a16z Podcast - a16z Podcast: What Software Developers (and Therefore Every Company) Need

Episode Date: January 6, 2016

The old constraint when it came to technology was hardware -- how many CPUs can I get my hands on. Today, spinning up compute can be done from any smartphone with an AWS account or something similar. The current constraint is software. And since software is written and operated by people, tackling that constraint comes down to making people as informed, enabled, and efficient as possible. Three CEOs and co-founders of three companies that serve software developers -- Chris Wanstrath from GitHub, Jeff Lawson from Twilio, and Ben Uretsky of Digital Ocean -- take part in a conversation with a16z’s Peter Levine about the needs of software developers. What are the emerging platforms, ecosystems, and tools that help developers succeed at what is increasingly the most important job in any company – writing and running software. The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation. This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments and certain publicly traded cryptocurrencies/ digital assets for which the issuer has not provided permission for a16z to disclose publicly) is available at https://a16z.com/investments/. Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

The content here is for informational purposes only, should not be taken as legal business tax or investment advice or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund. For more details, please see A16Z.com slash disclosures. Welcome to the A16Z podcast. I'm Michael Copeland. The old constraint when it came to technology was hardware. How many CPUs can I get my hands on? Today, spinning up compute can be done from any smartphone with an AWS account or something similar. The current constraint is software. And since software is written and operated by people, tackling that constraint comes down to making people as informed, enabled, and efficient as

possible. Three CEOs and co-founders of three companies that serve software developers, Chris Wanstrith from GitHub, Jeff Lawson from Twilio, and Ben Yuretsky of DigitalOcean, take part in a conversation with A16Z's Peter Levine about the needs of software developers. What are the emerging platforms, ecosystems, and tools that help developers succeed at what is increasingly the most important job in any company, writing and running software?

Peter Levine begins the conversation. We wanted to hold this panel on the sort of future of software development. And one of the interesting sort of rationales behind the panel was I often look back and software development has software development prior to, let's say, five years ago, there's always been software development as long as there's been computers. But it feels to me like over the past five years, software development has become a market in and of itself. Like there are many companies that get built to serve software developers. And if you look at a number of early stage and mid-stage companies, they often start out selling to the developer.

And that's a very unique phenomenon in sort of the industry today as different from several years ago. And what we wanted to do was bring up companies, GitHub, Twilio, and DigitalOcean, and companies that have that particular focus. So maybe I'll start out with Chris on that end. Chris, what do you think software development has been around forever? Why is this, you know, what's so new and exciting about now? I think software. I mean, and Drieston a couple years ago had the software is eating the world piece.

And I don't think that even begins to touch the transformation that we're seeing. I think you could make the case that the 1900s were all about hard. And so when you wanted more processors, when you wanted more bandwidth, you just built bigger factories and more semi-lines. Now we're getting to a place where you can, on your phone, go get CPUs and bandwidth, and the real constraint is software. And really, you can't order more software. What you have to do is get more people to create software because it's all powered by humans. So I think that's what's really interesting about our businesses and a lot of the transformation right now is we're going from a world where we're trying to create more physical goods to widen the amount of processing we can do, to we need to figure out a way to get more people. people involved and make the people that are currently involved more productive because ultimately

the world is thirsty and hungry for software and the only way you can get more of it is by creating more developers. Yeah. So Ben, you guys at DigitalOcean focus on the developer. What's unique about building a company? You guys have deliberately focused on, you know, sort of serving the developer. What's unique about what you do versus others and how do you sort of, how do you service them? Well, just to kind of piggyback on Chris's point, the social aspect is very important to note here. So building communities of software developers turns out is a very good way to engage a much broader audience. And one of the things that we've done is not only focus on software developers, but build a community at the heart of DigitalOcean. So to date, we've published 1,500 tutorials that talk about how to leverage open source,

software, we're educating developers worldwide on how to write code, how to deploy applications, and we're drawing in 4 million unique visitors a month to this community, and it's growing month over month. And I think it just speaks to the excitement that people have over the ability to create from the ground up. You can literally take a vision and deliver it. That's the power of software, and that's why it's attracting so many people. Obviously, it's a great job market to be in, and it's also very much a level playing field with the internet. So a software developer, somewhere in Asia Pacific, is technically as capable and as powerful as a software developer in the U.S. So one of the things that we realize is this intersection

of the developer, the software that they're creating, and then ultimately what we do is provide the infrastructure that it runs on, and bringing very much a social aspect to what we're deeming in the near future collaborative infrastructure and really empowering software developers to figure out the best way to run infrastructure in the cloud, because at the end of the day, what they actually care about is the application and not the resources that are behind it. So we're trying to make it as transparent as possible. You know, when I see the three companies up here, I just want to kind of frame it out a little bit. GitHub is where you build software, develop and build software, digital ocean.

