20 Jahre Ubuntu - Interview mit Gründer Mark Shuttleworth

I was very inspired by the work of open source communities and specialists. A lot of what I had been able to achieve in the late 1990s was only possible because of open source. So I wanted to give something back to the communities that had supported me, and I also wanted to increase the opportunities for others to do their own kind of entrepreneurial or educational or scientific magic with open source. I thought that if we could solve the usability challenge of Linux back then, then many more people could benefit from the contributions of open source engineers, and those contributions would have bigger impact in the world. So it seemed like a nice challenge. I was also a little insecure about whether I had been just lucky as an entrepreneur or whether I had been quite savvy about how I'd navigated the complexities of that opportunity. So I wanted to do something that was commercially very difficult, and building an operating system and then giving it away and then making it commercially successful, that seemed to be very difficult. It was interesting for lots of reasons, became a bit of an obsession. And I think if I'd known that I would end up this bald and this gray this soon, I might not have done it. But I'm on reflection kind of very grateful for everything that I've learned along the way and all the fun people I've got to work with, and to a certain extent the impact that we've already had is great.

I studied Information Systems, which is kind of like the business side of technology, and I was doing that and finance together at the University of Cape Town. While I was there at university, the internet arrived in South Africa in universities initially. So I became really interested in the internet and how it would change business, change societies, change culture. And I started working as a student. I started Thawte actually just as a consulting company to help local businesses get on the internet. So this was really how I learned the underpinnings of what today is IP and the internet and so on. Someone gave me a stack of floppies, which was Slackware, and so that was really transformative for me. I started to work with Linux. I couldn't study Linux because the professors weren't interested in that, but I started to use it to deepen my understanding of the internet and help companies get online. And I imagined I was looking for a business opportunity which you could do from South Africa. South Africa didn't have a lot of money and it didn't have a lot of bandwidth, but this area of cryptographic identity and establishing international secure communication, that seemed like an opportunity for various reasons. So I pursued that as a student and then grew that after graduating into a certificate authority. Along the way I had to essentially build all the business systems behind that, and all of that was done on Debian. I became a Debian developer as part of that process because there were various things that I wanted to use that weren't debs at the time, and I thought if I contributed them that'd be one way to give back. So I finished that kind of journey as a not very deep Linux kernel expert, but certainly with an understanding and appreciation for how a Linux system is put together and how Debian as a community worked. Those things were kind of real to me when I sat down to think about what to do next.

Well, actually I did do that, so that was the sort of starting point. In fact, at the time, rather naively I thought that Debian would just win on its own. When the other major distributions started to go proprietary, I thought, well, that's game over. Debian is technically much better, it has a much bigger community, and it has much better technology for packaging. At the time, I still think debs are better than RPMs in many sort of fundamental technical ways. So I sort of assumed that Debian would just win because it had that technical and community merit. And it was a hard lesson for me to watch. I saw the Linux distributions that went proprietary actually being very successful, and at that time Debian was struggling a lot to make releases, and it seemed kind of stuck. So my first efforts were to encourage Debian to do certain things, and I just found that what I initially interpreted as kind of resistance to change, I now think about slightly differently. I think about it as Debian has a set of values that are very true to Debian and actually very important and very unique. But at the time, I found that I couldn't convince people in Debian to move in the way that I thought they needed to move to get back in the game. And so on that basis, I thought, well, why not start the other side of the coin to Debian, which I still feel very strongly. Ubuntu and Debian are two sides of the same coin. We've consistently made choices to stay attached and joined at the hip, but we're able to bring something into the Debian ecosystem that hadn't been there before. And I think that approach has worked very well. Debian still very much owns its story, has a very clear set of values that can't be pushed around by me or anybody else, and I've come to really appreciate that. On the other hand, I think the existence of Ubuntu has punched that ecosystem right to the forefront of the broader kind of understanding of open source. And I say that not to detract from some of the very cool things that are happening in Arch and Nix, which are also interesting, but I think the Debian and Ubuntu story is still very central to how most people think about using Linux in production, for experimentation and production.

