00:00:00 David Egts: So, Gunner, I saw that the latest version of Susa Linux Enterprise Server has switched from App Armor to SE Linux. Gunnar Hellekson: Yeah. Incredible. Incredible. David Egts: Yeah. And and so my guess is that Dan Walsh is like, "My job is done here. I'm out. I I don't I have nothing to do anymore. I've I've mission accomplished. There's probably like a mission accomplished banner. He's standing at a podium on an aircraft carrier with a banner. We ought to get him on the show. Gunnar Hellekson: We ought to get him on the show. Hey, hold on one second. There you go. Daniel Walsh: Hey, that's a bad analogy there on the mission accomplished thing, but But I am I I have an announcement to make. Gunnar Hellekson: He's on the show. David Egts: Yes, you've heard it here first, people. Daniel Walsh: I'm uh I'm retiring in two weeks. So or at least going down to uh I'll no longer be a full-time employee at Red Hat. Gunnar Hellekson: man. 00:01:08 Daniel Walsh: So thank you. Gunnar Hellekson: Congratulations, Dan. David Egts: Yeah. Daniel Walsh: It's been after 40 years of working and programming and stuff like that. It's time to take a a step back. Let all the youngans take over. David Egts: Yeah. It's It's the end of an era. This is This is amazing. Gunnar Hellekson: It is the end of So Dan, is there is there a Well, first of all, let's get this out of the way. David Egts: Yeah. Daniel Walsh: Yeah. Gunnar Hellekson: Are you retiring for any particular reason? Like, are you are you being driven out? Daniel Walsh: Yeah. A AI is is forcing me out. David Egts: Took his job. Daniel Walsh: No, no. David Egts: Yeah. Daniel Walsh: This is uh purely by choice. Um no, I've uh uh gonna be 65 at the end of December and felt like a good time. I have five grandkids and like to play with them, but I'll still be around. Gunnar Hellekson: Oh yeah. Daniel Walsh: I mean, I'm planning on I'm sure I'll be I'm going to retire on a Friday. 00:02:01 Daniel Walsh: On a Monday, I'll be checking out what happened on GitHub over the weekend. David Egts: and saying it's not your problem. Yeah. Daniel Walsh: I'll just have my wife yelling at me, "What the hell are you doing in that office again? Stop working." But yeah, I have an agreement with Red Hat. I'll still be a a consultant for them. But I've only been I've actually been on partial retirement since the summer using up vacation time working about 20 hours a week. And so but from I don't want to do any more meetings. I don't want to be on Jer. I don't want to be uh making plans for 2026 and 2027. Just want to fool around a little bit, answer some questions, you know, accept LinkedIn requests. Gunnar Hellekson: Yeah. David Egts: Mhm. There you go. Daniel Walsh: At Red Hat, every quarter we have to uh make a your your quarterly plans and you say, "What do you want to what's your goals for next quarter?" And I always put it in retire. 00:02:52 Daniel Walsh: So, I'm hoping I'm still alive. Gunnar Hellekson: The final box. David Egts: Yeah. Gunnar Hellekson: The final checkbox. David Egts: Where do you see yourself in five years? Gunnar Hellekson: Check off. David Egts: Yeah. Daniel Walsh: That's You get to my age, everything is five, you know, I'm like the old Russian thing. Everything's fiveyear plans. Don't want to don't want to plan too far ahead. Gunnar Hellekson: Yeah. Well, so so Dan, if you if you look back on your on your 40 years, I think I mean, if you were to ask anybody at Red Hat, they would say, well, there was the SE Linux Dan, right? David Egts: Mhm. Gunnar Hellekson: And then there was container Dan, right? Daniel Walsh: Yeah. David Egts: Mhm. Gunnar Hellekson: And then recently we saw AI Dan, right? Daniel Walsh: Yeah. You skipped over you skipped over boots, Dan. Gunnar Hellekson: Um, is that Oh, that's right. Daniel Walsh: So, but so I can go on on to a couple of I I actually uh if you Google there's a YouTube video I call it my Red Hat obituary. 00:03:35 Gunnar Hellekson: I forgot about Bootsie. That's right. Yeah. Yeah. Yeah. Yeah. That's right. So, but how do you Yeah. How do you think about your your eras? Daniel Walsh: So there's a whole talk that I did actually at both DevCons DevCom check and DevCom US that I I go through my entire career at Red Hat and stuff but pretty much as a young engineer I got assigned to um basically security but my my area of security was always around controlling what processes did on a system. So when I first got assigned to SE Linux, it was basically, you know, grouping a bunch of processes. You're a web service, you're a database, and you do what those do. And eventually that rolled into containers, which was basically just the next level of really isolating those things. And um so you know, really I started with SE Linux. When containers came along, I basically added SE Linux to make them more secure. Um and then eventually we came to boot z which was basically about installing the operating system and treating the operating system as if it was a container. 00:04:41 Daniel Walsh: Um and then AI happened and um so and AI I'm I'm all about contain you know containerizing the AI. Um so I actually working in the last year and a half on a project called Ramal Lama which is all about running AI locally inside of containers on your system whether it's a laptop you know running Windows or David Egts: Perfect. Daniel Walsh: Mac or on Linux and basically putting your AI into containers. Um, I like to talk about my my big story now is basically when I started at Red Hat or maybe even a little before then. Um, when people first started using Linux, they went around grabbing random crap off the internet and just doing make installs as root onto their system and, you know, hoping that they weren't hacked. Uh eventually Red Hat came along and created Enterprise Linux which really about packaging up software and allowing customers to get software from trusted sources and then taking care of the security of that software. About 10 or 15 years later all of a sudden Docker explodes on the scene and everybody's ra grabbing random crap from the internet and running it as rooe on their systems inside of containers. 00:05:46 Daniel Walsh: Uh but no one knew what any of the software was. Eventually we developed u you know trusted container images and signing and build materials things like that. Um and again Red Hat was at the forefront of a lot of that um and then making the container security great. So then we go 10 or 15 years later, we come till now and uh people are going out and grabbing random crap, random AI off the internet, running it as rooe on their system. And uh uh then then hooking up MCPs, which is all about, you know, handing an API to an AI, and the AI says, "Oh, if I'm going to talk to this API, I need your secrets. Can you give me your secrets?" "Oh, yeah, sure. Here they are." So now we have random AI sometimes from China totally untrusted sources running on your system running inside a software that you downloaded you know again from the internet running it as root and uh and then we're just having it handing MCPS and so my last effort has been all around let's look at containerizing some of this stuff put it into put the AI into containers control the MCPS basically look at rolebased access control which tools in the AI I world their APIs are called tools and basically only limit the tools right if I might have full access to open shift but don't give uh the AI the full access that I have maybe maybe control what 00:07:09 Daniel Walsh: the AI can do and start to look at how you control software so uh but I'm sure 15 years from now there'll be some other thing and everybody be grab grab crap from the internet running it as rudiment system so I don't think I'll be around then for to handle that so maybe I'll be in the nursing home at that point and still doing GitHub pull requests. Gunnar Hellekson: Yeah. Well, no. You'll be in the nursing home downloading random crap off the internet and installing it on your VR glasses, right? Daniel Walsh: Exactly. David Egts: Yeah. Daniel Walsh: Screaming. David Egts: Yeah. Daniel Walsh: I'll be the old guy screaming at the cloud. Stop doing that. So, Gunnar Hellekson: That's great. Is it So, what that I like how you kind of wo all of those things together. Is there any um what do you think drew you to it? Is it just kind of like you started on like uh process isolation and then just kept doubling down and doubling down and doubling down or was it like or is there something like unique to this problem space that get that gets you really interested? 