think Cloud9 was the first cloud-based IDEs. In your browser, you could go to a IDE that just loads up.
The thing is your computer is not ... In that model, your computer's not actually doing any work, because every IDE, cloud IDE you're going to, is spinning up a VM in the cloud that your browser is connecting to.
The problem with that approach is there's a ton of latency. Every keystroke has to be sent to the server. You have to get a response from the server. Someone's got to pay for this VM, this virtual box that you're connecting to.
Turner Novak:
For each developer.
Eric Simons:
For each developer that wants to do this. The thing about something like Google Docs that has like a billion users, right? Can you imagine if they had to spin up a VM for each user every time you opened a Google Doc, right? There's not even a billion VMs on the planet, but there are a billion end user devices that are connected to these things.
If you can leverage the end user compute, there's zero cost to the provider, Google or Figma or us, right? There's no servers we have to spin up.
Turner Novak:
Oh, yeah. You're not paying for AWS to run and store. It's literally just on their device in the browser.
Eric Simons:
Bingo, because it's using your CPU, your memory, so there's no latency. It's a way better experience, right? You can have an incredibly permissive free tier because you don't have to worry about people going and using your free tier to mine Bitcoin on your virtual boxes and stuff.
Turner Novak:
Oh, does that actually happens often?
Eric Simons:
All the time. They're mining Bitcoin, they're hosting phishing sites, they're using it to DDoS know other sites or whatever using your infrared on your IP. Then you get banned from AWS. It's a whole whack-a-mole game.
All these things that have been in the space now and previously, other than us, have had this huge issue where they eventually just have to nerf the hell out of the free tier because it's this never-ending cat and mouse game.
The problem is if you nerf the free tier, you're limiting how fast ... The fastest way to grow is word of mouth. You want to travel at the speed of talk, and that's me sending you a link to a Google Doc or a Figma, and then now Bolt, and it opens up. You don't have to log in. There's no credit card. There's none of this stuff that you have to do, and it just works, right? So that's the magic of the experience, which is really the key insight that the company was founded on.
Turner Novak:
Interesting. Well, I guess speaking of the magic, so you officially ... Bolt, I think you tweeted it, it was sometime in October. You were like, "Hey, this is a thing." I think you probably started working on it a couple months before that. Just take us inside the last couple months, what's been going on in your world?
Eric Simons:
Yeah, yeah, yeah. Totally. Yes, we had the idea for Bolt like a year ago actually. We built a prototype back in February of '24.
Turner Novak:
What happened with it?
Eric Simons:
The frontier models at the time weren't good enough to actually make it usable.
Turner Novak:
What does that mean? They could not ... The output that I was getting on the visual interface it was just kind of sloppy, didn't work properly or ...
Eric Simons:
If the code was even correct enough to even execute, because you put in a prompt, you get an error, and you click fix it, but you just keep getting errors, so ...
Turner Novak:
Probably say like, "Fuck this. I'm not ... This is a waste of time. It doesn't work."
Eric Simons:
Waste of time. Imagine charging people money for that, right? It's like, who the hell is going to put in their credit card? The value prop wasn't there, right? It was too sloppy, it was too inaccurate. The designs it outputted were not good.
Then in May ... We put that project on the shelf. We spent two weeks building a prototype and we're like, "Okay, cool idea, but this stuff isn't there yet." Back in May, we got an early preview, the Sonnet stuff. We were like, "Okay, this changes. We think this might change everything." Right?
Turner Novak:
What was the difference with Sonnet that you'd noticed from prior versions?
Eric Simons:
From the first time we tried the API, it spat out code that not only worked for Zero Shop but was beautiful. It was like, oh my God, this is insane that this is so good.
We spent June kind of testing out the API. We were like, "We have to do this." Then so July 1st we greenlighted Bolt, and then we launched in October 1st. It's like what? Three something months we spent building the thing. It was like 90 days I think we spent building Bolt.
We had no idea what ... The other thing too is a lot of developer tool startups, us included, especially the cloud IDE sector, it's not been easy days in this. We were talking about spinning down the company at the end of last year.
Turner Novak:
Really?
Eric Simons:
Yeah, the Bolt was the last ... We tried a whole bunch of different ideas. I'm like, "It's really cool technology we made here, but we have to" ...
Turner Novak:
Make money, I guess, commercialize it in some way. We need a product. Yeah.
Eric Simons:
Exactly. Deep technology plays require commercialization for them to make sense. We're seven years in on this thing. Developers loved our tool, but it's getting them to pay money for it, getting them out of their local environment just was really tough. We were like, "If we can't figure out how to make this work we're going to have to start figuring out how to spin things down, whether that's get acquired," or whatever, right?
Turner Novak:
Which you probably had ... There's probably some fundamental value of we've created this stuff. Someone would've wanted you probably, right? The team or ...
Eric Simons:
I would hope, right? Here's the thing is the tech is really cool, but it's like there was no clear breakthrough application that was obvious ... Because heck, that's why we were going to try and get acquired or whatever, right? It's like we can't figure out how to make money. It's like, some other company ...
Turner Novak:
Yea, you have to commercialize it somehow
Eric Simons:
They're like, "Oh, well, we'll totally buy you and this, because we think it's like if we can't do it" ... That was this challenge. We were like, "Shit, we cannot figure out."
Turner Novak:
I did not realize all this.
Eric Simons:
Yeah. I think it got out in the press a little bit with the fundraising announcement recently, which I have no shame talking about it because this is startups. This is it, right? This is like, you kind of backs against the wall. You got to try stuff and hustle.
Turner Novak:
Yeah. A lot of this stuff is revisionist history where somebody is some genius product person that crafts a perfect thing, and the company works over.
Airbnb was classically making cereal to just get money to keep operating, and Instacart, I think Apoorva tried 40 things. I forget the number, but there are stories. He literally just tried tons of different ideas. You finally get something to work. That's less exciting to talk about, way less glamorous.
Eric Simons:
Totally. To me, it's ... I'm actually the most ... I love talking about that because, to me, last year was this ... Was a year of great challenges.
I think that just on multiple fronts. Last year I had my firstborn, my daughter, in April. We had to figure out what we were going to do with the company and try and turn things around, or whatever.
Both of those are ... I think you had said that you have kids too, right? Anyone else listening to this that have kids know exactly what I'm going to say here, which is first months of life, first year of life is ... I think I imagine it goes beyond that.
Turner Novak:
It kind of gets easier, but in a different way, I guess.
Eric Simons:
Yeah, especially the first months I think are ... Then going into it, I was ... I had talked to enough of my friends that I'd had kids where it was like, "This is" ... Basically kind of prepare for war, right? You have to go into it really understanding that it's going to ... To be a great parent during those times takes a lot of time. You're not on a lot of sleep. It can be very stressful.
One morning, maybe within a week or two of my daughter being born, I just woke up, it was in April. I was just like, "I'm going to do an Iron Man this year." I've never ran a marathon or anything in my life. Then the one I signed up for was in October, so it was six months or something after.
Turner Novak:
Okay. Piecing together the timeline here.
Eric Simons:
Yeah. Yeah. For me, it's kind of like ... I guess I've always found when I'm dealing with great challenges in my life, it's having equally great physical challenges helps balance me out.
Last year was nuts because it was these three huge things. I had no idea what goes into training for an Ironman. I found out most people spend years training for these things, so the Ironman actually happened three weeks after we had launched Bolt, when everything was just going up and to the right. I was like, "Yeah, the timing here is insane."
Turner Novak:
You probably added a million in revenue while you were doing the race.
Eric Simons:
Yeah, yeah. I think probably around that, honestly. For a day ... Back then those days was crazy. Yeah. Last year was this kind of wild gauntlet.
With Bolt, we put it online. Our expectation going into the thing was like, man, it would be cool, because at that point, the company was at less than a million of ARR after seven years of work. We're like, "Man, this would be great if we could get ... After a couple of months we were at another couple hundred grand to the ARR. That would be" ...
Turner Novak:
Because that would've been ... After just years of slogging through, that would've probably been huge for morale of just like, hey, it looks like we finally have something that is kind of working.
Eric Simons:
Totally. Absolutely. Yeah. I tried to temper the team's enthusiasm and then their expectations around, hey, what should we expect here?
This guy, Pete, on our team, our dev ops guy, the second day after we launched, I think, I can't remember how much ... What the ARR ad was, but it was like, hey, we added like 60K of ARR or something. He was like, "Man, this is crazy. This is going to be huge," or whatever. I was like, "Listen, launch day ... On launch day, people are excited. They're going to put in their credit card."
Turner Novak:
Yeah. Half the people are going to churn. It's like, it's all just VCs trying the new products.
Eric Simons:
I've seen this movie before the past 15 years of entrepreneurship. I know this movie. I know how this is going to ... You know what I mean? Which is I think the rational perspective. Just if you look at the sheer hits on bat that entrepreneurs take, that's typically the tech crunch of initiation and then whatever.
Turner Novak:
What did you do? How did you go into this launch? It wasn't even a big thing. It was like, oh, we built this. Let's just ... You tweeted about it, right? and just kind of shipped the tweet and ...
Eric Simons:
Yeah, we put out a tweet thread. That was it. There's no blog posts, there's no press article. Nothing. It was just a tweet thread.
I think we put a good amount of effort into our tweet threads. Tomek on our side makes these awesome videos that really do a great job with the Screencast showing how it works and stuff.
Yeah, it was a tweet. It ended up just going viral. I can't remember. I think it's, I don't know, maybe over a million views or something like that. Millions. I don't know, but ...
Turner Novak:
Did it escape onto other platforms? Did it hit Hacker News? Did it hit Reddit or any YouTubers, influencers talking about it or ...
Eric Simons:
Yeah. Hacker News, Reddit. I can't remember. I'm sure someone might've posted it on either of those. It wasn't from us at least if it did end up there.
Turner Novak:
That's true virality then, when it's other channels. I think that's how I've heard people describe it before is if it starts to escape more than one channel, and you have other people starting to talk about it, I think that's technically the definition of viral.
Eric Simons:
I think so. Yeah. I've done so many launches of things over the years where this was the first one where it just continued to grow outside of our ... What we were putting pressure on in a meaningful way, right? We had lots of YouTubers and people on TikTok that were picking the thing up.
Turner Novak:
Do you know why it resonated so much with people and why were people using it and getting value at it, and why were so people talking about it? Could you pinpoint that?
