The Stack Overflow Podcast

Zero to MVP without provisioning a database

Episode Summary

Ben and Ceora talk with Sam Lambert, former VP of Engineering at Github, now CEO of PlanetScale, a serverless database platform for developers. Sam tells Ben and Ceora how obsessing over the daily lives of developers helps PlanetScale deliver a product developers love to use.

Episode Notes

PlanetScale is built on Vitess, the open-source database clustering system that runs at colossal scale hosting YouTube, Slack, and GitHub.

A familiar theme: Big cloud companies aren’t set up for independent developers. Sam and Ceora discuss how serverless can get projects—even businesses—up and running quickly.

Choosing the stack for a new business? Tools like Netlify can scale with your product, so you don’t have to change your architecture as you evolve.

Staging environments should be a thing of the past. That’s why PlanetScale enables database branching.

And finally, a question from Law Stack Exchange: Can satellite images be copyrighted?

Episode Transcription

Sam Lambert People laugh at the term serverless, of course there's servers there. But I think the cloud has drifted very far from end user value. And it's actually, I think, a horrible mess in a lot of places and you log into tools, you have to provision servers and clusters, and VCPUs and all of this kind of complicated stuff. When really, you have an idea, and you know how you want to get going and build something, you might have the next Slack in your head, the next idea for the next tool that will change the way everyone works, and you want to get going. And so it's super important to us that it's a serverless experience. It's an experience that's incredibly easy and intuitive.

[intro music]

Ben Popper Keep your code compliance ready for SOC2, ISO or HIPAA compliance audits with Rewind backups for GitHub. Rewind offers automatic daily backups, and rapid recovery for GitHub with audit reports, Amazon S3 Sync, security assessments and more. Learn more at rewind.com/for/github. 

BP Alright, everybody, welcome to the Stack Overflow Podcast. I am Ben Popper, the Director of Content here at Stack Overflow. And I am joined as I often am by Ceora Ford from Apollo GraphQL. Hi Ceora.

Ceora Ford Hi, hi, super happy to be here today.

BP Happy to have you here. Our guest today is Sam Lambert. He is the CEO of PlanetScale and software engineer, angel investor, former VP of Engineering at GitHub, and expert on topics like open source, and serverless databases. So Sam, welcome to the shop.

SL Thank you so much for having me. It feels amazing to be on the Stack Overflow Podcast. After years and years of using the tool. 

BP We are a venerable institution. 

SL Definitely. I'm very happy to be here. Thank you.

BP So Sam, I think I'd be remiss in a podcast like this not to ask just sort of right off the bat, to be VP of Engineering at GitHub must be interesting. It's the tool that to my mind is used by, you know, the vast majority of engineers, or at least well known by them, what's it like to do engineering at the tool that everybody uses for engineering?

SL It's awesome. It's wonderful. And for many, many reasons. One being the community, I think the main reason being the community and user base, developers are incredibly creative, wonderful, spirited people. And that's actually why I came to PlanetScale, to do pretty much the same thing, which is build tools for developers, because I love the audience. And at GitHub we've got this really unique view. And I think there's something special about GitHub that's in everyone's hearts, like everyone has their first programming. Not everyone, but a lot of people have their first programming moment on GitHub, or they discover a tool that saves them massive amounts of time. And how get this like you'd be out and about, you go to a conference and you speak to people and people would even say things to you, like GitHub was the first tool I ever paid for online. There was this emotional connection that the software development world had with the platform. And that's incredibly special. And I still interact with it a lot often, and there's like, you know, projects for home automation stuff that's hosted on GitHub has a whole community around it. And it just feels like such an incredibly special place and still does to this day, and being there was awesome. There was also a big responsibility as well. When GitHub was down, people were very, very unproductive. And, you know, package builds and people couldn't deploy their application. And we saw massive amounts of traffic would go to the three major clouds from GitHub, which was people deploying their apps, and people started to feel very unproductive when GitHub had issues.