is where you deploy software, and Twilio is how you communicate between software. So maybe on that, Jeff, how do you guys look at sort of startup companies versus legacy companies in terms of their development processes, and are there differences that you see? Yeah, I mean, I think every company is becoming a software company or else they're not going to compete effectively. And every company is figuring this out. What's fascinating is that software in most companies has moved from, you know, the back office. It used to be you had ERP and there were people with spreadsheets and that was the extent of how software was used in your business to now software being the business, right?

And you see this with banks, right? Bank of America is now closing branches and moving it to the app. And I think the fascinating trend that has happened is that, you know, it used to be when we started Twilio, people said, well, you know, why are you going after? developers. Developers aren't a market. Developers don't, you have a checkbook, they don't buy anything. Right. Right. You heard that too? Sorry, there's no money in developer tools? I heard that. I heard that. Yeah. The interesting thing is, is people who said that were really thinking back to this age, like 1990, right? When every software that a company used, you know, costs, you know, three million dollars and took three years to install, and it was a big deal and you bought all your software from Oracle and Microsoft and a few others. And in that

world, you know, I like to call high-stakes software. CIOs bet their careers on some big Oracle implementation, right? And in those days, yeah, the developer didn't cut the $3 million check. Absolutely true. They just worked with what someone told them to do. But the big thing that has changed, similar to, like if you think about, well, software as a service came along in the early 2000s, and you had essentially multi-tenancy changed the game. Because now you could cost effectively on board thousands or millions of customers and every deployment wasn't a one-off. Well, that's now migrated to the developer, where multi-tenancy and APIs have come to play so that a developer can come in self-service provision infrastructure or any other form of component of a software application. And they're not spending millions of dollars.

They're spending a dollar, right? And it's not taking 36 months. It's taking 36 minutes to build the prototype. And so this is putting the developer now in control, right? Because while other parts of the business might be debating which vendor is going to take the three-year contract, the developer, in less time than it took to schedule that meeting, has built the prototype and is showing it off. And I've literally had developer customers of ours say, thank you so much, you made me into a hero. I walked into a meeting with the CEO of my company and some code that I'd hacked together the day before in this demo.

I made everybody in the room's phones ring. right to show off what we could build and people are like oh my god this amazing like i didn't even see the the 17 million dollar budget for this and the developer's like you know yeah i built it yesterday uh and that's the thing that's changing and we always like to say that you know the the prototype is worth a thousand sales calls right and that's what the developer has that's the magic that developers now have is that they can just while everyone's debating the merits of this and that and which vendor and someone's out playing golf with someone else you know the developers just like, here, I solved the problem. And that's happening at companies big and small.

Yeah. Yeah, that's, it really is a transformation in the industry along those lines. Chris, you know, by the way, all of these folks up here are all the founders and CEOs of their companies. So it's really interesting. I mean, they all started when the company was just them, and they've built very large organizations around them. So Chris, along those lines, like you guys started with this thing called Git, and then you added Hub and became GitHub. And like, what is that? And what's Git and what's Hub and kind of, how did it all get started? Yeah, I read a blog post once that was like, all you need to do is find a Unix tool

and make it into a website and you'll be rich. So I was like, okay, well, Git is a Unix tool, we'll just make it into a website. No, not really. But Git is a version control tool that was written by Linus Torvalds who created the kernel. And it's super good, super fast. It sort of unlocks a lot of different workflows, and it assumes that there's a network connection. And that's super fascinating to me personally, and I think to every developer, in that there's almost this paradox where we are working on computers all day, but it's very hard for us to work with each other. And this is why you see, I think, things like Twilio and even Slack blowing up is like, finally there's an easy way for me to work with someone sitting right next to me.