00:08:06 Daniel Walsh: Yeah. Well, I mean, I'm I'm a simple guy. I like I like living at the at the operating system level. Anytime people wanted to get me into like even the early years of Open Shift, like I don't understand those high level REST APIs and stuff. I always understand processes and and you know I dug deep into that and I was good at it and you know I got a reputation for it and uh a lot of it just you know uh back in the actually back Dave and I years ago when we used to work in federal sales a lot of this was led by the US government right US government really cared about security and so they you know they drove David Egts: Mhm. Daniel Walsh: SE Linux and then they they drove namespaces and you know a lot of this stuff long before Docker um was all about you know running multi-level security top secret secret things like that and uh so I found that Gunnar Hellekson: Yeah. Daniel Walsh: interesting um got to go to interest places you know I've been I've been CIA NSA Pentagon I've been you know so into talk to very important people and u in my uh one of the interest things if you see my um my orbituary I had an obituary very uh uh one of the first jobs I got uh when I was coming out of college, I I flown down to the National Security Agency to interview for a 00:09:29 Daniel Walsh: job. I was a mathematics major and I went in for the job and they they wanted some of the new math a lot better than I did. But anyways, I went in and interviewed for three days. started out the I first time I was ever on an airplane and I flew down to Fort me and they told me to take a limo to the Fort me and I was, you know, 21 year old kid. I didn't know what a hell of a limo was. I walked out and there's basically a a big black car sitting out there. I said, "Well, that's what they must mean." And uh so I took that and that probably was the first black mark on my uh when I handed in something that was twice as expensive because I didn't take an airport. Um and uh so I didn't get the job months later they didn't give me the job. So eventually when I started working Linux I got my payback you know so now you're gonna have to deal with me for the next 20 years. 00:10:23 Daniel Walsh: So but I I just liked it. I just enjoyed it. Um and um you know I was able to grow and move into other areas. Gunnar Hellekson: Yes. Daniel Walsh: Obviously I you know did SE Linux for 10 years and then did containers and SE Linux and then eventually did operating systems with containers and SE Linux and now AI and containers and mess Linux. So it all it all flows together. Gunnar Hellekson: Yeah. Yeah. Yeah. Daniel Walsh: But the the funny thing about uh Suzie going to SC Linux now, I think a lot of that has to do with containers because what one of the cool things that I actually came up with back I think it was 2008 I filed a patent for uh MCS security and what what we were looking at that time was we had virtual machines and so SC Linux was all about type enforcement and basically saying a database has Gunnar Hellekson: Okay. Daniel Walsh: to you know be a database and um a web service has to be a web service and the process but if you're talking about a VM it's like okay you're a VM well what happens if I have you know VM can write VM data and I can read VM data well what happens if I have two VMs or three VMs and so when I we developed MCS which was multicategory security which was basically a a form 00:11:38 Daniel Walsh: of MLS multi-level security but we didn't care about the um we didn't care about the levels it just like we just a different this container has a different MCS label than this one. David Egts: Right. Gunnar Hellekson: Listen, you got two weeks. Daniel Walsh: So, it can only read stuff with the sim same labels. We patented that idea. I'm wondering if that patent is about to expire because back it's probably getting ready to expire. David Egts: Yeah. Daniel Walsh: Um yeah. Gunnar Hellekson: You just go refresh it on the on Daniel Walsh: So, eventually we use that same same basic idea for containers and everything else. And that's basically what's been protecting containers for years and years now. There's been MCS SC Linux separation and almost every exploit that's ever happened out of a container has been a file system exploit and that's where SC Linux really shines is is basically controlling you know container kind gets out of the container and says all I can read is container file T for my MCS label and if it tries to read any other container or any other file on the file system gets blocked and um so it was really really simple um And now we see SE Linux, you know, it's on Androids, now it's, you know, moving into Suz. 00:12:44 Gunnar Hellekson: Yeah. Yeah. Daniel Walsh: So, it's kind of interesting, but Well, Gunnar Hellekson: Yeah. Do you have a looking back on because you have built a lot that we just take for granted now. Um you should be super proud of that. David Egts: Mhm. Gunnar Hellekson: Um is there anything you wish you had done differently? Uh or like do you have any like oh man if I now that I've seen this for 15 years I wish I had for example making SC Linux easier to configure. Daniel Walsh: let's see. I I did I wrote all my blogs I wrote all my blogs originally on livejournal.com and that that that got bought by a Russian company. Gunnar Hellekson: I'm just Daniel Walsh: So, every so often somebody from somewhere comes in and says, "Can you move all your blogs off of livejournal.com to some other site?" And I'm like, "No, you do it." Yeah. There's probably like a thousand blogs sitting out there. Yeah. Yeah, I'll transfer them forward for you. Uh, so I I don't know and regrets. 