 I applied Linus with Sushi and beer last night to warm him up for today's performance, but I want to start by asking you to respond to something I read the other day in Business Week, and I'm going to quote a rather long quote about you, and I just want to get your reaction. Torvalds remains the lone official arbiter of Linux, guiding how Linux evolves. When it comes to the software that runs just about anything, Torvalds is the decider. What's more, Torvalds may be the most influential individual economic force of the past 20 years. He didn't invent open software, but through Linux he unleashed the full power of its idea. Torvalds has in effect been instrumental in retooling the production lines of the modern economy as Henry Ford was 100 years earlier. It's absurd that so much power has collected in one man. Wow. It does feel a bit absurd, and I'm not sure about the power, but I love open source and how all the credit goes to be. Wow. What can I say? I actually, I mean, realistically, the only power I have is to say no, and sometimes I do that in a somewhat colorful manner because I don't even write any code anymore. So realistically I get a lot of the kudos for these days being just the maintainer and manager of a lot of very productive people. Just so I'm clear, and his ego doesn't get too big with this kind of credit, the author does follow by saying he's five feet ho-hum-tall with a paunch. Hey, five-eight is just right. It's cheap and easy, but true to say his body type and gait resemble that of tux, the penguin mascot. Wow. Give me why I came here again. It's like a rose. All right, so let's talk about some interesting things that are going on in Linux. So every keynote that has preceded this one has talked about containers, containers, containers, containers, containers. This is the big hot topic that we've been seeing this year, docker, coral, all these different things. What is your take? What do you think about containers? What's going on there? What's your perspective on that? I'm sorry to everybody involved here in containers. I'm so happy that the kernel tends to be fairly far removed from all of these issues, all the buzzwords and all the new technologies. We end up being in the situation that we're such an infrastructure play that we only care about us working, and then how people use the kernel, I actually don't see very much of. And to be honest, I'm so focused on just the kernel that I don't even care very much. We see it when people need technologies from us to implement all of this. Obviously, see groups and virtualization if you do it that way. So we see that side of it. But at the same time, I don't get involved in the politics between all the different groups and all the infrastructure that goes on top of the kernel. And I'm really happy. I'm happy to have you. All right. All right. So let me take another buzzword, IOT, just because I want to cover every single buzzword with you. But as more and more things become connected and require networking and so forth, a lot of just increasingly smaller devices have Linux inside, right? And the question that a lot of people ask me, and I think they probably ask you as well, is how small can the kernel actually get, right? What's your perspective on that? So we've traditionally been in the situation that the kernel didn't have to shrink down. It was always the hardware that grew up to meet the kernel's needs. And the pressure to actually shrink the kernel is, I mean, it's something that everybody has always wished for. If you build a kernel these days, the fact that with all the modules, it's tens of megabytes in size is shocking when you think about where we started. And you could fit it all in the low one-meg area when you load it, right? Those times are long gone. And trying to get back to being a lean, mean IOT machine, I don't know, is going to be really hard. Is it because of the article statements about you? No. It's always really hard to try to get rid of unnecessary fact, because as every developer in this room knows, things tend to grow. And there are projects doing that realistically. I don't think we'll ever grow back down to the kind of sizes we were 20 years ago. We can certainly grow smaller, or that shrink. But I do suspect that if you want to work on some really small devices, you will have to end up looking at other alternatives. Other alternatives. So we covered IOT, we covered containers. Security is also on everyone's minds. Bruce Schneier yesterday gave a talk that I think sort of shook everyone up a little bit, considering just the ramifications of all these different cybersecurity issues that we see out there. Now I won't get into that with you in terms of policy or anything like that, but I would love to ask you around kernel projects specifically and then open source more generally, what's your take on security in the community, practices around security, et cetera, et cetera? So I'm fairly known for sometimes being at odds with the security community, because the security community tends to be very black and white, either it's security or it's not. And if it's security, they care deeply and they make a big brouhaha about it. And if it's not, they don't care at all. And what I see is security is bugs. And most of the security issues we've had in the kernel, and happily they haven't been that big, or some of them have been pretty big, but they don't happen that often. Most of them have been just completely stupid bugs that nobody really would have thought of as security issues normally, except for the fact that some clever person comes around and takes advantage of them. And the thing is you're never going to get rid of bugs. And if any bug you have can be a security issue. I mean, not all can do that, but you just generally don't know ahead of time. If you think of it that way, then you just know that bugs are inevitable. Security is never going to be perfect. And in the kernel, we obviously try to do our very best. We do try to be very careful about code. We have fairly strict standards. And when new people come around, it's sometimes hard to get into the kernel community just because if you're used to user space programming, in the kernel you have to be very strict in some respects. The whole limited stack, the whole odd memory management model, it's just different from most of their projects. But even outside of kernels, the only real solution is to admit that A, bugs happen, and B, try to mitigate them by, for example, having multiple layers of security so that if you have a hole in one component, the next component will catch it, hopefully. And I think open source is doing fairly well, but anybody who thinks that we will be entirely secure is just not being realistic. We'll always have issues. Yeah, but are there things open source projects can do to improve test coverage, threat modeling, what are some things that projects can do to make bugs at least, to your point, happen less often? I mean, what I'd love to see is anybody who does network connections to just have random packet testing kind of things. And I have to admit, we do a lot of other random testing for the kernel, but I don't think we're necessarily doing as well as we should on things like that. And it's different spraying random packets at a kernel machine, which people do, than targeting very specific protocols and trying to then randomize the stream that you send to network filesystem server kind of thing. So there's a lot of work involved, and it's tough. And I think being aware of the issues and at least thinking about them occasionally is the first step, and not everybody has gone to that point. OK. Well, we're running out of time here, and I know I only had you briefly, but I'll give you the last shot by asking a very open-ended question. Linux, 10 years from now, where do you see it? You know what? That wasn't how I started Linux. I'm a very plodding, pedestrian kind of person that I look six months ahead. I look ahead this release, and I know what's coming up the next one. I don't think planning 10 years ahead is necessarily very sane, because if you think about 10 years back and where Linux was 10 years ago, trying to plan for where we are now would have been completely insane. So I don't take that approach. I think open source in particular, one of the advantages is that you have all these companies that are doing, they're trying to make the next 10 years happen, right? And by using open source, they can try to push their own agenda, and they notice, hey, we need this in order for our next 10 years to happen. And that actually helps Linux be, even if I'm personally as a maintainer, not very forward thinking, I think the whole process is very forward thinking, because open source allows and actually encourages that kind of behavior. So I'm not worried about the next 10 years, and I'm worried about a stumbling and the bugs we find today and tomorrow kind of thing. Wise words. I will leave all of you with one more quote because I can't resist. You work at home, correct? Managing, you know, there's 1,500 people, most of which are now professional developers working at different companies and going into their office to do this. However, you work from home and I quote, I'll show up in the kitchen in a bathrobe and it embarrasses my daughter. Her friends will ask, what kind of hobo is your dad? Yeah, well, most of them get used to it, like I had the FedEx guy, he already knows me. I mean, if it's 2 p.m. in the afternoon and I open the door for a packet in my bathrobe, he doesn't blink anymore, so yeah, it's a good job, right? All right, well, thank you so much for joining us.