How Sentry's David Cramer Turned an Open Source Side Project into a $3B Business
David Cramer, co-founder and CPO of Sentry, joins host Dan Levine to share the unconventional journey of turning an Open Source side project into a $100 million ARR business now valued at over $3 billion.
From his early days as a teenage gamer to becoming a key player in the Django community, David recounts how a simple error-logging tool evolved into an essential platform used by over 100,000 organizations worldwide.
In this conversation, David also discusses the Open Source Pledge. Recognizing the critical need to support Open Source maintainers, this timely initiative from Sentry and others aims to create a sustainable funding model for Open Source, encouraging companies to support the often invisible labor powering modern software development.
- 2:04 His path from Midwest fast food worker to self-taught programmer
- 07:09: The origins of Sentry as "djanjo-db-log" and its evolution through various tech companies
- 10:41: Why design was a priority for an Open Source project, even 15 years ago
- 11:34: The decision to bootstrap and accidentally start a business through Heroku add-ons
- 12:34: Balancing a day job at Dropbox with the growing demands of Sentry
- 15:18: The competitive drive that led to raising venture capital and expanding beyond Django
- 16:27: Sentry's mission to be used by every developer and its unique approach to building a business
- 17:34: How maintaining an Open Source ethos drove Sentry's growth
- 21:14 Spearheading the Open Source Pledge to address fair compensation for maintainers
Whether you're an aspiring founder, Open Source contributor, or just curious about the intersection of community-driven software and commercial success, this conversation offers valuable insights into building a developer-focused company with a foundation in Open Source.
Featured: David Cramer, CPO and Co-founder of Sentry; Dan Levine, Host
Learn more about Accel’s relationship with Sentry:
- Why Accel is Supporting the Open Source Pledge
- Helping Developers See and Solve Quicker: Our Enduring Partnership with Sentry
- Our Series D in Sentry
- Our Series E in Sentry
- Our Continued Support of Sentry
- The Future of Open Source: Launching the Open100
Explore more episodes from this season:
S2E1 | Klaviyo’s Andrew Bialecki on proving a tech startup can be built anywhere
S2E2 | Webflow’s Vlad Magdalin on the biggest lessons learned from bootstrapping and raising capital
S2E3 | Syrup’s James Theuerkauf on building an AI-powered product that cuts through the noise
S2E4 | Sysdig’s Suresh Vasudevan on embracing a “challenger mindset”
S2E5 | CrowdStrike’s George Kurtz on building a generational company
S2E7 | Remote’s Marcelo Lebre on the future of the global workforce
S2E8 | Gamma’s Jon Noronha on how early-stage startups can challenge industry incumbents
S2E9 | Monte Carlo’s Barr Moses on creating a playbook for a product that’s never been built before
S2E10 | ConductorOne’s Alex Bovee on the critical ingredients of a high-growth startup
S2E11 | Netskope’s Sanjay Beri on building an iconic company through controlled innovation
S2E12 | Chainalysis’ Jackie Burns Koven on building trust in new technologies
S2E13 | Wonder’s Marc Lore on blending vision and execution
S2 Bonus Episode 1| How Linear’s Karri Saarinen is Redefining What Scale Looks Like
David Cramer (00:00):
I don't care if we don't make money, nobody else will. That's it. That's it. It's a zero sum game. If we don't succeed, nobody's going to succeed. We will burn the whole thing to the ground. We'll make it so good and so cheap that nobody will have the opportunity to exist. Welcome to Spotlight On where we examine the technology shaping our world through conversations with the people building it.
Dan Levine (00:20):
I'm Dan Levine, partner at Accel, and I'm thrilled today to welcome my friend David Kramer, trader of century. David, thanks so much for joining us. We've known each other for over a decade now. It's got to be. When did you join Dropbox? That's what it would've been,
David Cramer (00:35):
20 13, 14, 12, 20 13. I don't know. I think, oh yeah, 2012. I left at the end. End of 2014, beginning of 2015.
Dan Levine (00:43):
The same timeline. Same timeline. So we a long time. Before we get all of that though, we're here to talk about Century century's. The first investment I made at Accel in June of 2015 and then would continue to invest a bunch of times thereafter. But the company had, I think at the time, somewhere between two and five people. I think Arman might've been a consultant at the time, but not full-time. I
David Cramer (01:02):
Think there were people transitioning out of their jobs.
Dan Levine (01:04):
Yeah, exactly. Was right as that was happening. Half basically. Okay, so how many people work at Century today?
David Cramer (01:09):
It's somewhere between three 50 and 400.
Dan Levine (01:11):
Okay. So hundreds of people. Despite you not being a big company person, I think you're doing an excellent job, but hundreds of people. I think you guys announced a certain a RR milestone that you crossed a hundred million dollars.
David Cramer (01:21):
Yeah, that's a hundred million,
Dan Levine (01:22):
Which is unbelievable. A few months back. And how many organizations, you mentioned almost everyone uses Century. What does that mean from it's
David Cramer (01:28):
So many that it's hard to keep up? I used to pay attention to the number every single day of the week. And now because the number number's so large, I check in once a quarter. Best case it's at least 60,000
Dan Levine (01:37):
Paid.
David Cramer (01:38):
It's at least a hundred thousand total. Total organizations. It could be hobbyists, but it's like logos if you think about it that way. Yeah,
Dan Levine (01:44):
Totally. Which is unbelievable. So it's one of the most popular open source projects in the world and it's been incredibly successful as a company. So maybe give us a little bit about your background that led up to what became Century. Talk about how you got first into software, open source, where you were working, and the first project and project name that would become Century.
David Cramer (02:04):
Okay, we'll do the bridge version to keep it interesting. So Teenager worked normal teenage job from the Midwest, worked at Burger King. It's got to be a restaurant. Got to be a restaurant. Got to be a faster
Dan Levine (02:14):
Restaurant. Fast, fast restaurant.
David Cramer (02:16):
It's Red Robin teaches you useful things like customer service stuff. I agree. And how to work for money. And then I was a gamer. A lot of kids are, and I kind of was able to game of choice. Actually, I don't really know what year is this now? Yeah, so this is early two thousands,
Dan Levine (02:31):
But it's not World of Warcraft yet.