00:13:44 Daniel Walsh: I mean, the best thing I ever did was joining Red Hat. Um, early in my career, I worked for a digital equipment corporation, which actually had a similar feel to to Red Hat. Um it was um although it it went down quickly because I never saw PCs coming but um I never understood uh but basically it was a fully open environment and uh allowed you to move around and change jobs and things like that and um so Red Hat you know had a lot of those qualities you know when I started Red Hat there was 400 employees and I don't know it's 22 to 25,000 employees now. So, so it's uh but it's still, you know, I still still think it's a great place to work and stuff. So, even though David no longer works with us, you know, but um Exactly. Gunnar Hellekson: Yeah, it's true. David Egts: I'm there in spirit. Gunnar Hellekson: That's true. David Egts: Yeah. So when you know like talking earlier about AI uh you know you often describe it as like working with an intern and you know you're not writing the code you're letting AI do it. 00:14:46 Daniel Walsh: No. David Egts: It's doing a better job but a lot of times you got to tell it what to do and everything. Do you like I remember too it's like years ago you were the individual contributor and then there was a point where it's like all of a sudden you had a team of like grad students or minions and and did did that help prepare you? Daniel Walsh: I call them youngans. I call them youngans. Although I I think the first one is Eric Paris who he was in uh he worked on he was like we hired him to work on SC Linux kernel and he used to make fun of me because I was 40 years old and few couple of months ago he contacted me and said well he's over 40 now at the age he was making fun of me for being old. Um so yeah I yeah back back a few years ago I mean probably 15 years ago Red Hat uh had opened up Berno headquarters for engineering. David Egts: It happens to the best of us. 00:15:32 David Egts: Yeah. Daniel Walsh: We started hiring engineers like Frezy and um and Chuck Republic probably because they were cheaper and um so after a while it started to get everybody at Redhead started to get really old in the engineering world and I started yelling a lot at senior management saying you know we got to hire some interns here we gota you know every everybody's in their 30s or 40s at this point and um so we started hiring interns and Um, I one I I basically signed up for any intern I could get because um I used to tell people my strategy was to be Tom Sawyer. I'm always trying to get other people to paint the fence for me. And um so one summer I think I had eight interns come in and a lot of those interns now have been at Red Hat for years. I probably had tons and tons of interns over the years and a lot of them end up being great engineers and um so that's always been a a strong point with me working with interns and we were talking before the show for the last few months I've uh as of you know working in the Ramalama project and working with AI I've started playing a lot with cursor and a lot of people like well cursor is not fast enough curs and I said just you got to treat it like an turn. 00:17:00 Daniel Walsh: You basically give it its instructions, say go off and do this. And then I go off and do something else and I come back an hour later and I look at what it did and I usually, you know, e either I'm amazed that it did better job than I would have done with the the problem or I come back and say, "No, no, you did it wrong. Do this. Do this." And um this morning I spent probably two hours with it where I worked probably 15 minutes. Um just basically changing the prompt and yelling at it and uh then it it opened up a great poll request and um but sometimes I like to go dig into it as well and look around. But um you really with playing with uh cursor and collad and things like that you really see that you know the future is all going to be uh working with these agents working with AI and um you know really getting it to do the bulk of the coding because it's just it's so much faster than than we are. 00:17:58 Daniel Walsh: Um, and you know, uh, it it also I tell it to do things that I don't know how to do. And I used to have to spend hours going into Google Docs and, you know, googling, you know, the other I I've never touched a Mac in my life. David Egts: Right. Daniel Walsh: And I said, someone said, "We need a Mac installer for the Ramal Lama project." I was like, "All right, hey, hey, Chris, go build me a Mac installer." And of course, it writes all the code. And now I I'm going to all these people that said, "Does this look good to you? I don't know what the hell it's doing and start to, you know, ask people um you know, it's good." And then sometimes they come back, it's good. Sometimes they said, "No, you got to change these things." But um one of one of the uh stories I like to tell right now with AI is uh the main complaint I hear from, you know, really good engineers is, "Oh, it doesn't write good code. It's, 00:18:50 Daniel Walsh: you know, it doesn't perform well and things like that." And I like to tell people we're all assembly programmers at this point back in 1980s when the C language was just first starting to become popular and we're all saying no I can write you know I can write two instructions better than that and and then suddenly it's like okay everybody's writing C and even going further than that it's like suddenly when Python was coming along no I'm going to write everything in C and you know it's like no Python is much faster and much better at doing all this stuff and um so every every generation and you Now, you know, so I went from assembly language to C to Python. Um, and now we're going to go to English language and it's just telling the code to go write the, you know, you know, you write your prompts. David Egts: Yep. Daniel Walsh: Um, and um, tell it to to write stuff. Uh, I like to anytime anybody opens up an issue on a project I'm on, I my first stab it is, hey, uh, hey, cursor, go go go look at this issue and build a pull build me a pull request for it. 00:19:53 Gunnar Hellekson: Yeah. Daniel Walsh: And, uh, I've been doing that for the last three months and it's doing pretty damn good job. And again, it's like an intern. You just tell you tell it, you don't have to sit there and watch it. David Egts: Yeah. Daniel Walsh: It's just just, you know, come back, check in an hour later and see what it did. David Egts: But do do you think that your leadership of interns helped you with your prompting compared to somebody that was just the the pre-inter Dan handing him AI and then all you know like being able to articulate what you want and and verbalize Daniel Walsh: Yeah, I think that that's probably helpful. You know, one one of the fears of a of of AI to me now is that we don't have, you know, I think there's a a risk that we're not going to be hiring interns or way we're going to hire interns. So, how do you get someone who's has 30 years of experience that can just look at the code and say, "Oh, no, that's the total wrong way to do it." 00:20:45 David Egts: Right. Daniel Walsh: Or, you know, or be like me in the Mac, you know, where I don't know anything about Mac and generate me. I don't know if that works, but I got to find someone else that can look at it and say, "This look okay to you?" And and may maybe eventually it'll be a different AI. You know, I'll tell Cursor to go develop it and then tell Claude to review it. And actually, that's exactly what I do now is because I the pull request I worked on this morning written by Cursor. I put it up and uh we have Google's um review agent. Um what's Symphony? No, not Symphony. What's Google's like? Gemini. Gemini does the full review on and Gemini say, "No, this code looks like crap." And so now I go back to Curse and say, "Hey, Gemini doesn't like your code. Fix it like the way Gemini wants it." Yeah. And so I can get the AIs fighting each other. So, but you know, I think that's the future. 