And for me, that was what Git was all about, is I had this tool. I could easily see the history. and it gave me very Wikipedia-like control over my codebase, but there was still this huge hurdle for me to work with someone who was sitting right next to me. And so what GitHub did is it really created that hub. Is it built on top of open source software, all the existing workflows, all the open protocols,

but it gave you sort of a communication platform to coordinate and just really discuss what you're working on with someone, whether they're next to you or across the globe. So we call it GitHub, and it is really, we think, a hub for people using Git, or maybe that's what it was originally. it's evolved into much more. And maybe this is something that everyone learns as they've worked with the Internet for a long time,

but the communication part is the hardest part and the most important part. And the technical part, you have unlimited amount of time to figure out later. And so we're really focused on the communication part, but we're still super involved in the community. And so really what we've done is we've created a social network looking to Facebook, Twitter,

all the prior art around developers and developer workflows. And so we heavily make use of and participate in open-source communities, but we are our own business. and that's where Git and GitHub sort of collide. Jeff, how did you guys pick communications as sort of the area to win developers and why, you know, Chris just alluded to that

and maybe explain kind of some of the API interfaces and that that you guys have, but communications seems like a interesting place to start. Yeah, yeah, it wasn't like we said, we want to start a company for developers and, like, here are all the different things we could do. Twilio really came from frustration of mine of things that I had wanted to build.

So I'm a serial entrepreneur. I'm a developer. The serial entrepreneur had started three companies before Twilio and have been at AWS as a product manager early on. And when I left, I was looking at a wide variety of ideas, things had nothing to do with developers. I was just sort of thinking about, like,

what are problems that are interesting and I'd want to solve? And I had this realization that at every single one of my previous companies, at one point or another, we had needed communications. And not like, you know, the desk phone for the employee communications, but actually communications

that was a part of, of our application, part of our workflow, part of our customer experience, like something that was like deeply integrated into the software product that we were building. And at every one of those companies, I said, well, I'm a software developer, so I don't know anything about communications.

You know, let me go find out how to build this. And you go talk to the industry, and industry, you know, telecommunications industry. And they'd say, oh yeah, yeah, we can help you build that. We're gonna bring in all these lines from carriers to your data center. We're gonna rack up all this telco gear. And that's gonna be a four million dollar

project and it's going to take three years to complete with an army of professional services sign here. And I said like, first of all, we were startups. We didn't have the $4 million to spend on it. But I'm also looking at these people saying, hmm, yeah, I'm a software person. In the world of software, you're agile.

You are keeping a backlog. You're shipping every two weeks. You're iterating constantly to find the right solution for your customers. That's the speed at which software operates. And here are these people saying it's going to take three years to deploy this. And I'm saying, that's insane. You know, nothing in software takes three years, right?

Because you assume everything will be different three years from now. And there's no way you ever embark on a project that looks like that. And so after having that experience, three companies in a row, I said, huh, you know, this whole world of communications, which is one of the largest industries in the planet, has actually been stuck in this very hardware-centric, this physical, the physical items of communications, actually dictate what you can do in the world.

But if you think about it, Alexander Graham Bell made the first phone call 150 years ago by wiring together a speaker or a microphone and some copper wire. And, you know, in the 150 years that have progressed, everything in technology has gotten amazing. Right? We've invented vacuum tubes and transistors and integrated circuits, miniaturized it all. We did copper wires turned into fiber optics and satellites and wireless. And we've wrapped the whole planet, like literally, in all this technology.

Yet if you want to make a phone call, the only thing. change that's really happened in 150 years was dialing with an operator, to dialing with a rotary, to dialing with a push button. Now open up the phone app on your iPhone. Supercomputer in your pocket. It's still just the touchscreen representation of the same push button telephone that John F. Kennedy would be familiar with using. And that's insane. And that's because the world of communications has been essentially the tightly coupled to the physical natures of that network. And so what you could do with communications is all about,

yeah, like let's rack more boxes and pull more wires and all this. And we're saying, well, let's bring this into the world, the software. Let's make it programmable. Let's make it agile. And let's make it scale up and down and make it global and remove the social, you know, the political boundaries that telecom always has.

And just let's turn this thing into software. And that's really where we started the company. Yeah. Awesome. So Ben, And DigitalOcean, an amazing story you point to the growth. Recent studies show that you guys are the second largest cloud provider by number of server instances or instances that you guys serve out second only to Amazon. And even though I'm on your board, like I still find that hard to believe. So like, can you explain that?

