Episode 75
· 43:46
This sounds obvious because everybody says move fast. Everybody says move faster. But in reality, you could always move faster. That's because what I learned. I'll give you another example for.
Jack Bridger:Hi, everyone. You're listening to Scaling Dev Tools, and I'm joined today by Glauber Costa, who is the founder of Terso, which is a brand new database company that has been taking the Internet by storm. Great to meet you, Glavo. Thank you for joining.
Glauber Costa:Likewise. It's a pleasure to be here.
Jack Bridger:Could you tell us a bit about Terso?
Glauber Costa:Terso, easiest way to explain. Terso is essentially SQLite for production. SQLite's been around for a long time. It's one of the databases, I guess, that people trust a lot. It's one of the databases that people play with a lot.
Glauber Costa:You always see SQLite in every tutorial. You see SQLite in every getting started. Is is Ccolite on mobile, is is Ccolite on your toaster, your watch probably runs Ccolite, and it's just a really this database that is everywhere. I think web production work closer, a little bit more complex than that. So a lot of a lot of the things that, makes SQLite so great for those use cases, like the fact that it's just a file, like the fact that it's very simple.
Glauber Costa:It's not that they are shortcomings for the web workloads, but you need to build some more you need to build, like, more infrastructure to make that usable and and to make that a good choice for your production workloads. And that that's essentially what we're doing. We have a fork of SQLite called Libs SQL, that we announced in October 2022. And then Torsos is the managed service, that we we built after seeing how much interest Libsico gathered in in in weeks with it.
Jack Bridger:That's, yeah, that's really interesting. I think I used SQLite. It was, like, one of the first ones that database I used with Rails. I think it was, like, used to be the kind of default option than if it still is.
Glauber Costa:I've seen someone on Twitter the other day say that SQLite will be a lot more successful if you didn't have lite in the name. But keep in mind, it was created 30 plus years ago. So 20 plus years ago, it really was a database that could be used only for small things. But the hardware advancements that we had in the past 2 years are fantastic. But the human workloads didn't grow that much.
Glauber Costa:So so there are there are workloads that are growing exponentially in terms of data, and it's usually like things that have a timestamp attached to that, like machine events, generated events. But things that humans do, which is a large portion of the web, it didn't grow that much. I mean, if you had a product catalog 10 years ago, it will be more or less the same size, today. Like maybe it grew from 5 gigabytes to 10 gigabytes in size and people think it's a 2, you know, 2 x factor. But that's nothing.
Glauber Costa:You have USB sticks today for a couple of bucks on on Amazon. Not AWS, the actual Amazon. They can buy and and and there you go. You got a terabyte of storage for nothing. I just and you'll just attach to.
Glauber Costa:And so every like, it's not a database that is only for small stuff anymore. Clearly, it's not the database that is gonna be used for the petabyte scale and the most challenging, like Google scale, Facebook scale, we're close out there. But it's good enough for the web, it's good enough for most companies. There are many companies that are running their production on SQLite. And what Tersu wants to do is just to make this more common, you know, more flexible, easier, more doable in general.
Jack Bridger:This episode is brought to you by WorkOS. At some point, you're gonna land a big customer, and they're gonna ask you for enterprise features. That's where Workhorse comes in because they give you these features out the box. Features like skin provisioning, SAML authentication, and audit logs. They have an easy to use API and they're trusted by big dev tools like Vercel as well as smaller fast growing dev tools like Nock.
Jack Bridger:So if you're looking to cross the enterprise chasm and make yourself enterprise ready, check out Work OS. We've also done an episode with Michael, the founder of Work Work OS, where he shares a lot of tips around crossing the enterprise chasm, landing your first enterprise deals and making sure that you're ready for them. Thanks, Work OS, for sponsoring the podcast and back to the show. I know, Jamie is working with you now. He's been on the show.
Jack Bridger:And I've seen some of Jamie's tutorials he's putting out where it's like, he builds a database for, like, everything. It's like, every customer has their own database. And that seems like quite a a when I saw it, like, scrolling through Twitter, I was like, what? Like, every What? Yeah.
Jack Bridger:Yeah. Could you talk a bit about that?
Glauber Costa:Yeah. So as I was telling you before we started recording, I'm very flattered. We follow the with the fact that people think we are around for a lot longer. But Durso is actually today is we're still in Sabre. This is our 1st anniversary.