00:21:38 Gunnar Hellekson: Yeah. Daniel Walsh: I was my kids were in college right now studying to be a computer science. say you might want to go a little more in robotics or go a little, you know, a little little less just being pure programmers are or definitely know AI find, you know, because a lot of that a lot Gunnar Hellekson: Yeah. Yeah. Daniel Walsh: of kids coming out now are going to know AI a lot hell of a lot better than, you know, old fogies like I would know it. But Gunnar Hellekson: Yeah. Well, that's the other side of uh Dave of your question is right is like um you having the experience to know if the AI is doing the right thing or not, right? That's that's one part of it. On the other hand, what's that? Yeah. Yeah. Yeah, that's right. Yeah, that's right. Yeah. Daniel Walsh: Yeah. Gunnar Hellekson: And then Mhm. Daniel Walsh: Yeah. I think Yeah. I mean that's probably the biggest thing is basically you know things like the interfaces you know what is the what is a good API what is a good CLI what is a you know human interface and that's all things I don't think AI is going to be great with that in the beginning um but um I think we talked earlier AI's it's it's at the worst it's ever going to be right now so you know every 00:22:45 David Egts: right? The the taste having having the taste, you know, like like knowing what looks good. Daniel Walsh: every three months it just takes leap leaps leaps forward and um so revolution is here Good time to retire. David Egts: Yeah. Daniel Walsh: Get out before it takes my job. Gunnar Hellekson: Um, well, yeah. Do you have any do you have any advice for anybody who's who's like somebody who's brand new out of college or somebody who's going to like make a transition into the into the work? I guess is there um Daniel Walsh: Well, most kids I ever I mean I I talked to thousands of kids over the years. I call of course I call kids anybody that's younger than my kids. So I think my youngest son's 32 now. So being a kid is anybody in the age of 32 but um I usually used to be always with students was you know you guys have a huge opportunity to contribute to open source and you know open source was sort of like AI at one point or you know the interns and uh you know it's like you open up good good project I want to work on something hey go work on this thing and you know the people from the internet would come in and I said that would used to be the best way to that you know if you came to me looking for a job and you said here's here's my GitHub uh 00:23:59 Daniel Walsh: handle you can go look at what I've done um and that's probably the best thing a student can do now again adding AI into that and showing you know you could do you can do pull requ you you David Egts: Yep. Daniel Walsh: can do 20 pull requests you can just go to any GitHub issue any GitHub source and find 30 or 40 issues use your AI to go and fix it then massage it make sure it works and open a poll request. But you better understand it. You know, if you're like me and the max thing and you don't know, open up a poll request and your only answer, I don't know. You know, that's probably not going to be very effective. But, um, but yeah, so, uh, my other thing was I always wanted hard workers. So, you could usually tell pretty quickly with an interns like who's who's a hard who's coming to your office all the time asking questions and who's, you know, really going out and and figuring it out. So I I used to say I don't need a bunch of doctorates coming in and you know wanting to do high level stuff. 00:24:58 Daniel Walsh: I I also need people to write specs and you know work work on the grunt work that you know the day-to-day stuff that needs to be done. David Egts: Right. Gunnar Hellekson: Yeah. Daniel Walsh: But, you know, I still uh you know, still listen to you have podcasts all the time and I'm always amazed at uh where Dave I think Dave finds more of these bizarre things. So, I think he he spends eight hours a day work working his real job and then eight hours a day looking for the most bizarre things on the internet that he can find. Gunnar Hellekson: Yeah. Yeah, that's right. Yeah, I'm I'm waiting for Dave to actually develop an AI to go out and seek out these music. Daniel Walsh: Yeah. Do all this stuff. David Egts: Mhm. Mhm. Gunnar Hellekson: Yeah. Then bring it back. Daniel Walsh: Yeah. because you guys what are you up to about 250 shows now or something? How many shows? David Egts: Oh, what? Uh, this is going to like 280ish. We're getting to 280. Yeah. 00:25:51 Gunnar Hellekson: Yeah. Daniel Walsh: So you probably just tell the AI to go watch all your shows at this point and uh and basically develop develop scripts for uh the next 30 shows. Gunnar Hellekson: Yeah. David Egts: Yeah. Gunnar Hellekson: Yeah. Daniel Walsh: Yeah. David Egts: Yeah. We ought to have an AI do an episode. Daniel Walsh: Yeah. I have a um you know so I wrote the pod man and action book and um about four years ago I think it came out and so the publishers on they want a second edition of it and I Gunnar Hellekson: Yeah. Daniel Walsh: know writing a book is probably the worst thing you can do in the universe the story the reason I wrote the book is I always promised my mother I'd write a book and then my mother passed away so I I I got to write that book now and uh um but just it's it's grueling and then you're working at, you know, sub McDonald's pay pay scale when you're writing it, you know. David Egts: Right. Daniel Walsh: And um so now they're after me to write a second edition because it's missing all this stuff. 00:26:46 Daniel Walsh: And so I was n then all sudden I stopped playing with AI. I said maybe I can have AI read my book, tell it what chapters to write for me in my style so it looks just like the book. David Egts: Yes. Gunnar Hellekson: Okay. Yeah. Daniel Walsh: So I might sign up to do it and see if but see if I I can do it. Gunnar Hellekson: Yeah. Daniel Walsh: you you'll have uh the next book will be written by Dan Walsh and uh you know se several AI agents Gunnar Hellekson: Well, you know, that's one of the that's one of the best things I ever did was tell, in this case, Gemini, to go through all of my old blog posts from like the last 10 years, 15 years, and then deliver a description of my written style that I could use in later prompts. So you could take those thousand live journal posts and then have it go generate a virtual Dan author and then duplicate your duplicate your voice. David Egts: Oh yeah. Daniel Walsh: There you go. Gunnar Hellekson: Easy, right? 00:27:42 Daniel Walsh: That sounds that sounds like a great idea. I going back to being in college. If I was in college right now, I would be cheating so badly. you know, all the course that you had to take. I I can remember taking courses in college that, you know, I'd go up and look at the synopsis and I went to a liberal arts college, so you had to do a lot of, you know, non-scientific stuff. And if if they had more than a five-page paper in that course, I wouldn't take it because I didn't I could couldn't write a damn. But now it' be like it's so easy. It's like, oh, and you write this like a 18-year-old would write it. Like, okay, make sure you put some spelling mistakes in there. try to hide the fact that it's AI. So, but can you can you tell the AI to to write this like it's handwritten? Gunnar Hellekson: Yeah. Well, now so we're going back to basics though because my stepdaughter is now to in her English classes. 