Like, how can a startup actually come in and grow so rapidly? the face of, let's say, Amazon, Azure, Google, you know, sort of all the big guys who you would expect to really win and dominate the cloud world. How does a startup do it? Yeah. I mean, all those companies at one point were startups themselves. I know, but yeah. So, you know, from humble beginnings giants emerge. For us, you know, I started my first company in 2003. It was a managed server provider. And so I was already in this industry and, you know, having done that for eight years, you learn a lot about the competitive landscape.

And what emerged is, you know, the old world was all about dedicated servers kind of going back to Jeff's point where, you know, we're just racking and stacking more equipment. And then Amazon came out in 2007 and kind of really brought cloud computing to the front and center. And the big shift there is that as a business, you no longer need it. to worry about the metal. Those resources were now abstracted, and there was a provider who was able to deliver them to you

via a web interface, an API call. No more human interaction required in order to build out a massive web infrastructure. And so what we realized, you know, having been in this industry for nearly a decade, is that as Amazon had grown over time, their features that had also become much more complex because they were serving an ever larger customer.

And, you know, we kind of associate and feel that word developers ourselves, kind of the type of people that would use Twilio or would commit code to GitHub and actually do. Twilio is integrated into our two-factor authentication, and our entire engineering team is on GitHub. And so we, you know, as, I-5, Air 5.

You know, as kind of, I have this tagline, developer as CEO, right? And I think what we're seeing is these software-defined companies are looking at the world in a very different way. And the problem is that

Amazon Web Services or Azure, all of those businesses essentially started off in a very different environment. They figured out, like in the case of Amazon, they needed to ship software continuously. They needed to iterate in the e-com commerce space, and that's what actually led to the creation of Amazon Web Services.

They were finally able to productize this internal system and push it out to the rest of the world. The problem became that there was a huge disconnect with the user base. The customer was now forced to use a tool that was not technically built for their use case. It was Amazon's use case and then generalized for mass consumption. And what we started with is this notion that if, we believe the developers really at the heart of the ecosystem going into the future, we should

really put their needs first, which is why we focus on the product experience. You can, you know, sign up with an email and a password in 55 seconds. You've already provisioned your server. I mean, sometimes we have customers and kind of group studies that we've seen there. They still look at the screen and we're like, what are you doing? They're like, well, I'm still waiting for the server to be provisioned. And we're like, no, it's ready. You can actually log in and start using it. think what this speaks to is a lot of companies are engineering driven and it becomes very difficult to differentiate in this space because can you truly claim to be faster than Google or more engineering driven than Amazon and it's like it becomes this murky mess whereas no one is really

focused on the on the user and the experience that they're that they have using these resources and that's really where where we started first and foremost and we're trying to actually abstract this away because at the end of the day, you know, the customer doesn't care what lies underneath the application. What they want is a reliable service to get the application out into the hands of their users to iterate on their feature set to actually ease software development. And that's why platform services have become so popular. So I think, you know, going back to DigitalOcean's origin, it was putting that user perspective first and building the entire company around the customer.

And, you know, we are one of the only pureplay, I think we might be the only pureplay cloud out there. DigitalOcean does nothing other than infrastructure as a service. So we've been able to curate this entire experience from the API with a single call to create a server, a single call, take a snapshot, transfer it to a new region, to the web interface, to all the tutorials, support, and documentation that we provide. And what winds up happening, right, the reason why, we've become number two is when a user actually goes through the process and consumes that

resource, they fall in love with the service. It's exactly what they expected. It's performant. It's cost effective. And then they can't help but share that experience. So they'll go on Twitter, on hacker news, write a blog post, or even just tell a friend. And even to this day, over half of our customer signups come through a word of mouth recommendation, keeping our cost of acquisition really low and essentially creating a viral component to our business, which is really unheard of to think about, you know, a utility provider in some aspects to be able to garner a viral network effect. And I think it just speaks to, once again, you know, not enough companies are actually placing enough time on what the customer needs. And what we realize is that simplicity is a great

long-term differentiator and everything that we do is through that lens. And I think that's obviously what speaks for the traction that we've been able to attain. It's really interesting. There's a few unique dynamics to the developer market. So big companies who try to market to developers, like I often imagine in my head when a big company says, oh, we have a strategic idea. Let's go after developers, right? And they like send their marketing and sales teams after the developer audience. Like developers literally like will look at a salesperson and like imagine them in a clown suit. I really think it's like oh look Bozo just walked in

