Viktor 00:00:00.000 I think that if this KubeCon was in Miami, there will be couple thousand more people. 'cause kind of, uh, yeah, I wanna go to or make it Mexico move, move to Mexico. You'll have more attendance for sure, because everybody wants a free vacation paid by their company in Cancun.
Darin 00:01:21.329 Well, here we are. It is November of 2025. KubeCon us. Just finished up in beautiful, sunny, warm Atlanta, Georgia.
Whitney 00:01:31.565 You know what you're doing there, right? Yeah.
Darin 00:01:33.515 I am lying. It's what I'm doing, I believe, right? I believe that's the correct ferries. Uh, victor's here. Whitney Lee is along with us, just like normally as we do these cube con reviews. it was not warm and comfortable and sunny in Atlanta. It was like real wintertime. That normally happens in January.
Viktor 00:01:52.567 actually it was warm and sunny. It's just that it didn't last long.
Whitney 00:01:58.957 Yeah,
Viktor 00:01:59.662 Whole, I mean, I arrived on Friday, so Friday afternoon and uh, Saturday it was warm and sunny. I made a terrible mistake that I brought one of the two sweaters I have on that sunny day and I wore it, and then I was desperate. The rest of, uh, Atlanta, I had to take swags, which I never take just to stay warm.
Whitney 00:02:19.582 He was wearing branded, branded gear,
Viktor 00:02:23.992 Yeah, I scavenged over there, the ambassadors, uh, whatever, kind of gimme everything.
Whitney 00:02:32.257 But the thing is, it was colder than it ever is in Atlanta, Georgia. Like it was unseasonably cold.
Darin 00:02:37.912 It was unseasonably cold. Yes. It it's not as, it's not as bad as it's been there before I'd lived there, so I looked at it and saw it and it's like, oh, okay. This again, like I said, it's more January weather than November, but if you were to go back and take a look at the weather charts, there was this big polar semi vortex that dropped down just over Tennessee into Atlanta, and then it was nice south of Atlanta.
Viktor 00:02:59.497 What do you mean polar? You know, how far is polar circle?
Darin 00:03:03.562 Look live in the US long enough. You'll learn. Polar vortexes are a thing.
Viktor 00:03:08.497 Okay.
Whitney 00:03:09.442 Viktor might not have come had he known
Darin 00:03:12.112 Well, it shouldn't have been that way, but, okay. So me again. We'll, we'll get into the real stuff in a minute. How was it compared to Salt Lake? Salt Lake last year? Was it as bad as Salt Lake? Weather-wise? Yeah.
Viktor 00:03:23.842 salt Lake was okay. I mean, not, not Miami. Okay. But it wasn't terribly cold. I, it even snowed one day or something like that, and then it was still warmer than
Darin 00:03:34.882 Yeah. Yeah. When it gets cold in Atlanta, it's bad. I was there in the Snowmageddon of 1993 because that's when my daughter was born and we just had her a week before we walked out, held her in the snow. We had 12 inches of snow on the ground. That doesn't happen in Atlanta.
Viktor 00:03:52.162 the real problem, being cold when I'm in my hotel is perfectly fine because. Uber comes directly to the entrance. I spent half a second between exiting the Koland being in a car. Now, in the venue over there, there is quite some vogue between the, the exit and, and, and Uber, and that was unacceptable. This is a message to CNCF Fox, kind of like if Uber don't, uber cannot arrive to the, to the entrance. I'm not coming unless it's, uh, Miami or something like that.
Whitney 00:04:25.916 Victor's a princess. Who knew?
Viktor 00:04:27.926 Yeah. So what, So what, are you jealous Is, is that what's going on
Whitney 00:04:33.836 Yeah, you have a prettier tiara than I do.
Darin 00:04:39.896 Now, typically, we record these on sort of the last day, but there was some scheduling conflicts and we had to skip over it. So we're
Viktor 00:04:46.166 Sketch scheduling conflict was that Viktor got, had a, had a hangover and couldn't record. That's the scheduling conflict.
Whitney 00:04:55.256 I was debating whether to call you out. I'm glad you called yourself out.
Darin 00:04:59.441 I, I wasn't gonna call him out on it. I just wasn't. So anyway, we are actually recording this as I'm sitting right now, it's 1150. Central time on Monday the 17th, which means I'm gonna be editing like a banshee to have it out for Wednesday, but that's okay. I have the time now, just in case anybody's wondering. Yes, I did get laid off. but let's not talk about that right now because that's a sore subject. How was Cube Con? There is one thing I want to bring. So I, I'll mention it now, but let's not talk about it. Ingress. NGINX. So just remember, we're gonna get there.
Whitney 00:05:31.430 Oh yeah. I won't let you forget.
Darin 00:05:33.585 Yeah. But ignoring that, how was everything? Was it well put on? I mean, normally it is, right? It's a nice, large venue. So you're not fighting for space. I mean, how, how was the event itself?