Eric Simons:
Yeah. Yeah, totally. I think up until that ChatGPT moment where you try ChatGPT, it was like, holy heck, this is crazy, right?
Turner Novak:
This is really, really cool. Yeah.
Eric Simons:
Yeah. I think for a lot of people that tried Bolt, it was kind of like that for the first time since ChatGPT. That's what a lot of people said. It was the first time they had the ChatGPT moment since ChatGPT, because it's really ... It's mind-blowing when you just put in an idea for a web app and you just see the thing work, build it, and it's ready in anywhere from 20 to 60 seconds, right?
Then you just keep prompting it. You're like, "Oh, I don't like the background color. Make it this. Oh, here's an image to use."
Turner Novak:
That's literally what I was doing yesterday. I was trying it out. It was just like, make me a website for the podcast. I couldn't think of what to do. It was midnight. I was like, "I really want to try this out before we talk and try to go through it all." Yeah, I was like, "Change the background color."
The first version was not what I would've picked, but it's like you got something there. It's like, "Oh, it's a website." You're like, " I just literally said make me a website and it made a website."
It kind of reminds me of ... If you've used any of those website builders like Squarespace or, I don't know, any of these kind of simple ...
Eric Simons:
Wix. Yeah.
Turner Novak:
Yeah. You get a template, and you change the template, but it's literally you just talk to it and you're like, "Here's what I want you to change," and you just text it, like send someone an email.
Eric Simons:
Yep, exactly. When we started seeing this, the traffic just increasing day over day over day for weeks, we were like, "What is going on here?"
The people coming in are not ... We've been a developer tool for seven years, right? We have people coming that clearly are not developers using this. They're running into all these problems because we didn't really initially design this thing in mind with, hey, this is ... Non-developers are going to beat the crap out of it, right?
We're trying to fix stuff, whatever, but it was interesting because, yeah, what you just said. I didn't really quite realize what you just pointed out there, though, which is why this is so much better than Wix or Squarespace, why people are using it.
There's this guy that ... Yeah, I think a couple of weeks after we launched, I put out a tweet. This salesman from Dallas, not a developer, replied to the tweet. I can't remember what I had said, but he said, "Thank you and your team so much for making Bolt. My daughter's got a medical condition, so whenever she travels she has to get donors wherever she goes ahead of time lined up. I made her a site where she can send that along and find folks to donate once she travels."
It was a very touching use case of Bolt, but at the same time I was like, "Should I tell this guy that Wix and Squarespace exists?" We're not the first people to allow you to build websites and put them on the internet.
Really, it really bothered me. I didn't understand. It finally clicked with me because it brought me back to when I used Squarespace. The only time I've ever used one of these things was back in 2021 to make my wedding website. I had originally ... I'm a developer, so I was like, "Honey, I'm going to make the site myself. I'm going to go in the backyard, chop down a tree."
Turner Novak:
Build the table myself.
Eric Simons:
Yeah, exactly. I spend a Saturday building the thing out. I'm like, "Dang, this is actually" ... It's a lot of work to me. This is '21. There's no AIs. It's like I'm writing line of code, right?
Okay, next weekend I'll come back and do it. Didn't. My wife was like, "Here's a Squarespace account. Go use this thing." I did. That's what I found out. The interfaces for these things are complicated as hell, these site builder things.
Then you compare that against Bolt, which is ... It's a text box, right? You go to Bolt.new, it's like ... I think it's one of the things people love about it is that there's no marketing landing page, there's no all this crap. You go to Wix.com, it's like, dude, there's a bajillion things going on there, right? It's like ...
Turner Novak:
You said it, not me.
Eric Simons:
It's a text box. You tell what you want and then it gives you something back, and then you tell what you want again. You're like, "Yeah, not that. Here's the text."
Again, to the point where my mom made her own website for the first time at 73 years old. I think that that ... We see just this range of use cases of folks building for personal sites for their small businesses. Either there's a bakery that I know they use it to build their websites instead getting a web contractor because they were quoted like five grand for a bakery website. It's like, they bought our $20 plan, have a bakery website.
Turner Novak:
It's because instead of ... Then if you are working with an engineer or dev shop, whatever, you're probably typing up the specs and saying, "Here's what I want. Do this, do this." They make it. You come back and you're like, "Oh, actually can we make this change and use these images for this thing?"
It's basically just taking what you would probably communicate with an engineer and it just does it for you. It does the work of the engineer just very quickly.
Eric Simons:
A hundred percent. A hundred percent. The folks that are billing really ... The people have launched startups and made money off their startups off of Bolt at this point.
The people that are doing it, it's ... The people that are most successful with this product are people that have actually had to interface with development teams before. They're PMs, they're entrepreneurs, because that's exactly what they're doing. They're talking to the AI as they would one of the developers. They have a PRD, they give it to Bolt, and say, "Here's exactly what I want built. Oh, this thing, this little problem's happening, let me Google it. Oh, what if we use this API instead?"
You know what I mean? It's like instead of just writing a Jira ticket, they're hitting enter on it into a text box where it actually is coming back.
Turner Novak:
Yeah. Then you significant ... Instead of waiting like a week for it to ... Someone to get to the ticket and build it, it's like, to your point, 60 seconds.
You kind get updates too. It'll say, "We're doing this now." Then a new line will be like, "We're doing this. There'll be ... Almost like a check will appear next to the certain things, if I'm remembering right. You kind of show people.
It's like when you're using ChatGPT and it's like, "I'm thinking. Here's what I'm doing. This is the part of the process I'm in right now," of the model going through and searching or finding or constructing whatever you need.
I think it's interesting the way it's so intuitive. I guess one question though is couldn't people sort of use something like ChatGPT or Claude to kind of do this? What's kind of the difference? If I'm thinking ... If I'm listening this like, "Ah, Claude, I've used that to build something before," what does Bolt do differently?
Eric Simons:
Yeah. I think in Claude they've got artifacts, which is they kind of let you build little mini web apps. I think ChatGPT has a similar thing. I think it's called Canvas, if I'm not mistaken.
I think for small things those things can totally work fine. If you're trying to build an actual website or web application or mobile app that you want to deploy and have a domain on, or even just, heck, just deploy somewhere, right? That's where you need Bolt but again, you need to be able to install third-party packages on demand. If you tell it, hey, have Confetti pop out on the screen when the user signs up for my app, that's to install some NPM library for the Confetti thing. These other tools can't do that because they're limited. They're not running an operating system that can just do those sorts of things.
Turner Novak:
So you've just built the environment and you made all those kind of integrations with all the other different, to your point, the different libraries that you might need, is that why?
Eric Simons:
More or less, but it really comes down to, if you want to build real software versus just what Artifacts, or Chesapeake, Canvas can spit out, all real software is built in dev environments. Dev environments are operating systems with tools and package managers and libraries and stuff. You need to actually be running an operating system that can install those tools, install the dependencies on demand, and that's, under the hood, really, the big difference.
I think everyone wants to do this sort of stuff we're doing right now, but the two routes you have to go do it are, one, go have Cloud VMs that get spun up for every user, and if you look at Anthropic and OpenAI-
Turner Novak:
Is that what they do?
Eric Simons:
They're not, because ...
Turner Novak:
Oh, because it's probably too expensive, or logistically impossible, or?
Eric Simons:
Yeah. Expensive, security. Yeah, exactly. If you look at what's the number of VMs that one DevOps guy can keep an eye on, it's maybe a couple thousand at best. But it's like, okay, how many MAUs does ChatGPT have?
Turner Novak:
I think it's 500 million. I don't know, maybe that's a little too high, but it's one of the biggest products in the world at this point.
Eric Simons:
The number of DevOps guys you need just to keep an eye on all ... It just doesn't make sense. I think everyone, when you look at Artifacts and Canvas, it's all completely in browser, but they're just baking very specific component libraries and stuff statically, so if you try and do anything outside of that, it doesn't work.
Not that it's not useful, I actually use Cloud for generating graphs and stuff like that, but if you're trying to build a website for your podcast, you're trying to build a CRM, you know, et cetera, that's just not going to work.
Turner Novak:
Yeah. I think another interesting maybe parallel to what you're doing, another company that went from zero to, I think, a hundred million is the public number, I forget the time, but was Cursor. But it seems like it's much more focused on developers and very technical users. Is that a good way to compare the two products?
Eric Simons:
Totally. And there's a lot of overlap between ... 30%, 40% of our audience is developers using this. Developers will come to us for the agentic, hey, I need to go ... especially web freelancers and contractors, the arbitrage, if you search Bolt arbitrage on Twitter, they-
Turner Novak:
Bolt arbitrage?
Eric Simons:
Yeah. People are actually saying this because ... There's a guy, his name's CJ, I've actually gotten to know him a little bit over the past month or two, but he's a web freelancer, and he posted the most legendary ROI score I've seen today. But he had a customer that needed a web app. He spent seven bucks on Bolt in inference cost to make it. And he posted this and someone asked the question I would have had they not, which is, how much did you charge the client, though? $9,000.
Turner Novak:
Wow. That's the highest margin product, outside of committing fraud.
Eric Simons:
Yeah, exactly, without the FBI coming to your door or whatever. But the people that do dev shops, this is changing the game. This is changing the game. Because you can come here, say, hey, this is the name of the costumer, this is what they want. Here's a screenshot of a Figma or whatever, Enter. It takes you from 0.0 to 1.0. And then, Cursor is really great at taking from 1.0 to 1.1, 1.2, whatever. We see that's a lot of developers are using Bolt to really kick things off, go zero to 60 insanely fast, pull that out of Bolt, go to Cursor, build on it, bring it back in the Bolt to add big functionality pieces, flip it back.
The developer audience for us is still really more important than ever. What's just really cool is that we also now are seeing PMs taking their company's code base or whatever, pulling it in, instead of writing a JIRA ticket, hitting Enter into a prompt, it writes the code, they're able to commit changes back. That's the wild thing that's going on with a tool like Bolt.
And to your question, why is Bolt seeing this adoption? What other things would you use rather than do this? The challenge is, for a PM to set up a dev environment or designer, if you go and join Meta or Netflix or whatever, the first four to eight weeks of your full-time job as a developer is setting up your dev environment.
Turner Novak:
Really? That's crazy. I didn't realize it takes that long.
Eric Simons:
Yeah, it's set up your environment, learning how all this crap works. It's been built up over decades.
Turner Novak:
That's true. You have to get familiar with the code base and all that stuff.