Glauber Costa:So we launched the private beta just for a select few people in February last year, and then we launched a public beta. We never officially GA, but, like, let's call that point, like, maybe the best way to describe it, like, we we didn't have any monetization. So it wasn't GA yet. We launched a public beta in June, and then we started charging people in August. So in Tursu is a very, very new offering.
Glauber Costa:And one of the ways I explain what Tersho is, is that instead of thinking it's a database, you can think of Tersho, which it is a database, but if you help serve mental model, you can think of Tersho as a web server. Right? So you can send HTTP requests to that. And then there is a SQLite file in the web server. So you send your SQLite requests through HTTP.
Glauber Costa:It's the only protocol we support for network connections. And then we give you the responses back, and there's a SQLite file. And there was really only SQLite file. So every VM that we spin had a for every customer of Terso has their own infrastructure. So you spin a VM, there's a SQLite file there, and we would we would let you spend 3 VMs, for free because we wanna, like, from the start, we knew that, a good a good strategy would be very generous with that, and we can talk more later about what allows us to do that.
Glauber Costa:But, okay, you can get 3 small VMs. And then you have 3 SQLifiers, you have 3 databases, or because one of the things we're doing a lot was replication. So you can either have 3 independent databases, or you can have a database with 2 replicas and you know, however, however you wanna you wanna spin this. But people kept coming to us and asking us. So again, this is something that we heard from the community.
Glauber Costa:And and and look, I wanna clarify that at that point, we were already quite successful. So we were we were having users, you know, especially how early the service was. We were very happy to see the reception. We were very happy to see that we hit the right note. But but people kept coming in.
Glauber Costa:The one thing they asked the most, like the one feature request that we heard the most is, okay, if you're based on SQLite, and SQLite is just a file, why can't you let me create like 10,000 files? So short. Right? And then we figure out, look, it needs some changes to the architecture. Right?
Glauber Costa:Because we had this very simple thing. There's a web server, serve you SQLite file. So we changed the the architecture. So now you have routes and then each route is is essentially routed by the database name. Each different database name has its own SQLite file.
Glauber Costa:And then in September, I mean, we are one of the things I'm I'm super proud of is that we're shipping extremely, extremely fast. So again, remember, August is when we started charging people, like having the Scalar plan. And in September, we had a launch week where we announced that now you wouldn't have 3 databases on the scaler on the free tier, you would have 500, and and you you wouldn't have 6 databases on the the first paid tier, you would have 10,000. Right? So this is, I think, what, you know, everybody fell in love with.
Glauber Costa:That month that month that we announced this, we saw 300% growth in weekly active users in the platform. And we can do this once more because every database is just a file. If you are not using that file, it it doesn't use any resources. Right? So the file is just there sitting in the file system.
Glauber Costa:A lot of things that databases have to deal with is exactly like provisioning resources, destroying resources, scaling to 0. SQL like this, all of that very naturally. You have this web server, and then you you can have 10,000 databases sitting in the file system of the web server. If you have a request coming to the database, you serve from the file. If you have a request coming to the other database, you serve from that file.
Glauber Costa:Not using, not pay. So that allows us to be very efficient with what we offer.
Jack Bridger:Yeah. And I'm gonna be asking a lot of questions from, my friend, Joe from Delaware. Hello, Joe. Hey, Joe
Glauber Costa:from Delaware. How's it doing?
Jack Bridger:Who actually also said that I should I really should ask you on the show. So, yeah, big shout out to Joe. So he was actually asking, like, how come your free tier is so generous? Like, how does pricing work? So it feels like kind of a good segue into that.
Jack Bridger:Yeah.
Glauber Costa:So I think, first of all, we are we are shipping very fast, as I said. I think this has been so far critical and crucial for the success that Torsos has enjoyed and and hopefully will continue to enjoy. But our costs are now much higher than they should be because with a couple more changes that we plan to make at some point, I can put all my free tier users in the same web server, right, then just allocate databases for them. And this would allow us to be even cheaper. Now we're not there yet.
Glauber Costa:So what we what we have at the moment is that we have essentially VMs, but those VMs are fairly small, and we scale to 0, those VMs very aggressively. We're actually not that aggressively. But if you don't use the database for an hour, and here after, just a clarification, after this thing about, now you can you have 500 databases on the free tier. That means if you don't use any of them. So if you don't use any of your databases for an hour, we'll shut it down.
Glauber Costa:When a request comes, we wake it up automatically. And then it just has a little bit more latency. I don't love that model. Again, I think with SQLite, you could have something that really is like fully serverless, just file based, but it's where we started. It's where we are today.