and he's going to try to sell us some stuff I mean like that's how I always think about it like you have to be credible right you have to be authentic you have to be a part of the community you're serving and it's through that authenticity that you actually build your ecosystem your community and if you are

successful at doing that you know I often say the developer market is like you're doing customer acquisition like a consumer business like a massive funnel lots of developers coming kicking the tire like a consumer business but unlike any other consumer business

they have the ability to later spend like an enterprise because they know about your service and all it takes is the use case to come up at work like oh we need communications or we need service like oh I know how to do that let me spin something up with this tool

that I know and now suddenly that can grow into an enterprise size revenue stream for the company yeah interesting I agree we'll talk about business models in just a second here because I know many people often ask, how can companies that either run open source, support open source,

how do they actually make money? And we'll talk about that in a minute. But Chris, wanted to get to you on China. There were reports earlier this year, I guess a couple of months ago, that GitHub had a security attack from China. Allegedly. Allegedly. That's what I said, apparently, allegedly. Why would anybody care about hacking into GitHub? Well, why do you allegedly think that somebody would care? Why do I allegedly think someone would allegedly hack into GitHub? I think that, you know, we're entering this age where whether you want to write code or write text or write a poem or share a photo, a lot of people want to put something out there on the Internet and make it available to everyone.

And I think when maybe we're getting used to with Facebook, the idea of posting a comment or a photo everywhere, but it's still less sort of ingrained in the general psyche that you could post code for everyone, right? And when code is still seen as very much like infrastructure, as data, as something very tangible, maybe more so tangible than a blog post or an idea. And so I think what GitHub lets you do, and one thing we've always done really well and cared about,

is you just have a couple clicks and your stuff is up there. And just like the basic YouTube model or anything, a blog, you sign for an account, you can post whatever you want up there. And I don't know, I think that that has been a challenge for us in that a lot of really amazing ideas have been posted out there. There's never been more open source than there is today in the world. And I really believe GitHub is a big part of it. And I think a big part of that is, you know, it's the sharing economy. All of this code was sitting around on people's laptops, little scripts, three lines of code, a little enhancement they made to their JavaScript or whatever.

And before there was this arduous process for you to share it with the world. You had to publish it. And I'm not going to publish three lines of code. That's absurd. but I'm going to share it, I'm going to put it in my bucket or on my profile or whatever, right? I'm just going to throw it over the wall. And as that happens a lot more and more, you get a lot of great stuff that was never visible to the world before. But you also get weird stuff or bad stuff.

And so we often have to deal with people writing code specifically that's malware, people writing code specifically that helps a botnet. And that's a challenge that we have to face, in particular, regarding the United States' laws. And so I think when you start having a global community, and we are, only 30% of GitHub's traffic comes from North America, you start dealing with different mindsets and different governments that have a different perspective on what is okay to share and what's not okay to share. And when our mission is make it as easy as possible to share everything, that comes in conflict with the way certain people see the world.

So I think it could be a GitHub problem. We could get into conversation about China and GitHub and Freedom of Speech. But I think it's just larger issue happening on the Internet right now is that everyone's lagging behind. I mean, as we talk about the developer, honestly, I think everyone in this room, including us, we lag behind the concept of the modern developer, of the new developer. We don't have a clear understanding of what that is yet. It's happening so quickly. So the idea that a state government or a state actor would be understanding the nuances of sharing

information and what that means and what's harmful and not harmful, it's ridiculous. They're at least 10 to 20 years behind on that. So, yeah, we, I think, as a social networked, as a publishing platform, have to deal with the same sort of challenges that any publishing platform deals with. For us, it's particular to code. but I think that what we're experiencing related to China or whomever is not dissimilar from what Facebook or Twitter

or all these people are facing with, I think it's going to go a lot worse before it gets better. And it's not about hacking or stealing information or trying to get passwords from us, it's about suppressing speech. Yeah, interesting. So there's often, I've often associated the open source movement