Eric Simons:
But even just getting it running, getting it running on your computer.
Turner Novak:
That's crazy. Wow.
Eric Simons:
Again, you look at design tools before, you had to download these things, you had to install them, you had to learn how to use them.
Turner Novak:
Oh, yeah, get all the libraries, the custom fonts, the graphics packs and all that stuff, yeah.
Well, it brings up an interesting question. I don't know exactly how to ask it, but the quality of the code, how can I get comfortable thinking, ugh, this LLM is creating the code and my code is very good. I'm just taking the AI code and there's a PM that's just throwing it into my code base. How does Bolt work and interact with that question, I guess?
Eric Simons:
Yeah, no, it's a great question. I think a lot of what we see is, in companies, the initial use cases that are getting a lot of traction are prototype. Prior to this, if you wanted to make a prototype, if we wanted to make, let's say we're working Netflix or something and we want to make the next great app for Netflix or whatever, next great feature in the streaming app or whatever, the fastest way to do it would be, go to Figma and make a whole bunch of different screens manually, and link them together and all this stuff.
The problem with doing it purely in a design tool, though, is that it's not real. There's always this gap between what's in the Figma and then what's in the code. Because when you actually are using the thing, you get to click on it and feel it. That's when you really know. There's a whole feedback loop that has just happen around that. Often, design does this whole thing, okay. Development does this whole thing, great. Now, we have to do this whole thing between all of these groups, and this whole handoff process has to happen.
But what's changed is that, typically, before Sonic 3.5, it was more expensive to prototype and code, obviously, because it takes a while to write good software and developers cost a lot of money, da da da, so everything design was the fastest way. That's no longer true.
With a tool like Bolt, you take your component design system ... Netflix and every other major company has buttons and forms that every app that they build uses. It's all laid out, their design systems. They put it into Bolt and they say, we're making the next great part of the Netflix's screaming app. It's going to look generally like this, it needs to do this, da da da da, Enter. And you spend an hour prompting the thing, and then you have a live version of it that you can use, and test, and in real time, tell it to make changes to it.
Turner Novak:
And you can send a link to Reed Hastings, the CEO of Netflix and be like, here's the thing, this is our new ad product, advertisers are buying ads to Netflix now, the self-serve ad tool that we're going to ... So then, let's say I get Netflix's self-serve ad tool as an example. Reed Hastings is like, yeah, looks good, approves, we're going to ship it in a couple months, whatever Netflix's timeline is, what would probably happen from there on the product and the engineering side?
Eric Simons:
At that point, they take the code out of Bolt, probably. They scaffold out all the pages in UI they need, they make it look great or whatever. They might even add some functionality in. They take that, they push it through their get repo, they're probably committing things there, whatever have you. They will continue to come back to Bolt if they need to go and scaffold stuff or whatever have you. But in this incentive, the prototyping sense, it's really basically get us zero to 80, zero to 90 or whatever on the application, and then we're going to go and take it from there. That's a common one.
Obviously, we see cases where people are building, again, full CRMs and startup apps, et cetera, all the full stack things from zero front to back. It's totally possible to do. When you look at big orgs, they typically have a lot of very custom things, though, where the primary value is just way faster prototyping, feedback loops, et cetera, with software, where they normally would've had to do it all in design or something. That's one thing.
The other one is, companies have to spin up a lot of marketing landing pages and just public website pages, and those are things where it's code quality. Yeah, you want to be performing and stuff, but you're not talking about ... With Netflix, the code that's transacting you buying and viewing ads, it's pretty important. You don't want to get that wrong, because you have to report how well that thing did at earnings, and if there's bugs in that and you're misrepresenting ...
Turner Novak:
A billion dollar mistake, yeah.
Eric Simons:
Yeah, it matters, right? So code quality, that's really, really high stakes stuff. And a lot of mainline apps typically have these things at big companies.
Your marketing landing pages, it's like, oh, no, it's using a CSS selector that is wrong or whatever, but it doesn't have any tangible impact on anything. You see this delta, but then that allows your marketing teams to just go rip, because it's like, okay, well, they don't need to learn how to use web flow now. They don't need to bother developers for time. They can just go prompt their new pages on the website or whatever have you. At existing companies, those are the big use cases we see.
Turner Novak:
So then, just generally speaking, if you take this click and drag it forward, if you think about six months, 18 months, a couple years, how good is some of this AI cogen going to get and how are some of these roles going to change, a product role on Netflix? We can think of new examples or maybe we keep playing around with the same one, but how might that look differently going forward as it keeps continuing to get better?
Eric Simons:
Yeah. I was actually kind of skeptical up until I saw the Sauna stuff. I was of this mindset, I don't know how much better this AI, not cogen specifically, but AI LLMs in general, I don't know how much better they can generally get, you know what I mean? Where it's really, really replacing workflows or seriously augmenting them.
Once I saw Sauna, I was like, right, this should have been obvious. In retrospect, you can do really good reinforcement learning on software. If you wanted to train an LLM to be really good at law or something, it's not deterministic, meaningfully. There's a lot of things that are subject to opinions of the judge or society at that time. There's just so many variables.
When you look at software and if you have this thing punch out code and you try to run, it either runs or it doesn't. It's very deterministic. And you can do that a crap load of times per second. So you can actually, if you focus your guns on making the best LLM in the world at writing code, it makes sense conceptually that you could make a lot of progress by doing that. And that's exactly what Entropic did and is doing.
Now, the other labs are gearing up to do the same. Specifically cogen, I think that, and there might be other things that I'm just not aware of because they're outside of my specific domain, they also fall into this, hey, this is deterministic, and therefore you can really reinforcement train against it. But I think software is going to be really the first thing that is the breakaway success case here for all of us, because you can just keep making these things better and better, better the more compute you throw at generating, you're permitting every type of app that could ever exist and throw it into the training data.
Turner Novak:
You talk about breakaway, breaking out, I don't know where we are in the story where, a couple weeks in, but it became evident, holy shit, this is kind of working. Maybe we were in the end of October. What is going on on your end as a founder? You went from, I might shut the company down to, holy shit, this is really working. What is going on in your mind, in your world, and for the team?
Eric Simons:
Every day was different than the day before because when we launched it, it was pretty barren. We didn't have basic functionality like persisting chat messages and stuff. It was missing ...
Turner Novak:
It would lose some of the input that you'd given to it.
Eric Simons:
Yeah. It would store your project files, but all of your chats would be gone to people. People were like, what is wrong with this thing?
And then, the other thing too is we launched our SaaS plans, our personal SaaS plans for StackBlitz, which Bolt piggybacked on, they've been nine bucks a month since it's seven years ago or whatever. So when we launched this again, we didn't expect this crazy demand.
People were coming in, and I can't remember how many thousands or whatever people were signing up. And nine bucks of inference doesn't get you ... If you're going to hammer this thing, that gives you a day of usage or something, you know what I mean, if you're really hammering it? And that's what they were doing. But there's no other way for them to upgrade, to get more token tokens to use. Our team, it felt like 300, if you've ever seen that movie.
Turner Novak:
Oh, the movie, yeah. You have an army of users attacking, and you're just trying to deflect the bugs and build the features, yeah.
Eric Simons:
It felt exactly like that. Three months before we launched, we were being crushed by existential void and dread. And then in the months after the launch, and honestly, still today, we've been getting crushed by demand. We're a 15, 20 person team. And I think, in month one, we had it at least a quarter million. In month two, in November, we had, I think, 4,000 people that used the product. And 20 million of ARR. Across 20 people is, you have tens of thousands of customers. We have no support team. It was insane. It truly felt like that scene where you have these 300 guys surrounded by 10,000 people.
And all the folks on our team just flipped a switch and just did the impossible, because we were able to keep the service online. We got to the point where we blew out Entropic's GPU capacity at that time. Yeah, Dario emailed me, he's like, listen, we're going as fast as we can. This is the fastest thing we've ever seen grow on this platform at this point.
It was just, every day, new challenges that were coming in. Yeah, I think, after month one, going into month two, I was just waiting for the morning where I woke up and it just stopped, because every other launch I've ever done, that's how it goes.
Turner Novak:
You just have this cliff and it just nosedives?
Eric Simons:
Yeah. And it just never came and it hasn't come. Once I realized that, I was like, okay ... Up until that point, I guess, in my head, I was like, this is a short-term thing we're growing into, you know?
Turner Novak:
Oh, really?
Eric Simons:
Yeah, because how often does this happen? Look at all the other launches that happen every day, month, or week. I've done this for 15 years. I've been building startups and products, whatever. I've never seen anything like this, personally. Finally, it clicked out for me. I was like, this is the new normal. This is it. At that point, we were like, we need to go raise money, because in order to scale to this demand ... And a lot of the pieces, the story about the guy who built the website for a startup, it's like, oh, wow, this is better. This easier than Wix.
It was like, oh, my God, there's all these no code tools that only exist because, how else would you let non-ticket people write code other than some drag and drop thing. That's gone now. That entire category's gone, because AI can write good code from text. All these became very clear, holy crap, the entire software world order is getting rewritten.
And Bolt was really the first product that came out and really cracked it and really proved ... And now, there's a lot of other companies that are coming in and doing this in various ways, but it was like, this is the moment that we've been working for for seven years.
Turner Novak:
Yeah. Well, congrats. I don't think I've ever actually said that, but it's pretty incredible. Maybe I mentioned earlier. I think it's probably one of the fastest, most sustained, maybe aside from ChatGPT. I think ChatGPT got to 100 million users within a month or two. But it's on a comparable scale of, to your point earlier, it's like, holy shit, this is an insane moment.
Eric Simons:
Thanks. Yeah, all credit to our team because the amount of work and just how high pressure especially the first two, three months was ... Our core folks have been working, we've all been working together for five plus years. I don't know many other ... I think other startups way have been challenged with just the level of trust you needed to make that magic happen.
Turner Novak:
I think an interesting segue and something to think about is you mentioned, you said, okay, we probably need to raise money because we need to figure out how to just manage this demand and work with it to keep growing. We were talking before of, you selectively disclosed, hey, we hit this specific milestone. You said it publicly. How did you decide what you shared about how well things were going publicly?
Eric Simons:
Yeah, it's a good question. I'll try and rewind my brain to the various points. We shared two numbers. One was, after month one, we went from zero to 4 million a month. We actually did more, but we undersold what we had done.
Turner Novak:
You just said four even though you were a little further?