Whitney 00:05:45.126 Uh, it was well organized. You know how to get from one place to another. It was really big. Like there were, if you went from what, there were two different buildings, and if you had two different things to see in the two different buildings, it was a long, long walk. And I'll say too, that. All the sessions I gave, well gave and attended both, like both that none of them ever had a full room and sometimes it was even less than 50%. like in Europe. Sometimes there could be lines or definitely like every room is at least 75% full. It didn't feel like that. Overall, cavernous getting from place to place, cavernous in the rooms big.
Darin 00:06:19.757 Whitney, since you've been doing more events stateside than Viktor has, does that seem normal right now for us?
Whitney 00:06:27.617 I don't think it's accurate to say I've done more events stateside. Viktor
Darin 00:06:30.947 Well, in the past year, probably in the past year, you have, even if it's smaller, smaller events you've done.
Whitney 00:06:36.808 in general, I think, uh, US events have died down and, and Europe events have picked
Darin 00:06:41.353 That's what I was trying to get to because that's what I've been seeing from other places,
Viktor 00:06:45.103 That's probably because you folks know something that we did not realize yet. Whatever that something is. I mean, usually, you know, usually tech comes from us now, China as well. So kind of we need couple of years to pick up. So if you are. If you're going down with, with, uh, CNCF, Kubernetes Cloud Native type of events, that means that, uh, we just need to get the memo.
Whitney 00:07:08.829 I talked to an organizer of a US event during KubeCon, and he was salty for sure, and he was saying the reason US events have slowed down is that Americans are too comfortable and they've gotten lazy. That's his take on it. What do you think about that?
Viktor 00:07:25.717 I think that if this KubeCon was in Miami, there will be couple thousand more people. 'cause kind of, uh, yeah, I wanna go to or make it Mexico move, move to Mexico. You'll have more attendance for sure, because everybody wants a free vacation paid by their company in Cancun.
Whitney 00:07:46.612 Next year is Salt Lake City, but the year after is New Orleans. Do you think is, is
Viktor 00:07:51.727 Oh, oh, oh, oh.
Darin 00:07:56.347 Does that mean in two years we won't be recording this until we're actually releasing it because he'll be so hungover
Whitney 00:08:02.962 the bars don't close until 4:00 AM there,
Viktor 00:08:05.752 Oh yeah. Yeah. I mean, you know, witness is a problematic one here, right? Because she actually does go to talks, right? And she has something to say about Cuon. But I can record before the Cuon, right? My story is gonna be the same anyways. Yeah,
Whitney 00:08:21.682 we'll just get some sound bites from him and you and I can do the live recording.
Darin 00:08:25.477 Right. how was the show floor this year?
Whitney 00:08:28.775 Viktor, you are on the show floor.
Viktor 00:08:30.668 Yeah. I spend most of the time on shop floor. It was typical cube. I mean, uh, my all observation really is that. I have impression I cannot prove it with numbers, is that the number of ex exhibitors who are company behind CNCA projects is going down, down, down, down, down. Right. I, I could hardly see actually, Hey, companies behind CSA projects, companies using C Ns A projects somehow. Yeah. Plenty. Right? But I'm talking about companies that are formed on top of project, you know, that pays for maintenance like Abound. Yeah. But there are others, but near matter for example. Yeah. Companies like that. I, I have impression that that's going, going down, which is a bit concerning for me.
Darin 00:09:20.842 Okay, let's step into it. So the big news that I saw at least from afar was ingress X being, they're classifying it as deprecated. Okay, great. No big deal, but being deprecated in four freaking months,
Viktor 00:09:35.735 You know, if you follow the N NGINX, I mean this, this was long time coming. So I think that this is just acknowledgement of the reality kind of that that exists for years now.
Whitney 00:09:47.593 As far as I know, there are two people who maintain that project and neither one of it does they, they both do it as volunteers. Neither one is paid to do it. Yeah.
Darin 00:09:56.568 That's the big thing, right?
Viktor 00:09:58.410 that's actually, I feel somehow related to what I was commenting before, right? If companies behind the projects that are trying to monetize on top of projects are going down.
Whitney 00:10:11.120 Uhhuh.
Viktor 00:10:11.775 Then the num, then paid maintainers who actually do keep those projects alive are going away and projects are going to die slowly or fast depending on the project. Right. that's the danger.
Whitney 00:10:26.284 I talked to one of the maintainers and he said. That 40% of Kubernetes clusters use ingress NGINX worldwide. It's just a huge number. But, but they don't have a clear path. The Gateway, API and they're
Viktor 00:10:40.234 FF five. F five is jumping in. I can tell you that right away and I know know how, I know that I got an email only two hours ago.
Whitney 00:10:52.214 F five is a company.
Viktor 00:10:54.094 F five is a company behind NGINX, so there is NGINX Ingress and Ingress NGINX, and nobody knows which one is which.
Whitney 00:11:03.049 why does F five want to pay for Ingress? NGINX, the one that's supported by CNCF.
