Tobie Morgan Hitchcock, Co-founder & CEO at SurrealDB

0:00Tobie Morgan Hitchcock:We are back. Welcome. I am here with Tobie. Tobie, give me a quick pitch into who you are and what you're building. Hey, great to be here. So I'm Tobie, co-founder and CEO of SurrealDB. We're building the kind of next generation database that will be suitable for traditional applications, but also is suited to AI applications and agentic memory as well. Amazing. How broad are you targeting? You know, I read that this is sort of like a general database that you're trying to build it for almost anyone. Is that right? it is i mean it's it's hard to be to to launch a database in the general sense and not be used by people you know building traditional applications so we've got we've got clients who are migrating older applications traditional applications and then we've also got people building new completely new agentic workflows agentic memory back ends with surreal db i'd say for us, the applications of today are the agentic memory, the AI applications, knowledge graphs, things like that.

1:01And that is where a lot of our users are using SerulDB. But generally speaking, SerulDB can be used in any application.

1:07Tobie Morgan Hitchcock:And you launched back in 2021. And when you launched, was the vision to be the database for a lot of these agentic applications? Or did that come about just as AI blew up in the last few years? I think I'd love to say that I kind of foresaw that and that actually the need for multi-model data in agent or AI workflows was what we needed. But generally speaking, we were building, when we launched in kind of early 2022 and then we took off on GitHub in later 2022, we were just building applications or a database for modern applications, I would say, applications that wanted real time and wanted to be able to query data in different ways.

1:48Now, that had a very good application in AI apps and AI agents, so it made sense. But I don't think that was truly visible back in 2022.

2:01Tobie Morgan Hitchcock:Yeah, it makes sense. And you touched on, I guess, was it your breakout moment on GitHub in late 2022? Can you talk about what happened there and kind of what enabled you to take off? Yeah, so we open-sourced the database, I'd say probably April or May in 2022. And we were just building. So we built a significant amount. I think it was hundreds of thousands of lines of code. We'd rewritten it from Golang into Rust. And we were really trying to push the boundaries of what's possible in a database. And I don't think many or any database has tried to bring as many data models together in a single database, a single platform, and a single query language.

2:38And we were driving forwards towards you know what we wanted to to be a kind of stable version one and in august september i think it was august of 2022 um we were featured on fireship and we were we kind of took off on on hacker news and and it kind of just blew up i think in one day we had 2700 stars uh alone and over the course of a couple of weeks we i think we got 12 000 stars and we had developers starting to use us and organizations starting to build on us and we weren't even at a version one at that point not for another year did we actually get to our version one but i think people saw what we were trying to build the ambition of what we're trying to build and and what that would mean for the applications that developers and organizations were trying to build themselves amazing and can

3:26Tobie Morgan Hitchcock:you touch a bit on you know what you have done differently or the problems that you were solving that enabled people or that kind of meant that people loved what it is that you've built I think the biggest thing is just trying to bring all of these models together. If you look at a traditional tech stack, all the way back to kind of 2015, but actually it's even more important nowadays because the same issue is happening. So if you go back to 2015, people were building on time series databases and graph databases and relational and document. And the big term back then was big data. And really nothing's really changed.

4:02People are now building with vector databases and graph databases and document stores. And, you know, it's slightly different terminology, slightly different use cases, but it's the same thing. People are using workflows and pipelines to kind of piece this data together and make it work in the application. But nowadays, I think that's even more important because people are trying to get the best insights and the most accurate insights into data for real-time chat, for real-time decision-making, for real-time understanding of whatever the user is wanting to do. And so being able to query different models of data, whether that's graph relationships or whether that's document attributes or whether that is file data itself or some kind of time series temporal analysis over data, in a simple way that's performant and enables you to analyze all that data in one go and then bring the results out, has always been important, but it's even more important for today.

4:58So I think it's just the ambition of what we were trying to build, the simplicity of how we wanted to build it and enable people to deploy it, and doing that with a kind of developer mindset, making it really easy for developers to get going, to use it, to hack on top of it, and things like that.

5:17Tobie Morgan Hitchcock:Amazing. Can you touch a bit on, I think you raised a big round last year, was it series a was it as series a 20 million dollars you know what enabled you to to unlock that big round i guess yeah it's a good question i think you know if you're looking at the infrastructure space you always have to have a large amount of capital up front in order to build because these things take time they take time to build they take time to gain the trust of users and they take time to kind of win those customers over. For us, we had an incredible amount of early traction from the open source community, and that is obviously a big proof point.

6:00Over the course of those two years, kind of when we launched our seed round to our Series A, we had large organizations, Fortune 100, Fortune 1000 organizations using us, deploying us in many different ways. Some people deploying us in kind of large scale deployments, distributed highly available deployments other companies deploying us at the edge or embedded in in devices on device so because of that because of the way that people were using us it's kind of showed a proof point of of what people wanted from Sururidby what people needed and therefore you know what the potential was so that that's probably the biggest the biggest reason those there's proof points from the community and uh and just the different that the wide variety of of use cases that people were starting to build on sort of or had been building on sorority b

6:51Tobie Morgan Hitchcock:already and were there challenges in raising money both that the series a and the seed given that you are building you know a highly technical infrastructure business which i imagine a lot of vcs may not have familiarity with and it may be very different to the typical businesses that VCs invest in? It's a good point. I mean, look, there's definitely VCs that invest in infrastructure and VCs that don't. And those that invest in infrastructure know it very well. And those that don't invest in it typically don't understand what's required or what's necessary. Infrastructure is one of those areas which has incredibly successful outcomes.