David Cramer (02:32):
World of Warcraft wasn't quite there yet, but that's key to the story. So it was in whatever was popular. Counterstrike Ultima Online was kind of like the imaginative thing and EverQuest was that day and era. But World of Warcraft was important. I was in games, but I also liked programming. I didn't really understand what that meant yet. And I found I could kind of take my love for video games and just figure out how to build these websites. And I was basically trying to build databases of content for the game. So it'd be like you're finding items in a game, can I look 'em up in the database and where the Warcraft is? Kind of what was coming out shortly after that and actually led to my first real job post Burger King era where we were effectively data mining the game files and building a big online database about it. And that company was called Curse, which is I think part of Twitch, which is part of Amazon. I don't know the Cascade.
Dan Levine (03:20):
Oh, I didn't realize all that stuff. But I do remember I was a wow player, world of Warcraft. Wow. Player. And it used to be
David Cramer (03:24):
Curse gaming.
Dan Levine (03:25):
You would look up any item and the SEO was cursed
David Cramer (03:29):
And it was like mods and everything else. It's like an interesting era. We actually, there's some people that I know in San Francisco that I know from that era that were just mod authors and stuff. So it was like a gateway drug almost for a lot of software. It's like self-taught software engineers. The thing I learned doing was basically MIRC scripts, which was their own little language in this chat room. And then I learned PHP quickly after that.
Dan Levine (03:49):
Very cool. Okay. PHP first. That's a standup thing.
David Cramer (03:52):
It was the best thing back in the day.
Dan Levine (03:53):
Were you Tack or Lamps Stack?
David Cramer (03:55):
MM
Dan Levine (03:56):
Was that Mongo Macintosh? So lamp was the original thing. I worked at Burger King. You think I could afford a Mac? Oh, that's fair. Yeah. I got my Macintosh when I went to college and I remember I heard about Lamp and it was like Linux, Apache, my sql, PHP. And I was like, Linux, I really hope there's an easier way to do this. And then I found this map. Somebody had packaged it up for Macintosh. I was like, yes, thank God. Okay, so PHP is your
David Cramer (04:18):
First language. Yeah, and I was on the lamp, but I think I just had a virtual server or something somewhere.
Dan Levine (04:23):
Okay. So you start doing PHP. When do you start dabbling in Python?
David Cramer (04:27):
Shortly after that actually. So I worked for this little tiny site for about a year doing World Warcraft stuff. And then the curse guys who were all 20 years old, 90, none of us are serious business professionals.
Dan Levine (04:36):
And you're in the Midwest and where are they?
David Cramer (04:37):
I'm in Nebraska. They are in Germany. There's basically French people that were living in Germany for some reason. I don't remember why. I love it. But Curse had already existed. So a year into this thing, they knew who I was. I was also just as agro then as I am
Dan Levine (04:48):
Now
David Cramer (04:49):
Competitively that is. And they're like, Hey, why don't you come join us? And so I moved to Germany, worked on Curse for, actually it wasn't that much longer, maybe two years.
Dan Levine (04:58):
And how old are you? Are you like 18?
David Cramer (04:59):
So I think I moved to Germany when I, I moved before I was drinking Age in America. I remember that much. And I think by the time I came back I was drinking age.
Dan Levine (05:08):
Got
David Cramer (05:09):
It. So I was probably just turned 20, moved there, came back when I was 21. It was super fun. And that's also what introduced me to San Francisco because they moved the company here a year later and I came out here with them. It was just as expensive back then. Mind you, we made a lot less money, but
Dan Levine (05:23):
It was not as expensive just for
David Cramer (05:25):
It was
Dan Levine (05:26):
Felt.
David Cramer (05:26):
Yeah, it felt as
Dan Levine (05:28):
Expensive. Yeah, that was crazy.
David Cramer (05:28):
I remember actually it feels cheaper now, to be honest with you. In some sense, even excluding my century success, it feels a lot cheaper. The
Dan Levine (05:33):
MA lattes are expensive, but broadly it doesn't feel
David Cramer (05:37):
That bad. It's more affordable if you work in tech than it used to be. Yeah, this is
Dan Levine (05:41):
Kidding. Okay, so at Curse Python or PHP,
David Cramer (05:44):
That was Python. So that's when I picked up Django and I was really involved in the Django community. What's
Dan Levine (05:48):
Django? Just describe something.
David Cramer (05:49):
Jengo is I think still probably the largest Python web framework that exists really good. It's not so modern anymore, but I was big into that. Met a lot of, that's basically my entire career with Jengo. And actually the reason I left Curse is because they wanted to switch to.net and I'm like, I don't want to do this. And I joined Dropbox because they were using Python. It was basically Dropbox or Instagram are the two choices for me. And I'm like, I'm going to go somewhere where I know the thing, I'm good at it.
Dan Levine (06:14):
Oh, we jumped right? We've gone out of order if I recall correctly. So you're at Curse.
David Cramer (06:17):
Yeah. Yeah. That's fair.
Dan Levine (06:19):
This is an important stop.
David Cramer (06:20):
I worked in a bunch of startups, but it was like curse. I went to a company called Discuss, which back in the day it was the little comment widget on basically website. Website.
Dan Levine (06:26):
It was D-I-S-Q-U-S
David Cramer (06:28):
Discus.
Dan Levine (06:29):
And many people thought discus. I also thought
David Cramer (06:31):
That
Dan Levine (06:32):
It was the best startup we ever, because the first time you heard it called Discuss, you realized that was obviously the correct name. You were like, oh, obviously the comment widget is discuss. So
David Cramer (06:41):
What they would tell you is that's unintentional. And I think it's a lie.
Dan Levine (06:44):
That's got to be a
David Cramer (06:45):
Lie. It's got to be a lie. I'll talk to Daniel. It's got to be lie. Yeah, Daniel swears, at least back in the day, he swore on it. Maybe he was fooling us all. But I for discuss for two and a half, three years, something like that, still Django Python. It was fun. We were a small company. What years are we at now? It was 2008. Got it. Probably 2008 to maybe 2011 ish. Got it.
Dan Levine (07:04):
And this is where you start to work on the open source project that would become
David Cramer (07:09):
Actually Century existed right before that. I don don't remember when 2000,
Dan Levine (07:13):
Well it wasn't Century before then,
David Cramer (07:14):
Right? Yeah, it was called Jengo DB Log. Yeah, exactly. Random open source. Yeah. Yeah. So I guess the history of Century is random open source project I built built many random open source projects, renamed it off of a StarCraft unit at some point. That's what we were naming things.