Glauber Costa:And I think we're gonna move to to a model that allows us to have an even lower price point. But here's the thing. Like, we offer so much, exactly because we can be very efficient with how those resources are are being used. SQLite is a very efficient database. That is why we we picked SQLite to to build our previous product.
Glauber Costa:And then again, we forked it, and and we saw the the reception. We decided to compete with people at the company. But it's a it probably the most efficient database there is. It runs on toasters, it runs on smartwatches, it runs anywhere. So it uses very little resources to do what it does.
Glauber Costa:We don't have very complicated Kubernetes infrastructure. We don't have any of that. We just have a couple of VMs running web servers. So it's very cheap to operate. Being very cheap to operate, it is very cheap to which we can pass that along essentially as free tier.
Glauber Costa:Now, storage, I think we overdid it a little bit of the if if we were well, we're not gonna backtrack now because, you know, just it's not just not worth the But on on storage, I think we overdid it a little bit. But the reality also is that it's a Powerwall, we offer so much storage for free. But very few people actually use it. Most of our users don't use all the storage that so that's part of the reason why we don't backtrack on that. First of all, it's just like, hey, look, I'm just let's be generous.
Glauber Costa:And it is what it is. And we're gonna work to reduce those costs in in other places. But also, people create those databases. And most of the web workloads are not that big. So you you come and and and you use maybe a gigabyte, 2 gigabytes, 3 gigabytes, and we offer you 9 gigabytes.
Glauber Costa:So some databases actually use close to that, but most don't. And
Jack Bridger:Do you have to think about, like, if you because, like, not just, like, in terms of the costs and stuff, but, like, if you're trying to get people to, like, upgrade and you're providing, like, so much, is it, like, is that something that's
Glauber Costa:Yeah. Yeah. So so absolutely. I just, it's all but, at my previous company, I wasn't a founder. This is the first company that I founded.
Glauber Costa:But I worked at a at a startup before that for years in the database space as well. It was an open source database called Cilla, which handles, you know, the opposite of SQLite, the petabyte scale, the incredibly high volume. Still was an open source database. So you always have, we always had that issue when selling the database. Yes.
Glauber Costa:We wanna make money, but the reality is that the open source is so good that some people just use it. Like, you just you just use the open source and and don't pay us because, again, the the the high the the open source version itself is extremely high quality, extremely trusted. It will run it it was good enough for for most people. So you always have that issue. I I don't think this is an issue you need to Tersu.
Glauber Costa:Every database will have that, every open source project will have that. You wanna make the most because you know that at the end of the day, your community growing is the most important thing there is. That's why we take VC money because, you know, you wanna accelerate that process. So you want that community growing, but you wanna create a pipeline of people who are actually sustaining the company. So this always exists.
Glauber Costa:This this is not a problem that is unique to Tursue. But look, our free tier, our free tier is very generous, but our scaler tier is even more generous. And and we're gonna and we're gonna be launching a new plan, in in next month that again, it's gonna be even more Even
Jack Bridger:more generous.
Glauber Costa:And that and that will and then what we try to do, is that we will add things to those plans that are not related to the amount of requests that you do. Yeah. But are related to the fact that you are doing something with that that is more serious. And I'll give you an example. On the free plan, you have your your databases associated with your GitHub account.
Glauber Costa:Yep. And that's it. So if you are a person that is doing like your side project, if you are a person that is really just doing something simple, that's good enough. But the moment you are doing something more serious in an organization, you want other people in your team to have access to manage this database. And this is not available at all in the free plan.
Glauber Costa:So again, the free plan is, is how you I understand why so many indie hackers are flocking to Turso again, because the the limits that we offer and how generous the plans are. But you are at that point, like just yourself. So that's fine. But small companies, they are already a little bit different. If you're a small company, maybe you don't hit the limits, even on the free tier, but you want other people in your team to have access.
Glauber Costa:So you're paying the $29 a month that we have. And our new plan, it that's not gonna we haven't announced pricing yet, but should be announced in in a couple of weeks. We'll have SOC 2 compliance, we have HIPAA compliance, we have a lot of the other things that like a people in in larger organizations may want. So that that is one of the ways to do it. Because, look, most of the time when you're dealing with a vendor, it's not just the service that you want, it's the relationship with the vendor.
Glauber Costa:In case something goes wrong, it's the help, it's the support. So you are unable to charge for those things, they're not really just a request. And the good thing is that our higher up plans, they're fairly high margin, right? Because again, then the cost of, the the cost of, actually running the service is very low.
