Zuckerberg’s AI Power Play, DevOps’ AI Reckoning, and Cloud Native’s Comeback

Happy Thursday, every... Oh. Happy Thursday, everyone.

I had a little false start there. I was all hopped up to do this, but- Jumping the gun again ... jumping the gun on a Thursday.

That's how eager I am. No, it was your first time. It's okay.

It's your first time, Alan. Well, yeah, I'll learn. I'll get better.

Anyway, it's Techstrong, gang. It's Thursday. It's noon.

What a great show we've got for you today, and we're going to jump right into it because we've got a lot to cover. Let me introduce you to our gang today. First of all, the man in black, Mitch Ashley, joining us.

My friend Mitch. Hey, Mitch. How are you?

Wonderful to be here. Thank you, sir. Takes one man in black to another man in black to say that, but good to have you.

That was a movie, right? Yes. Mm-hmm.

" Oh, yeah. We had the first one. And then also joining us, though, it's always a pleasure to have him on down Austin way, R2, Robert Reeves.

Good to have you, Robert. Hello, hello. Nice to see you.

Thank you. I'm more morally gray today. That's okay.

The world isn't always black and white. Sometimes it's shades of gray. Not "Fifty Shades," but...

Let's move up to Silicon Valley where we have the one and only, with his mic turned up, John Swartz. Hey, John. Yes.

How are you? Yes. I'm good.

I'm amplified. You're amplified, exactly. You've reached the electronic age.

And then back in his Adirondack hideaway, it sounds like something like Al Capone had or something, our own Mike Vizard. Hey, Mike, good to see you. Good to see you.

It's 60 degrees here, so enjoy the heat dome. Yeah. Excellent.

So, as usual, there's all kinds of stuff going on in the world, a lot of it having to do with this AI theme, that's the master theme, it's the age of AI. Wasn't there a... No, that was the age of Aquarius.

Close. But in any event, Mike, what are we leading off with today? Is it- Well, Mr.

Zuckerberg had a no less than four-outlet media blitz talking about how important it is to put AI at the edge, and we need super intelligence for all, and we're a little too maybe overly wrapped up in the cloud. And while I might conceptually agree with that, I also feel like, John, you follow this a little more closely than I do, but is this like one of these stock market stunts where I put something out there- Oh, yeah ... and I'm trying to juice the stock?

Because frankly, between you, me, and that wall, I'm finding a hard time figuring out exactly why Meta is actually relevant in this AI conversation. Yeah. They're chasing from behind.

And so Mr. Zuckerberg is pretty good at messaging to Wall Street at least. And there are two quotes that he gave, one that he gave to The New York Times, and one that he wrote in an opinion piece in The Wall Street Journal.

" Okay. " So he's basically arguing the defining question of the AI era isn't whether super intelligence emerges, but who controls it. He thinks everybody should have access to it, and not just a handful of labs.

In other words, not just a handful of labs of his competitors who are far ahead of him. He also, during his media blitz, not thinly veiled, but he took some shots at Dario Amodei and Sam Altman for their doom-laden messaging, argued open access across the labor market, et cetera. So cutting through all the BS, and again, this is Mark Zuckerberg, so take what he says with a grain of salt, is his openness argument principled or purely competitive positioning?

I think it's basically Meta trying to play catch up. And this whole defense of personal intelligence isn't strictly an ideological shift or a cynical trick, it's just platform economics. And I know Alan has some thoughts.

He did a column on this. And I'm going to pass it to Alan because I'll come back later with some thoughts I have that I think will piggyback off of what Alan says. I guess you're anticipating.

So look, there's a couple things here. Number one, I think one thing to note is Zuckerberg is acknowledging and stating pretty plainly that he does believe this path we're on right now does lead to super intelligence. He didn't say if we don't get there anything.

He said super intelligence is coming. And again, all of these AI snake oil guys are, what's his name? Sam the Sham said that we had reached the singularity last week.

This one's saying we've got the super intelligence, and it belongs to all of us. " So that's number one. Number two, look, Zuckerberg took a page out of the Chinese menu.

And I don't mean one from A, two from B, by the way. This is something we learned in law school. If you've got the facts on your side, argue the facts.

If you've got the law on your side, argue the law. If you have neither on your side, baffle them with b******t. And what the Chinese did is they didn't have the leading AI.

They were behind, but they said, "We've got something better. We've got it cheap. We'll give it to you for free.

" And they captured a good share of the market Meta's doing that. They were giving Llama away for a long time now, and they realized that Llama just wasn't up to in the top three of the models, and they gave it away. Now, to show you Zuckerberg's no fool, once, he spent about $14 billion hiring brains and talent to try to rebuild the Meta AI machine and then came out with a paid model.

I think it's Spark, right? Meta Spark. Mm-hmm.