BP Yes, Ceora, I don't know about your experience. But speaking from someone who doesn't code a lot, but yeah, has been living sort of in the developer world for the last two and a half years, feels like Stack Overflow, GitHub, Vim, or Emacs, these are things you could say to any developer, they have sort of an emotional experience. And for at least Stack Overflow, or you know, your IDE, often it's both positive and negative. You know, you love it, or you hate it. Yeah, I don't really get that same sense with GitHub, that they're people have like a sort of daily use and utility and satisfaction of a but also a daily frustration. Am I getting that wrong? Like, is there some cohort that has like a chip on their shoulder about GitHub, that's not the way it feels to me.

CF I don't think so I think more so people have a struggle with Git. But GitHub makes Git much user friendly, in my opinion. I know when I was first learning how to code, I used to always think to myself, Git is like one of the most frustrating things in my life right now. But I've never felt that way about GitHub. And the cool thing about GitHub is that like, no matter what side of tech you're on, or like what language you build with, or if you're a front end developer, DevOps engineer, whatever. We all have a shared experience with GitHub, which is really cool. There's not a lot of developer tools out there that you can say that for where like, everyone in tech basically has to at least know about GitHub, and probably it's very likely that they use it, too.

BP So Sam, you were sort of mentioning that, you know, this was a great experience you had, obviously it's something yeah, that's, you know, many developers know in love. What are you doing now and why did you choose to do it? What's the tool you're building out PlanetScale or the platform? What do you want developers to utilize it for?

SL Yeah, so PlanetScale is a serverless database platform, our mission is to give developers a database that is so simple and seamlessly easy to use that you do not need operations teams. That's why we call it a serverless database. I mean, people laugh at the term serverless, of course there's servers there, but I think the cloud has drifted very far from end user value. And it's actually I think, a horrible mess in a lot of places and you log into tools, you have to provision servers and clusters, and VCPUs and all of this kind of complicated stuff. When really you have an idea, and you want to get going and build something, you might have the next slack in your head right, then the next idea for the next tool that will change the way everyone works, and you want to get going. And so it's super important to us that it's a serverless experience. It's an experience that's incredibly easy and intuitive. However, the back end of our tool is powered by the very popular open source project, Vitest, which is the database layer and orchestration system that was built at YouTube to scale YouTube. So it's run colossal scale, it powers 100% of Slack's database, back end, all every single Slack message in the world is stored in Vitest. Similarly, with a lot of hyperscale, sort of customers of ours used Vitest power very, very large websites. And so this is the thing that really excites me about what we're doing is there's very rarely the tools you pick in day one of your company life are the ones that you that stay with you forever. Your language, the language, the framework they do. But Heroku, for example, just incredible platform, absolutely incredible. So many companies founded on Heroku, they eventually ran into scalability or pricing scalability issues, and had to move away. The thing that really excited about PlanetScale is when you sign up, and you provision and we provisioned your database in less than 10 seconds, it was a really great competition that someone ran actually on YouTube, and they stack them all up and we came first, for the first revision. The thing that excites me is you're provisioning something so quickly, and you're getting the back end stack that uses to power Slack, the largest chat platform on Earth. And that, to me, is really inspiring the fact that people will use these tools that will last them for the lifecycle of that company. And I don't think we've seen anything like that before.

CF That's interesting. I never even thought about that side of things.

SL I think it's a tooling, a lot of tooling is coming in that way, Netlify, again, more tools that you can, you can easily use at the beginning. And then there's this stack, that kind of now will scale very, very, very well. And it's kind of a new era, I would say.

CF I don't know if you know, this, Ben, but serverless has like a special place in my heart, because that was like my niche when I was first getting into tech. I used to like write a lot about serverless, and talk a lot about serverless. And give talks on serverless stuff all the time. And one of the things I love about it, like you mentioned, Sam, is that you can get something like a project up and running relatively quickly, even a business up and running relatively quickly. And like you said, the really cool part is that these tools scale with your product. So it's not like once you get like a certain size, you have to like change the whole architecture of your product to like fit the scale. I just think that's so cool. I love anything that makes like building stuff more accessible to everyone, whether they make it like easier or quicker or like make scaling easier. I love that kind of stuff.