I was able to see clearly that I could be very bold as an entrepreneur if I could build on top of open source, but it was really hard to build on top of open source. In those days it was just really hard to get it onto a laptop, it was really hard to set up and administer. And so I thought, look, if we can solve that problem then millions and millions more people will have access to open source. If they have access to open source, I don't know what they'll do with it, but I'm sure it'll be interesting. And I think really we have seen now today many more students have access to open source. It's a normal part of most computer science or other curricula. There isn't perceived to be a lot of friction associated with that at the level of getting a single server up and running or getting a laptop up and running. So it was a way of saying, look, we have this beautiful thing that open source contributors are contributing to, but it's not reaching as many people as it could reach, so let's solve that. And that way this thing's going to get better as fast as all the brilliant innovators in open source can make it better, but all of them will then have a bigger impact because more people will be able to consume what they're doing.

Making it easy, making it pretty, making it just work. And today their work continues. If I think about WSL for example, it's a really interesting way for the masses to get access to open source. For Mac users, we make Multipass, right, so that you have a Mac, you just install Multipass and then with one command you get an environment and you've got open source at your fingertips. I think it's tempting always just to look at the things that you can control, but I think you also have to think about all the people who are in other environments, and sometimes they don't have any choice about being in those environments, and do some work to enable them to get access to open source as well.

There are so many things that didn't work out exactly as I thought they would work out. The original business plan for Ubuntu was Netflix and Spotify and Dropbox and so on. The idea was that in a connected world you could shift the sustainability model away from licensing software, which is difficult to do in open source, towards services, and then a free platform would make sense. And that has happened. If I look at Apple today, you don't pay for Mac OS. I don't think you necessarily even pay for Windows upgrades these days. The services story I think has moved to the center of the economics for platform software. But we completely failed to do that. I thought we needed to do a Linux in part because if we tried to do that on Windows, then we were likely just to get crushed by the Windows people, absorbed into Windows, or have a Windows version built into Windows that did what we did. I didn't see iOS and Android coming, which are the things that ultimately made it possible to do Dropbox and Spotify and so on. That was the first big mistake. And the second big mistake is I underestimated how expensive it would be to build something like a Spotify or a Netflix or Dropbox. I mean, I thought I had a lot of money at half a billion dollars or something, but it turns out all of those things each consumed like tens of billions of dollars, so I would have been dead in the water. So I had to really rethink things, and that's why we ended up largely anchored around an enterprise play and so on, which isn't very innovative. Red Hat had done that before, SUSE had done that before. It was a bit of a punch in the face for me, right, I was just wrong and failed on a bunch of those things. So we had to then regroup and reshape and build a business around the enterprise story for Ubuntu. I think we continue to invest back in Linux for human beings, but we're doing it from a position where the enterprise is the commercial tent pole, effectively, and that is the sort of the Robin Hood part of what we do.

Definitely, but not permanently. Back in 2015, 2016, I literally nearly lost my shirt. Things were very desperate. I had to be pretty ruthless in what we kept going with and what we stopped doing. And so we had a very strong kind of pivot to the cloud, IoT as you said, enterprise, the server essentially. Nowadays, private cloud, Kubernetes, public cloud, you know, these are all the big reasons why large companies adopt Ubuntu and engage with Canonical, which funds us. But once we got profitable, we again turned around back to the desktop and started investing more in the desktop. So I hope people have noticed that the last couple of desktop releases have had quite a lot of energy and sexiness, right, like interesting things going on. Some people like it, some people don't like it, but at the end of the day I think that team is now again a focus for investment and is growing. And that's not commercial, right, that's mainly because I would love the world to have a really wonderful set of choices for free software desktops. We've also started working more recently again with the KDE community, and essentially doing pathfinding exploration together as to how we can get the diversity of Linux to not essentially hold us all back, rather to be creative and fun and innovative but not as fragmentary as it can be.