New Spark, yeah. So that's a paid model, and of course, now it's having a tough time. It and Grok, let's face it, it's clearly a two-horse race in terms of the American models, OpenAI and Anthropic, and that's why he took a shot at Dario.

But at the end of the day, he's playing the classic underdogs game, and he's using all the underdog tools. We're open, we're cheap, we try harder, we're nicer. Those Anthropic guys, in spite of saying they're nice, are pretty mean.

These are all classic underdog kind of maneuvers as he tries to keep Meta relevant. So I get what he's saying, and I don't disagree with it, but I just don't see Meta as the vehicle for delivering it. If I'm going to put AI and super intelligence at the edge, isn't that going to be driven by, I don't know, Google, Samsung, Apple, Microsoft?

I got one word for you, OS2. OS2? Yeah.

Okay, great. Right. So if you're going to make the rational argument, we should have all used OS2.

Wow. Oh, God. But Microsoft out-marketed them.

" It's like almost like a placeholder strategy, right? Openness is the challenger strategy. It's not a permanent corporate doctrine.

I'll cede this floor, but I may not. I'll mention something later. No, John, you opened the door.

Keep going, man. Okay. All right.

So their recent shift to close- John's a real teaser, Mike ... commercially. Yeah, I know.

So when Meta's behind open sourcing, Llama provided immediate talent distribution and goodwill, but now that the CapEx costs for Meta alone, $145 billion in 2026, the calculus shifts towards monetization and moat building. And I think what they're doing is they're, as Mike said, they're trying to strain to figure out a way to become relevant in this space. I don't think they will be, but I've been wrong about them so many times before.

But I think if they actually do achieve some sort of frontier dominance, I expect his rhetoric will quickly shift from providing access for all to protecting proprietary safety standards and recouping their capital expenditure. I just think he is shifting in the wind like he always has, and this is what makes him, I think, as slippery as Sam. And he's probably met his match, at least for now, with OpenAI, to name just one.

You're right. But John, you said a word there, CapEx, and that's another important thing. Watch Meta's quarterly results.

They might be negative free cash flow because of the build-out. And look at the footnotes, because where they're sticking a lot of this CapEx spend is off books. They're calling it out.

Look, we all know it's a house of cards. The five of us and every single person that's watching this video has been through this before, and we know what happens with a bubble. So the trick is not to predict the bubble popping.

Trick is ask yourself, how can you take advantage of the bubble right now? If companies are not pricing appropriately where they're losing money, use it all. If you're getting cheap AI right now through the big two or whoever and everybody's trying to win, take advantage of that.

Customer company, take advantage of this pricing. Play them off each other. Get as much as you can and get the good deals.

Go upstack. Upstack. So the other thing to consider- With Llama 2, that- Go ahead ...

that capital investment, is it a strategy or is it a hedge? It's a hedge against, I want to build AI. Well, he might wind up becoming a data center provider or a GPU provider.

Well, that's possible, but that goes to, AI loves the term drift. I think this is strategy drift, which I think is normal for Zuck. That's kind of what he's in, a constant strategy drift mode, and he's trying to find center of where is the center where he can land that he's not going to be trying to topple Anthropic or Microsoft or Google or whoever.

So is it the edge? No, I think, and yes, we'll have glasses that'll be connected to the edge. Okay.

But that's not his next dominant business. I can't believe that. No.

So Mitch, follow this up with me for two seconds here. Why is he going this route when, as far as I can tell, Meta is, they're an ISV, just like every other ISV. So why not just build smaller models, shove them in your app, and take advantage of all the other models that are out there?

Why fight this fight? It makes no sense. Well, think about their business model.

They're primarily consumer driven. With the online presence they have with Facebook and everything that they've acquired So they're going down the IoT glasses, et cetera. Well, what does that naturally push you to?

To the edge. And what's AI going to do at the edge? So my glasses are not going to run an LLM today, not for a while, maybe ever.

But I think that it's a natural, where do you extend off of where the strengths of Meta are, as opposed to pivoting off into something very different. And you can argue whether the data center business is very different because they're an online presence, but they also aren't in the, "Let me host your applications and run your AI in our data center" either. So that's why I think the data center business is, one, not losing out on if there's an opportunity to participate because there is such a drastic need, and it never bubbles.

The other is it's a hedge, so I'm not caught with my shorts down because everybody else has consumed the AI capacity, and I don't have any. No. Wait a minute.

That's part of it, but go ahead, Robert. Well, this AI capacity thing is really... I'm not understanding what Facebook is doing with this stuff, because aren't they working on a $10 billion deal with Anthropic for Anthropic to lease AI capacity?

Isn't there-- There was some recent talk about this, right? I didn't imagine that. No, that was Groq.

That was Elon. That was Groq, yeah. This is as the AI world turns, which is my shimmy says at 2:30 today.