as also being sort of developer oriented, meaning that the companies that build open source are the ones that are going to have the most developer traction. And it's kind of interesting. We have company, I mean, Twilio is not an open source company, and DigitalOcean sort of touches on it. GitHub, of course, supports all open source, but isn't really an open source company. So maybe we can just talk about the business models. I've often, you know, we're investors in two of the companies up here, and, you know, part of of our analysis in investing in DigitalOcean and GitHub was the way they monetize the use of

open source, right? Open source is a very challenging business model. And if you don't get it right, you are chasing a downward spiral in terms of, you know, everyone racing to zero with open source. And yet here we have these incredibly profitable companies that are two of them, and then we'll talk about Twilio's model, but two of them that are... applying we're not profitable? No, no, no. I'm implying that you're not open source, yet you're appealing to the developer, and so how does that work if you're not open source? So we'll talk about that in a second. But Ben, maybe, you know, how do you guys make your, you use open source, and yet you monetize it? How do you guys do that? Yeah, absolutely. I mean, at the heart of what

we're doing is virtualization, and today we're using KVM prior to that. We used Zen. And it was a very, you a very concerted effort to ensure that we don't bring some proprietary close-sourced software, A, because there are license fees involved, and we just wanted to really deliver the most cost-effective solution to our customers, and then B, ultimately, we actually want to have full control over the environment. Now, we haven't made any modifications to virtualization at this point, but the good news is that in the future, if we wanted to go down that route, we have the capability to do that. So that, you know, that kind of open source capability is really what empowers DigitalOcean from, from its origin. But I think what's even more interesting is that

the virtual machines that run on our cloud, the majority of the software that's loaded on them is also open source. And so it's, you know, it's providing this underlying foundation for everyone else to run a successful open source project that makes DigitalOcean successful because if you take a look at the more complex projects out there like, you know, Hadoop is pretty complicated, Cassandra, some of the big data stuff. I mean, we offer a lot of like one-click installations

that really simplify and reduce the barrier to entry into those environments. And so you're able to load this open source software onto a open source cloud at essentially, you know, with a pass-through in terms of cost. So I think we're very complementary to open-source. A lot of the work that we're doing right now

we're actually contributing back. So in the metric space, we're getting heavily involved with some of those open-source projects which we're using internally and contributing back. But even more importantly is, you know, I think if there was a closed environment like if we were running VMware for instance it would

I'm not sure exactly the right way to like phrase it but there's a community that's created around the open source and they they go hand in hand right so they're I don't know I can't really finish the thought but they're self-reinforcing in nature and if we were running a VMware cloud it wouldn't make sense to run open source

on top of it which is why I think VMware's V-Cloud initiative hasn't necessarily hit the traction that they wanted to see. So going back to Chris's point, you do have to stay authentic and original with the developer community. So if they're writing open source and your cloud is open in nature, and actually I think this will be our differentiator over the long term, not that I think about a little bit more, Amazon is known as a closed cloud with, you know, vendor lock-in is kind of written all over about in the press.

and the idea for us is that we actually want to empower that open source movement because it's not the proprietary implementation around DigitalOcean, it's the full package, it's the full experience that will actually allow us to succeed over the long term. So we want to empower open source projects to become as useful, as powerful as they can on top of our infrastructure, and if you want to rip that out and take it to Amazon, you actually can. And we think by staying open that we will actually win more customers in long term because they'll choose us for the right reasons rather than being locked into that environment.

So, Jeff, I'll jump to you now. Ben just made the point that developers follow open source. If you're an open source developer, you want something built on open source. You guys are not open source yet you have this incredible developer following and you're selling your, you know, you sell to developers. so that feels counter almost to the momentum in the industry right now. How do you guys, how does that happen? I think we're actually very similar to DigitalOcean and GitHub in that we use open source software extensively.

Sometimes we release code back to our customers, usually non-core assets that we've built that we think the community will enjoy. But fundamentally, one of the things that we are all doing, as our businesses is taking open source software and actually making it so a developer doesn't have to figure out how to operate it. And, you know, there's been so much open source software written and it's very sophisticated, very complex software

and using that software, you know, just writing a thing like that exercises Cassandra or whatever, well, that's gotten pretty easy, right? But actually operating it, scaling it, making it reliable, redundant, is harder than ever because of the nature of global, real-time, always on cloud applications.