I can think of a couple of different groups that we actively invest in meeting their needs. One is the person who wants a secure private desktop, because they either don't have the resources to use the commercial offerings, or they find them difficult, or they're worried about privacy. So that is the real Linux for human beings story. I don't care if you're never going to build great software, I don't care if you're never going to pay us money. I want you to have a wholesome, secure environment where you can explore what's possible in compute. Then you have the developer audience. They're probably not customers of Canonical, but I think they could do interesting things in the world. I want them to feel like they can take a laptop, put Ubuntu on it, and then they have access to literally all the best stuff in open source, in a way where if they're successful, if their ideas work out, they can go to production and really believe in it. And then you have the enterprise, and enterprises tend to care about things like compliance, conformance, auditing, standards of various forms, management and things like that. So those three groups are related but different. I would say we have teams that would say their mission is one or the other of all of those, and then we have some teams who just focus on making the compiler as great as it can be or making the kernel as great as it can be, and they don't think in terms of those different audiences but they're ultimately working for all of them.

Well, I have to give credit to both Android and Chrome OS, because they have both fundamentally changed what we think of as what Linux can do in the world. If I go back to 2004 when Ubuntu started, neither of those things existed, and they've both done really, really well. I would like to go back to the core desktop itself. I think the opportunity is still there to build beautiful, secure, usable stuff. We had a first go at that with Unity, and again, I nearly lost my shirt doing that, so I have a bit of PTSD from that experience. But on the other hand, I think we kind of owe it to the community and the open source world to keep investing in the desktop experience. The nature of that is a little fragmented. KDE is an important desktop, GNOME is an important desktop, MATE, each of these environments have their communities. I don't want to exclude them from what we're trying to do. But yeah, I would genuinely love to see a winning, stylish, beautiful, enviable, pure consumer desktop experience for those people that I talked about in the middle, the Linux for human beings crowd, the people who aren't developers and they aren't corporate, they just want something that's great to use.

I can say it's tens of people who work on the desktop now, and that covers lots of different aspects of it. I can also say it's growing, because now that we're profitable I enjoy investing in that. That team is very passionate and they represent really the face of Ubuntu. When you say Ubuntu to a lot of people, they literally think, I can picture it in my mind, and for those people that isn't Ubuntu desktop. We also have teams that work on the cloud story and also have teams that work on the server story and the IoT story, but mentally for most people, Ubuntu is the desktop.

I think it's just part of being part of the open source community that you try to publish things in a way that they can have an impact outside of what you imagine they could have. Upstart, for example, there was a Canonical engineer who was very passionate about that problem, this was back in the SysVinit days. And so he was able to make the case that we should do Upstart, and we did Upstart. In fact, Upstart is still being used in Chrome OS. So I think that's part of what's great about open source, that when you have passionate people who have technical insights and understandings, and they can find a space to explore that and be successful, then that thing can have a life beyond necessarily the company or the project that it started in. So I think that's a nice aspect of open source. Not everything that you do that way is going to be successful or is going to be the end destination. There's always creative destruction. Sometimes you do things and they win for a while and then you have to abandon them in favor of better things that have come from somewhere else. Generally, I've tried to set the tone that we should be willing to ship other people's stuff if it's the best, but we should also be willing to lead if other people aren't attacking the problems that we think are interesting. It's complicated in the community because you'll be criticized for not innovating and criticized for innovating equally, sometimes by the same people. But I do think it's part of the responsibility of being one of the platforms that lots of people engage with and play with, and to a certain extent again, being one of the commercially successful platforms, even though we're much smaller than Red Hat and SUSE, being commercially successful I think increases the responsibility on us to do something with that that might have a broader impact. So I always tell the team, in 20 years time Linux is going to look fundamentally different to what it does today. The last 20 years have shown us that Linux will change radically over that sort of period. So if we just froze and just did what we do today, well, then we'd be obsolete in 20 years time. So we have kind of two responsibilities. On the one hand, we owe it to our existing users to keep the platform that they depend on stable and unsurprising. And then we also owe it to that same user base to be successfully surprising in inventing the future. And those are sort of directly in tension with each other, those two responsibilities. But we carry them both, and we have to carry them both as well as we possibly can. And sometimes that's controversial, sometimes we're successful, sometimes we aren't, but we can't be shy about taking a deep breath and trying to invent the future.