Jack Bridger:That's that's really helpful. Hopefully, that answered one of Joe from Delaware's questions. But the next one, I think I think it's a really good one. So Joe thinks that your CLI is, like, is really good. And I've tried it as well, and I agree.
Jack Bridger:It's like, you kind of it walks you through what what I can say is most CLI kind of, like, quick starts, I get stuck on something, and I'm like, okay. I did something wrong. I need to go back and see where did I do that. And it's like, you have quite a lot of steps, and I I don't think I mean, there's maybe one that I was just being silly, but that was, like, super good, for, like, in terms of how fast it is to get running. And this is Joe's question.
Jack Bridger:How did you build, like, such a kind of good CLI experience?
Glauber Costa:Time for the truth. Time for the truth because half of it was intentional.
Jack Bridger:Okay.
Glauber Costa:The the other half was was an accident. It was a happy accident. So let let's start with the intentional part. If you look at my background and if you look at my cofounder background, we're not web developers. We worked more or less together, a lot together by the end of the period, but more or less together for 10 years in the Linux kernel.
Glauber Costa:And then we joined a database company, which again was a very backend oriented, high very specialized NoSQL database, not something the web developers are usually reaching for, and for another 8 years. So and this is my career. That's it. Like, those two things. So, we really, like, are CLI people.
Glauber Costa:Right? So just that that that that's the first thing. Like, just we never, never we we're not very comfortable with with GraphQL interfaces, with web interfaces as the founders. So, like, we, for us, like, if if I was building a tool to appease a developer like me, it will be a CLI tool. Right?
Glauber Costa:So so that that that's thing number 1. So, yeah, I think this is the the intentional part. We said, look, we we understand this. It was a risk at the time because the big the biggest thing that we thought is, like, Thurso is very oriented to web developers. Will that crowd even like this?
Glauber Costa:Maybe maybe they will just say, hey, there's no and there were a lot of people in the company thinking like that, if there is no UI, people may not use it. So it was a risk. So we we figured that how do we balance that risk is by making that CLI, as as as as good as as we can. So we we come from this background. We use CLI for for everything, and we knew that we needed to make it as as as as good as we could, if we were not to have a UI.
Glauber Costa:Now the accidental part comes from this, from the fact that we didn't have a a UI at all. Right? So that that is the part that we didn't necessarily intend. But as I alluded to earlier, this company was doing something else. For around a year, we were doing something else that was also targeted at front end developers, in or application developers, more correctly, you're saying.
Glauber Costa:In in just so the audience knows, it that how related, it was, it was essentially we didn't wanna be a database company that's not the play that we wanted. So we were not positioning this as a database, but but it was essentially a way to do, like, global persistence, from TypeScript. So you write your TypeScript, we will try to figure out which SQL which SQL queries could, be used to store the state that you're referring to, in in TypeScript. So we had a compiler that that will do this for you. And then we had SQLite, that we plan to run on the edge.
Glauber Costa:And the idea is like if you're to play a global application, then you can just go in, without worrying about SQL, just by running TypeScript, you get your stuff there. And as I mentioned, we put this fork of lib SQL, with fork of SQLite, convicts SQL out in October, we were using SQLite, and we were a little bit frustrated with that. So that's why we put the fork out. As we saw the reaction to the forklift, I don't care like that. I think we're better off just pivoting the whole company.
Glauber Costa:But here's the thing. We had already spent, and and this previous product, because it was designed for front end developers, we spent a lot of money, by hiring people, hiring developers, hiring agencies, etcetera, to have a fantastic UI. So we had this very beautiful UI with a product that nobody used because we we didn't get a lot of usage from from that product. We were hoping we would, and it was not something that we were giving up, but we we was again early stages, like, oh, we've been doing this for only a year. That's fine.
Glauber Costa:So but but we started from this mindset because we're targeting this web route. We need to have a UI that is visually stunning, that is very nice. That that's how most products are. So we pivot. And and we we find ourselves in the situation that look, the market is terrible.
Glauber Costa:That that's just reality, right? It's just a so this was like the end of 2022, beginning of 2023. There's a crisis, everybody's laid off everybody, all the big companies' valuations dropping by half. So we we we understood that we didn't have a lot of time. That's what I think they like to say, okay, so we are essentially burning 1 year of money by doing this.