BP So for someone Yeah, who's not as familiar with this, what's the difference between going to a cloud provider and saying, you know, I want to spin up this or that project in the cloud, and you're gonna sort of be my back end, you know, it's on demand in the sense that like, I could pay more, I could pay less and serverless. Like, what's the distinction between those two things?

SL I would say AWS now has ended up being a great place for operators, to set up infrastructure, install infrastructure, I would not say AWS was set up for consumption by developers. I think if you have some very simple use cases, you can get going. But it's very difficult. Like we compete with a couple of RDS offerings, and I've used them extensively. And my background is in databases. So actually, the reason I came to PlanetScale was my backgrounds and database. So I was GitHub's first database engineer, it could have been I spent a long time, even I get confused and stumble around on some of their tooling. And it's just extremely complicated. And for the cost, it just doesn't, the value isn't there. In the modern world, and Ceora, something you just said made me think that smaller, smaller teams are going to start to do bigger and bigger things. I think we're going to start seeing 10 person teams worth hundreds of millions of dollars. And that's extremely valuable in terms of tooling. If you're doing anything serious on AWS, you have to have teams of people managing for you. There's no way that you're doing that yourself.

CF Like I said, I got started in tech with like the serverless cloud stuff. And like getting started was very difficult for me because if you're trying to do anything. I don't want to say simple, but I feel like it overcomplicate things a lot. It makes things take a lot longer than they necessarily should. And it could get easy to get lost and to make mistakes into, you know, all of a sudden you have an $1,000 bill when you were trying to do something very simple. So I love like abstractions that make it easier, like cloud technologies and serverless are like great options for us. But it's so complex sometimes that it, it's like, is it a great option for us? But a lot of companies now are taking some of the things that AWS tries to do. And like, Okay, this is a good idea. But here's how it should be done. You know, I think that's cool. 

SL I think just time has moved on, people demand better taste in the tools that they use. I mean, look at phones, for example, used the business phone used to be the BlackBerry with the big keyboard and all that really horrible to use enterprise software. And people like rebelled against it, they wanted to take the iPhone to work, they actually wanted to have joyful user experiences or work. Look at Slack, I mean, look how people communicate. And before then, you know, Slack has this playful nature to it that kind of has had deep penetration in the enterprise, we want to use great tools that's happening now, in the software stack. People want to use tools that make them excited, and productive, and that are enterprise grade. And that's something that we're very, very happy about. And people break their brains. We have some phenomenal back end engineers, and we did a GitHub as well. And you see people getting stuck using these tools, when it should be way, way simpler. And I think people feel some shame sometimes that they can't understand this stuff. And I think there needs to be a reversal, as we shame on the cloud providers and that they can't produce something it's human usable or inspiring to use.

BP And so Sam, I guess, you know, if somebody were to pick up PlanetScale and start to play with it, what would you recommend as sort of an entry? Are there tutorials? Is it something where there's specific kinds of projects or companies that you'd say, hey, that's a good fit. If you're doing that, you might want to check us out? How open ended? Is it and what are the best onramps? 