Dan Levine (07:29):
The gamer thread stays live.
David Cramer (07:30):
Yeah, the gamer thread stays live.
Dan Levine (07:32):
I think a very fun, playful, delightful StarCraft unit. The century's like a very protective unit. It's like delightful but very savage ultimately.
David Cramer (07:42):
Yeah, actually I don't know why because we always tried to name things. We also had a few others. We had NIUs, which was the NIUs one, which was our sort of distributed communication library for Redis that we make. That makes sense. We had something called Hi that I don't remember what it did. We had a few of these things, mostly of the discuss era. We started naming things. So when I joined discuss they were using Jango DB log. It was not good. First couple of weeks I fixed it all at some point. What was the
Dan Levine (08:05):
First feature and what caused you to make Jango DB laws?
David Cramer (08:08):
Somebody asked me how you would do this. Hey, I have errors in my app, like my Jengo app, how would I log them into a database, show 'em on a dashboard. And I'm like, I dunno why you would do that. Let me, here's an example of how I would do it. And I whipped up literally basically the MVP, which probably would achieve product market fit still to this day that it was like a jengo middleware, it captured the airs, put 'em on django's admin kind of panel. And it had a lot of the context that we captured today. And I'm like, oh, this is kind of neat. I just kept fiddling with it over the years and adding a little bit more from a feature point of view, more polish, all this stuff. And it's like one of those things, I'm always the person who likes to hack on stuff. I want to have a project even to this day. And so that was just one of my projects and I'm like, I just want to keep tinkering with this.
Dan Levine (08:46):
And not to date you here, but I have to ask a funny question, which is now we think about where do open source projects live. They live on GitHub is right around the same time as GitHub coming about. It might have predated GitHub.
David Cramer (08:59):
It predated I think because GitHub's first office was also below discusses first
Dan Levine (09:05):
Office. Right, exactly.
David Cramer (09:07):
And I think it predated it, but it was on, I think it started on Google Code.
Dan Levine (09:11):
It was on Google Code, was it IRC? And so people were talking about it and the JA community starts to pick it up and so it becomes open source. But pre GitHub, which became this great place to shoot Open
David Cramer (09:20):
Source. Yeah, yeah. We moved to GitHub I think not actually super quickly, but fairly quickly in the grand scheme of history. They were Ruby people, so it was like That's true. They also loved Dubstep, which was popular at the time. Let me tell you, we're one floor below them in this old building, small building, there's like 10 of us on each floor. Five every day. You just hear it. It's like that.
Dan Levine (09:41):
Wow. Big dubstep Bakers top short. Okay, so now Jango DB log, when does it become named Century?
David Cramer (09:47):
I think it was like a year in after joining discuss. I don't quite remember, but this is when StarCraft two came out, which is a super fun era looking out like San Francisco. This is Peak San Francisco in my opinion. It's like all the companies are small. None of us make a lot of money, but the corporations have some money. So there's meetups, free pizza, free beer. There was Justin tv.
Dan Levine (10:06):
Yeah. Well because Justin being Twitch are just starting. So they start what is gaming after Dark? Another thing you mentioned that I just wanted to dig on is, so Chris is a designer and the founder of Century. I think one of the things Century has always been great at is this incredible design and creative sensibility. Obviously a lot of that is you from an idea perspective, but Chris, a lot of the execution side. Why did you think this side project, open source project within discuss where you self acclaimed, only wanted to work on it when it was part of your job, needed a designer to work on it. Because I think that's way ahead of its time. Even now, I try and encourage open source projects that they should care about design and people don't believe me. And this is over 15 years ago.
David Cramer (10:41):
I mean, I care about design. I'm not good at it by me any means, but it's something that's important to me. Visuals, aesthetic, all these things. A lot of us do care about this. We just don't have the skills to make it good. And so whenever I have a project, the first thing I try to do is socially pressure a designer to help me on it for free. Ideally I can't afford to pay 'em even my side projects. Now how do I get somebody? And to be fair, it's a little easier these days and to do it, not pressure somebody to help me, but it's just something I cared about. And to be fair, it wasn't like cutting edge design. It wasn't time consuming design. It was like, oh, here's some colors that are reasonable. Here's some round corners. Also harder back then though, to be fair, to do any of that super hard.
Dan Levine (11:18):
Yeah. Okay, so now C is underway. You discussed, you end up going to Dropbox. We can kind of fast forward through some of this and then you decide at some point I want to start a company around this open source project. And there's a few things in there that you should include about how that's
David Cramer (11:34):
Accidentally
Dan Levine (11:35):
Happened.
David Cramer (11:35):
So we bootstrapped the company and that was this trigger of, we had a bunch of traction in the Django community. Basically if you use Django, you use sent back in the day. And we had a lot of traction and somebody kind of convinced me, they're like, Hey, there's a guy that worked at Heroku who was a PM and he's like, Hey, you should launch an add on out of this. They're trying to money off of add-ons. And I'm like, oh, that'd be cool. People can run it on Heroku. I can make a little bit of money. That seems great. That's not actually how it works by the way. You had to build a SaaS service and then sell it via Heroku. It turns out I got tricked a little bit, but I'm like, okay, maybe it'll make, we joked at the time, beer money as in there's not a lot of ambition. We're just like, oh, it'd be great. It needs to pay for itself. It doesn't have to pay for us, but we don't want to be spending money on a server bill every month. And so that was actually a requirement from day one. So it was not free. You had to pay for the cloud service. We had customers, it was actually profitable that first day, but that was what triggered the financial side to start the business out of it. And then we did that for, I don't know, what was it, like two or three years.
Dan Levine (12:34):
And you clear you're working jobs while doing that. I'm
David Cramer (12:36):
Working jobs, so I'm spending even more time nights and weekends and stuff. Again, I'm like mid twenties or something. And I think I started that at the tail end of discuss, continued in Dropbox. But eventually I get to Dropbox and it's a pretty, they're going through their hire everybody in their mother phase kind of, which is just, that's all I got in the door. Yeah, same. But it's kind of a miserable phase to live through of a company because just like new faces every day, you're spending so much time hiring people. I don't know, I'm not a good big company person either. I just want to build stuff. I'm a great IC, I'm high throughput, but if people get in my way, I'm very low tolerance of it.