Glauber Costa:It was a hard decision in Celica and and going to start again from scratch. How can we win? And we understood that the only way of winning is, like, we had to move extremely fast. So this thing that, like I said, we're very proud of doing this because we we understood that this was the only chance that that we as a company have. And if you building a UI is not the way you move the fastest.
Glauber Costa:Right? So the the the the CLI allows you to move a lot faster because you don't you don't it's a lot easier to change. It's a lot easier to evolve. So in December, we we decided, like October, we released the fork. In December, we said, hey, look, I mean, just this this is it.
Glauber Costa:This is getting like this guy. This is nothing. This is just a fork that we announced and didn't really do a lot of work. And it has more GitHub stars, it has more people talking about it than our previous products. So, our previous product had around a 1000 GitHub stars after a year, which again is not bad.
Glauber Costa:Starting to see, like, the first production users, could not bet, maybe we were at the beginning of the curve. But Lip Synco had 1500 GitHub stars in 2 weeks. Right? So we figured, okay, this this sounds like a lot more explosive than than what we were doing before. And we were not just in the mode of building IUI, because we're thinking, look, it's just, it's gonna this will delay us putting this out there in in the market.
Glauber Costa:So it has to be CLI. So that that is the part that I think was accidental. If we were starting this today, maybe we would have gone in the traditional thinking of flow, the UI is the most important thing. So you need to feel that this very, you know, good UI, very intuitive UI. But we were not going to do this.
Glauber Costa:So we knew it was gonna be CLI. We understand CLI very well. And we understood that like all of those PLG things that products will put in in in the WebUI, we have to find a way to put them in the CLI because we want we want that, we want the good quick start. We want to guide you. We want to show you the features.
Glauber Costa:But we can't do this in the yuan, so we have to do it in the CLI. So it was a little bit of, like, I think, us being the right having the right background to do it with the fact that because of, you know, macroeconomic reasons, we would not be able to beat the milestones that we wanted if we were building a whole UI before we put the product out.
Jack Bridger:That makes sense. That makes sense. You kind of used your skills, used your natural kind of mechanism, and then also
Glauber Costa:I mean, I'm happy people opted because, again, that was a concern for us.
Jack Bridger:Yeah.
Glauber Costa:Lots of people were fairly concerned, in in, like, okay. So just brought this in and fly. Will people use it? Will people adopt it? And in in fairness, we heard a lot of early feedback in the feedback form that we needed a UI.
Glauber Costa:So look, and we did it in in August, we put the UI out together with the scaler plan. So I think today, a lot of people are using Thirso through the UI. But once you start doing like a good CLI, you kind of want to keep the momentum. And so like we can invest in the CLI, it's still, there are a lot of features for us that we do CLI first and then exposing the UI, Again, because doing CLI first allows us to iterate, a lot more rapidly. But we have a UI today.
Glauber Costa:And we also wanna, of course, like, we it's not that I disagree that the UI has to be polished and etcetera. As I said, I think if you are starting today without the money constraints that we had, and without the time pressure, maybe we'll just have taken longer, and build a good web UI from the get go. But we didn't have the choice. And let me say, hey, look, just so if if it's gonna be CLI only, it's better be the best CLI, experience you you have.
Jack Bridger:When when you say about investing, because I feel like it's it was obviously that, like, you're good at it and you've done it in the past, but, like, you I feel like you you also probably spent more time than most companies thinking about this stuff. Like, what if you want to, like, make your CLI better or you wanna add something, like, how do you how do you make sure that the kind of quality is there?
Glauber Costa:I don't I don't know if this can be explained. I think it a lot of that is just like you've done it so much and you understand, like, the the functionality has to be there.
Jack Bridger:Like a taste, kind of.
Glauber Costa:Look. You can book a meeting with me through our CLI. Right? You can leave feedback to Terso through through our CLI. So 11 it it was really I think a lot of a lot of the times, and and as I said, maybe this would have been how we would have landed in a different universe.
Glauber Costa:You believe that your product has a UI. And then the CLI is is a way to automate some tasks, is a way to get some stuff done. But for us, the CLI was how you use the product. Right? So, we had this this this thing that, like, everything that you could possibly want to do must be doable by the CLI.
Glauber Costa:All of those PLG mechanisms that companies put, like, for example, and we want it in the beginning. Now, I think we focus a lot more on just, like, a 10,000 databases, database per user, and and such, because it's been more successful. In the beginning, we focused a lot on data replication, which was something unique that SQLite didn't have. We still focus a lot on that. It's it's just that, you know, this other use case is is slowly overtaking, replication as the main reason people come to Tersa.