And so while building software is getting easier, operating it is getting harder. And that's actually where the as-a-service world is adding value to its customers, which is to say, you no longer have to worry about how to scale it, how to make it resilient, reliable, secure, et cetera,

you just get the benefit of it. And you get the benefit of what open source software or proprietary software. You get the benefit of what it promises, but we're the experts in scaling it. That's why you pay us. We're going to aggregate the expertise. Like no one in the world is probably better at running GIT than GitHub, right? So if you want to install it yourself, great. But you're not going to run an instantiation of it as well as a service provider who that's all they do.

And they build up that expertise. And when you think about the economy is a scale that go into operating. these services, when you do the math, you're like, yeah, it just makes sense. I'm going to pay someone who's an expert who's operating at global scale. They're going to do it better and more cost-effectively than I can do it myself. And you know, you saw this. A lot of companies jumped on NoSQL bandwagon, you know, in the 2010, 11, 12 time frame, you know, and stood up their own Cassandra clusters because they read the vendor arguments about how fast it was. And like at the beginning it sounded good. And then, you know, boom, node died, hit a bug, all the

data was deleted down for 36 hours and it's like everyone realized oh yeah like there is no free lunch like operating cloud scale software is still really hard yeah yeah and you know in some ways like we're just taking open source and we're making it multi-tenant and we're making it scalable and we're making it cost effective yeah so chris let me you know jump over to you most people think github is an open source company and i remember i wrote a blog a couple of years ago and titled Why There'll Never Be Another Red Hat. Like, I really, I'm not a big believer in sort of the traditional open source models. And I was very worried.

I had just joined the board of GitHub, and I was, like, panicked that Chris was going to be, like, all over my case about, you know, how can you be, you know, throwing open source under the bus and not supporting it and all that from a business standpoint. And I showed him the blog, and he's like, oh, yeah, this is great. This is exactly, like, we don't, we don't release anything open source. and I'm like, hmm, GitHub is all about open source, but you don't have open source, like, help me to understand that. So, you know, how do you guys, I mean, I know, explain to the folks here how you guys monetize what you're doing and what, you know, how you feel, like, what's happened there?

Yeah, I mean, I have been criticized and have been in this conversation for many, many years now, but I grew up as an open source developer. I have some, luckily, I'm like a little bit too old to be part of the generation where everything on Facebook, but I, like, wrote papers in college about how open source everything is the way to go that I'm completely embarrassed by now. I mean, well-written papers, well-researched, but I wouldn't put them in front of this group. So I grew up as, like, I was installing Linux, and when I was

young, and I was going to install fest, and Microsoft is evil, and, you know, Bill Gates is the Borg. I came from that world of open source, but at the same time, you're watching Red Hat, you're watching Microsoft, you're watching Apple, and it's like, I'm a capitalist. I think capitalism fuels innovation, and I think it actually fuels altruism as well. And so, What I want to do is make an impact in the world, and I think that history speaks for itself. And really, what history says to me is pragmatism. It's the best tool for the job. And there are cases in which open source is the best tool for the job, and there are cases in which it's not.

And ultimately, at the end of the day, I am not a crusader for the cloud, and I'm not a crusader for open source. I'm a crusader for you building the best software possible, and you shipping the best product to your customers possible. You having an awesome software development team. Part of that is having a software development team that is empowered, that has autonomy, that has agency, that's able to use and make decisions about the software that they're using, able to share, able to communicate openly. And part of that is about making money off of the service you're providing. And so for us, I think GitHub really is the sweet spot there.

It's not about open versus close for us. It's about open and closed. And I really believe that that's something that we embrace that has opened doors for a lot of companies where it's okay for you to sell proprietary closed source software and at the same time send a pull request to one of the biggest open source projects in the world. It doesn't make you a hypocrite. It doesn't make you anything like worse or it doesn't make you a fraud.

In fact, some of the biggest closed source companies in the world are built on top of Linux, the most massive open source project in the history of the universe. So for me, the whole discussion, the whole philosophical discussion, it's not anchored in the right sort of mind frame, which is what's best for the end user. And ultimately, I think developers are lazy. And this is the thing. Like, Twilio's awesome because I'm sure there's a million Twilio competitors that are open source, but, like, they're difficult.

I don't have time for that. I don't care. I just want to use Tuolio. And DigitalOcean is the same thing. You've yourself been in hosting for over eight years. Like, what's different? DigitalOcean is simple as hell to set up. And really, that's what I think it's about.