(13:12):
And so it was kind of this frustrating time period for me where I had sent that was super high demanding. I'm on call, it's an infrastructure project. It's tedious to manage. It's also bootstrapped. So I'm running thin margins on the whole thing. And at the same time I'm building Dropbox is, it was like a thousand server distributed test infrastructure, which I was almost exclusively managing and building. And so it was a lot of two high stress, I guess jobs per se. And I think it was a year and a half. I'm like, okay, I got to do something about this. One of these has to go. And I'm like, what do I do? Blah, blah, blah. And VCs as I do always harass you even if they're not interested. So hey, you should consider raising funds.
Dan Levine (13:46):
I love harassing
David Cramer (13:47):
People. Absolutely. But I'm like, yeah, Dropbox is fine. Different error you, this is still when we got paid, but probably 50% of what salaries are now. And so it was still, you're like, okay, well it wasn't like I was giving up a half a million dollar a year cash comp or something like that. And so I'm like, well, we got to a point where we had this bootstrap business, we could kind of pay ourselves. We couldn't quite match our salaries from drop. Chris was at GitHub at the time, but we couldn't quite match our salaries. But we got really close. I ended up dropping my salary like 120 k. I think I paid Chris a little bit more. He had a family, but that was pretty close. I was making around 150 at Dropbox and I am like equity, whatever. Who cares? I've been this startup game, it doesn't make any money. So I didn't really care about that, which in hindsight was fine. And so I'm like, okay, let's do it. And so we made that decision, kind of kicked it off. And then from there we're like, okay, what do we do? Do we go raise money? Do we keep riding thin margins? Which is, it's stressful bootstrapping with thin margins, which is, I dunno why we chose this, but it is very stressful.
Dan Levine (14:47):
Well one of the things that we always talked about in the beginning was, and you kind of to this a little bit, so Century is this open source project that starts in the Django community. People use it in Django. I think from the moment we started talking about the company, you already had the sense that it made sense to also do at a minimum broad Python and JavaScript. You were always very adamant about that at a time when it was not so obvious and now all of a sudden you're supporting multiple language environments and frameworks and you're like, there's a lot to be done here that makes sense. But we wouldn't be able to do very quickly if we just bootstrapped on
David Cramer (15:18):
These. Yeah, I think we were looking at this and I think we had hired, I think we could afford to hire three of our friends basically. So we hired Ben Armand and this guy named Matt.
Dan Levine (15:29):
You have much more impressive friends than
David Cramer (15:31):
I did at the time. That's what happens when you're not just out of college or whatever.
Dan Levine (15:34):
Ben wrote a book on third party JavaScript and Armon is the creator of the Python framework. Flak, just so that everyone knows these are useful
David Cramer (15:41):
Friends. It was kind of one of those things where, well, we can hire a few people, but then we're out of money and we again, we're scary close to margins and we had a bunch of competitors that were not bootstrapped. They had raised plenty of
Dan Levine (15:50):
Money
David Cramer (15:51):
And I think, I don't know that I was fearful, but I was competitive. I'm like, I do not want to be beaten at any cost. And so I'm like, let's go raise money. Let's be more ambitious. I think we did it honestly, not knowing what we could do. Like well, we got this little open source thing. Honestly, most people didn't take it seriously at the time, Dan, you were the first person that took it seriously. I think everybody else is like, this isn't a big business, there's nothing here. Or they're like open source. That doesn't make any sense. How are you going to make money and all this stuff? And mind you, we had 2000 customers, like 600 k in revenue when we did this. And so anyways, we ended up raising money and that kind of allowed us to one, solve some more of the infrastructure problems.
(16:27):
It allowed us to expand sort of the language support and we were really big. Our entire thing was do what you want. That was kind of the premise of the company. And one of the things we want to do was, for me particularly is I wanted every single person I ever met to be using Century. I wanted to have every conversation. If they're peer mind be like, yeah, we use Century. That's basically true today, mind you, which is phenomenal. But that is a really, really different objective than most people starting businesses. It's not monetary driven. It's like, no, we just want to build a great tool that helps everybody. You got to make money doing it, of course. But
Dan Levine (16:59):
When you started, my sense from what you've told, but you should correct me if I'm wrong, is first of all, we don't think about this as interesting now, but at the time you were using a bunch of open source technology yourself, like you learned on PHP, you were presumably using Linux, Apache, maybe you started using Nginx or something at some point, and your databases I suspect were open source for all these things. You pick up Python and Django, which are all open source and naturally you build these little tools. I'm curious, did you think at all at the time that Jango DB L was open source that that was a consequential decision at the very, very, very beginning? Or is it just like, why not? This is whatever was.
David Cramer (17:34):
Yeah, everything was why not? And I still kind of believe that I have a lot more thoughtful informed opinions about how to do intentionality is what I have these days. But back then I'm like, I dunno, this is just what you do. You just build software, who cares if it's open source, why wouldn't you be? And it was permissive. We were BSD license, which is crazy permissive and I think it was always just one of these things, I had the luxury of working on it. Nobody cared. We were all like kids basically. And so nobody had big company enterprise opinions about the scariness of open source. And also it's one of those things, it's interesting because once you go down the rabbit hole, take century, even when we started commercializing, we had some closed source on the cloud service, but the product was open source and the rule was you had used this
Dan Levine (18:12):
Look on billing software and stuff.
David Cramer (18:13):
It's like billing. There's not some conspiracy here. But the rule was you had everything had to be open source in the sense of we cannot not rely on anything proprietary. So even if there was a good solution in the industry to solve, one of our infrastructure problems was not allowed to be used. We had to make it adapt because otherwise you couldn't distribute. It's like all of a sudden nobody could use the software. And again, our goal ignore open source, it wasn't make more open source. It was like we want everybody to use the software. And even back in the day, I don't know which version of the government shenanigans were going on at that point in time, but it's like there's certainly sanctions against somebody or companies in poverty that can't afford anything. And so you had companies all over the world using Century. Then Russia, Yex uses Century, actually I don't care. Russia can't get me and we're fine. So I don't think that's confidential, but they've used it for ages. I've never met any of them. I just know they use it
Dan Levine (18:58):
And they use the open source century,
David Cramer (19:00):
They use the open source century and now we're legally not allowed to sell them or anything. And back in the day, even if we wanted to, it would've been too painful to do. And that's cool to me that politics aside everything aside, it's cool to me that tech can be neutral in a lot of ways that we can just be like, yeah, you can use our tech, that's great. It's great that you get value out of that. We help other tech companies build better tech ideally, and it progresses humanity for it. And to me that's kind of like the reverse engineered almost. I feel good about what I do think I'm building technology that helps us progress as a society, but so that open source thing and everything and also was always like I just wanted more people to use my
Dan Levine (19:35):
Stuff.