Glauber Costa:So you create a database. And look, I wanna show I don't wanna wait for you to go to the documentation. I don't wanna wait for you to hear about the fact that Tursa has replication. I wanna show you right there. Right?
Glauber Costa:So just, so in in an onboarding process on the YBY, this is how it would happen. Like, you would log in and then you create a database now. Do you wanna create a replica? So we wanted to bring all that experience through the CLI. And so that that that's what we did.
Glauber Costa:It's like, we had this understanding that this cannot be a slice of the product. This is the product. And I think just going in that mindset helps a lot. Right? Because now we don't have those excuses of, this is really not a good path for the CLI.
Glauber Costa:Your thinking is, I need to find a way to make that flow make sense in the CLI. And and and there you go. Right? And you look at the company, other companies, for example, I met with Fire, they have a CLI, which is a way to create a project, is a way to get started. But once you do that, like this usually, that's it.
Glauber Costa:I mean, that's the role of the CLI. But for us, the CLI has to do everything. Every there are some things that don't really make sense on the CLI. So those are web only. Like, for example, when you create a when you sign to swipe a credit card, you have a Stripe integration.
Glauber Costa:There's not no way you can do this with
Jack Bridger:the CLI. CLI.
Glauber Costa:But the CLI would take you to the web page. You can start the process also. We're always thinking whatever is doable in Trisha has to be done on the CLI.
Jack Bridger:Yeah. That's okay. That that's cool. You're like, did you ever see that, Justin Khan, the one that founded Switch, and he was like Yeah. He's like the wearing the live stream thing before, and then it was like I don't know.
Jack Bridger:Kind of somehow feels similar that you're just like, you're so CLI. Like, it's it's it's good because you're just living, improving it, and putting it.
Glauber Costa:No. I I I got fairly involved in the CLI in the beginning myself. And one of the things that I was doing a lot was exactly, just those kind of onboarding mechanisms. I know we created our app we created a database. Now it's time for you to create our app because I got very personally involved with that.
Glauber Costa:And I think another thing that helps as well is that from beginning, we wanted to search the API first. We actually do not even CLI first. So our CLI was also very simple, because most of the time what it's doing is calling every, is calling a bunch of APIs. Right? So this is one of the things that Jamie has been doing very well, is showcasing the platform API.
Glauber Costa:Not by accident, when he joined, I did tell him like, this is the main thing that we should be talking about, like, you know, he's doing this very well. So you can create databases, you can destroy databases, you can manage your users, you can manage your tokens, you can do all of that from the API. So the CLI is really it it also makes it simpler to to to use it, to to code it.
Jack Bridger:Yeah. It makes sense when you put it like that. One of the things that I kind of wanted to circle back to the, as you were speaking, just you were talking about, like, kind of the economic climate and stuff like that. And I wonder, like, if, part of the reason that Tersos, like, hit home with people so much is that, like, you're able to do it more cost effectively at a time when people are, like, being more cost effective. And it's, like, something that I feel like doesn't get taught about very often among dev tools.
Jack Bridger:Do you think it's been a factor?
Glauber Costa:Maybe. It it's hard to say. Right? Because look. If think if we were in a different macroeconomic environment, we would just have kept the free plan, like a free only offering for a lot longer.
Glauber Costa:Right? So that's the thing, like we, we had to start like putting plans and putting some restrictions quite early on. I think because of that in in a different world with a dessert world of, no interest rates and free money, I think we would have had the luxury of just being growth mode only free for years, which is what what a lot of companies did. Right? No, I don't know.
Glauber Costa:It's hard to say. I think the, yes, there's a lot of people that flock to Terso because of how generous our free tier is. But what we noticed from this crowd is like that they wouldn't be paying a lot for other services anyway. It it's mostly like solo builders, free free, you know, side projects, and etcetera.
Jack Bridger:But some
Glauber Costa:We actually get a fairly decent amount of yeah.
Jack Bridger:Yeah. I was gonna say, like, in their heads, sometimes, like, I BV about is, like, they don't probably think of themselves as like, okay. Yeah. There's no one using it right now, but, like, in a year, there's gonna be 100 of 1,000 of people. Yeah.
Jack Bridger:Tesla is gonna be cheaper.
Glauber Costa:What what what I what I mean is that, like, the kind of applications that this crowd building is definitely the crowd that adopt the Turso the first and and like from from the get go. If it wasn't for Turso, maybe they will be paying like $5 for another solution and for Turso, they could get it for free. Yeah. Right? So it's not because the workload doesn't demand that.