Viktor 00:11:09.909 Well, the question is really when they chose to. Brick on opensource version. Why they didn't jump into, at that time existing Kubernetes NGINX or whatever it's called. That's a real question. Uh, somehow related with, uh, disappearance of companies, they probably could not do that while still maintaining business on top of it, right?
Whitney 00:11:37.354 Yeah.
Viktor 00:11:37.824 That, that's my wild guess.
Whitney 00:11:39.655 But my understanding is even if a company jumps in to maintain it, it's still out of the CNCF now.
Viktor 00:11:46.055 Uh, yeah. Yeah. I mean, and nobody's jumping to maintain it. Uh, I can tell you that in advance because there is only one company with interest that's a five. Uh, uh, because other, there are other rees and, but they're not engineers based, right? Uh, and F five has their own stuff going on. They have no interest in it.
Darin 00:12:04.821 I saw somebody from, uh, the company behind traffic commenting about all of this. I don't know what's gonna happen, but let's, let's circle back to, you didn't see as many, I'm gonna call 'em vendors on the show floor that are behind products.
Viktor 00:12:22.168 as maintainers. Yes. So there, there are plenty of companies that are behind products that simply use those projects without ever, contributing to them that yes,
Darin 00:12:32.119 so if we have sort of this maintainer body count going down, does that really mean for the ecosystem?
Viktor 00:12:38.949 slow and painful, that no, no. So this is all bit. Throwing random theories here, right? I, I feel that we are going towards the, the, the world in which big corporations like Microsoft, AWS, Google what not, are the companies that can pay people to maintain those projects, which is fine. Nothing against, that would mean that there are, there is a limited number of projects that will survive, right? because that, those must be the projects that they're directly. Really monetizing on Kubernetes to begin with, right? Uh, Google is going to continue doing something with, uh, Istio because, you know, it's part of their offering. but for everything else, it's gonna be tough If I'm right and I might be wrong,
Darin 00:13:28.257 But it feels like you're close to, right. That's the problem though.
Viktor 00:13:31.677 My theory is that simply the next round, uh, none of those companies are profitable. Let's start with that, right? I mean, outside of AWS or the world, none of them are profitable. And my theory, and this is not hard to guess, is that none of them is getting next round of investment until, unless they say ai, and then everybody says AI and then nobody gets investment.
Darin 00:13:52.481 Well, that was gonna be my next question. Is it because of, well, let me, it's a two part question. So first, first off, how much AI did you see feel here? And secondly, is it because of the AI investment that we are, you know, nobody's paying attention to the other stuff. That's sort of the underpinnings of everything right now.
Viktor 00:14:10.265 It is not that nobody's paying attention to other stuff. I think that a lot of people pay attention from the user perspective to other stuff. Uh, I would actually say even that, hey, if you want to, uh, have low attendance in your talk, then you say AI and you will see how much people care about it in our area.
Whitney 00:14:29.240 I think they're getting tired of it for sure.
Viktor 00:14:31.828 not only tired, but I think that AI was. Did not yet reach the adoption threshold amongst operations. I'm gonna call it, you know, SRE, DevOps. Uh, I think that it did not stick yet. Whether it'll be or not, that's a different thing, right? It's some other areas, like if you go to a marketing conference, I'm pretty sure that everybody marketing now is using ai. Or if you go to Pure developer application conference, you will see a lot of that, right? But our people are kind of. Uh, no. Right. Uh, doesn't work.
Whitney 00:15:07.183 we care about safety in stability. Yeah.
Viktor 00:15:09.848 Yeah. Yeah, because
Whitney 00:15:11.263 not,
Viktor 00:15:12.073 when Joe does it, it's safe.
Whitney 00:15:13.851 yeah,
Darin 00:15:16.623 How much was the push for AI though at the conference? Was it as big as it was last year? Because I think last year it was like, oh, it's the new bright, shiny, was it as bright and shiny still this year?
Viktor 00:15:25.766 I dunno about talks, but in the expo, uh, there was a lot of ai.
Whitney 00:15:30.191 And it was in the keynotes. It was pretty heavily featured too. But this time it was more about like AI conformance and AI safety than it was about, uh, you know, the, the magic of it. I,
Darin 00:15:43.228 Once we realized that it's not really magic, but it's just really fast, ETL, it doesn't matter anymore.
Whitney 00:15:48.482 it is just a tool. But they are putting together a new working group, an AI conformance working group that's part of Kubernetes, and they're trying to establish, rules so that your, platform can be AI platform certified or some such thing.
Darin 00:16:04.783 Oh yeah. That's not gonna go sideways at all, is it?
Viktor 00:16:08.398 Uh, that it always produces great results. Amazing. Yes.
Darin 00:16:12.283 uh, I mean, it's been one thing to have SIGs around. Okay. SIGs around core things and so, but SIGs around how people are actually making money. that definition of safety, it's like safety for Viktor in Europe is very different than safety for me and Whitney that live in Texas, right? 'cause Texas, Hey, as long as you ain't hurting anybody, you're good.