Eric Simons:
Yeah, I think we were at five or something like that, but we didn't want to ... Because again, at that time, I was like, tomorrow could be the day that it just starts going down, so we should probably have a little bit of buffer on whatever the public number is. Entropic did a case study on us where the title of it was Zero To 4 million In Under Four Weeks, or something like that.
Turner Novak:
We'll throw a link to that in the description show notes for people to check out.
Eric Simons:
It was a really excellent case study that they did. A lot of case studies you read, they're just like bland and it's just puff stuff. They did a really dug in on what was interesting here.
Turner Novak:
Wow. 99% reduction in development costs for Bolt users. That's an insane number.
Eric Simons:
Huge. This comes back to that guy, CJ, the $7 inference to $9,000 thing. This guy, Paul, who built his CRM on Bolt, he got a quote for $30,000 six-month timeframe to build that CRM. He paid $300 on Bolt and had it done in two or three weeks. The 90% number is a consistent number we see across the use cases of people that are building real stuff here. Yeah, it was nuts.
And that's when the pieces were coming together at that time. I think we went live with them because it was just, this is crazy. I think, at that time, we were starting to think, maybe we go raise money at some point or something.
Turner Novak:
So for a while, that wasn't really a thought?
Eric Simons:
No. We were just so slammed just keeping stuff online. It was hard for me to imagine stepping away to go raise capital, honestly, because it's such a small team stretched across so much.
And then, by the end of month two, when we saw just an unbelievable acceleration at month two, we went from five to 20 or whatever. I think it was Thanksgiving week, my chief of staff flew out here because we were like, if you're going to raise money, the end of the year thing always sucks because no one's raising money past week 2nd of December, right? So it's either we raise money now, or in the new year, we go and raise money. And I was like, we should just do this now, because I want to hit January first gunning.
For that reason, and also just that zero to 20 million in two months number, I think I went on Sarah Guo's podcast, No Priors, and that's where I think we originally dropped that info.
But yeah, strategically, the first zero to four, I think that if you're really confident in your numbers, it's fine to share this stuff. I think a lot of people are very cagey about it, but I think it's totally fine to share.
On the fundraiser thing, this is an interesting one, though. The company strategy today, we took a very unconventional approach to announcing fundraising stuff where we actually had raised our Series A a year or two before we had announced it. And same with our seed round. We had this thing where we would only announce our funding after we'd already raised the next round.
Turner Novak:
Okay.
Eric Simons:
Back in 2022, I think, we announced our seed round which had actually happened in 2019.
Turner Novak:
Oh, interesting. Okay.
Eric Simons:
And the reason for this is we wanted to fly under the radar until the moment came where it was like, just hit the gas. Because the deep technology plays, the value tends to accrue to the first mover so long as, one, the market opportunities really don't take our foot off the gas. And a lot of founders, I feel like you want to let go and get all this press, oh, we raised a seed round, da da da, you know?
Turner Novak:
I can feel really good to have people texting you and being like, congrats, and you're like-
Eric Simons:
Yeah, you made it, man. But actually, all that is bullshit, right? It's just like, no, you didn't. You sold a piece of your company.
Turner Novak:
Yeah, you own less, and you have to actually move faster now, or else ...
Eric Simons:
Totally. Totally, right? To me, earlier in my entrepreneurial career, there's this ego thing where it's like, oh, how much you raised. When we did staff, I'd gotten to that point where I realized this shit doesn't matter. And if you focus on that stuff, you're screwed, because you're going to be focusing on the wrong things.
So we floated under the radar. Our competitors must have been just like, God, why won't these guys die? They only raised a seed round. And none of them took an interest in what we were doing because they were like, those guys are nothing. They have nothing. They are nothing.
Turner Novak:
Yeah, no one cares about these Cloud or browser operating system thing. It must not really work.
Eric Simons:
A hundred-percent. And the other thing too I'll say to our competitors is, you should just be behind the closed doors doing the hard work and honestly doing it. And the thing is, if you don't go make a whole fuss about raising money and stuff, if it doesn't work out, it's not this whole thing where you have to be ... You didn't go and make a whole fuss about it, right? It allows you to be logically, intellectually honest about what you're doing, the actual opportunities in front of you. And I think it just takes discipline to just set that stuff aside.
Turner Novak:
I was going to say, I had a similar thing with my ... I did a first close for my second fund in May of 2022. It went exactly as you could imagine, trying to raise money in May of 2022. I'm just glad ... I don't even think I've ever said how much money I've ever raised. Banana has this much AUM, because it's been up and down. One week is amazing. I'll get great news. And then, one day, I'll get two companies shut down or whatever. It's not up into the ride.
So I also, similarly, it was almost like his public exam. I kind of became an influencer accidentally. I probably shared a little bit too much in 2021, but it's interesting, this whole build in public concept of, who does it actually benefit? Is it actually helpful to be super vocal about certain things? I'm not sure. And it is tough when you make a big deal about something. I don't know. There's certain things like, oh, man, I wish I had not said that publicly, because it doesn't look good now.
Eric Simons:
Totally. It's a balancing act, right? By the way, we did our A right at the same time, you close your fund, so I know what you're talking about. That was insane. We started that process the week that Putin invaded Ukraine, which is when all the indexes started sliding. Just great timing. This is awesome. Worst fundraising experience ever.
Turner Novak:
What was that exactly? What did you raise on? You raised a Series A. Take me inside that one.
Eric Simons:
Yeah. I don't think I've ever talked about this one publicly, because we had never announced we'd raised the money. And then, by the way, we are so scrappy at the company. We had the majority of that capital still in our bank account when we did our Series B, two years later, whatever.
But at that time, we had enterprises that were adopting a StackBlitz with this enterprise product.
Turner Novak:
So this was just engineering team, you'd have 1,000 engineers at, say, Netflix or whatever, that were all using the StackBlitz terminal?
Eric Simons:
Yeah, they're using it for ancillary use cases and that sort of thing. And what I noticed with that is, I think, during the 2020-2021 era, enterprise companies, they were very hungry to buy almost anything. And then, after 2021, with the market reset, the appetites changed completely, for very good reason.
Things were under a lot of scrutiny for the business value they provide. And the problem with the StackBlitz product is developers weren't using it how Cursor is today being used by developers, where it is the thing they're using. It was a tool they're using to prototype a little thing or whatever. PM's use it because they don't know how to code. There's none of what we're seeing with Bolt where you take frontier AI, merge it with our technology stuff.
But there's early signs of, we had some very good customers that typically, a seed company would not be able to get, just because developers did love our products. We were able to land in a couple of great orgs. So we raised the mind to go and scale that stuff out, et cetera. And we just ran into the lot. Again, it came at that moment where the macro economic environment completely changed. The business we had been in was no longer as high-growth lucrative as it was compared to even six months earlier than that.
And we were in this long grind where we were really trying to invest in our GTM machine and make that stuff work. And we did great work. That's the thing. That's the hard thing about startups is, you have to find fulfillment in just doing good work day in, day out, regardless of the challenges and whatnot, and really try to maintain an optimistic perspective on things even when things are not working, right? And especially if you don't see clear ... what we saw with Bolt. I knew, we didn't have that ... the market wasn't pulling the product out of our hands, money wasn't just getting thrown into this. It's like, "Why are we going to go and spend a crap load? Why are we going to go spend all the money we raised in the series A if we're not seeing that? We need to be working smarter and taking very smart bets and just really going deep and iterating, et cetera.
Once we see it, sure, let's crack it open." But I think it's easier to lie to yourself and be like, "We raised a series A, we just need to go hire a head of this, a VP of this, a da da da da." And then you end up with this bloated company that's not doing anything. Anyways, that's the hard thing. You have to just be comfortable.
It's a lot like training for an Ironman. You have no idea if you're going to be able to do it. On the weekends, if you want to actually do it, you're sitting in a bike saddle for six hours a day every weekend. You got to find enjoyment in that, or at least find peace with it.
Turner Novak:
Yeah. I was going to ask, and this is actually a question from Alex on the Chain Smokers, Mantis, I know they're one of your investors. He was like, "Just ask him how the heck did they stick with it for so long?" Was there a certain thing, and maybe you already described it a little bit, but how do you think you and the team just kept at it despite things not quite seeming like they were working yet?
Eric Simons:
Yeah. I mean, we were getting to the end of the road. First and foremost, it's always been fun to work here. We all have a lot of fun. I think that a lot of people say that, they're like, "We love working here. Our startup is so great."
Turner Novak:
"You have a great culture." Yeah.
Eric Simons:
Yeah, yeah, all this crap. But the reality is the original founding engineers and my co-founder and I, all of us just have become really good friends. My co-founder and I have been best friends since we were... We grew up down the street from each other in Chicago, have been building software and startups ever since. And it just happened, our founding engineers, we met them because they were actually users of the first version of Stack, it was in 2017.
Turner Novak:
Oh, nice. Okay.
Eric Simons:
They've been working with us remotely ever since. That core crew has just really... Then's some other people that joined up through 2020, 2021 as well that we've all just stuck together, we just have fun working with each other. And we're working on cool technology. The technology of StackBlitz, Webcontainer, was mindblowing. When we announced that in 2021, it was an undeniably great technical achievement that was for the first time ever done in a browser. And so people were really stoked about it. And so to me, I think I would say this for everyone on the team that obviously we're running a venture-backed company, but why we're getting up in the morning is just because we like building cool stuff. And why my co-founder and I started the company was that at our previous company, we needed in-browser dev environments, basically. And we looked out there for what was out there and I was like... Our previous company was this educational company, kind of like Pluralsight when we're teaching web development, basically. It's like a lifestyle business.
Turner Novak:
I always love that word, lifestyle business.
Eric Simons:
I'm on a VC podcast, so I can't be like, "It's a startup." Because it was making, I don't know, like 10K month or 20K month. So something that it was not an up and to the right hyper growth, so it was like... You know.
Turner Novak:
Lifestyle business. It's always funny. And so many VCs will be like, "Yeah, that guy, it's just a lifestyle business." They're making $50 million a year. It's way more than you've ever accomplished your build. People just dunking on the lifestyle business. And meanwhile, all your portfolio companies have no revenue and there's turmoil. But I have respect. I have respect for anyone who can make a viable, sustainable company, build something that people want.