Glauber Costa:It's not something that you pay 1,000 of dollars a month. Now today I had a call, again, we were a lot more mature. So we had a call with a prospect that's spending $8,000 a month on their database solution. We believe we can do it, you know, 2 or 3. So it's a big saving.
Glauber Costa:Now I think those things start to become more of a factor. But in the beginning, like, look, is the difference between $5 that much? You know, at the end of the day, people want it, people people ask for it. Yeah. But if you really have no option, and if you have to just pay $5 to use a service that your side project depends on.
Glauber Costa:You pay more than that for Netflix. So it's not a big deal, and and people end up doing like, my wife runs a side business, and she pays, like, $30 a month for a bunch of tools that that she uses. Like, it just doesn't need a database.
Jack Bridger:Okay. I was gonna say, is Tarsow a part of that?
Glauber Costa:No. No. Just that she does a company. She uses, like, Adobe stuff and etcetera.
Jack Bridger:Okay.
Glauber Costa:So so she pays for the for those tools, look, a small amount. All all of those companies and all of those offerings would have starter tiers and things for hobbyists and etcetera. So is is the difference between 10 and and 0 that much? Not necessarily. So I don't I don't think there was a major factor.
Glauber Costa:But again, that I think the fact that we can, due to our cost structure, just offer this at such a low price. Of course, it helps adoption. Of course, it helps adoption. But I I don't I don't think this is because of the macro environment like, because interest rates are 0, I'm not 0 now. I have to stop being 5 or I'm gonna spend 0.
Glauber Costa:This happens maybe it's gonna happen more now. I mean, it happens more, you know, at the later stages where you're really talking about a $10,000 bill, right, just or a $5,000 bill.
Jack Bridger:Yeah. I don't I I don't know. Like, I feel like it could be because, like, I'm in, like, the kind of, like, a lot of the indie hacker Twitter I see, like, select level scio, you know, like the like, a lot of what he talks about is, like, how his costs are, like, really low, and he's got, like, a really simple stack. And I show,
Glauber Costa:like levels. Io makes 1,000,000 of dollars a year. I mean, like, at this Yeah.
Jack Bridger:This And he's still like he's like, and I spend barely anything on it. You know?
Glauber Costa:But but, look, 0 that's that's exactly my point. $010 doesn't really change its cost structure.
Jack Bridger:But but maybe it's, like, where you kinda know, like, okay. As it scales, if it's cheap now, it's gonna be cheaper going live.
Glauber Costa:Sure. Yeah.
Jack Bridger:Right? But Yeah.
Glauber Costa:In his levels, I owe the and that's, I think, the central part of my point. Is he doing this because that's how he operates, or is he doing it because of the macroeconomic environment? He's not doing it because of the macroeconomic environment. So just
Jack Bridger:That's true. He's always been that.
Glauber Costa:That's exactly my point. Like, this this the the the initial crowd that that came to Tirso and of course, we're very happy about that, came, and I I I I that that I I fully acknowledge. I think the fact that we could offer so much for so little, right, was definitely a factor. But is it because of the macroeconomic environment? No.
Glauber Costa:It's because that that crowd already has the mindset, and they would have paid $5, but, like, zeroes of matter. I just
Jack Bridger:You're probably right. Yeah. It's interesting, though. I still think it's, it's it's very interesting.
Glauber Costa:Well, people talk a lot. I think in this community, people talk a lot about, oh, how much can I do with a $5 VM on digital ocean
Jack Bridger:versus like, it's fun to talk about, like, you know, build everyone loves it when someone could build, like, a $1,000,000 startup with, like, a $200 a month?
Glauber Costa:Yeah. As high margins as
Jack Bridger:it can be. That's the dream. Yeah. Klava, I think that's what we've got time for. But the last thing I just wanted to ask you is if you've got one kind of takeaway that you would give to other DevTools founders, early stage employees.
Glauber Costa:Move fast. Yeah. Again, this was the saving grace of. As I said, the reason this sounds obvious because everybody says move fast. Everybody says move fast.
Glauber Costa:But in reality, you could always move faster. That's what I learned. Because you have those thinkings like, look, just a I'll give I'll give you another example. For for launch, we were saying, like, we're gonna offer you, you know, 10,000,000,000 rows red per second per per month, and that's free. But we didn't have the back end infrastructure.
Glauber Costa:And I spoke about that in a different podcast as well. We didn't have the back the back end infrastructure to stop your reads, or even track your reads at the API layer. We just did not know. Right? Just how how many reads you were doing.