00:28:29 Gunnar Hellekson: All the assignments have to be handwritten, right? Get all that technology out of here. Daniel Walsh: Oh, you have the AI write it all. And then they'll be they'll be sitting there writing down the AI, you know, transcribing the text into It's uh but kids don't know how to write cursive anymore. Gunnar Hellekson: Yeah, that's right. Yeah, that's right. Daniel Walsh: So I don't know they're printing things. Gunnar Hellekson: Yeah. Daniel Walsh: I don't you know it's all I'm sure it's all blue books and everything else. Gunnar Hellekson: Yeah. That's right. Daniel Walsh: It's but in some ways it's you know back back when I was in college it was oh you kind of use a calculator right? Gunnar Hellekson: Yeah. Daniel Walsh: It was like, "Oh, yeah. You Gunnar Hellekson: Yeah. David Egts: right? No spell check. Daniel Walsh: had to do you had to do all Yes. Yeah. If I didn't have spellch check, it's like I wouldn't, you know. Um, now now I get mad at spellch check when it doesn't fix it. You know, it makes me go click on it." It's 00:29:20 Daniel Walsh: like, you know, the word fix it. But, um, yeah. So, it just it's a different world, right? Gunnar Hellekson: That's great. Daniel Walsh: You know, now it's like, can you make a cogent idea? Can you you know can you can you get the the AI intern to to to write something useful? Gunnar Hellekson: Yeah. Daniel Walsh: That's where ought to be graded on you know in the future but My Gunnar Hellekson: Yeah. Yeah. Yeah. Yeah. That's right. So, what Okay. So, you're going to consult with Red Hat for a while, give you give yourself kind of a long goodbye, which is fine. Uh but then when you're at some point you're going to run out of steam and you're going to okay I'm I'm I'm done here and then what you're uh Daniel Walsh: kids my kids made me buy golf clubs this summer. So I So I took out golfing at 64 years old. David Egts: Wow. Daniel Walsh: So I'm I'm really really pathetic. But I took my wife out one day and the two of us were, you know, she was like far worse than I was. 00:30:18 Daniel Walsh: So it made me feel like I was like a professional golfer. So um I have five grandkids. Um I live at the beach. I have a beach house which is an attraction to drag the kids here, you know, so they come, you know, my my my children comes kicking and screaming because their grand my grandkids all want to be here. Gunnar Hellekson: Yeah. Daniel Walsh: Um, so I spend a lot of time with them and u be doing some volunteering stuff. Uh, hanging out at the senior center, you know, who know? Um, but I I I would love to go to uh I love to travel, but uh I also might I'm hoping to hook up with the high school and see if they have a robotics club or something like Gunnar Hellekson: Yeah. Daniel Walsh: that. I think that would be kind of cool to to to play that play those kind of games. David Egts: Oh yeah. Gunnar Hellekson: Yeah. Daniel Walsh: But if I was, you know, 18-year-old Dan going to I would be robotics. I would I love the stuff that Tesla's doing. 00:31:12 Daniel Walsh: I love the iRoot. uh your Boston Dynam I think that's the the coolest stuff and to see to see your code actually physically embody would be really really cool um even uh actually I even got fired up so I again I have Gunnar Hellekson: Yeah. Yeah. Yeah. Daniel Walsh: this little tiny project called Ramal Lama and we were talking uh Ramal Lama is a u AI project but the there's a great in the world of AI there's two uh major tools for running AI engines local locally outside the cloud. Uh the two projects are Red Hat's invested heavy heavily into VLM. Um which we're investing a ton of money and VLM is really about sort of a centralized AI where you have thousands of people talking to an AI sort of what open AI is uh you know basically building your own open AI where you have thousands of users doing it. Uh the project I'm working on uh uses another really popular internet project called Llama CPP. Um, and lambda CPP is really about running AI on your laptop or running it on an edge device, but it's more one-on-one AI where you basically um, you know, want to run random models, say on your laptop, but even I think it's going to be really important on edge devices like I have a camera. 00:32:34 David Egts: Yep. Daniel Walsh: Uh, I want the AI to notify me if, you know, UPS man is delivering something or if a guy's walking up with a gun so I can get out the back door. Um, so, um, you know, so there's thousands of projects like that where you're not going to take, you know, a 100,000 video streams and hand hand it back to the centralized brain and then have the brain come back to you and say, "Hey, uh, you know, someone has a gun at the door." Um and so I think there's going to be lots and lots of projects and you know you get into rivos and in vehicle operating systems AI running locally. Um so anyways CPP is is all about that. uh VLM has been concentrating on large um you know very powerful Nvidia GPUs and that that's that's sort of like where the money is where Llama CPP has been concentrating on low-end crap um you know GPUs everybody because David Egts: Mhm. Daniel Walsh: everybody has like oh I have a laptop that has a GPU from eight years ago that I use for my video gaming can I run you know can I run uh Deep Seek on it and um So, Llama 00:33:42 Gunnar Hellekson: Yeah. Daniel Walsh: CPP does a really good job at that stuff. Um, so anyways, my my little long story here was we had some people at Red Hat who were playing with Ramal Lama and showing it to the NASA and NASA has uh a computer on the space station that's running real 7 that has some ancient GPU on it and they came to me and said, "We want to run some AI on this." Gunnar Hellekson: Cool. Daniel Walsh: And they tried BLM. of course BLM, you know, expect something a little a little fresher to run on and um so they tried Ramal Lama project with Llama CPP and uh they got it up and running. So, uh, my my little last project is going to be running on the space station for, um, AI models, uh, locally. Gunnar Hellekson: That's awesome. Daniel Walsh: And again, you know, so, uh, I I'm trying to tell Red Hat all the time to look at, uh, you know, there's two ways here, right? There's, you know, obviously the place the way the money is is going to be the centralized AI, but a lot of a lot of our customers have really old s***** uh, systems that they want to run AI either 00:34:46 Gunnar Hellekson: Yeah. Daniel Walsh: in CPU only mode or anything else. and the Lama CPP does really good and there's different projects in Red Hat that are picking up on it. Gunnar Hellekson: Yeah. Daniel Walsh: So, um so but that's that's what I'll be fooling around with after I retire and um and then sitting on Slack and stuff and answering questions about how do I do this with SE Linux or or sometimes coming Gunnar Hellekson: J Daniel Walsh: in yelling. I I found someone wrote a a blog on how to do VLM inside of a container and they obviously had SE Linux disabled. So I got to send them a note and saying, "Hey, give me a colon Z at the end of your pod man, you know, line there to make sure that that was labeled correctly for people, you know. Yeah. Gunnar Hellekson: Do you feel like on the now that you know the the AI work uh and you we're on Llama and Llama CPP Do you feel like uh we're ever going to live in a world where people are able to do inferencing on kind of basic hardware? 