David Cramer (19:35):
It was almost like attention seeking. I don't know how you describe it. And open source is a great way to increase distribution and so it's just a natural thing. Even a Dropbox they let me work on, I basically open sourced everything I did at Dropbox and I don't think anybody really cared. They're probably just like, why?
Dan Levine (19:50):
Yeah, it seems like there might be some headaches, you
David Cramer (19:51):
Could get some more tickets, but I don't know. It's rewarding. You can share your work, talk about it. I dunno.
Dan Levine (19:58):
So initially you do it, the company throughout its history, you're exactly right. We always were like, everything should be open source because people should be able to run Century on their own. When did you start maybe getting more intentional about what open source meant to Century, do you think?
David Cramer (20:15):
That's a good question. I think at some point we had talked like, okay, we should probably put it on a real license like Apache two. And we're like, yeah, we don't really need to though. And I think we only wanted that because you don't know what you don't know and well, there's some protections in around patents and things like that and those might matter. We just didn't know. We have no patents, we don't even care. So there was that and we're like, yeah, it's not really worth changing the license. And then this actually wasn't even that long ago, 2018, maybe somewhere in that era we became much more intentional and it was just triggered by external forces, not venture capitalists. Like everybody conspiracy theorists think
Dan Levine (20:48):
This is important because this is actually really, this is one of my two most famous hacker news comments ever that I'm very proud of, which is you guys made some nice change, which we're about to talk about I think. And somebody was like, it's because of the venture capitalists. And I with my hacker news account was like, this is my moment. And I replied, I am the venture capitalist. They didn't even ask me about the change and I probably would've lightly disagreed, but I'm fine with it because they run the company. That's how it should be to be fair, which I think
David Cramer (21:14):
Is great. I think that's great. I actually had a conversation with another founder about this. They're like, oh, we're worried our investors will not like if we go open source. I'm like, who cares? Yeah, exactly. What do they know? In all fairness, they the money. Money, yeah. So we had some outside pressure from other companies trying to distribute or basically commercialize their software. They're not part of, we almost exclusively build Sentry and have for the longest period of time and we go out of our way to pay literally anybody we can that contributes to sent, contributes to a core project we use or something that was the same then same. Now
Dan Levine (21:46):
I'm going to spend a little time on this because I want to talk about this at some point, but this seems like a natural time, but then we'll come back to this, which is Century's a little bit unique. There's lots of different types of open source projects in terms of the distribution of how people engage with the project, who's committing and how that works.
David Cramer (21:58):
The contribution model.
Dan Levine (21:58):
Yeah, contribution model. So if you think about Linux, right, Linus builds it, but there's a lot of companies that contribute to Linux now. There's Core Committers that work at a bunch of different places that have a vested Linux
David Cramer (22:08):
Plus a huge ecosystem of stuff that runs on it
Dan Levine (22:09):
And a huge ecosystem that runs on six. So it's quite distributed,
(22:12):
The nature of Linux, but that's not quite the case for Century. I mean as long as I can remember, first of all, the people at Century have committed the vast majority of the code, but you've also had people who are on the outside contributing things like SDKs or small patches. And you have always, as long as I've known, been overly positively inclined towards those people. It's like how can we compensate them? Can we have a consulting agreement? There was send them free shirts if they want to do meetups and people, there was a zeal for Century, but the percentage of commits or the amount of code was primarily done by people who worked at
David Cramer (22:43):
Century.
Dan Levine (22:44):
It wasn't open source in the sense of tens of thousands people. Yeah, it was not
David Cramer (22:48):
Community built is how I describe it. There's some people that helped way back in the day, contributed a little bit here and there. We had some customers contribute here and there, but it was mostly written by me in the early days. And if you've ever seen me on a keyboard, I'm insanely productive. And so it was mostly that SDKs a lot of third parties, contractors, whatever, some open source people who just wanted to contribute back and that was great. Another
Dan Levine (23:08):
Thing you mentioned that I want to come back to set this up is, and you've never been somebody, and I don't think either of us have been somebody, we've never cared that people use open source century and don't thrilled by that for the most part. So you mentioned the index, right? You think it's quite cool. I remember early on, but Uber was a huge user of the open source and they would have a lot of requests and thought processes around that, which is great, but never once did you or I think, oh, we have to get Uber to pay us money.
David Cramer (23:33):
I did try to get them to pay us.
Dan Levine (23:36):
At one point they did some support contractors.
David Cramer (23:37):
We actually charged him $5,000 an hour to have an engineer show up and he wasn't allowed to touch a computer.
Dan Levine (23:41):
They wanted like an insurance policy. It was very confusing. We never said,
David Cramer (23:45):
But yeah, we couldn't get them to cloud was what was important. And we're like, it's okay if you're not going to go to cloud, we're not going to force it. And there was a point in the early history where we're like, that's now the absolute truth you want. Great. Eventually you're probably going to come to our cloud service if the product remains great because it's a pain in the ass. It's infrastructure. You don't actually want to run it. And it is
Dan Levine (24:03):
Monitoring infrastructure. The moment it's the most valuable is when you least want to be worrying about the infrastructure of your monitoring tool, which is what you remembered hearing about and discussed, which is part of your experience. That's part. And then the other thing is, and I think we probably both agree this not in a nefarious way, but the fact that Uber uses Century, even though it's open source is much better for Century over long periods of time. The open source product and the company then if they use the competitive product and all these people who work at Uber who then think, I'm going to go start a new company or go work somebody else, I probably want to use Century. That's the funnel. And I don't really want to run this myself. I might as well use a very reasonably priced SAS tool. So never, there was never any moment where we said, this really bothers us or we can't believe these people are using Century for free. That was never
David Cramer (24:41):
Any issue. We actually, because people are going to be people early in the history of company, we've had to lay down the law that you are not allowed to sell open source accounts. If they're using self hosted, that's their choice. You don't get a harassment. We had to lay that down and go to markets like that was off the table completely. But yeah, you pointed out a really critical thing, and this is what I tell people about open source. I have this article, I wrote my blog, ccra mr, but open source is not a business model, it's distribution, all this other stuff. But that Uber story is critical for people to understand because Uber, yeah, they paid us a little bit, but not for cloud services. We couldn't convert 'em. We gave up and they opted, they wanted our help, they just wouldn't go to cloud.