Glauber Costa:We consider a delay and and we we had those very aggressive timelines, very aggressive timelines. So I started to see chatter inside the company, or maybe we should delay the launch because it's gonna take us this much to to build this. And then we thought about it, and we thought about it, and I think we got very good at at doing this here at TRSO. And we realized, wait a moment. What happens if we just don't build this?
Glauber Costa:What happens if we just don't do it? Well, people people will want to know. Right? We're saying, like, we're gonna charge you per rows red, and people wanna know how many rows they read. Sure.
Glauber Costa:But you can actually get this from the client, from because you you you have the data. The the CLI can just connect you to database and and ask, how how much you read. And this is super trivial. You can do it in an hour, but we we cannot we cannot know, and we cannot block you. So our thought process at the moment is like, how many people are gonna complain if I promised you a 1,000,000,000 rows read per month, but you are able to do 2.
Glauber Costa:And we understood that the number of people will be 0 because like, you complain the other You know, you will complain in the opposite direction. Right? You promised me a 1,000,000,000, and then you locked me out after 5,000,000,000,000, you complain, but like, if I promised you a 1,000,000,000, and if you did 10,000,000,000, who cares? Right? So in the in the name of shipping fast, we said, don't do it.
Glauber Costa:Like, just just don't do it at all. Right? So so, there are, and the CLI is another example. Right? So I feel I think a lot of people have this understanding that, which is the right understanding, our product needs to be polished, our product needs to be loved.
Glauber Costa:And then because you have this mindset, you end up taking a lot longer to ship, you know, because you have your idea of, like, what the product polished product is. But because of our time pressure, we had this thing that we knew that off if if we didn't put something polished out, nobody would use. But we also were forced to think of other ways of making the product polished that wouldn't sacrifice the timeline. We discussed the WebUI example at Lance, I just gave you another example, about the restriction on Rose Red and I could probably give you 10 more. Alright, so we kind of got very good at adopting this mindset of like, okay, really, what is the minimum thing that we need to put out there?
Glauber Costa:Let's put the minimum thing out there. And look, it served us wonders. People are still surprised that we're so so young. And we are already, like punching way, way, way above our weight, the service got a lot better. But also the service got a lot better in the right way, because we didn't build our useless stuff, because there was a product out, you are hearing from users from day 1, as an example, like the 10,000 database, this is what I want, this is what I need, this is what I would like to do.
Glauber Costa:Right? So like, again, sounds obvious, sounds something like the, so that every startup founder will say every startup founder will hear. But I think the way I will put this advice is like, no matter how fast you think you can ship, you probably can ship faster. Right? So once you think about it and you're designing your way, like, oh, this is the fastest, ask yourself a little bit more.
Glauber Costa:Okay. But is this an even faster way? Like, where where where can we scope this in a way? Like, what kind of creative solutions do we have to ship it even faster? It pays off.
Jack Bridger:That was great. Yeah. That's that's really good. I I feel like I need to reflect on myself. Thank you.
Jack Bridger:That was very good. Glover, thank you very much for coming.
Glauber Costa:My pleasure.
Jack Bridger:If people wanna learn more about you or about TURSO, where should they go?
Glauber Costa:So first of all, there's not that many interesting things to learn about me. Definitely don't spend too much time on that. I am Ggl CST on Twitter. I spend a lot of time there. So definitely, you know, get in touch.
Glauber Costa:We're also on the tour. I'm also on the torso Discord server all the time. It's just as I said, I had a very boring career. So I mean, there's a lot about me that I believe people will be interested at. But if they, Thurso on the other hand is the other way around, they definitely recommend people get to know about Thurso.
Glauber Costa:We are at Thurso. Tech. So definitely take a look at that. You will find there the link to our Discord server, our Discord server at the moment has I think over by now way over 2,000 people. Again, we're also on Twitter all the time.
Glauber Costa:We have a great documentation, that Jamie is spearheading, at the moment. And and look, any questions you may have, just just ask on our Discord server. We will be happy to help you.
Jack Bridger:It's awesome. And I have heard that you are very open and responsive. That was, that's what I've heard is people can just DM you.
Glauber Costa:I try my best. I try my best.
Jack Bridger:That's awesome. Thanks, Glover.
Glauber Costa:My pleasure.
Jack Bridger:Thanks, everyone, for listening.
Listen to Scaling DevTools using one of many popular podcasting apps or directories.