00:35:44 Gunnar Hellekson: Uh I mean because that's the dream, right? Is to be able to do uh you know. Daniel Walsh: I it's it's interesting because I um right now there's you know Nvidia is the you know $4 trillion company and it's like okay but they're kind of you know everybody in the universe is trying to do something without Gunnar Hellekson: Yeah. Daniel Walsh: GPU or or build their own GPUs or and then you see uh you know uh Google's coming out with their TPUs and things like that and it's all about basically how can I run these AI models better on David Egts: Yep. Daniel Walsh: on localized hardware. Gunnar Hellekson: Right. Daniel Walsh: Um, so I always see Nvidia as like, you know, there someone is going to have a breakthrough and all a sudden Nvidia is going to be sitting out there, uh, you know, on their shrinking margins as they're not no longer the king of the hill. Um, but even, you know, with Ramal, we can run AI if you have a a decent system with a lot of memory, you can run it with CP only. 00:36:32 Gunnar Hellekson: Yeah. Daniel Walsh: It might look like a tickotate, but if you you um there's there's a project in inside of Red Hat called uh Red Hight uh I don't know if I'm talking Gunnar will probably kill this if but uh there's David Egts: Yeah. Daniel Walsh: Red Hat Insights or Red Hat Lighteed whatever the latest name is and um there's been effort you know that's basically a centralized server that you connect to the internet and you can ask any question about Red Hat you Gunnar Hellekson: Light speed now. Light. Yeah. Daniel Walsh: know it's basically has all the mods of of Red Hat and you can you know the goal is to take a log file out of uh you know some systemd service crashed and you can hand it hand the log file to the AI and say hey how do I fix you know the systemd or here's my journal or here's my kernel um crash you know how how do if I handed this to a Red Hat genius you know how would they fix it and the AI comes back and says you know set this flag you have the wrong permissions on this file are, you know, you need to update to this version of the kernel. 00:37:36 David Egts: Right. Daniel Walsh: Um, and so that's really, really cool, but a lot of our customers are isolated from the internet and they're not able to um or or they're on a system that can't reach the internet at all. Um, so wouldn't it be cool if they could just walk around with the a Mac um running AI locally and they could just, you know, copy the log file to the Mac and just ask the Mac and have the Red Hat. Gunnar Hellekson: Yeah. Daniel Walsh: And so that's a a you know a really nice use case for running AI locally on on either either purely on CPU and if the thing is like a ticker tape par you know it takes two minutes to David Egts: Uh-huh. Daniel Walsh: you know give you the answer you get the answer that's that's the thing you're after. So running AI locally in those type of situations and on again fairly old hardware um is is is you know a really really cool use case when you don't have access to the the super brain Gunnar Hellekson: Yeah. David Egts: Right. 00:38:41 David Egts: Well, the other benefit too is that you're pushing the the um uh the AI computation onto the customer system. Gunnar Hellekson: Yeah. David Egts: So your cost to provide that service goes way down because you know the the cost is being burdened by the person running the laptop or whatever their local hardware is. Daniel Walsh: All right. David Egts: And and I think that can make the that that AI as a product a lot more cost-effective. Daniel Walsh: Yeah. Gunnar Hellekson: Yeah. Daniel Walsh: I I I like to I define AI as being um your choice on AI tends to be like buying transportation, car transportation. So the most expensive car transportation is, you know, Ubering or Lift Lift driving around. And that's sort of what everybody's doing in in open AI world right now is, you know, you're paying by the token. Um, and then eventually, you know, you want to get to, well, rather than renting tokens in cloud AI, I want to uh, you know, maybe rent a GPU and I can run something like Open Shift with Red Hat David Egts: Right. Daniel Walsh: AI on top of it. 00:39:48 Daniel Walsh: And to me that's that's sort of like u renting a car for a couple of days, right? So it's you're paying for a car rental and that um and then the next level down is basically get the AI um locally on your uh you know on your hardware and that's like buying a car, right? So that now you can you know um so and each one of these obviously there's values in all of those. We all do every one of those at different times depending on our our use cases. Um, but you know, so that's the choice that people are going to need to make. And you know, what Red Hat's concentrating right now is on the second two, right? We're we're in the, you know, Herz Hertz rent a car business and we're in the uh, you know, Ford selling them help helping to sell those GPUs. Um, and you know, our our job is to make those AI, you know, depending on where you're going to run your AI work as well as possible. Gunnar Hellekson: Yeah. 00:40:42 Daniel Walsh: In my case, I'm I'm looking at when you have really, you know, crappy GPUs, I can help you out. Uh or no GPUs at all. Gunnar Hellekson: Yeah. Daniel Walsh: And if you want centralized AI, then, you know, you look at Open Shift AI and Red Hat AI is is the way we we would tell you to do that. And you know, that's the case where you have, you know, decent GPUs, you know, Nvidia or AMD GPUs. Um but I think this college changed to go back to on and maybe Gunnarin knows this. Um I as cloud vendors develop their own GPU replacements like TPU are they selling those or are they going to be lockin so that you know if I want to use Google's TPUs and if TPUs become better than Gunnar Hellekson: Yeah. Daniel Walsh: GPUs is is a company going to be able to buy those or are they just going to be only available for rent it's Right. Gunnar Hellekson: Yeah. I mean, well, as long as GPUs are scarce, they're going to keep them, right? Uh because they can make way make way more money renting them than they are, you know, selling them outright. 00:41:41 Gunnar Hellekson: Um, and uh, I feel like until I mean it seems like the big constraint, I'm not a chip guy, but I understand the biggest constraint is like the manu the actual manufacturing of the chips, right? Um, and as long as there's one ASML and one TSMC and then Samsung, uh, like as long as that's true, then uh, we're always going to be constrained. Daniel Walsh: Thank you. Gunnar Hellekson: Um, now but another interesting thing that I learned recently and I don't remember where I heard this but like maybe it was Brian Stevens told me the the AI CTO at Red Hat uh that average GPU utilization is like 15%. So, we've got all this hardware, you know, people hoovering up GPUs and spending an extraordinary amount of money getting all this hardware on premise and then they can't actually get productivity out of out of them, right? Um, just because uh and I can't imagine they're like wanting for work. It's got to be there's some uh that's a software problem, right? David Egts: inefficiency. Gunnar Hellekson: Uh the software is not good enough to to maximize the the hardware investment, right? 