7:30If you look at some of the biggest tech IPOs, a lot of them have been infrastructure or data based. But, you know, you need a lot of capital to get to that point. You need to have a lot of success. You usually have to go open source. There are lots of risks to it as well. So, you know, it's a high reward, but also high risk investment. I'd say, you know, was there a struggle for us? Look, we were incredibly fortunate to have that early open source traction and the significant use cases from large organizations and large deployments. But having said that, it was also 2023, 2024. It's not the best market that there's been in the last 10 years.

8:13And I think a lot of startups have struggled over that time. But we brought on a new investor who really believed in what we were building. And actually, they had a database background. the founders of that VC firm had a database background themselves, which was really important to us. And also, our seed round investor doubled down and invested in our series as well. So, it was, you know, it was right for us. But generally speaking, infrastructure can be a challenging area for VCs generally.

8:47Tobie Morgan Hitchcock:And, you know, it's amazing that it worked out, I guess, and that you found those investors who have got the experience and the conviction in this space. I also wanted to touch on a lot of the AI hype that we're seeing at the moment. So we're seeing a lot of hype, crazy valuations. And I'm curious, do you see that happening at the infrastructure level as well, especially as you're kind of building a database to enable a lot of these agentic tools? It's every single infrastructure company is different, right? You have those who go out there and they get their first three customers and they go slowly um they open source it you have others who don't open source and go straight for enterprise i'd say that mark that that model is typically harder especially nowadays um and then you have those who go fully open source and go bottoms up and then try to uh try to work out down the line how to how to go after the enterprise i'd say we're kind of sitting in the middle somewhere where we've got that open source traction we've got significant amount of enterprise adoption and enterprise use um so you know we're kind of beneficial you know we benefit from sitting in the middle and and uh that's great for us i'd say generally speaking uh every infrastructure company is different and how they raise money when they raise money what they do with that money uh is dependent on who the the founders are which market they're operating in um who they're selling to is it bottoms up is Is it developers?

10:14Is it enterprises? So it's hard to say generally, but yeah, if you look at any of the 20 databases out there at the moment or more, they've all done it differently and they're all continuing to raise in completely different ways.

10:28Tobie Morgan Hitchcock:And can I ask, are you thinking about fundraising at the moment? I think as a company, as a startup, but also as an infrastructure company, you've always got to be thinking about fundraising. Our Series A was completely pre-revenue. We didn't have any revenue at that time. and it was purely about building out more of the product and of the database. But, you know, as a startup, you've always got to be thinking about raising your next round, what are the targets to get to, what does success look for the round that you've just raised and to get to the next round. So I can't say we're not thinking about it, but also you've also got to focus on the product and building the business and progressing forward to that next stage.

11:08Tobie Morgan Hitchcock:Got it. Okay. Sorry. I'm also curious, your background is as a software engineer. What advice would you give to engineers at the moment about the tools that they should be using to help them with AI? It's a good question. So I'm actually self-taught. My undergrad degree was in French and I wouldn't say that's very useful to me nowadays. and then I taught myself software engineering and computer science and then I went and did a master's in computer science. I'm a big believer in teaching yourself. So that's the first thing, not learning from university courses or books, just teaching yourself and hacking at things.

11:54I think that's really important in today. Today is the way that people are building applications and learning AI today. It's all about hacking. It's about using as many tools as possible, using as many different LLMs, chatbots, integrations, platforms, testing, building, hacking. I think that's really important. So I wouldn't say there's one tool. You know, personally, I use lots of different models, lots of different tools, and they're constantly evolving. I think, you know, it's safe to say, and I don't think this is news to anybody, the market, the industry is moving incredibly quickly. and one day Google could be leading and Google's model could be leading and then the next day OpenAI or Claude's model could be leading with regards to coding.

12:41I think you've always got to stay on top of it, ahead of it, and just try to use and learn and hack with as many different applications and tools as you can. And I think by doing that, that gives you the best way of learning what's missing, where the gaps are and where you can build or contribute or start something yourself.

13:00Tobie Morgan Hitchcock:And do you see AI as posing a threat to engineering jobs, civil engineering jobs, or do you see it as an enabler? I think we're at this stage where it's an enabler in one way, but then it's also changing the landscape in another way. So if you're an experienced engineer, and I guess it also depends on what product, what platform, what tech you're building as well. If you're building websites, front-end applications, things like that, obviously this is a complete game changer now because this is able to build it very quickly, add features and iterate very quickly. If you're building something like SerialDB, it's an infrastructure project, it's a database, you have to have that experience with regards to how does it scale, how does it distribute, how does it ensure data is consistent and durable.