Eric Simons:
Absolutely. And I got to tell you, it was such an important experience for Albert and I to do because we had tried some venture-backed stuff before, things through the educational thing. And unlike most startups, it just didn't work for whatever reason, just no product market fit, whatever have you. But it was cool to do things through where we were... We bootstrapped that from, we were broke. I mean, literally just we bootstrapped it from nothing.
Turner Novak:
You had to make money, you needed it to live. Yeah.
Eric Simons:
Yeah. Had to. And to cut our burn down, we ended up living at a fraternity live-out house in Seattle. I don't think I've ever talked about this publicly either, but we were at this frat live-out house living out $100 dollars per room per month sort of thing to keep us afloat. It was pretty ratchet, this place.
Turner Novak:
Both of you in one room, $100 bucks a month?
Eric Simons:
Yeah. Yeah. And so that's the cheapest deal ever. And we were in our early 20s, so it was kind of fun to go to those parties and stuff. It's a frat house, right?
Turner Novak:
Oh, so wait, so you were living in a frat house?
Eric Simons:
Yeah, it was like the fraternity live-out house. There's a big frat house and there's a little house next to it where some of the elders would get to move to or whatever, and they would occasionally rent it out to people. One of my friends was finishing school there. He got us in for $100 bucks month thing, and we're there just teaching web-developing courses while whatever you can imagine frats are doing every day of the week in the background.
Turner Novak:
Was it online teaching classes?
Eric Simons:
Yeah, yeah.
Turner Novak:
Okay. Well, at least weren't having moms come into the frat house learning how to...
Eric Simons:
It made recording our screencasts hard because it's like, "Hey, so guys, what's the party schedule this week? 10 AM? Okay."
Turner Novak:
Yeah, certain time windows. "When are they all going to be hungover and passed out and no activity?"
Eric Simons:
They're like, "You guys want to come out with this?" "No, no. Got to record a Node.js screencast." Anyways, yeah, so we were bootstrapping and it taught us a lot about how to... Man, when you are literally living hand to mouth, you understand how far a dollar can stretch. And so we took those lessons. When we started StackBlitz, we took all those lessons, and then our company is just... The capital efficiency of this thing is nuts. I mean, it's crazy.
Turner Novak:
So how do you be capital efficient? If I'm a founder that, let's say I raised a couple million bucks and I'm like, "Wait a second. I don't have to spend all this money?" How would you advise someone to just say like, "Hey, here's what you should probably think about as things you maybe do or don't need to spend money on, how to make the dollar last a little bit longer."
Eric Simons:
Absolutely. I mean, certainly at the early stages of a company, I mean, typical things that people will spend money on that are dumb, like offices. Not saying go to live out of a frat house, but go rent-
Turner Novak:
Yeah. A nice office can be good.
Eric Simons:
Yeah, rent an apartment or something or do remote. I mean, that's a great way to save money. I think especially in the Bay Area, it can work in some cases, but I mean, there's phenomenally talented people around the globe, and if you're a remote company, that's a huge advantage, right? Because the comp expectations in SF because it's so expensive, the living here is just insane. Don't spend money on PR agencies and crap like that when you're a seed company. I mean, there's just so many things that people spend money on and spend their time on, going and talking at conferences. Especially now that we've raised money, I can't believe how much stuff hits your inbox when you go raise money or whatever.
Turner Novak:
Really? What's been the most surprising?
Eric Simons:
The sheer volume of stuff that people try. And they're all nice people and stuff, but it's like, "Can you go to dinner? We're getting this group of people together for a dinner." And, "There's this conference, you can come talk at it." And some of stuff is like, "Okay, is there business value to going? Is there other people there who would be prospective customers or something?" But 99% of the stuff is just stuff where it would feel good to show up to it, and it's like, "Oh yeah, I'm a founder or whatever."
Turner Novak:
Yeah, I'm LARPing as being a startup founder.
Eric Simons:
Yeah. Yeah, it's all bullshit, right? The reality is if we're going to go and hit our ARR target this year, it's me in front of this desk sitting with my team or a remote company, and just grinding and doing the work. If we raise money, the last thing that I should be doing is going and signing autographs. Not that anybody even wants mine. But metaphorically-
Turner Novak:
Autographs.
Eric Simons:
... going and wasting my time at some dinner of people. It's like, "Dude, if there's people there that want to do business with us, I'll get on a Zoom call with them for 20 minutes." But I feel like a lot of people... I just think there's this reward mechanism that seems to happen where you're in Tech Crunch or this or that, and it's, "Oh, da, da, da, da." 99% of that is crap. I mean, it's not actually how good stuff, great stuff gets built.
Turner Novak:
I guess I'll say thank you for coming on my podcast. I mean, assuming there's been a lot of people who want to have you on, I'm just glad I'm in the top 1%, I guess.
Eric Simons:
No, no, no, no. I love your stuff. I think podcasts are great because it's fun to sit down. We're having a real conversation here. We're actually talking about real stuff and it's the stuff where it's... Just going to networking events and stuff like that. That's the stuff that's not a good use of a founder's time. And you get that regardless of the stage that you're at. When you raise a seed round, you're going to get invited to stuff like that. You're going to get invited to tons of this sort of stuff.
Turner Novak:
So you personally filter of, "Will this be valuable for Bolt? And help with new customers? Maybe from a recruiting perspective, maybe from a capital fundraising perspective, maybe a learn something new or something." Is that generally the framing that you would use to just filter this stuff out?
Eric Simons:
Totally. And yeah, 100% percent. That's always been my... I learned this in earlier companies of, these things just weren't good time investments. There was no ROI pan out. It's hard because some things it's like... But I think I would just recommend founders be default, "No." Because your focus is the most important thing. And let me put it like this. If going to meetups and dinners and all this sort of stuff truly was an indicator of building a great company, there's no reason that Bolt would exist and be growing as fast as it is, because I never go to this stuff. But again, it's very unconventional because I think that it's important, obviously. On the flip side, we run go to market machine here, so running events, you want people to come to them and it'd be cool. And that thing with other companies or they want to build relationships. And so some of this stuff is good and important, but you just want to be selective with your time.
Turner Novak:
Yeah. And then continuing the capital efficiency, living in interesting places, also, this one is public. I guess I'll let you say the story. But yeah, at one point, going back to the early days, there's a point you were living on a dollar a day, you were living in a certain company's HQ, I guess. What was that story? And then how recently was that?
Eric Simons:
Yeah, no, that was a long time ago. It was a long time ago. This is when I first came to the Valley.
Turner Novak:
Oh, so this is 15 years ago, kind of a thing?
Eric Simons:
Oh yeah, yeah. This is like 2011, 2012 I think, and I was 19. And I originally got out to the Valley because Jeff Ralston, he started a, it was basically a Y Combinator, but for K-12 education startups. And they actually ended up pulling it back into YC. And then I think Jeff was the president or the CEO of YC or something.
Turner Novak:
Yeah, I was going to say, he was a partner there, at least, for a while, I feel like.
Eric Simons:
Yeah. And that's how he ended up as a partner there is they pulled Imagine K-12, the name of that thing back in. So back after high school, I was working on a learning management system because Blackboard sucks, it was a terrible experience to use.
Turner Novak:
Oh, yeah. I used that in college. Yeah. I remember that.
Eric Simons:
My co-founder and I, Albert and I, were good buddies and we built a better version for our school and we were like, "Hey, this is cool. I'd much rather use this as a student." And so we got into Imagine K-12, came out here, and back then it was the OG YC deal where you get 20K and that's it. So we ran through that in three months in the Bay Area or whatever. And so we ran out of money. So Imagine K-12... At that time, AOL was trying to reinvent themselves. I think they sold to Verizon or something.
Turner Novak:
Yeah, $3 billion or something like that. Or maybe, no, I think they got bought by Yahoo and then Verizon bought Yahoo or something. Yeah.
Eric Simons:
Yeah, that sounds about right. And so at that time, they were trying to turn around the company and so they're like, "We need to get young startup blood into the building." And so they started renting out office space to startups and stuff. And so Imagine K-12 had gotten an office at AOL. It was a super cool building. It was down in Palo Alto off of Page Mill and El Camino. And so all of the companies in Imagine K-12 had access cards to into AOL's headquarters. So at 19, I had a backpack with all of my earthly belongings and I was-
Turner Novak:
"What do I do with this?
Eric Simons:
... out of money. And we would go there and they always had Cup of Noodle. They had the little snack board, like Cup of Noodle and this and that. And then teams at AOL would order food and there'd always be leftovers or whatever, and they'd be in the fridge, and they would get thrown out at the end of the week if no one ate this stuff. And they had a gym which had a shower in it. You could even run laundry there.
Turner Novak:
Did they have nap pods and beds and stuff like that at that point?
Eric Simons:
They did. They did. They had a room called Nap Quest.
Turner Novak:
Nap Quest.
Eric Simons:
Cause you had AOL and Map Quest, so it was called Nap Quest, which had one of these, you could go in there and sleep. So as you can maybe imagine, that became my home for a couple of months. I had infinite energy as a 19-year-old. It's hard for me to imagine I did that because now I'm like, "God, there's no way I would ever do that." But-
Turner Novak:
I mean, I'm married with kids. Trying to have my kids living in AOL headquarters is impossible.
Eric Simons:
Yeah.
Turner Novak:
There's a minimum burn rate. Your burn rate definitely goes up with certain phases in life. And it's like, I don't know, your kids, they want dance classes. And it's like, "All right, this is a certain price that I got to pay every month."
Eric Simons:
Totally. And even just from a personal comfort perspective, the idea of living out of... At that age, your standards are a lot lower. So I loved it. I actually really loved being there because I was coding 12 hours a day. And I got away with it for a while just because the security guards worked 12-hour shifts. So the morning guys would see me there, they'd be like, "This guy's working hard." The night guys would be there, they'd be like, "This guy's working hard." So there wasn't overlap between the night and the day schedules typically. So I became friends with a lot of those folks or whatever. So I was living on a dollar a day because food and housing and whatever was taken care of basically. And every now and then, I'd go to McDonald's and I'd get this back... This was when the dollar money still existed.
Turner Novak:
Splurge.
Eric Simons:
Yeah.
Turner Novak:
Like an actual dollar, yeah.
Eric Simons:
Literally an actual dollar. And yeah, ended up building that startup. And the K-12 space is brutal. So it didn't end up working out at the end of the day for that one. But that was actually what ended up leading us into things, like a web-based course and stuff. That was the origins of how that can be. And then ultimately, how Bolt StackBlitz came to be, because that was on the end of that thing.