(25:16):
And we're like, well, we're not going to sell you anything. Everybody at Uber went to work somewhere else. All those companies bought Sentry, bought, sent Cloud services. That's beautiful. Who cares about Uber? It's great that they used it. We won. Or rather the business got the revenue it needed. We could justify that. And that's one important thing. The other important thing, and I have vivid memories of saying this because I was so frustrated with investors when we were raising the seed. I'm like, people would be like, how are you going to make money? Blah, blah, blah. I'm like, I don't care if we don't make money, nobody else
Dan Levine (25:43):
Will.
David Cramer (25:43):
That's it. That's it's a zero sum game. If we don't succeed, nobody's going to succeed. We will burn the whole thing to the ground. We will make it so good and so cheap that nobody will have the opportunity to exist. And there's some intentionality there. To be fair, I could easily defend why open source was such a compelling strategy, but less so than when we eventually changed licenses and whatnot. Of course.
Dan Levine (26:03):
So now we come back to this moment where this external force comes in. Who, and I'll say this part, but you won't have to, their kind of view of this is open source is kind of for Patsy almost and almost what's the most parasitic thing you could do? You could take a lot of more than open source. You could commercialize it better while having contributed
David Cramer (26:20):
Nothing. Do you want me to say the companies? No, I don't. There's a company that likes to sell on-premise software. They sell a bunch of open source basically some is their own, some is other people's stuff. Their entire model is based on the belief that people are going to run things behind the firewall. Our entire model is cloud is the future. We know that there's a blend obviously, but cloud services are every company's
Dan Levine (26:40):
Solution. And if you want to run behind the firewall, you can use open source,
David Cramer (26:43):
Which is wonderful. And so they were trying to distribute Century and they wanted to integrate it more closely, but they didn't give back to the project. They didn't pay us anything. They didn't do anything valuable to anybody else other than they were trying to basically add commercial value to the thing they sold. And I don't care what the license says to me that's not okay.
(26:59):
And this is the fundamental argument in the open source ecosystem. It's like it doesn't matter if the license says you can do something, that's fine. We made the choice of what the license says, but that also doesn't mean we have to be okay with that kind of behavior in the industry. And again, century single source, we fund everything. You helped fund some of these things. And so we changed the license and the license is basically a big middle finger to those kind of companies. It said all the users, Yandex included, can all run it for free, no strings attached. Same software. We sell. Anybody that wants to commercialize it in any shape or form that looks like commercialization from a cloud service or subscription services point of view, good luck. We have lawyers too. That was the day we went from open source to technically proprietary or source available, which we now call Fair Source. It's kind our agenda behind this whole thing. And that was with the
Dan Levine (27:42):
Which by the thing, importantly, things were eventually open source, but at the moment released they're not, which is a
David Cramer (27:47):
Critical good second, lemme talk about it. So we switched to the business source license, which TLDR after four. It's variable, which is the complicated part about it. But after four years at most I think we chose three years. The code becomes whatever other license you want. So we chose Apache two and you have sort of a restricted use clause. That was a little complicated. So last year, somewhere in the last two years, we built our own license called the functional source license, which was originally going to be the fully functional source license. So that like FS, you can figure out what that acronym means. Mostly we were tired of all these arguments, which got rid of all variability. It hard coded everything. It said after two years this has to convert to MIT or Apache. It has to be a permissive license. Most importantly, and the only restriction is the cloud services restriction. And our intent was like make it easy for legal teams, protect the people that want to run it and use it, make it so eventually it goes open source. It's almost like the feelgood safety net in case the company fails, but also because then people like there's trust and that was our solution and we're still working on that. And then we wrapped all that up more recently in this sort of initiative branding called Fair Source, which the idea with Fair Source is it encompasses a bunch of these values
(28:57):
And the overlap between open source and fair source. And I'm going to ignore free software, which is a completely different thing. But open source and Fair Source is fundamentally the freedom that we take away with Fair Source is the commercialization freedom. We don't believe everybody should be able to commercialize anybody's software no matter what. There's arguments on both sides of that to be fair. So it's like users can still use it. It's the real software. It's not open core, it's not part of the software. It protects the business and it's actually mandatory to convert. I think we actually made it mandatory to convert to a permissive license after a period of time, which is a big deal because permissive licenses have a lot more technical freedom, a lot fewer restrictions on use.
Dan Levine (29:34):
So all this happens. But one of the most important threads in all this is at its core, at its roots century was an open source company. It important to you, you were using open source technologies, we're huge fans of open source century, even if from a technicality perspective, which we would acknowledge that's not technically the license we have today.
David Cramer (29:50):
Yes. However, so much of our stuff is actually
Dan Levine (29:52):
Built. That's right.
David Cramer (29:53):
That's important. Open source, permissive. We actually power every competitor in the industry with our open source tech. We distribute some qualification and stuff like this.
Dan Levine (29:59):
That's exactly
David Cramer (30:00):
Right. It's all over the map what we do these days. But it's almost all some variant of open source or some of the values of open source.
Dan Levine (30:07):
That's right. That's right. But now you're doing a new thing. Tell us about the open source pledge.
David Cramer (30:11):
Oh yeah. So the most recent thing is the pledge. So from a business point of view, I actually think it's important for everybody to understand. There is ROI from a business, which is how we can justify these things. It's just marketing. And so I want people to understand that we actually have a business agenda with it, but it also is a nice ethical thing.
Dan Levine (30:26):
Well first, what is the open source
David Cramer (30:27):
Plan? It's again, same with Fair Source. We have this agenda to better the open source ecosystem, especially because we think there's a lot of these decisions we made that pair really closely together. Our approach to market share works really well because of the distribution of open source or open source licenses. That also allows us to give back, whether it was through contractors or US funding random projects because we often do these with no strings attached to now us in the last few years, just straight up donating to our dependency tree in the past, we donate to Jengo Software Foundation, all this stuff. Now, we literally directly donate to people that write libraries and some of those people can start businesses that are really successful, but then there's a lot of 'em that can't. And there's obviously a sustainability conversation in the industry of NPM is the famous target where there's infinite dependencies, nobody knows what they do.