Whitney 00:16:40.331 Wild West.
Darin 00:16:41.206 It's literally the wild west still, I mean, okay. A little bit tongue in cheek, but there is a, I think a great truth in that because Europe in general is a lot more risk averse than the US
Whitney 00:16:53.501 Hmm.
Viktor 00:16:54.456 That's absolutely true. Yeah. Which can have both positive and negative. Just to be clear, uh, there are a lot of problems with that as well, but it's true. Yes.
Darin 00:17:02.931 Yeah, the positive is you guys have more people show up conferences than us don't,
Viktor 00:17:07.003 Alive.
Darin 00:17:09.208 Alive. Yes. That's, we'll keep going from there.
Whitney 00:17:14.196 Also related, tangential to ai. I heard more about Spiffy and Spire in this conference than I have in a long time. Spiffy Inspire is being kind of touted as, yeah. Yeah. They're, they're the, bell of the ball, if you will, because with Spiffy Spire, you can automate workload identity for your AI workloads, and then put, Policy in place about what it can and can't do in your system.
Darin 00:17:36.922 You have anything to say about that Viktor?
Viktor 00:17:39.007 No, no, that, that's one of the big missing pieces, or mostly missing pieces, right? Because okay, one thing is capabilities of something, right? Oh, AI can do this or that. But then when we get to enterprises, you know, there is a bunch of things that most people are really bored doing, but without which you cannot sell it, And securities is definitely the top one, right?
Darin 00:18:03.809 How far do you think we're really at before we have good solid enterprise AI workloads running on Kubernetes?
Viktor 00:18:11.326 like infer, inference.
Darin 00:18:14.022 Yeah. Like 30, 40% normally accepted. Are we still a couple years away from that?
Viktor 00:18:19.977 I think that most of it is Kubernetes, right? Kubernetes is heavily used both to train models and for inference because what, what else is there? Now agents could be a different story, right? I think that logical for agents, logical is Kubernetes. But you can argue, hey, if there is Lambda like feature in AWS and then kind of it get gets up when, when I need it and goes down. Yes. but inference training, that's Kubernetes.
Whitney 00:18:51.503 and a lot of work's already been done in terms of integrating GPUs into Kubernetes clusters and, and in terms of dynamic resource allocation, I don't know. I, I feel like it's getting
Viktor 00:19:01.713 Yeah, and you know, I feel that actually the necessity of using something like Kubernetes and Kubernetes is arguably the only thing. Only player is much bigger now than it was before. Because before, okay, so I have five servers that are right now not doing anything. Oh my, oh my. I'm not gonna go bankrupt. And uh, right now is we have the whole warehouse of the size of, uh, uh, big airport not doing anything right now. That's, that's a problem, right? You need that dynamic allocation of, okay, oh, you can do now this and that. Moving around distributing stuff.
Darin 00:19:44.603 Is it At the end of the day, we're just always gonna have servers running somewhere in the back room, and that's it.
Viktor 00:19:50.328 As servers are always there, the question is how you use servers, kind of what is the utilization of them, right? If you go back 20 years ago, you can easily come to conclusion, and I, I, I'm not exaggerating because I've seen those things that, like 20% of your server capacities used on average. Right? Really, and the goal is to get, I dunno, 80%, 90.
Darin 00:20:14.321 Let me phrase it this way and we'll see how y'all react to this. For people that don't know how to code today. They're using AI to help code. Once you learn how to code, or once you have it coded, you can run it on Kubernetes. You've got a lot of waste in between those things. If people, if enterprises, let's put it that way, if enterprises think that's gonna be their answer, we're gonna see a lot more enterprises wasting more money than they've ever wasted before, versus. What I think would work is, I'll give you my use case right now. My daughter has a project for me to do. This is not enterprise scale by any imagination. I'm working on the project. I'm doing it in Go and Rust because I need multi-platform static binaries. I don't know how to write go or rust, but it's getting me there. I can read Go and Rust, but as I'm having it write, I'm learning how to read it. I'm learning to manually edit when I have to. Key point when I have to, and then I'm initially probably gonna run it on like Google Cloud run, right? Just something like that. In theory, once I get it right, it'll probably run just fine on a single server with a, a single static binary 'cause that's all it needs. So I'm talking about if, if you think about like a funnel at the top, I'm using all this stuff, but at the end of the day, once I get on the bottom of the funnel, I've got a nice, clean, deterministic, everything else and I don't need a big pool of things. Is that what we're gonna come back full circle to that or we headed back towards, I'm gonna call it like micro computers, like just a step this other side of mainframe in that cycle.
