 Why is security such, you know, it's becoming important for you? I mean, it's important, right? I mean, one thing is Linux runs the world, right? So we want to make sure that things are secure, like your device, you have your whole life history on your phone, right? Contact information, all this stuff. You want to make sure that that's secure, that that's something that you can trust, that you can trust other people don't grab. I mean, you have access to your bank account on your phone, right? I mean, so the basic underlying building blocks of what those things rely on, you want to make sure are secure. And I said earlier with Linux, I think it's also a matter of education, people are more aware of this, but it's also a matter of accommodation, that that awareness combined with reliance or use of this technology in the world is taking over. And that's true, you know, the whole world is running on, you know, Linux and open source technology. But at the same time, there are so many layers. Yeah. So which one of those is the one that you get, you know, the most mad and angry, you know? So the whole spectrum meltdown problem, what made us so mad in a way is we were fixing a bug in somebody else's layer, right? I mean, taking ownership of our own problem is great. We have security bugs, our problems will fix them. But when we have to fix the problems of hardware or other issues, it's frustrating, right? Now the goal of a kernel is to paper over all the bugs in the hardware. It's just that usually the bug isn't in the CPU. I write drivers, so we're always working around drivers issues. It's fun if you read device driver code or sometimes like one famous line is like, do this, do it again, hope it is going to really work. The third time it's going to try, it works. So I mean, it's stuck, right? So that's working around bugs in hardware. But now the bugs in the processors. One also interesting thing about the whole spectrum meltdown is the complexity of that black box of a CPU is much, much larger than it used to be, right? Because they're doing in order to eke out all the performance and all the neat things like that, you have to do extra special tricks and things like that. And they have been and sometimes those tricks come back to bite you in the butt and they have in this case. So we have to work around that. But every single layer of the stack, silicon all the way up, there are issues and you have to be aware of your only focus on your issues, right? So I focus on what the kernel can do. And as long as we can make that layer the best we possibly can, add additional security and additional layers will be good. Do you think that things are getting better now? People say things will get worse before they get better. Things are getting better, yeah, all the time. So we're doing more and more testing, more and more builds. So a big problem was the last round of security patches that we had as we worked on it on our own performance and then we pushed them out to the world because we got them barcode and then we found all the problems with those because when it hit real-world testing and it made us realize how good and how much we've relied on the infrastructure that we have built up over the past years to do this testing and to make sure that we don't have bugs before they hit other people. And so when you're forced to remove yourself from those things that you relied on you realize we've gotten a lot better. We're doing more testing. So there's a really cool project from a number of Google engineers called the SYZ bot. So it's a fuzzer. So it goes and it exercises the kernel and it tries to tweak parameters to it and send bad data to it. And I saw a presentation by one of the developers. The presentation was the world is on fire and from his point of view he's finding hundreds and hundreds of bugs in the kernel and it's like what is going on and because we rely on this and everything and I get to step back and remember the kernel as is works really really well. But when you start exercising the crazy paths and the crazy paths and the crazy paths and that are weird stuff. The fringe of the whole thing there's problems and there's always problem but now we're testing them and we're finding them and his layer is like he we fix all the bugs on one layer or one wrapper and then his tool can go one level deeper and then we find more bugs and then we find the fix those and then he goes one level deeper and we find more bugs down there. But we would have never found those last level bugs because nobody would have ever exercised those paths unless we fix the levels of above. So as we get deeper and deeper in the stacks and figuring out we're finding more and more bugs which is better. We're finding things that have always been there just no way has ever exercised those code paths in the kernel otherwise it would have crashed and we wouldn't have noticed it so or just wouldn't have gotten there. So the testing and the development effort that's going on is better than it ever has and it's finding lots of problems but that's not to say that the world is on fire it's just that we're doing really really good at finding problems. Exactly and some of the cool stuff they're doing is like we're adding so I watch all the patches that go into the kernel to find out what they're going to stable kernel so we've been adding a bunch of new networking protocols and new stuff has been coming out in the working stack and these developers these testers are finding and fuzzing these network stacks are being added to the kernel before they get in a real release. So you see the patches go in you'll see a whole bunch of bug fixes that go in which is great because traditionally those bugs would have never even been caught until years and years later when they're in production and stuff like that. So this infrastructure we have is catching things at an earlier stage because it's there which is awesome to see so you look at all the bugs they're fixing but it's like wait you're fixing bugs with codos out of last week and nobody ever else saw at all. So our infrastructure of this testing and validation and fuzzing is getting really well really good.