(31:13):
There's a lot of security concerns, they're not maintained, et cetera, et cetera, et cetera. We were inspired actually when we hired me, our CEO many years ago. At some point he came from VMware, bigger company stuff. He turned me on to the pledge 1%. We had talked about maybe there's a way we can do that with sort of all ideal of mind, the open source thing. We're like, okay, what if we could do the 1% pledge for open source? And that kind of kicked us off early on. We were like, what if we pledge 1% of revenue and something like that? Then could we get others? And they're like, oh, but then they got to share how much revenue they make and they're probably not going to do that. We played around with a lot of ideas, but our goal is like, can we give more money to maintainers? And this is the business ROIs, it's marketing. You actually get brand recognition from it. And brand awareness is the greatest thing you can have in a business like ours. We like to talk about liquid death in companies like this. They're the greatest.
Dan Levine (31:59):
Talking about liquid death,
David Cramer (32:00):
One of the greatest brand marketing companies of the last decade. They're just selling basically water in a can. There's nothing more to it than anybody else, but they're great at
(32:09):
And they just do stupid marketing. So brand marketing I very much believe in and us doing a good thing with open source is just brand marketing from the business ROI. It's actually not that much money in the grand scheme of marketing budgets. And we're like, what if we could take that, turn it into a program, social pressure, I'm not going to use the word bully yet, but social pressure other companies into also doing this because it's good and it kind of needs to happen and they actually get a lot of benefits out of it. They actually potentially de-risk a lot of things and they get the marketing benefit. And so open source pledge, which I think formally launches sometime early October, is our attempt to productionize a version of the program we were running internally that we thought was pretty successful, which the TLDR R is like for every full-time engineer on staff, you have to give at least $2,000 a year to your dependency tree
Dan Levine (32:54):
Is we, which just to be clear for engineers that are based in the Bay Area is a relatively small
David Cramer (32:57):
Percentage of their compensation. It's
Dan Levine (32:59):
Nothing here. It changes based on where the engineers are based. So in different places it's complicated, but in the US it's a relatively small percentage of what you spend. I remember, I still remember when I came out to the Barry in 2009, there was a period of time maybe in 11 or 12, you remember this where I remembered around Asana and Quora when they were recruiting people, they recruiting websites would talk about the benefits you would get joining the company.
David Cramer (33:18):
And
Dan Levine (33:18):
I think it was like $10,000 for the setup of your choice. You'd get a fancy chair and two fancy monitors and a standing desk, 10,000 for however you wanted. And this to be clear, that would cover five years for one developer. And we've had a lot of inflation and developer salary
David Cramer (33:33):
Six. And the great thing is, and there's a lot of challenges with it. If your company's in another country, that's a lot of money, but it actually doesn't matter. Our goal is to not say, can we get every company to do it? Our goal is to get every US technology company of any significant sale like series B is a great target area to join the same because if we could even get a hundred million dollars, which in the grand scheme of all these companies, that's actually not that many companies. That would change a lot of people's lives and that could make more full-time people that just work on useful open source. What I did in my career was always just build useful stuff. I built a lot of open source libraries, no strings attached. Some for the businesses I was working at, some not. I didn't get paid anything for that.
(34:11):
And what that means, I'm not saying I should have, but what that meant is there was not really a recognition or incentive for me to keep doing it, let alone to maintain all those, which I did not maintain all of those. And that's okay, but we're like we could change it. And so that's kind of like the agenda behind, and we actually talked at one point, the original inspiration for this of the $2,000 thing was like what if we could do an employee perk where we let every employee do that? Well one, they're not going to do it, which is a problem. Two, there's no vendor that makes that reasonably easy for us to do. As a perk. You have that for charitable matching and stuff, which is great, but I'm like, come on GitHub, come on, please help us out. And then OpenAI came and now nobody cares.
(34:48):
And so yeah, we're just like, it'd be cool if we could get a bunch of people on. We've done okay so far. We've got a bunch of startups. We're still working on it, but I'm like, if this is the last thing we kind of do at Century, that's like a really big hurrah. That's also cool. We're finally a stage of company where we actually have a lot of influence and power. People will understand we exist, we're big and we can do some cool things. Now we can actually be very, very helpful to people beyond just giving them our software for cheap or free or whatever.
Dan Levine (35:12):
Well, I think just obviously I'm a biased fan of entry, but to give you guys credit for this, a lot of people would point out the challenge of sustainable open source and we need to do something to change the funding model for open source. A lot of people have pointed this out and there's some wonderful startups that have done things around that I think enable Century amongst others to give directly to maintainers in certain ways, although relatively small volumes today, but they're doing their best. But very few people actually put their money where their mouth is.
David Cramer (35:36):
That's actually
Dan Levine (35:37):
Do it. By the way, there's some companies like big companies like the Google's and Facebooks of the world with their own open source efforts, whether it's Kubernetes or Google on Linux or Facebook React, they are PyTorch or llama. They put in a lot of work. But for a lot of companies it's non trivial. You don't have an open source product of your own to be investing in, but you fully rely on open source. Almost everyone I know is running Linux servers for any company, your programming languages, which we take for granted that programming languages are open source, but they have not always been that way. It's still in the finance industry. You can find them. They're not databases. And so where C'S been fortunate enough to build a business model that can coexist nicely. It's hard or for other companies, but one way they can do it is something simple like this where for every engineer at least $2,000 a year, you do open source. It's still a fraction of the cost of that engineer for the year.
David Cramer (36:25):
It's literally a rounding
Dan Levine (36:26):
For most companies. And it's the investment you would want to make in tools you're using anyways. You want them to be more sustainable. And I always think to myself, when I think about the pledge, I hope one of the things that happens is people say, this is a really important piece of software that we use. There's not even an obvious person to give the money to. That's alarming. And I think that's a non-trivial circumstance at open source sometimes where you're like, you rely on some package but the maintainer does not even working on it anymore or is incentivized to focus on fixing bugs. And you finally think for the first time there was no to give the money to for a thing we deemed was so
David Cramer (36:55):
Important. Yeah, I think there's a lot, and actually you touched on something that I think is super relevant here because one challenge we've had with this whole thing is a lot of people do give in great ways to opensource already, like Facebook or Meta as an example, but it hasn't solved the problem. And so this is actually super important. We keep going back and forth internally on this. So Meta has done phenomenal stuff. You can't say Llama by the way, it's illegal to call it open source. Yeah, it's not good. But they've done phenomenal stuff for open source
Dan Levine (37:19):
React, the Alpaca cousin thing.