Viktor 00:21:56.627 The question is really for what? Because the, the, I think that the attempts to have heavy workloads, AI related, are disappearing. And when I say heavy workloads, I mean training or inference. I, I think that that's going away because companies are realizing we cannot train our own models. We, we cannot do that. Right. I, I mean, we can try and then, um. Couple of hundred millions later, we'll realize we are still short. Kind of like, oh, we just put a hundred million into this. We were under, uh, we, we did not have even fraction of the budget we need. What should we put a trillion into it? Or, or billion or should we just give up? And they give up? Right? So what you will be doing as a company is potentially is go down to the agent level. Right can basically, okay, so my model models are trained somewhere else and Tropic does it. It's running somewhere else. AWS does it, or you know, Azure, whomever. And the only thing left for me is to create an agent. If I do that, and that's Google Cloud Run. You don't need anything special for that.
Darin 00:23:08.536 Okay. Sorry, I, I sent this down a really bad path. We should be talking about KubeCon. Um. But to me I was, it feels like there was some consolidation this year, not consolidation of companies, but AI wasn't as big of a buzz, or at least not in the whizzbang way. It was more standard operational, I guess is the better way to say it, right. Of how we're gonna deal with it from that perspective.
Whitney 00:23:29.037 Yeah. Securing, um, observing and then the infrastructure underneath the bigger pieces.
Darin 00:23:35.148 in that statement, it almost feels like AI has now become boring.
Whitney 00:23:40.096 I hope so.
Viktor 00:23:42.006 No, no, absolutely not. No, no, no. A hundred percent no, no, no, no, no. Far from pouring. I think that the problem is different. Uh, problem is that we had quite a few atoms in last couple of years in, let's say cloud native companies to, to kind of put it in in a box. Were doing something. And that failed miserably simply because. Okay, so just putting a wrapper around GPT, that's not something that people are going to buy, especially since most likely next photo, GPT will have that wrapper baked in. Kind of like, come on. Right. So quick hacks did not work. Thank God. Right? And. Yeah. Now companies are trying to figure out, okay, so I know that I need to do something with ai. I don't know what, because I have no skills in, in my company, I'm talking about startups now. Right? So zero people experienced with it an amazing competition. Like if you look at Cursor Canada, that company did not exist few years ago, and now it's bigger, that single company than all the startups in our landscape combined. That's how big it is. And so we are trying to figure out what to do
Whitney 00:24:58.596 bef I see tremendous, uh, success with. AI in terms of chat bots like chat, GPT or in terms of coding assistance, but I'm not seeing where it's making a big splash anywhere else. What am I missing?
Viktor 00:25:14.219 so what you're missing is that, who make those coding existence companies that is not in our landscape. Right.
Whitney 00:25:21.359 right. Yeah, exactly.
Viktor 00:25:22.439 Okay. So, and now something, something amazing happened with AI that haven't happened in many, many years, and that's that you realize that actually developers can be buyers.
Whitney 00:25:35.170 Hmm.
Viktor 00:25:35.845 Because developers for not buying anything for the long, long time, right? Kind of who was in charge of that 10 million bill to AWS, not developers, right? Who's buying this, uh, 500 KA year license for whatever, not developers, right? and AI vendors realize that developers are probably the first adopters. Right. And they are,
Whitney 00:26:00.015 Yep.
Viktor 00:26:00.805 I can sell it to them. And if I can sell to developers, why would they ever sell From the external perspective, why would they even bother trying to sell to, to you people ops? Because I actually did statistic, this morning, on average, there is anything between 10 and 30 developers for every ops person, right? So kind of, it's a, it's a minuscule market. Head wise, if you're selling something per seat, it's, it's in significant market, right? Uh, so they're not, it's, it makes sense for a company like Tropic to say, okay, so I'm going to target 90% of the company, not 10, and they're buying it before they were not buying it. So it doesn't matter who you target, they're buying it. So it needs to come from us in a way, or that kind of, once you sell it to all developers, let's think about other people, right? and they're thinking about other people. Marketing is buying it. salespeople are buying it. Just that didn't trickle down to us or it'll come from us. Uh, hypothetical us right now. The problem with hypothetical US is that if you're a small company, and I've seen that in my previous companies. Three choices say, no, I'm going to continue my business as usual because that brings me some revenue. Today
Whitney 00:27:19.477 Hmm.
Viktor 00:27:20.827 I'm going to go pull down and ignore the revenue today and go with all 50 developers I have go into this new thing that will kill me because I cannot survive a quarter without any revenue, no matter how insignificant that revenue is. And third option, which is what kills most of the companies. I'm going to go in both places. And then you cannot maintain the, the factory that is producing some revenue because you don't have any enough people. And you put insignificant people in this, in this new thing. And don't build anything tangible.
Whitney 00:27:53.182 so you're not presenting a good solution? What's what, a, is everybody doomed?
Viktor 00:27:57.781 No, it just takes time until it trickles down to us or we or somebody will. Figure it out, uh, kind of way. I, I suspect that, and this is going to put me to trouble, uh, I suspect that the innovation in this space will come from new companies, not from existing ones. That's what I suspect. Somebody who does not have a baggage from before, right? So there is no IP to protect. I'm just company to seeded. Today I'm seeded very well. I'm seeded more than CC companies because everybody puts money into AI and I'm going to do it from scratch. 'cause when you start from scratch, actually you go faster. It counterintuitive because everybody thinks, oh, I have this thing that is already giving me a jumpstart. But very often it's not.