SL Oh, that's a great question. So we have tutorials, when you sign up and login, we'll actually walk you through the product and how to use it, we have a few concepts that are really new in the world of databases that we're really, really excited about. So one thing that we enable is database branching. So staging environments are a big thing, right? Again, another thing that people have to manage and set up so that you have this area to test and play and you know, create your new feature. We don't believe staging environments are a thing of the future. In fact, we think they're a thing of the past. We believe that environments, isolated environments for your database should be as logically cheap to set up as a Git branch. So we enable database branching, which gives you a copy of your database in an isolated branch, they again, take around 10 seconds to create. And the idea is, you know, you're about to go and create your feature, you create a Git branch, you create a plan, scale branch, you connect your application there, and you have a fully isolated working copy, so that you can build your application. This also enables cloud based development. So these branches are in the cloud, which means you don't have to mess around making your local copy on your laptop, like consistent are work, which is again, just another huge and annoying promise or someone tweet even this morning that they had an idea, they went to go and do something and they lost their whole day on local development setups, right, I think this is a thing in the past. We've seen GitHub Code Spaces, which I think is going to be revolutionary for this future. And PlanetScale is built to be very long on that future of the idea that we all do cloud development pretty much exclusively in the future. And we enable that really natively. The other thing we do is we allow you to change your schema through a branch. So if you want to manage your database schema, you want to make additions and or change it, you would make it on a branch. And then you say you want to deploy the schema, we deploy that into production for you fully online, no locking, no outages, it's very rate limited. That means we solve a immense amount of pain, I speak to organizations that take months and months to get schema changes into production. And we can now make that a very, very seamless experience. So I would recommend if you sign up, learn about how we see that like this mental model that we have, which is very much around a database that comes around the software development lifecycle with you, rather than being something that's there at the very end for you to like, deploy to and break probably. And yeah, so I think we're starting to see people use branching is a hell of a lot. People getting really excited by just cutting out branches, getting a copy of the database immediately in the cloud, no screwing around. And that's the thing that really excites people. And if you join up, sign to the product, that that becomes very clear and immediate for you to start playing with.

CF I have a quick question too that I'm wondering, I think so far through hearing you describe your product. And like the process of building the product, it's very clear that you put the developers experience first. And I want to hear more about like the process of that. I think other products, like they have a great idea. But the execution isn't necessarily what developers actually need are looking for. But I think it's the reverse, or not the reverse, but the opposite with PlanetScale. It seems like you've been very intentional about creating a product that developers are actually going to enjoy. So I want to hear more about like, how you went about doing that or like when you What's the process like for when you have an idea for a new feature for your product, making sure that it's something that developers not only need, but will want to use and enjoy using.

SL That's a really awesome question. So it comes from many, many things. First of all, we obsess over the daily lives of developers, and I think a lot of folks in the room infrastructure space that are building infrastructure tools, they think about the needs of the infrastructure too much without the balance of how people interact with it. And if you overdo it on the needs of the infrastructure, you start to get comfortable with passing on too much complexity, like you think, oh, you know, we're solving a hard problem back here, databases are really difficult. So let's pass on some of those problems to the user, because they'll, you know, they're, we're solving something for them, they'll be grateful that we're doing that, and they can adapt some pain. So our tolerance for pain and our users going through pain is extremely low, we refine and refine and refine. And that's when it leads into our culture. So we've regularly delay project because they're just not there yet. And we run through them continually as a group, refining them with our engineers, we're not really a very product management led company, we've hired a kind of a profile of engineers that have built really, really great products for developers in the past. And they have a really good level of taste as an engineering team. And so we spend a lot of time really thinking about how we would use this product. And we ask questions like, why is this two steps when it could be one? Why isn't this instant? And if you have that bar, and you ask those questions of yourself first, and then do the engineering work to make that real, then you end up with something great. So for example, database provisioning, you only do it once, right? And I, when I was signing up on using competing products, it was like 25 minutes that you have to create a database. And people probably happy with that, I'm sure. But we asked the question, why couldn't it be straightaway? Why can it be instant? Why can't you have that magic feeling? It's that feeling of you know, when you unpackage, your iPhone, it's half charged, and you can use it right away, you want to get going, you want to build, you want to you want to do things. So we ask ourselves, why a lot. And we don't kind of settle on poor answers. No, just because or it takes a long while we do the work to make things speedy and quick and, you know, involves having an amazing engineering team that really care about our users and giving them the space and time to build things in the way they would like to consume them.