So when open source solutions can provide that, that's great. And when they can't, that's great too. And I think ultimately it's about software development and not open and close. And I think four years from now, there won't be a single business on the planet that only uses closed source or only uses open source. I think even the companies like Red Hat,

even the MongoDBs, even the Hortonworks of the world, they're using some ERP software or some Intuit software that is itself closed source, right? And it's disingenuous to say, and I'm not accusing anyone, but there's no only open company. It's a mix. And really it's about what is the best for your business, your developers, and your customers right now. That's what I believe, and that's what we've tried to push with GitHub. And that's why the business model has been open source is totally free and closed source costs money because we've always believed that the more we can get people using open source, the more we're going to have closed source businesses making money, profiting, and ultimately

contributing back to GitHub. So that's sort of what we've come from on that. Chris, I'll extend something that you just said. You said it's about software development, right? And I think that's one of the key things is that developers love building. It's like it's inventing. It's the human spirit of building that's existed

forever. But once it's built, they don't want to operate. Right. That's just annoying. That's busy work because what happens? You deploy it. And then, you know, three months later, later, you know, a pager goes off and you're like, oh, I got to deal with this now.

That's annoying. And so what a lot of companies in the as-a-service space you're doing is essentially removing the non-fulfilling, the thing you usually get yelled at by your manager for or your customers buy, and the thing that's annoying and getting rid of that so that all you do is have that joy of building. And once you've built it, it's someone else's job to operate it. And it's that separation of concerns that I think has really come to a maturation in the last five, ten years

because you can do it now, because of multi-tenancy, because a lot of the things that have advanced that allow us to say, yeah, we can, you know, ship that code to somewhere else and let it run in a way that many customers can run on the same infrastructure without affecting each other. And that's like, you know, joy for the developer, right?

I don't have to worry about it anymore. Yeah. I remember my last company, we were installing, we needed, you know, trouble ticketing, you know, for our customers, so we're installing in an open source mantis. if anyone remembers that, right? And, like, we needed a blog.

So I'm installing WordPress, and we needed a CMS, we're installing Drupal. And we had this collection. I had one server that had, like, 40 open-source PHP applications on it. And they were all in various stages of, you know, too many versions old and getting exploited because they were insecure or not scaling right.

And, like, starting Twilio, it's like, we need ticketing. We use Zendesk, right? We need, you know, CMS, or we need WordPress. We use WP. engine. Like, you need, and everything is now as a service so that we don't have to operate it. Great. There's still a big obsession, I think,

with code. And I don't know why, I mean, we're a part of it, but writing code is the easy part. Running code, maintaining code, debugging code, scaling code, that's the hard part. People ask me, I'm in this fortunate position. I've been a developer my whole life, and I no longer write a lot of code. People will say, do you still write code?

And I do. Like, it's fun for me. Like, a hobby is, like, some people play guitar for money. Some people play guitar on the weekend, and they're different. But my idea of writing code now is, is like a little snippet that no one will ever use but me, or a plug-in from my editor that no one will ever use but me. So my answer is no, because now I think of writing code as debugging it, getting code review, running it,

watching the exceptions. And so much of the software development lifecycle is not writing the code, but in maintaining the code, in making sure it's bug-free, and running the maintenance on it. And I think that's really the next thing. I think that GitHub is actually going to get less interesting in the future. And it's going to be less about I can grab an open source project and run it, and more about where

do I run it? What are the systems it integrates into? And so GitHub is really awesome, I think, for many reasons, but it's going to become more of a commodity, more of a sort of table stakes that, oh, yeah, there's this place where I can get all this code. That's not, I'm not trying to hire for that. I'm not trying to hire someone that can download something from GitHub or write some software that integrates from Trilio. I'm trying to hire someone that can sort of manage this holistic process and run the servers and see when things are wrong, and in fact, debug them when things go to crap at 2am. That's the real business value you get out of it. And I think in the future, that's going to be the thing that

schools start teaching and the things that we start hiring for more and more is not the aptitude test around what Java functions do you know and do you know the difference between static and dynamic, but can you debug a piece of running software, how do you scale a piece of running software, and how do you work with someone else across the planet when you're both working on a piece of software that's going wrong?

I think that's the real thing that we still haven't figured out yet. Yep, great. Chris, Jeff, Ben, thanks a lot, and thank you all. Thank you.