Whitney 00:28:49.668 So basically we need to be a. help make rules between ai, which is non-deterministic, and developers, which are non-deterministic. Like, how can we prevent bad things from happening? Like maximize output, maximize good output, let's say, and minimize disasters. Does that fall to ops?
Viktor 00:29:11.673 Yeah, that falls to us creating some kind of rules. Uh, it's not all two things. First, kind of how do you do the things you need to do without messing up everything because, uh, does not care about our internal company rules and how do we teach it to do something tangible in a way that makes sense. The problem. That I suspect is that the rules that we have in a typical enterprise company, you know, the company you don't wanna work in ever, uh, are simply are not going to be to cut it. And we are going to realize a couple of years from now, it's kind of, okay, so we have this heavy process that takes three days for something to reach production
Whitney 00:30:02.103 Mm-hmm.
Viktor 00:30:03.633 and. Even if you put AI on it, it's still going to be three days because it's, those three days is actually only two hours of actual work. So kind. You cannot make it faster without changing the rules. when you say, okay, so I spent two weeks on this, I spent a month on this. Three days is not much. You can live with that, right? But now you have developers say, oh yeah, I just did it during the meeting.
Whitney 00:30:29.667 Yeah,
Viktor 00:30:31.377 So one, once, once you, once the, time that it takes non-productive time becomes actually larger than the time it actually took to, did do it. We will see trouble.
Whitney 00:30:44.681 I've heard like we shouldn't be looking for developer productivity, but de development productivity and more of like the problem of our systems being able to take in so much code now that like 30% more codes being generated, uh, we need, as ops folks need to figure out how to handle that extra input.
Viktor 00:31:05.265 correct. The problem is that very often ops cannot do anything right? Because,
Whitney 00:31:13.785 Who, does it fall to?
Viktor 00:31:15.070 because becauses are bound to rules made by somebody nobody knows When. Right. If you go to a typical enterprise, I dare you to find me ops and to be able to, that the person that will be able to explain to you why we do things this way, I dare you to find that person because it doesn't exist, right? It's simply, oh, we have to do it like this. I don't know why, and, uh, I spent my whole career making it. Work like this. I
Whitney 00:31:48.080 It.
Viktor 00:31:48.450 we do it like this, but we have to do
Whitney 00:31:50.685 written on these stone tablets
Viktor 00:31:52.830 Exactly. Exactly. It's, it's the only way you can find out why we are doing it like this is that if you bring the whole construction company to lift off the building where you're working to check the foundation because it's buried somewhere there with dead bodies. Right. Those, those reasons why we are doing it like that.
Whitney 00:32:13.170 Which takes you back around, to your point earlier, is that it's gonna have to, like real AI innovation will come from companies that are younger that don't have these heavy processes in place.
Viktor 00:32:23.850 Yeah, but true. Those are actually two separate things, right? Uh, one thing is how do we, how do you jump on a new wave
Whitney 00:32:33.960 Mm-hmm.
Viktor 00:32:34.680 without being completely stopped or slowed down by your existing ip?
Whitney 00:32:39.665 Mm-hmm.
Viktor 00:32:40.395 If you look at waves, whichever wave you want, let's say Kubernetes, right? How many companies, you know, excluding big giants like AWS, how many companies you know that actually built something tangible, useful, good, that had the pro uh, a product before Kubernetes. This is there Now. of you can answer
Whitney 00:33:04.125 Uh, somebody who had something good before Kubernetes
Viktor 00:33:07.725 and. And No, no, the still is around that I, I know examples. But you say, Hey, this amazing x like service mesh, uh, whatever is, uh, is, is a great solution for Kubernetes. And it was built by company that is not giant,
Whitney 00:33:30.025 Cloud
Viktor 00:33:30.435 AWS that, uh, cloud Foundry. That's giant. Come on.
Whitney 00:33:34.860 Oh, okay. P? Yeah, it is
Viktor 00:33:36.975 Yeah.
Whitney 00:33:37.650 know Pivotal. Yeah.
Viktor 00:33:39.060 Yeah. Maybe they, they were already VMware before that happened, so,
Whitney 00:33:43.350 Okay. I.
Viktor 00:33:44.820 right. It's always new companies emerge. New startups that go a hundred percent for this are going a hundred percent Kubernetes. Right. I'm not trying to fit my existing business into Kubernetes and going a hundred percent in Right. And you can see that all around, right? Unless it's a Microsoft and then you say, oh, I'm just gonna put thousand engineers on it, and some, some three will. Happen. Right? I I don't even know what's the subject? What did you ask, Darin?
Darin 00:34:14.385 I don't remember.
Whitney 00:34:15.300 about?