00:42:39 David Egts: Yeah. Daniel Walsh: Yeah. Yeah, just getting getting things to to the point where they can, you know, do sharing and things like that. I mean, that's a lot of it right now is you run one workload on top of one GPU and it's like, okay, how can I get how could I get three or four workloads running? Gunnar Hellekson: Yeah. Yeah. Daniel Walsh: So, a lot of it's, you know, the stuff that was developed in the operating system many years ago. Gunnar Hellekson: Yeah. Daniel Walsh: And actually when I've gone to CBL talks in the past it's like oh we figured out memory sharing and we figured out take the stuff that was developed but they even call the kernel which I I find funny when they talk about kernels like are you talking about the real kernel or you talking about the AI kernel and but yeah a lot of the stuff that they're doing in AI right now is just taking best practices Gunnar Hellekson: Yeah. Yeah. Yeah. Daniel Walsh: of what was been done in the Linux kernel and and adding them to adding them to the way they share the GPUs. 00:43:29 Gunnar Hellekson: Yeah. Daniel Walsh: and stuff. Gunnar Hellekson: Yeah. Yeah. I mean, it's funny how complicated uh it's kind of funny how complicated the whole space is and you listen to experts talk about this for like 15 minutes and it's like, "Oh my god, I need a coffee." It's like like it's so you know esoteric like you're saying like memory sharing like distributed cluster management like these are you know relatively obscure topics in computer science are now suddenly like extremely relevant and everybody's Yeah. Daniel Walsh: I'll go and um oh one other a little anecdote I have talking about edge computing. Gunnar Hellekson: Yeah. Yeah. Daniel Walsh: So we have a demo in the computer in the AI world. There's a thing called multimodal. Uh there's multimodal and multimodal which multimodal is basically being able to easily switch from say granite to llama 4 or deepseek and then there's multimodal which is basically having an AI both being able to be promptable and to David Egts: Yep. Daniel Walsh: watch like a video stream at the same time. So, we have a really cool demo where um we were running AI locally and with the camera, your basically camera on the laptop and you walk in front of it and u it it'll say, you know, a man 00:44:29 David Egts: Okay. Daniel Walsh: wearing a red shirt just walked in front of me. Of course, one time I did it and it said an elderly man just walked in front of I said we need some real heavy security on this real improvement to the security u on this AI. David Egts: Ouch. Daniel Walsh: Um, but you know, it was again a great use case for the AI, you know, and and you could prompt it and, you know, prompt the AI, what do you see now? Or, uh, what is what does that man's shirt say? And, uh, what are the people across the room that lipre the people across the room uh, type things and stuff. Gunnar Hellekson: Yeah. Daniel Walsh: So, um, but again, it's it's all about a, you know, in this case, it's AI running locally because you're not going to centralize those streams. You know, uh, you know, at least you're not going to have 10,000 laptops sending video streams to centralized AI and and prompting it and it would probably melt the the centralized GPUs as well. Gunnar Hellekson: Yeah. Yeah. 00:45:34 David Egts: What? Daniel Walsh: You know, we've got all that data coming in. Um, so I won't miss Jiara. Gunnar Hellekson: Yeah. Yeah. That's great. Um All right. So, um what are you what are you gonna miss? Are you gonna miss any of it? Daniel Walsh: Uh, I'll miss Yeah, I'll miss actually the interns. I miss the young people. Uh, one of the reasons I I like to work with young people is I think I don't know, you know, I like to steal their energy, right? David Egts: Yeah. Daniel Walsh: Yeah. I feel like uh Voldemort or something like I want to like the life out of them. David Egts: Yeah. Energy vampire. Yeah. Daniel Walsh: It's just it's just fun to work with young enthusiastic people, right? Coming up. Um, you know, I used to love working with my kids were young, working with young kids and and you get the same kind of joy when you you get a very enthusiastic 28, you know, 20 year olds coming out of school and they're so excited and um, so I'll I'll miss that part. 00:46:36 Daniel Walsh: You know, a lot of a lot of people I've known for 20 years of Red Hat and so of course I'll I'll miss that. U, you know, the the tough part when we all went to COVID was, you know, working remote and so you start stop seeing people, right? Gunnar Hellekson: Yeah. Daniel Walsh: It used to be lunch lunch times and you know and Dave used to come to Boston area and you know take me out to dinner at night and stuff like that you know but you know it was always David Egts: Yeah. Yeah. Gunnar Hellekson: That's it. Daniel Walsh: these gettogethers and stuff. So yeah, the the camaraderie and um you know, one of the things in my career I always wanted I I really love sports and so you know back in the day you when I wanted you know darker pissed me off and I wanted to basically uh ruin their market cap you know things like that. So that competitive nature of like oh you know now it's like okay how how can we compete against cloud cloud computing right cloud computing is gonna ruin the world how do we how do we get back to open 00:47:23 Gunnar Hellekson: Yeah. Daniel Walsh: source so that type of competitive stuff is is probably so I'd say youngans and uh competitiveness of it you Gunnar Hellekson: Yeah. David Egts: Yep. Well, no, I remember Dan uh like early on uh like when you know Gunner and I would make our pilgrimages up to the Westward office and we'd walk the halls. Gunnar Hellekson: Hey. David Egts: I'd buy donuts, stick them in the break room to, you know, uh pavloanly uh uh uh re-educate people or or uh uh you know, get them going and everything. I, you know, you guys have always been so inclusive and welcoming of us of you know, where like I I felt like when I started it was like, oh, there's that guy from sales, you know, and it's like, how's he going to ruin my day? uh to I just remember, you know, going to uh you know, go in the lunchroom, holding my cafeteria tray, looking around, I see you waving your hand, wave waving me over, and all these other guys are like, "Yeah, this is Dave. He's he's works with the government and he's he's okay, you know, and and so I Yeah. 00:48:37 Daniel Walsh: That's an excellent impersonation there. David Egts: Yeah. Daniel Walsh: You didn't drop enough hours in the conversation there. David Egts: But it's hard. Daniel Walsh: Actually, I always liked working with sales. I loved working with sales and sales engineers because you they're the ones that come to you with the real problems that you know, customers are seeing, right? David Egts: Yeah. Well, and in the reality, Jack. Yeah. Daniel Walsh: So, you know, to me that was always, you know, I always like to work. Um, as I said, I like to be in front of customers. I like to talk to them, find figure out what their real problems are. David Egts: Yeah. Daniel Walsh: Um, you know, my wife, you know, back at the Red Hat Summit, she people always want to get selfies with me and my wife's always, you know, she's rolling her eyes, what the hell are you doing? And and I always say, well, they're they're not supermodels asking for selfies with me. So, uh, but I mean, those are the people at at, you know, summit and stuff like that. 00:49:25 David Egts: Yeah. Daniel Walsh: That's where you get the really good conversations. You know, people come up to you say, you want to, you know, why why don't you do this? That's a great idea. And I always tell him that, you know, send me an email about that because my my memory, you know, 30 seconds later will be what what you said is going to be evaporated. Gunnar Hellekson: Yeah. David Egts: No. Yeah. Daniel Walsh: But um so yeah, love working with outsiders because that's where, as I said, all the all the great ideas come from. Gunnar Hellekson: Yeah. I I know I'm grateful for the amount of time that you spent in the field uh back when I was a lowly sales engineer and I remember dragging all kinds of deals. customer would say multi-level security and be like oh we got to get Dan to talk to these guys and these you know these conversations I know I know you did and uh and a lot of and most of Daniel Walsh: I hated multi-level security. Gunnar Hellekson: these you know most of these opportunities had you know they had hair on them and there was it was most of them were pretty ugly and but you were super patient with them and that same intern energy also you brought to the customer conversations right just like walk them through the problem calmly explain to them why multi-level security was ridiculous this and explain them why you know the value of type enforcement. 