Turner Novak:
We jumped around the story a bit, but you mentioned you raised the Hidden Series A around 2022, and you did raise a series B. I remember there was a Bloomberg article, I think, that was like, "Yeah, Bolt might be there in conversations." So what actually happened with the series B? You obviously went from zero to five to 20 million revenue, and you started to think, you said, "Okay, I should probably raise money at this point." How did you approach all of that?
Eric Simons:
Yeah, good question. Because of what was going on with the rapid growth and our team was so small, my chief of staff and I, we were the customer support team basically. And that inbox, when we went to go raise money, that inbox was at 500 or 600. I mean, we were getting destroyed as is. After the couple week fundraise process, it was like 1200 plus. It just kept accumulating. So going into what we knew that was going to happen, we were like, "This is a dangerous thing that has to happen here because we're going to have to take our eye off the ball. Balls are going to drop. It has to be time boxed. We have to go to time box the process because we can't go and spend a month or two or whatever on this thing. We need to get it done and move on."
And so what we ended up doing, was rented out, there's this really awesome coffee shop. Actually, shout out to them. So there's this coffee shop called Lighthouse Coffee, I think, in South San Francisco. I actually found out about it about a tweet because someone was like, "This is the most slept on coffee shop in San Francisco. It's beautiful. It's never crowded." And it's true. It's amazing. It's fantastic. And it's a little outside the city, which is why it's not just crammed every day. So amazing. So anyways, they have a boardroom you can rent there.
Turner Novak:
Is it the Lighthouse Cafe or the Lighthouse SF?
Eric Simons:
Maybe The Lighthouse SF. It's in South San Francisco. That would be the one, I think. And yeah, it's modern, it's just beautiful what they've done with the thing. And great food, great coffee. I've been going there for a couple of months and they had a boardroom because we're a remote company, we don't have an office. And so rented to the boardroom there, and then just met with a whole bunch of... From our investor network, we got introductions to a handful of folks that were interested in chatting and brought them there, met with them, and made a slide deck and stuff.
Given where we were at, is was understandable. I feel like a lot of entrepreneurs, you do want to have the process move quickly, but often they'll be arbitrarily, they're like, "We're closing Tuesday." I mean, there's this artificial thing, but with us, it was like, "Listen, we have to get back to this thing because if we don't, we risk ruining what's happening here." Anyway, so that's how we ran it and got to meet a lot of great folks. But yeah. You ever seen that video of Lady Gaga when she was doing interviews and she's like, "You only need one person in the crowd to believe in you." And have you ever seen this video where-
Turner Novak:
No, I've never seen this.
Eric Simons:
Okay, you guys have to watch this because it was a interview with Bradley. Her and heard Bradley Cooper did A Star is Born together or whatever, right?
Turner Novak:
Oh yeah, that was a good movie.
Eric Simons:
A wonderful movie. And so in the lead up to... Because the celebrities have to go and advertise this with interviews, so very clearly, she had a script of something very inspirational to say. And it was like, "This movie is great because it shows that you only need one person to believe in you," and dah, dah, dah, dah. So someone took a montage of all of the times she had said that, back to back to back to-
Turner Novak:
Oh, on the whole press tour, she kept saying the same thing? Oh, amazing. Okay.
Eric Simons:
That's what fundraising feels like. There's true meaning. That's the hard thing about this stuff. Genuine storytelling, you're like, "This is my view of reality." But you have to say it over and over and over and over and over again. And that's what fundraising is. Yeah, I think if you want to do it well, you have to get really good at... It's acting, channeling your emotions over and over.
Turner Novak:
And so it sounds like the best way, the way that you met people was through existing investors. You're like, "Hey, we want to raise some money. This is the deck, the game plan. Who do you think we should meet?" Is that how it went?
Eric Simons:
Yeah, totally. And I think every time we go and fundraise, it's always an opportunity to bring new people onto your team. So it's great to go and chat. In this round GV, Google Ventures, Eric over there, he co-led the round. I've loved working with him for the past four or five years. So if you have investors that are down to lead the thing, participate, heck yeah.
It's always good to just see, "What other people can you add to your team?" So to speak, because basically what you're doing. And so just ask our angel investors and the other VCs, "Hey, what other folks do you think are really great that would be really great for the challenges that we have to go and tackle?" Because for us, we were an enterprise motion. And then overnight, we woke up with a B2C company. And so it was like, "Okay, actually, it would be very smart of us to go and find folks that have done this before, that have actually helped participate with scaling companies, that have done a Prosumer B2C bottoms up motion." And we had saw a little bit of that in our existing investor network, but we were under indexed. So going and talking to a lot of people, a lot of high quality people, is important, I think.
Turner Novak:
Okay. Yeah, I think Emergence and maybe GV co-led, I think of Emergence as the enterprise B2B software, or is that not true?
Eric Simons:
Yeah. And Joe, I talked to Joe there for years, and so I'm stoked that I finally get to work with him because I really love Joe. And so yeah, he's the one that co-led the thing with Emergence. And they've done Enterprise, but it was interesting, they've done a lot of Prosumer, with Zoom,
Turner Novak:
I guess, Zoom. Yeah, makes sense.
Eric Simons:
When I talked with Joe over the years, a lot of the stuff from Zoom, there's a lot of corollary with this adoption, how we're trying to... As we look at the B2B side of this, it's like, "How do you pull people into the business application of this?" And et cetera. Lot of work to be done there. But yeah, they have an incredible amount of knowledge in house of just how to really build great Prosumer bottoms up motions.
Turner Novak:
Makes sense. When you announced, I think the number and the announcement was 105 million, I think. Was that the A and the B combined together?
Eric Simons:
Yeah, A plus B.
Turner Novak:
But it sounds like you didn't spend much of it. So you've got $105 million at least that's sitting there. What are you going to do with it? What's the game plan now?
Eric Simons:
Not spend it. No, I'm kidding. I think it's at this point in a company... Sarah, one of our investors, I can't remember what CEO had said this, but she had referenced it one time where it's like, "You have to know when to lean in on a company." I remember I called her one day before, I think it was the day before my Iron Man. I was like, "This is the lean in moment for us. It's right now, basically."
And so when you're at this point, you need to begin deploying capital to grow faster. There's going to be loss-y things that happen, but it doesn't matter because the rate of growth is so insane that you're dealing with. If you don't have a high rate of growth, to go and just start deploying millions of dollars or whatever when you have 10 million or something in the bank, it might not be a good idea. It's very, very risky. It likely might not work out well. But for where we're at, we need to hire a lot of people. We need an actual customer success team. We need to increase our engineering head count by 2, 3, 4X or something. Really invest in our go-to-market engine. And because we have two or three people there today. If we had a physical office, there's entire sections of the building. They're just empty.
Turner Novak:
You need to do an add-on to the building, add a new wing and add a growth team, add a customer success, expand engineering by 4X.
Eric Simons:
Yeah, totally. I mean, for the scale that we're operating at, we're so... This is what I say to the team is, "If you look forward a year, what would be one of the reasons we fail?" It's like, "We try and stay too scrappy because we don't scale up to actually really match the demand and make sure we're providing a great experience." Because you have to unlearn some lessons based on... Because this is an extraordinary new context we find ourselves in. What got us here won't get us there. And so we have to go and find things that are smart to invest in where we have to be okay with more lossiness in bets being made and stuff, and scaling out our leadership team.
Turner Novak:
So it sounds like you're hiring some customer success people. How many people are you bringing on board in customer success? Just roughly. All these numbers, how many people are you thinking about hiring across all these categories? I'm just asking because there's probably people listening that are like... If they've made it this far, we're an hour in and they're like, "This sounds really cool. I'm a customer success. I want to work for Eric."
Eric Simons:
Right, yeah. So let's talk about opportunities. So I think at customer success, we're hiring probably half a dozen people, at least, in there, at least. On engineering, people that work on AI engineers or that do full stack TypeScript, et cetera, we're going to be hiring probably a dozen people just there over the next six months or something. Same thing. People are great with product, like PMs. We don't have any PMs at the company right now. That's a bit of a problem.
Turner Novak:
Is that you? Are you the product person?
Eric Simons:
Yep, I'm kind of chief product guy at the moment, so that's a role that's top of the docket to get filled. Same thing, a head of engineering, as we're scaling out these teams. And on the marketing side, we have a whole bunch of roles listed. If you go to StackBlitz.com/Careers, you can see all of them. But on the GTM side, there's a ton. I mean, we have growth marketing, I think we're about to add an influencer, marketing manager, et cetera, et cetera. Rev ops. Yeah, there's a whole... For the scale of revenue we're doing, typically you have a year or two to grow into this, but we're...
Turner Novak:
Yeah, usually you hire and then you scale up into it. It seems like you grew into it in a couple of months and now you're like, "All right, we got to backfill almost this, that we grew too fast for."
Eric Simons:
And it keeps growing. So it's this cat and mouse, we're catching up to this, and then we're like, "Okay, well we actually need..." So now we're making hires where we're like, "Based on where we're going to be there, we need a CFO, we need this." Because we got to skate to where the puck seems to be going. I've never experienced this sort of challenge in my career.
Turner Novak:
So you're sitting in your seat of you want to bring some people on, you're adding to the team. What is an ideal new... Do you call them Bolters? Do you call them Stackers? StackBlitzers? I don't know what the acronym is or the wording for people, but how do you think about people that you bring on the team? What do they usually look like? Are there any characteristics or personality traits that work best at Bolt, or?
Eric Simons:
Totally. Yeah, really good question. I mean, I think the most important thing to me is always working with people that are really genuine, because I think that says a lot, that word. When I met my wife, and I said this in my vows to her, but she's just the most genuine person I've ever met. And for the first months that we were dating, I couldn't figure out if she was the most genuine person I ever met, or if she was the most believable sociopath that really had me fooled. And I was like, "I know there's got to be something that's not right here," but she's just the most incredible, genuine person. What's great is when you work with genuine people, there's none of this political crap, this ego crap. You can have real conversations and make sure that we're just doing great work as a team, and you build real friendships. Case in point, the folks, the original five, seven people that have worked here, they've been here for five plus years. It's because they're good people. They're not egotistical. They're very open to feedback, they're open to learning things. You know what I mean? It's just kind of the perpetual learner mindset. Just very chill. Very chill in the sense of how we work with each other, but also the people on the team are very intense about doing great work and really pushing themselves to the limit. A lot of the people on our team have been either like semi-pro athletes or semi-pro whatever, musicians or whatever. I mean, the extracurriculars are nuts here. I think that to me is like it all starts with that, great people, smart, hungry down to learn, down to join 300 in protecting Sparta or whatever.