Darin 00:34:17.385 Is it Christmas time yet? I, I can't remember. I've lost track. Um, I can't think of a good company that ever did what you were trying to explain. Nobody. I mean, I, I, I, I really can't think of anything.
Viktor 00:34:27.325 Yeah.
Darin 00:34:29.280 Oh, okay. I, I think I'm sad 'cause I was hoping to hear something that's like, dang, I'm, I really wish I would've been there and I haven't heard one thing that y'all said that I wish, yeah.
Viktor 00:34:46.020 No.
Darin 00:34:46.260 don't feel like I missed anything.
Viktor 00:34:47.790 It was good. Uh, I enjoyed it to be honest. Kind of,
Whitney 00:34:50.730 Yeah.
Viktor 00:34:51.660 uh, was just as good as any other cuco, give or take. parties were just as good as parties are. Cuco are, you know, by the fact that I canceled the next day. was a typical and I enjoyed. It's great.
Darin 00:35:05.186 What do you think Amsterdam is gonna look like in, when is it March, April timeframe?
Whitney 00:35:10.631 March, late March, I think.
Viktor 00:35:12.671 Yeah.
Whitney 00:35:13.541 I think it'll be bigger than this one. And I think I really, I, I think it's at the same venue it was at before, which was a really nice venue with natural light in the expo area and red carpet.
Darin 00:35:24.199 Natural
Whitney 00:35:24.694 what you wanted to know. Yeah,
Darin 00:35:26.449 random things there. Whitney.
Whitney 00:35:28.474 that's, that's why there, it's easy to remember. I, I think Europe ones are, are generally more excited, like better to attend. And therefore more exciting and also just, um, European cities are more walkable and there's more stuff to do right by the venues. So as a rule, I prefer the European Cube coupons.
Darin 00:35:48.574 Do you think there's gonna be any major pivots technology wise, between now and that end of March for Amsterdam? Or is is Amsterdam just gonna be Atlanta light
Viktor 00:36:00.302 It'll be a plant.
Whitney 00:36:00.862 At least. It's a heavy V.
Darin 00:36:02.432 you think? Atlanta heavy.
Whitney 00:36:03.767 Yeah. It'll be more than, it'll be bigger than
Viktor 00:36:05.417 Oh yeah,
Darin 00:36:06.182 No, but as far as you know, what changes are gonna be happening in the, basically the next five months that's gonna make it feel like, okay, I really should have been in Amsterdam.
Whitney 00:36:17.115 Victor's, predicting a slow and painful death of the CNCF ecosystem. So I think according to Viktor, that's what we have to look forward to.
Viktor 00:36:27.080 on slow. So you don't notice this,
Whitney 00:36:31.635 One project, two projects, fight in the dust, but then 10 more creeping up.
Darin 00:36:37.305 Yeah, I, I, I don't know. It's, uh, so to me, uh, alright, you've stuck your neck out. I'm gonna stick my neck out. Uh, it is now, no, November 17th, 2025 By November 17th, 2035, 10 years. C NCF F will no longer exist.
Viktor 00:36:56.707 It'll exist, but it'll not be such a huge and profitable organization. It'll be like Linux like. Linux still exists and Linux Foundation still exists and we all use it and we all need it, but we are not going necessarily in tens of thousands of people to conferences. Right. ' cause it's simply there. It's it's boring and that's fine. Boring tech is is perfectly fine. It just, it's not good for, uh, startups. You don't innovate on kernel kind of thing. You're not seeing new operating systems popping up. Oh, in this conference, we just got 20 new operating systems
Darin 00:37:35.246 It just feels very, very sad to me.
Viktor 00:37:37.586 that something becomes boring. You, you retired, man,
Darin 00:37:43.091 I am pseudo retired.
Viktor 00:37:44.216 excited, excited by being boring.
Darin 00:37:47.621 I know I, I like boring, but also I need those, uh, stock market returns to still be very strong to help me carry through, carry me through my retirement. Um,
Viktor 00:37:56.205 So good news, good news. All those that I predicted will disappear, uh, are not on a stock market yet
Darin 00:38:04.515 Ah, okay, good.
Viktor 00:38:05.355 be. So we haven't invested in any of them. No
Darin 00:38:07.545 Okay, good. good. Good, good, Uh, whew. Alright, so next up is Amsterdam. Have they already, they've closed call for papers. That was like end of last month, I think end of October,
Whitney 00:38:20.790 That that was day zero call for papers. End of last month, they closed the Q con proper, uh, call for papers the month before I think.
Viktor 00:38:29.655 Which is very confusing because I have muscle memory. That tells me the first thing you do when you return from Cube Con is that then you submit
Whitney 00:38:41.490 CFPs also.
Viktor 00:38:43.755 to it.
Whitney 00:38:44.850 Yeah, and like being at Cube Con gives me ideas about things to submit, and then I would have those ideas and be like, oh. I guess not, I guess I'm not gonna do that. So I feel like maybe you're missing out on like, like we should be learning from each other first and then putting together talk ideas.
Darin 00:39:01.870 Instead of doing cult talk ideas.