David Cramer (37:21):
But
Dan Levine (37:23):
I learned all the cousins of the llama. Oh no. Yeah, it's fun. I think it's a aka. Anyways,
David Cramer (37:28):
Anyways, react is one of the most useful technologies we've had from a broad open source in recent history. Those problems are not problems. Great funding, great maintainability, Linux, Linux Foundation, ludicrous amounts of money that they rake in. Those are not problems we're trying to solve. The problem is, and so those are great, is we want to recognize companies that do that. Those don't qualify for the pledge. And that's really important because what we're trying to solve is not those problems already actually in a good spot. And again, we don't want to take away from that, but the pledge is very much like you have a lot of these dependencies from this random person that maintains a bunch of stuff because they like it. They're passionate about it. You could reward them, you could tip them. You'd be like, buy 'em coffee kind of thing. That's how I think about a person. I'm like, this is a tip jar. Make people feel like some recognition for the amount of personal time they put on all these random things that power our silly websites and stuff. So much of it that's unfunded and is never going to be remotely big enough. That could be funded
Dan Levine (38:17):
By
David Cramer (38:17):
Any big company or a vc. And then a lot of it's, yeah, it's trivial to rewrite, but we're not rewriting. We're still relying on it. And so that's our main goal is really to focus on the people that have no other support mechanism in this whole thing.
Dan Levine (38:29):
You said tip Robert. I just want to point out, this is an idea you can steal from me for this project. You should create an animation that's the square register turning around with the tip screen. That's that moment where you're like, wait,
David Cramer (38:43):
When do we
Dan Levine (38:43):
Start?
David Cramer (38:44):
We should do, we should. When you get Stripe on board and flips around everything you buy these days, how do I not tip because I'm not supposed to tip here, but somehow you are guilt tripped into tipping these things.
Dan Levine (38:55):
See
David Cramer (38:56):
It's just like, oh, in the checkout there's like a how much do you want to tip to open source? We actually, there's been people talking about an open source
Dan Levine (39:01):
Patch. Carbon emissions does a version of this,
David Cramer (39:03):
Right? Yeah, that's right. And Stripe enable.
Dan Levine (39:04):
I think we could talk to Patrick.
David Cramer (39:06):
I'll call. Yeah, that would be a clever way to inject this. You, it's like a checkbox. You enable it.
Dan Levine (39:10):
Just get to open source. Okay, so I have one more question before I let you go, but while I have you, so we're doing this whole open source thing. You're one of the foremost open source creators and company founders of the generation, which is kind of an incredible thing, but I hope you appreciate that and take stock of it. But I have to ask because I'm sure people want to know, what do you think the next 10 years of open source looks like? And I know it's a hard question and I didn't prep you on it, so you're off the cuff, but what are some big thoughts you have about open source for the next decade?
David Cramer (39:36):
I dunno. I think it's complicated. Everything always has to adapt.
(39:40):
Failure is always rooted in stubbornness in my opinion. You need to be stubborn, you need to have conviction in your decisions, but there's always a lot of noise in the industry that you got to see pass. And the reason we're doing what we do is because we think there are serious problems that are not going to magically solve themselves. Licensing is one of the particular modern challenges that everybody's talking about. I think it's going to continue to evolve. I think in the best case, a lot of open source stays roughly what it's always been. There's more and more of it, but there's a lot more things controlled by big companies these days. It's a very different era than when we grew up in the industry. The ML thing's weird and wild. I don't know what's going to happen there. I don't even know if it ends up being related to open source because I'm like, it's cool you got some open models or whatever.
(40:23):
They're going to end up calling. They're still working on the terms here. I'm like, what's that do for me? I don't have a million GPUs, so I actually don't know. I think a lot of what we have will keep looking the same. I think there are some very serious problems with things like the NPM ecosystem that I don't know if they're solvable problems. I think the best way to solve problems, this is my career lesson as an engineer, is always change the constraints. Change the problem you have to solve. And if you told me the problem with NPM is there's 99% of packages that are unmaintained and a risk of security, I'm like, well then you should have 99% less packages or something like that,
Dan Levine (40:53):
Right? Yeah. There should be. There's a time-based detonation on a package. Nobody touches this for a long
David Cramer (40:59):
Time. Just so the argument in other programming is like standard library is more of a thing. And so that to me is probably the most serious concern of these days. Pros and cons of both models. But it's like fundamentally, if you ignore the ML world, which has a lot of open source stuff going on in it, and I don't know anything about it, but I'm basically an expert on RAG now.
(41:19):
But I do think this sort of the ecosystem of third party software we rely on has exploded. I don't know. I would love to see, it's probably hard to measure, but if you saw a chart even of our own of, it used to be a little bit of third party software and then Postgres and some stuff now, and it's like we have 10,000 different vendors and that's probably true. I bet we have 10,000 different vendors at Century. If you treat every single individual author of Open source as a vendor of code, we rely on, and that's insane to think about. Yeah, it's a lot. And you can't fix that problem.
Dan Levine (41:51):
What at a high level, it's a good thing that that's possible, but at some level you need to figure out how to grapple with the new world you've brought. Right.
David Cramer (41:58):
Here's a prediction. I bet somebody will find a way to monetize if JavaScript doesn't change one of the most important ecosystems. If JavaScript doesn't change, I think somebody will find a way to monetize almost like a curated third party standard library or something like that to where you've got all the utilities baked in for the most part of all these common concerns. You don't have to plug and play a bunch of different vendors. Right? Because that is one of the biggest issues of that ecosystem and it's a real challenge these days. And so, sure.
Dan Levine (42:26):
David, thank you so much for joining us. I really appreciate it and I'm sure I'll see you around soon.
David Cramer (42:31):
Yeah, this was fun.
Meet your host
Daniel Levine
Developer Tools
Growth Stage