Turner Novak:
I actually have someone, I think maybe after recording I'll send your info, but she's awesome. I would love to hire her personally, just I'm not really in a position to hire, and I don't think she'd really want to do VC, but I've been trying to get her a job at one of my portfolio companies, but it sounds like, I don't know. Well, we can talk about it later, but I think I have somebody that might be a fit for one of the things on the site. We'll throw a link in the notes too just for Stackblitz-slash-careers, stackblitz.com/careers, and people can find it. But then, so you're hiring a bunch of engineers, product people, product's going to change, you're going to build some stuff. How are you thinking about it just over the next couple of years? What does the roadmap look like? What should we be on the lookout for?
Eric Simons:
Yeah, good question. I mean, I think we're going to be investing pretty heavily in the B2B side, really enabling companies to use it for the use cases I was talking about, like prototyping and building their public websites and even building their own apps. AI cogen is just going to keep getting better. I mean, over the coming months, I think we're going to see some new stuff coming out that every time model updates come, we saw this with the second version of Sonnet at the end of October, we flipped that on the same day. Our conversion rates cranked double digits.
Turner Novak:
So the product is getting better as the models get better essentially?
Eric Simons:
Absolutely, and I mean our product, we ourselves are doing tons to make the product better, but the second the better model comes online, it can be a 10x multiplier. And that's what happened before when Sonnet 3.5 came out, it was 10 times better than the previous Frontier stuff at doing cogen, and what's going on with Cursor, us, all these cogen things, it is all because of Sonnet 3.5. You take Sonnet 3.5 out of the equation, none of this stuff exists today. And so it's like, I think, and now all the other Frontier Labs, they get it. They're like, "Okay, we want to be best in class here."
Turner Novak:
Which is incredible for you, just benefiting from them making their models better.
Eric Simons:
Yeah, totally. I think there's going to be so many other different types of things that come out too as a result of this. Yeah, I think the next six, 12 months we're going to see more models, which means more capabilities, which means ... And so for us, when we look at B2B, Cursor is the most successful thing when you look at a business adoption perspective right now, but that's largely because it's a very developer-centric tool. We want to crack non-developer use cases, and developer use cases of course, but you have marketers, PMs, what does it mean to live in a world where the Fortune 500, the PMs are writing more Jira tickets into Bolt and hitting enter than actually assigning it to the devs, where the PMs are actually coding?
Turner Novak:
Plus it feels like a lot of devs, you hear it's not their favorite thing of making small little tweaks and it's probably not the best use of their time either.
Eric Simons:
Exactly. It's like you want to assign this stuff to them that is actually intellectually challenging, that an LLM is not going to be good at. And what are PMs great at? They're great at crafting great products. All the little details of that experience where they're happy to be going and ... Now, I'm kind of more of this, I'm like an engineer, but I'm also really more of a product guy. I would much rather ... I can write the code just because I have sheer force of will since my buddy and I were 13, because that's the only way you could do it. But I would much rather tell a thing to do it than just get the result.
Turner Novak:
I think you mentioned before too, you actually learned to code from reading those books or whatever, the learn to code books back in the day before ChatGPT, Google, "how do you do C++?" Or YouTube videos that would just teach you.
Eric Simons:
There was nothing. Yeah, I mean, this is back in mid-2000s. My co-founder and I were 13. Albert and I were 13. And now there's so many great resources to learn how to code. There's Code Academy and this and that, whatever. There was nothing like that. And so for our 13th birthdays, we asked our parents, like, instead of an Xbox or whatever, we're like, "Can you get us these O'Reilly books?" Those things were like 200 bucks a pop. It's an Xbox in a book. And it was like, if we really want, this is the only way. That's how Albert and I learned how to code, is he and I both got books, we sat next to each other and just ... If it was without Albert, I don't think I would've been able to learn, because it's like setting up a tech environment is terrible. Most people churn out at that step, because it's like getting stuff just working, it's not even about writing the code, it's getting the environment going so you can even write the code. And then when you write the code, oftentimes the environment craps out for a billion different reasons.
So it was like Albert's always been more technically proficient than I am, and so it was like being able to really learn with him and from him was huge. He and I always, as we've been building Stackblitz and now Bolt, we're like, "We're building the tool we wish we had when we were 13." And have for everything else we've built since then.
Turner Novak:
One thing I was super curious about though, business model, it seems like you make money right now. I hit a paywall where it's like, if you want to make more changes, you got to upgrade. So I was like, all right, it's going pretty well. This is pretty cool. I'll pay the 50 bucks, get more tokens. So how do you guys make money right now and just how do you think that's going to change over time?
Eric Simons:
Yeah. Well, yeah. Right now it's entirely Inference. And I think we were the first people to figure this one out with the Sonnet 3.5 stuff where, prior to Bolt, I feel like with Copilot and all these other things, there was kind of this mental model of, oh, these AI tools, you should charge for it like Netflix where it's like 20 bucks a month all you can eat, but if you eat too much, we have to slow them down because it costs too much money for us, but it's like low price and you can just have it. It's like that.
And the problem though is now that cogen is so good, the ROI, the 99% cost reduction, anyone who knows the true value, the true cost of building quality software, they're like, "I want more." And this is exactly what happened the first week we launched. They burned through their nine bucks of credit and they're like, "Let me give you money."
Turner Novak:
And so were you gating people and slowing things down initially?
Eric Simons:
Unintentionally, because we only had one plan, it was like nine bucks a month and you get some Inference because we just thought people were going to casually use this thing. We didn't think ... but people were like, "I'm trying to build my product here, man, I went through nine bucks in a day, why won't you take my money?" So that weekend or that week, I was like, "Team, we have to ship multiple different tiers. I will work the weekend and stay up with whoever is going to do this with me." And everyone did and we flipped it on. And that was really what we're seeing now with a lot of other stuff in the AI cogen space, people are pricing based on different tiers. So on Bolt we have like 20, 50, 100, 200, and then you can also just buy Inference add-on as needed. You look at Cursor, they have usage-based billing. Everyone's starting to do this thing where it's like, okay, actually because the expected ROI of Inference is so high now, people obviously are going to spend more money, they're going to want to spend more money to use it.
And so I think we're going to continue to see a lot of that as our revenue stream is selling Inference or selling our agent, but we're also going to be getting into the game of just integrating hosting and databases and stuff. We're, like, if you're a developer, you can go sign up for all these things individually, but within Bolt, it just makes sense you'd have centralized billing where you can add on hosting from Netlify or add a database from Superbase and it just works. Same thing with domains. Imagine coming to Bolt and it's like, "Hey, I want to make a website for my VC firm. Here's the name of it. Oh, the .com's available. Do you want me to buy that for you? Yeah, sure. Go ahead. Boom, live." And there's nothing that you had to do there, right? You're chatting and you now have a .com domain with a website to put into it. You have a database backing it or whatever. Within the next six months, that's what you're going to be able to do on Bolt.
Turner Novak:
That's pretty awesome. Yeah, I was going to say I saw the Superbase button. I remember seeing, I didn't click "Deploy" because I got stuck on a YouTube API integration. It was 12:30 last night. I'm like, "Shit, I got to go to bed." My wife was watching, there's a show on Netflix, it's called Glow. It's this women's wrestling league from the '80s, and it's like the show profiling the league. Have you come across this show at all?
Eric Simons:
No. No.
Turner Novak:
It's like a crazy show. I don't even know. It's basically all these struggling actresses that join this wrestling league and it's like a show about them putting together this amateur wrestling league in the '80s. It's like a recent show from a couple years ago. Anyways, my wife's been watching that the last couple nights and I'll just chill at the table watching the show. And it was like 12:30 and I was like, oh shit, I got to go to bed. I got to wake up. I got to be awake in seven hours, so we got to go to sleep. So I got stuck on the YouTube API integration. But that seems like something you could probably automate eventually, right?
Eric Simons:
A hundred percent. And I think that's what our team's like, we're just focusing on this long tail of all these different APIs you want to use, just making it seamlessly work without you having to interfere much with it. And we've done a lot of that with the Superbase stuff and like 3D models on the web. There's a ton of stuff we do and look at the next year. It's like there's just going to be even more of that.
Turner Novak:
So just continuing to make it go more and more of like type in a prompt, it just does what you want, and just continue to just dive deeper and deeper into making that even more fluid and seamless?
Eric Simons:
Absolutely. We're like, well, you can come to this and you can literally say, "Hey, make me a reasonably sophisticated application in zero-shot." It provisions the infrastructure and it has a real working version on the first shot. And sometimes you can get that today. Other times it takes a couple of extra prompts or whatever, but it's having it be kind of Slack where a lot of the magic was the integrations. I think a lot of that's going to be like this too, where it's like, oh, I need the YouTube API, I need it GPT 4.0 to do duh-duh-duh-duh. I mean, it's just having this built in and really just making it magic and smooth.
Turner Novak:
Maybe one more big question I wanted to hit on. It seems like you guys have built a pretty good community. You kind of mentioned you have all these people, if you go on YouTube, examples of people that are a hundred thousand view videos of how to use Bolt, blah-blah-blah. How did that develop? And was it an accident? Did you do it on purpose? And how should I think about building a community around my product if I'm a founder? What do you think you did right and just how should somebody think about that?
Eric Simons:
Great question. I think for Stackblitz, we spent a lot of time over the past seven years building a community of folks. If developers are using our product ... and it's like when we launched with the tweet, it was like we had, I don't know, 20,000 or 30,000 followers on the account, and people would generally like the stuff we would put out. This went totally mega in a way nothing else we'd ever done before, right?
Turner Novak:
Yeah. It's almost like your community lifted and pushed it forward because you had spent so much time nurturing it and it truly resonated. And so it wasn't like you had to manufacture this launch. It was literally like you built the muscle of just here's something and it's stuck and people went with it.
Eric Simons:
And I would recommend anyone who's doing startups, build community along the way no matter what, because exactly what you just said, it is useful to have a distribution mechanism. When you have something that's really going to work, the message will get out if you're cultivating community and building great stuff and helping people be successful, duh-duh-duh.