Whitney 00:39:03.850 exactly.
Viktor 00:39:04.780 is very fine. Call speakers, kind of oh, oh, oh, you know, couple of gin tonics or juices, whatever you're, you're having, and, uh, you figure out the talk. That's how it works.
Whitney 00:39:18.775 I did have people tell me that at the party you were trying to be cos speakers with anyone who came within five feet of you.
Viktor 00:39:25.705 It's a blur, Whitney. It's a complete blur, bigger blur than my typical blur. It wasn't good. And, uh, I publicly apologized to everybody on that party for whatever I might or might not have done.
Darin 00:39:41.740 So one thing I've held off to the end here, somehow or another, Whitney, you as A-C-N-C-F ambassador we're all over the place. You were tagged along with a first timer, right? You guys, you weren't going to explain to me what happened there because I was, I'm just curious.
Whitney 00:40:00.325 uh, CNCF has a new marketing team and they asked if they could follow me around and make posts about what I'm doing, but then they didn't follow me around all that much. So I think it's just a couple of random posts. uh, we'll see if they use more. I like, as I was taking videos and stuff myself, I shared them with the C Ncf F internally, but they haven't posted anything. I think they didn't realize like what a stone cold weirdo they choose to, to lift up. And they're, and they're like, oh God, what have we done? And, uh, they're hoping it goes away now.
Darin 00:40:34.326 Wait, you mean we actually have to, she, we actually have somebody that's providing content to us and that means we have to work. Did I say that out loud? I'm retired. I can say what I want. Um.
Viktor 00:40:44.316 I imagine a meeting where they started with, Hey, can, can we ask Viktor? And somebody said, hell no. Go for witness. She's nice.
Whitney 00:40:53.526 Yeah, that's, I seem sweet, but I'm not gonna do anything I don't wanna do. I'm kind of, uh, persnickety.
Darin 00:41:00.156 What? No, not you. Uh, what, what do you wish you could have done that you didn't get to do?
Whitney 00:41:07.749 I don't know. I, I pet the puppies. What else? That's the cherry on top. I, there are people who I, I meant to see who I didn't see, like maybe I should have scheduled that, but like in terms of, I don't have any fomo, like I missed out on anything really. Do you Viktor.
Viktor 00:41:23.555 No, not really. I mean, I, I always approach Cube Corner. I give up on kind, I need to see person X or y kind of whatever happens, happens, kind of whomever passes the same hallway at the same time, is is whom we are going to meet?
Whitney 00:41:38.600 I did my best where I'd arrive at 7:00 AM for platform coffee and like be engaged all day until the parties at the end, and then go home and sleep and do it again. So. I did my best. No regrets.
Darin 00:41:52.822 Viktor would've had less alcoholic beverages.
Whitney 00:41:55.707 Oh, I just, yeah, do, do you wish that
Darin 00:41:59.007 I wish that for him.
Whitney 00:42:02.997 just peace out when he gets a little toasty.
Darin 00:42:08.787 We love you, Viktor. Just, just lay off a little bit. If not for yourself, for us. Come on.
Viktor 00:42:15.322 I don't remember doing anything bad.
Whitney 00:42:18.102 No
Darin 00:42:18.822 the problem. Enough said. Okay. So overall, Atlanta 2025. Thumbs up. Thumbs down.
Whitney 00:42:31.032 thumbs up.
Viktor 00:42:32.262 Thumbs up.
Darin 00:42:33.132 Thumbs up. Okay. predictions for Amsterdam. what what do you think is gonna be the, so tell me what the key theme was for Atlanta. And then tell me what you think the key theme is going to be for Amsterdam.
Whitney 00:42:48.382 Atlanta, it was settling into the boring bits of ai, making it operationalizing ai and then I'd say there was a good, a healthy dose of platform engineering in there too.
Viktor 00:42:58.501 Yeah.
Whitney 00:42:59.403 I don't see why Amsterdam would be any different.
Viktor 00:43:03.303 No, it's probably the same.
Darin 00:43:05.858 It'll just be, it'll be the second chapter in the book.
Whitney 00:43:09.063 Mm-hmm.
Darin 00:43:10.113 With, with a couple of gloss ops.
Whitney 00:43:12.363 more granular details of how to do it consistently at scale.
Darin 00:43:16.443 Yep. Okay. Well there you go. There's our review from the Atlanta 2025 Cube Con. Uh, if you got any comments, head over to the Slack workspace, look for the podcast channel. Leave your comments there. Whitney Viktor. Thank you so much.
Whitney 00:43:31.675 Thanks for having me. We hope this episode was helpful to you. If you want to discuss it or ask a question, please reach out to us, our contact information and a link to the Slack workspace or at DevOps paradox.com/contact if you subscribe through Apple Podcast. Be sure to leave us a review there that helps other people discover this podcast. Go Sign up right now at DevOps paradox.com to receive an email whenever we drop the latest episode. Thank you for listening to DevOps Paradox.