00:50:34 Daniel Walsh: I try to tell sales people that there's 10 there's 10 of these systems in the world, right? But but the funny thing is multi-level security led to containers and it led to uh you know NS for separation and the virtual machines and things like that. Gunnar Hellekson: Yeah, that's right. David Egts: Yeah. Daniel Walsh: So it did do good things. Gunnar Hellekson: Yeah. Daniel Walsh: It just you know the whole low-level stuff. Uh but uh my proudest thing the best thing I ever did at Red Hat was when I figured out that you know SC Linux coloring book right because I I tried to explain SC Linux for like 10 years David Egts: Oh, yeah. Daniel Walsh: and as soon as I did the coloring book it became simple to explain it to people. David Egts: Yeah. Daniel Walsh: So to me that was like the the breakthrough of my career was coloring book. Gunnar Hellekson: Yeah. Daniel Walsh: Um, yeah, Moren Duffy did all the all the great artwork and that stuff and just explaining it as cats and dogs and I followed it up with a couple other ones and um I have a um a little 00:51:13 Gunnar Hellekson: Yeah. David Egts: Yeah. Gunnar Hellekson: Yeah. Shout out to Marine Duffy, right? David Egts: Yeah. Gunnar Hellekson: Yeah. Yeah. Yeah. David Egts: Yeah. Daniel Walsh: story. I have a niece who's a uh uh she's a basically daycare and living nanny and she takes care of a kid who's six years old in the West Coast. He's a bit of a savant. He's uh uh really really unbelievable at mathematics and he's in kindergarten and um so he um I met him. He was actually my niece got married and he was like the ring bearer at the wedding. So he came up the day before and he met me and he he he he thinks I'm like the this this guru and he he actually sits down and watches my videos on online. Um, but anyways, he was the only six-year-old, I believe. David Egts: He's not the only one, Dan. I'm telling Daniel Walsh: So, one day I got a text message from my uh from my niece and she says uh um this kid uh wrote out 0 01 10 11 100 and and nobody in the family knew what the hell he was writing. 00:52:30 Daniel Walsh: And she sent it to me and I said, "Oh, that's binary mathematics." And then I then I had to spend like the next, you know, basically hundred lines of of text message trying to explain to her what it is. And uh um so um he sent me she sent me a picture of him now. He has all my coloring books in his room like a up in a shrine in his corner. David Egts: Oh, that's awesome. Nice. Gunnar Hellekson: That's so cool. Daniel Walsh: So uh he actually he also went uh he went out his pie for uh for Halloween. David Egts: He's not the only one. I'm telling you. Nice. Daniel Walsh: So he's walking around with his pie. Gunnar Hellekson: Okay. Daniel Walsh: So this kid this kid would probably be like one of these super super geniuses at some point. But he's a really nice kid. Gunnar Hellekson: Yeah, that's great. David Egts: Yeah. Daniel Walsh: Yeah. David Egts: No, you want him for an intern, right? Gunnar Hellekson: Yeah. Daniel Walsh: Exactly. I'd hire him right now. 00:53:15 Daniel Walsh: I'll have him do have him do my cloud cloud co coding. David Egts: Yeah. Daniel Walsh: I'm not sure how how long you can have a kindergarten work though. It's probably you probably after about two hours of labor, you probably have to shut them down for recesses. Gunnar Hellekson: Well, great news, Dan. It's actually illegal to pay him. So, you got Yeah. Daniel Walsh: There you go. David Egts: So Dan, I guess two things. Um, one is for uh Podman 2. Who's going to play you in the movie? Daniel Walsh: Wilfred Brimley, I guess. No, Tom Cruz. David Egts: Without the mustache or with with the mustache? Daniel Walsh: Yeah. You know what? I I did have a must I I had a mustache in a bid, but it made me look like 80 years old for a while. So, I don't know who's a real old guy. David Egts: Okay. Daniel Walsh: Tom Hanks is looking rather old. Actually, I just saw um Russell Crowe in he is a movie that just came out and uh he's looking he's looking rather old in that movie. 00:54:17 David Egts: All right. Gunnar Hellekson: Yeah. David Egts: Yeah. Daniel Walsh: So, Russell Crow, he's in uh Nermberg. David Egts: Yeah. Yeah. Daniel Walsh: Just saw that the other pretty good move. Gunnar Hellekson: Yeah, that's right. David Egts: Right. Yeah. Yeah. Daniel Walsh: So well I mean it's where I have mine right now is dshow.com but I think I David Egts: All right. Then I I guess the other one is is uh you know for everybody that's listening, you know, it's who should they be setting their uh uh homepage to? What? Oh, take two. That's a 404. Daniel Walsh: was gonna say DG.org. Gunnar Hellekson: org.org. Daniel Walsh: No, the problem reason I changed I was gonna say my ask me again and we'll do it and I was gonna say it's eventually it's going to be dshow.ai, right? Gunnar Hellekson: Okay. David Egts: So Dan, where where should uh where do you have your uh uh homepage set to and where should everybody else be at setting their homepage? Daniel Walsh: Well, right now I mean obviously it's dshow.org book, but I would figure in the next few weeks, you know, once AI has reviewed all your videos, it's going to be dshow.ai at that point. 00:55:27 David Egts: Yep. Yep. AI will create the content and listen to it. Yes. Daniel Walsh: That sounds perfect. Gunnar Hellekson: Well, Dan, listen. Thank you for spending the time with us. I really appreciate it. David Egts: Yes. Gunnar Hellekson: Um, and also, thank you very much. Daniel Walsh: That's fun. Gunnar Hellekson: I think both Dave and I owe a good portion of our Red Hat careers to you. David Egts: Yes. Gunnar Hellekson: actually um for a variety of reasons. David Egts: So many good memories. Yeah. Gunnar Hellekson: Yeah. Um and uh man, I'd really appreciate it and I would say I'm gonna miss you, but uh I'm actually not going to leave you alone. Uh even after you That's right. Daniel Walsh: I'll be sucked by be like Michael Cordleon suck. David Egts: Yeah. Gunnar Hellekson: That's right. David Egts: Yeah. e either for Red Hat or for uh a future podcast episode. Gunnar Hellekson: Yeah, that's right. Daniel Walsh: Well, as my wife says, well, he's listening to those two idiots again. So, you guys are my shower companions. David Egts: Nice. Daniel Walsh: I have you playing on the on the speaker in the shower. David Egts: Hey, I can't unsee that. Daniel Walsh: It's like my wife's always make fun. David Egts: No. Daniel Walsh: I can never take a shower without I go nuts if I can't find my speaker and my cell phone. I can't I can't spend that five minutes without you listening to a podcast at 1.5 speed. So, yeah. David Egts: TMI. Gunnar Hellekson: Those are some luxurious showers, Dan. I mean, our episodes are like 30 minutes long. It's like, you're doing great. David Egts: No, not at all. Gunnar Hellekson: Yeah, thank David Egts: All right. Thanks, Dan. This editable transcript was computer generated and might contain errors. People can also change the text after it was created.