Great example of this is Joel Spolsky, his blog, Joel on Software, I think it was, tons of programmers would read that back in the 2000s or whatever. And so when he launched Stack Overflow, he wrote a blog post about it, but that was what seeded Stack Overflow's growth, because he's just making content. He's a smart guy. It's not like fluff stuff. It's like real stuff. He's just talking about software and what he learned in his time at Microsoft and how he's building Fog bugs and whatever, right? So when the time came when he launched Stack Overflow, just boom. And so I think, from a distribution point of view, that's the value of really building community.
And so up to the launch of Bolt, that's been key. We've been investing that for seven years. I think did a really good job there. When we launched Bolt, so much stuff was going wrong where it's like there's no chat history, you can't buy more than nine bucks of Inference. Now we fixed that, but there's all these bugs that are happening. It was just endless. And the problem is you have to ... It's early. We didn't have a support team to answer all these questions. And so what we did, I can't remember what week of the story this was, maybe it was week three, I was like, "Team, we all got to get on a livestream every week where we're just answering questions, telling people what we're doing and whatever."
And it became this kind of staple of the first two, three months there where we got to meet people, we got to really know them, and we've invested in our Discord, and now we have just half a dozen, dozen people that are just these amazing folks that're just helping other people be successful with the product, working with our team getting feedback to them, and hundreds or thousands of other people that are also helping people that are not just directly manning our Discord or whatever, creating content, helping people get unstuck with stuff.
And I think for them, when we saw people doing this, we would boost their tweets, we would put a spotlight on them, like, "This is awesome. This person's great." And so they're able to build a presence and actually know that we cared and appreciated about the stuff they were doing.
And so every week at our office hours, we would do shout-outs, shout-outs or spotlight or something. We would all pick the top things that we saw that week that we wanted to give a shout ... We'd send a T-shirt to the different people that got picked or whatever. Each person on our team would pick multiple people or whatever. And just it made people know that we're not just some corporate thing. It's like we're just people trying to make a thing and we're like ... I remember one of the livestreams, I was like, "Listen, if you're watching this, we need your help. There's more people that need to understand how this stuff works than we can possibly tell it to. If you see people on Twitter or on Discord or whatever asking questions you know the answer, please help them. And everyone who's doing this already, thank you. It's building this community." And that's what we saw. It's just people helping and us telling them how much we appreciated it. And I think for AI tools it's really important because when you get that input box, it's like how do I best use that?
Turner Novak:
Yeah, it can kind of be overwhelming. It's so simple, but it's also like you don't know what to do and you don't know what the capabilities are.
Eric Simons:
Absolutely. And especially a tool like Bolt where if you haven't done a lot of product building or web development, whatever, and you come in, you're like, "Build Google for me." And so people are like, "This thing doesn't work. You can't do it." And it's like, "Yeah, but let's talk a bit about that and how you actually ..." A really good example of this is Midjourney. This is one of their big moats, is that they have built an incredible Discord community where people are sharing their prompts and they're like teaching each other how to use the product successfully, which then further informs the Midjourney folks on how to better tune their models and whatever.
And so for us, I think we've done a really good job of that. After week two, we were no longer the domain experts of how to best use this product.
Turner Novak:
That's crazy. So your customers knew it better than you did?
Eric Simons:
Now we've brought some of these people on contract to work with us so we can actually meaningfully have them do an R&D for us and go, "How do we bring these approaches back into the product and how do we help the community, et cetera?"
So yeah, especially for these AI things, you basically want to create a community where you can crowdsource the knowledge because you're not going to be able to figure it all out yourself. You're not. You have to create a collaborative place where people can go and experiment and really be rewarded for doing cool stuff.
Turner Novak:
Yeah. You also open-sourced everything too.
Eric Simons:
Yeah, yeah. The core version of Bolt we actually open-sourced because, again, we didn't think it was going to be this huge thing. And also at the time we were like, when we went to go build Bolt, there's nothing out there that's close, that's anywhere close to being open source. And we were like, this is kind of a gap in the market. We feel like we ought to ... it felt uncomfortable because we're like, "Is this a terrible idea?" We're like, "Listen, we have to continue making this the best product." I think GeoHot had a podcast one time, he's working on Comma, this is like maybe a year or two ago, he said something like, "The only reason to be afraid of open-sourcing your stuff is if you think that you are going to run slower than the competition." Something like that basically, right? The only reason you should be afraid of open-sourcing your code is basically if you think that you can't iterate the fastest. Yeah.
I don't think that's universally true, but it is an interesting point, though, thinking, "Okay, you can actually ..." Open-sourcing stuff, there's a hu- ... And we saw this with Stackblitz. We're one of the core investors, and Veed, which is the build tool that's kind of taken over the web development world.
Turner Novak:
Oh yeah, I've heard of Veed before. Is it the green logo, I think?
Eric Simons:
Yeah. I think it's got a little purple V-looking thing with the lighting bolt in it. And we've released a whole bunch of open source ourselves over the years. And releasing good open source software is a huge tide-lifter for the company that is actually doing that work or investing in that work. And so when we put Bolt out there, we're like, we think this will actually help us get people to adopt Bolt because they can actually see how it works and fork it and do cool things with it. And that's exactly what's happened.
It's actually kind of cool now where because we had open source Bolt while all these new models are coming out, often Bolt is one of the things that people are testing them against because it's ... what other production tool can you say, "Oh, DeepSeek came out. How do we kind of test whether it's useful in a real product?" Well, unless you're the company that owns that proprietary product to get the data, you can't.
And so a lot of the frontier labs are actually using, it's called bolt.diy. They're using bolt.diy to just plug in their models and see, okay, how well does Bolt perform with this new model?
Turner Novak:
So I can basically go to bolt.diy, looks like you have all the code. I can just take all the code, spin it up, put it in my own environment, and in theory I have Bolt existing and, if I wanted to, I could clone it and copy it or whatever. But it looks like what people do is they'll basically, like Anthropic is using it to test Sonnet 4 or whatever, the new ... I think the 3.5 is the current or ... so they'll test the new models they're working on on their own Bolt environment internally.
Eric Simons:
And there's a lot of these open source models where the open source community is just going and testing these things out. People are taking bolt.diy and stripping out parts of it and putting it into their own products and stuff. So you have enthusiasts, you have research labs, you have companies using this. And what we've seen is just, it's lifted the tide for us Bolt is, I think at this point, probably the most widely known, widely adopted whatever thing. And I think that the open source strategy has been key for that because people are adopting it in ways that the commercial version of the thing just couldn't, and it doesn't hurt. It has not hurt our ability clearly to go and build a business on the thing.
And that was from the get-go, it was an uncomfortable bet we made there, but we were like, we just think that this should be open source. Nothing else exists. It'll probably only help us, which it has, because when new models come out, people are like, "Well, are you going to implement this?" We're like, "Let's go look at the open source version." And usually someone in the open source world already tried it, already has all the data points on it. So we have this incredible amount of R&D happening against our product specifically, whereas everyone else has to figure it out themselves.
Turner Novak:
So it's kind of like if I was making a competitor, I built the same thing, whatever, and Anthropic comes out with a new model, and it might not be fine-tuned for turner.new or whatever, but it works in Bolt right out of the box immediately, and people are like, "Oh, why am I using the turner.new? I should just use Bolt."
Eric Simons:
Exactly. And so that's kind the downstream effect. The long tail of it is that if this continues to be the main thing, that some type of test environment that people are looking at as they fine-tuned things or training things or whatever, it's going to just continually just keep getting better for Bolt or things that look very much like Bolt, and that helps us and that helps anyone else that's using bolt.diy for whatever use case, right?
Turner Novak:
Yeah. I guess if you're building all these different integrations and stuff, you just generally want people to like Bolt and use Bolt and just have it be more of a common thing. Like, why would you make an integration for something that nobody uses?
Eric Simons:
Totally, totally. And I think for us too, it's like the people ... again, people are bolt.new are like, most of them are not developers, so they would never ... like, to go in and clone a repo, they're just not going to do that. Even a lot of developers, I have better things to do with my time. Get API keys and run this thing locally myself. And the production version's more reliable. We have whole things that are not in the open source version, whatever. So it's just really symbiotic, which is super cool because there's not a lot of examples of this in the AI world because I think there's a clear incentive for things to be kept closed-source, because then you can kind of market it as whatever.
It's like, "We're going to announce this crazy AI agent," duh-duh-duh-duh. You know these other things are open source, they can say whatever. They can make up terms, they can say, I don't know what, I don't want to talk on anyone, but these kind of handheld things, they made up these terms. It's like blah-blah-blah, and it turns out it's an Android device that does nothing basically. So when you close source, you can stretch reality way beyond the fricking universe, the known universe, but that's not real. You want to do the hard work, build something that's real, actually provides real value. And if you can do that, then you can open source core parts of it.
Turner Novak:
It sounds like kind of just to tie up the whole this community thing, it's like when people are using your product, you elevate and you almost, you just acknowledge them and you make them feel good. You incentivize people to further use and to further invest their own business around your business. Is that just generally a good way to just high-level summary?
Eric Simons:
Totally. Yeah, totally. They get equal value back out of it by being a part of the community, and if not greater, ideally greater. And that's key to building communities, is that people feel valued and they feel like their contributions are really valued and the spotlight is thrown on them and recognition is given and it helps them out. And then it shares knowledge, where, again, we are learning how to better build our product to match what people are discovering what works really well with these things.
Turner Novak:
This has been awesome. It's been fun hearing the overnight success story that took a decade or maybe 15 years or maybe, I guess, 20 years at this point, started at 13. Where can people find you? Just follow along. We'll throw some links in the show notes if you want to do that. But any like shout-outs just to close things out?
Eric Simons:
Yeah. Yeah. Totally. Bolt.new. Check it out if you haven't tried it yet. Hopefully your mind will be blown. If it's not blown, you should definitely tweet me. I'm at ericsimons40. Our Twitter account for Bolt is boltdotnew, not a period, but B-O-L-T-D-O-T.
Turner Novak:
Oh, dot the word.
Eric Simons:
... N-E-W. Yeah. They don't allow it in there, but it's kind of cute, just like boltdotnew, and yeah, I think that's the main ways.
Turner Novak:
Cool. And we'll throw a link to the careers page too, people want to check things out. Some of the links that we mentioned, we'll throw links in the show notes.
Eric Simons:
Yeah. Thank you for having me on. This has been a blast.
Stream the full episode on Apple, Spotify, or YouTube.
Find transcripts of all other episodes here.