13:55And that's not something that AI can necessarily give you right now. It can reason with you, but you have to be able to understand where it's going wrong. And I think that's really important. Right now, the need for experienced engineers who are able to look at the output or the results of what AI is generating, actually say, look, actually, this is great. This is right. But actually, this needs improvement or this isn't right. And obviously, this is incredibly important from a security perspective as well. but on the on the entry-level software engineer job space and the market as a whole you know that's that's completely changed in the last six months that's completely changed now AI the chat bots the LLMs the integrations can effectively do a lot of what the the junior software engineers and junior computer science graduates were able to do so I would say it depends what you're building it depends on what level you're at.

14:49But I'd say right now, the need for more experienced engineers who can understand and understand at pace what the AI is generating and critique it or improve it or outright say, no, that's not the way we want to do things. But that's going to change. Six months ago, we weren't at this point. Six months time, I don't know where we'll be. So will that change in the future? Absolutely.

15:11Tobie Morgan Hitchcock:And how much is AI impacting what you're building at the infrastructure level? Is it enabling you to build more and to do better things or is it changing the nature of the product itself? I wouldn't say it's changing the nature of the product itself yet. I think AI can definitely help iterate on certain things. But as I said, you have to have a really good understanding of what it's doing and how it's doing it because it can come across as very certain and very sure in what it's writing, what it's building. And it can talk about how it improve this and improve that but actually its implementation may not be totally accurate or totally performant or you know could be completely insecure so i think it's really important to to look at the output of of ai tools or ai in general and be able to understand it and critique it um that said you know what people are building with surreal db and on top of surreal db has completely changed six months 12 months ago people were still building applications they were building them in a traditional way and now people are building agents and they're using to be the memory source for agents and to reason about memories and history and context and facts and albeit the way that's really does that isn't really different i think the way that we have to integrate with different tools and different platforms on the market has completely changed six months ago there would have been a few python frameworks that you had to integrate with and now Now there are hundreds and there are more coming out every single day as tools build these frameworks for different workflows or pipelines or agentic methodologies.

16:49So I think the way that people are building on top of CeruleDB has completely changed and is continuing to change every day.

16:56Tobie Morgan Hitchcock:And how much do you have to prep or build in advance for things like that? Or can you not? Is it purely reactive? You see what people are building and then you try and adjust it to match them? Or are you trying to think, OK, six months down the line, we think people might be building these types of tools and therefore we need to make SurrealDB do X, Y and Z? I think with regards to the tooling and the integrations, you just have to build what's popular at the time. You know, these things can be very popular and then another one can come along that takes that limelight. I'd say in terms of the feature set that SurrealDB offers to enable those frameworks or that functionality that's needed, that doesn't really change.

17:35And I think, you know, we're already pushing the boundaries of the feature set and the functionality that developers want or need in a database. And being able to combine those features together to give the outcome that developers or organizations need, I think, is really important. So iteration on new features, I don't think, is really our aim. But making sure that what we've built, how we built it, is really stable, really performant, and can scale, I think, is our aim. But in terms of the tooling integrations, you've got to keep up and just continue to see what evolves and what changes every day.

18:11Tobie Morgan Hitchcock:Yeah, it makes sense. And I'm conscious we're running out of time. But one of the things I wanted to talk to you about was you founded this business with your brother, right? Can you touch a bit on what that experience is like? I can imagine that like any time you meet up as a family, it must just be work chat. So are you able to kind of separate your brotherly relationship from a co-founder one? I'd say, you know, my brother and I have worked together for a long time and we're very different. So he's not, albeit he can code, he's not an engineer. He's really a designer and marketer by trade, as it were.

18:48And our understanding of the business comes from different aspects. Jamie is very focused on the marketing and the outward facing aspect of the business. and I'm very focused on the tech and the direction that we go. And I think that is important because, you know, we don't have necessarily the same ideas

19:14over each other's own ideas, if that makes sense. And I'd say, but, you know, coming back to it, we work hard, we work a lot. And I would say, is there actually that divide between personal relationship and business relationship? I think when you're building a startup, most startup founders or even some of the early employees, early engineers will know this. It's very hard to have that divide between what is business and what is spare time. And it's much harder, even more so when you're founding that business. um so i would say you know when we when we meet up uh as a family or with our kids and with our wives uh yeah it's it's very much always focused around work still uh that might not be the best uh the best thing but then at the same time i think we both love what we do and you know we want to uh we have the same vision and the same uh dreams for for sorority being what it could

20:11Tobie Morgan Hitchcock:become so that definitely helps amazing no it's great to hear that you can sort of yeah yeah i mean i kind of agree when you're building something you kind of lose the boundaries between personal and work and you just kind of invest all your time in it uh but i think building with your brother is an amazing thing to do um but look toby thank you so much for joining a conscious amount of time but it's been great chatting i love what you're doing i love what you're building i keep i'll keep watching along uh and anytime you want to come back on or share some news just let me know awesome that's great thanks so much for having me it's been it's been great thanks toby take care see ya