BP Yeah, I think that's really interesting, what you're saying. It must be fun and sort of, it's a nice position to be in, I think maybe all three of us are, which is that you're working with developers to build tools for developers. And so you're kind of like always using it and dogfooding yourself. And you can say, would I like this, like, is this a good experience for me? So you know, like, the builders and the end users, or the consumers are kind of the same, which is interesting. Yeah. And then I was thinking about, like, to what degree do you think in 20 years, you know, everybody will be writing some code, or using some no code apps and have some knowledge how this stuff works, kind of, like, from my generation, everybody could build themselves a website, and you know, prior generations, that would have been like, pretty challenging. They could jump onto WordPress and figure that out, or now everybody, I guess, probably spends most of their time on their social media profiles. But like, you know, they all know how to edit video and music really well, right? Like, they've all learned those skills, which once used to be highly specialized. So do you feel like this world that the three of us live in now will become increasingly sort of mainstream, you know, when I look at like a statement from it was Google Cloud they they make, they're like, we're gonna hire 40 million developers in the next 10 years. It's like, when we talk about Stack Overflow, we're like, well, there's 20 million developers in the world, and 50 million people a month come to Stack Overflow to use code. So if you're going to add another 40 million, you know, like, that's pretty ambitious. But yeah, Sam, you know, you say PlanetScale. So I'm curious, like, what how big of a user base, you know, like, who do you think is going to be using this, you know, 5, 10 years in the future?

SL I hope many, many people, and I do hope that the barrier to entry for building software gets lower and lower and lower. And I think we're starting just across the industry. You know, I speak to a lot of folks about this, and we spent time with, you know, our partners that are doing this in the front end space, like Netlify. And, you know, they were always saying, well, when is the database gonna show up and do this, because it's such an important part of the stack. And now we're here and there's other tools coming in. And it really excites me that tiny teams now can build a really successful business leveraging loads of these tools. And the barrier of entry is unnecessarily high right now. And I think it's going to take a lot of people thinking through this, and it doesn't need to be this high. It just like, there's no good reason for it. And I think by us making the shift as an industry, more people will have access, which will be great for society, it will be great for all of us, and we'll just start to see more economic uplift everywhere, which will be really fantastic. And that's what gets me excited. And I think there's so many industries that are still left to truly adapt like high technology, and really transform themselves with technology. And the more that happens, the better the world gets, in my opinion.

CF I just want to speak to your point of like seeing more people get involved with development on all sides of the stack. My younger sister, one of her school projects, she's in her junior year of high school was building a website. So this is something that we're seeing like even students starting to do, and it's become so accessible to the point where like, she has no prior coding experience, but she's building projects like through, you know, some no code tools and things like that. So I definitely think we're gonna see a huge increase, like her assignment was to literally build a website, you know, I think we're gonna see a huge increase of people who aren't necessarily developers building like really cool stuff with tools like PlanetScale. So I'm excited to see where things go in the in the future. 

BP Gotta bring her on and see if she had a good experience. Was that fun or was it horribly frustrating? I need to know now.

CF We'll see, I have to ask her.

BP Don't help if she wants you to help her cheat. [Ceora laughs]

[music]

BP Alright, everybody. Well, thank you so much for listening. We really appreciate it. I am going to read out a fun question from across the Stack Overflow Network. Today, we are going to pick a question from our Law Stack Exchange. Can satellite images be copyrighted? So this person is asking, well, they're just a bunch of facts and GPS coordinates. It's a picture something we all know how to understand. And if you want to know how they can be copyrighted, there's an answer for you there. I am Ben Popper. I'm the Director of Content here at Stack Overflow. You can always find me on Twitter @BenPopper, email us podcast@stackoverflow.com. And if you like the show, please do leave a rating and a review. It really helps.

CF And I'm Ceora Ford. I'm a developer advocate at Apollo GraphQL. So if you're looking to find me somewhere else on the internet, I spent way too much time on Twitter. My Twitter handle is @ceeoreo_ on Twitter and I also have a website ceoraford.com. But if you really want to reach me, like I said, I spend way too much time on Twitter.

BP Sam, tell the people who you are, where you can be found if you want to be and where they should go to check out PlanetScale.

SL I'm Sam, CEO of PlanetScale. You can find me on Twitter @ISamLambert and you can check out PlanetScale at planetscale.com. We have the most generous free tier that's out there in the world of databases. You can get going, build your full project or company on our free tier. And then we'll take you up to hyper scale. And yeah, I hope you enjoyed checking out the tool.

BP Alright, everybody. Thanks for coming on and we'll talk to you soon.

[outro music]