It's a freaking soap opera. But- Oh, you know what? There was something I mentioned- No, go ahead, Mike ...

sorry, Robert. One thing I really want to quickly mention is I think what he's trying to do, he's thinking about this in political terms, in public relations fallouts. So in political terms, he's trying to present himself and his company as a white knight of sorts, because their reputation here is they're always on the wrong side.

They're always the black knights, right? So that's number one. And also he's thinking about the midterms, he's thinking about the public discourse.

Now I'm going to mention something that's going to happen in October, which is going to be a huge black eye for him, by the way. It's coming. The Social Reckoning is the sequel, and I'm serious about this.

The sequel to The Social Network is coming out in October, and it centers around Frances Haugen, who was the whistleblower. That's going to get a ton of attention. " These guys don't want to be commoditized, but they don't want to be beholden to anyone else, any government, or especially a competitor.

They don't mind using competitors' models mixed with their own models. Much as Microsoft's doing, right? Satya's doing.

But what Zuck doesn't want to do is have-- Because Anthropic, who's his foil right now, Anthropic's being accused of, "Hey, use my model. " Then they look at what your business is, and then they go rebuild it themselves. He doesn't want Anthropic or Open or any of these guys being the source for the AI that powers his next gen Meta apps, and then that source itself winds up being tomorrow's competitor because they've got to move up stack.

Then why is he offering all that compute? Look, let's back up. All right.

The reason why Coca-Cola pays a dividend is because they can't invest any more into the business. Companies that pay dividends get to a point where they've got market leadership, and they're able to, like, "Okay, we could take this cash and invest in the business. " In essence, that's what Facebook is doing.

" It is an indication that they overbought or that they don't have any good ideas. But they've all overbought. $8 trillion of overbought.

Yes. $8 trillion. So for him to say and come out as some sort of white knight leader or something like that for the entire world, great.

That's awesome. How do you try and lead your business to use those assets and return some let's do the damn job. Well, the growth's got to be there, and clearly the growth hasn't been there with Llama and now Spark.

But Mike, we're over. Well, I want to mention one more thing. The crazy thing too now is he's presenting himself as a white knight, and Anthropic is becoming the villain in Silicon Valley, and that's something that Al and I wrote about, and we'll- Yeah ...

probably talk about it in a future episode. Jargon doesn't have a villain. Yes, we're looking at that tomorrow, actually.

But let's go on to this next topic. All right. So it's this interesting world we live in, right?

There's more code than ever being created. It's starting to clog our pipelines. We're creating more code, but we're not necessarily shipping more software faster or better, or we certainly don't know if the software's any better.

But Mitch has a new report out talking about how people are starting to reinvest in their DevOps platforms to modernize them to deal with all this code that's coming through. Mitch, you can go into details, but I think if I read the numbers correctly, we are looking at a double-digit growth rate in investments in DevOps. But honestly, I wondered if that was low because everybody has the same issue.

So why isn't everybody fixing this problem at the same rate? What's going on? Well, the future always arrives slower than we expect, right?

So I think there's the where we're going and how fast we get there are the two equations. So yes, the market's healthy, right? People are still planning and are investing in software.

They're debating about how much of that is people hire versus software versus AI. That is under flux. But what we're seeing is in the data, and this is both looking at market data, but also our own research from practitioners and buyers, et cetera, is a couple of trends.

One is Essentially, AI has now infiltrated itself throughout the software development life cycle. Doesn't mean it's as present as it is in the code development and the front end of it, but we see companies investing in releasing products and people starting to buy. They're looking for how they leverage AI, not just in their development, but in their pipelines, in their CI/CD, in their testing, in their security, all of that.

The numbers are something like 40% of people are saying AI has substantially modified their production code merged in the last 90 days. That's the early adopters. You look at the folks that are doing it at scale, it's more like 6%.

I think that's even probably high at this point of how much AI code is making it into production. So, very much we're seeing the front end, not the long tail of this happening, but you're starting to see the tail of it emerge. The other big headline is governance.

We have code going into production that doesn't have the proper governance, it doesn't have the security, it doesn't have the controls that normally an enterprise would dictate. And this is including enterprise organizations, not just small, Alan and Mitch, let's go start our startup, or let's go work on our open source with Robert projects. These are people that are running businesses, large businesses, that have AI going into production, but they're doing it in advance of having everything in place that they actually need.

I think part of that, lots of reasons as you can suspect behind it, some of it is expectations of getting value out of AI, not just at the desktop, but in the production environment. The other is, until we have some big incidents, and we're starting to have a few, where AI was the source of the problem, security issue, the downtime, whatever it is, sort of the cost of doing it is not readily apparent. That'll get clamped down, kind of like your argument you were talking before about social media.

That reckoning will be coming. But it's happening. I think that's the big message in this data, the market growth and the buying data that we see is people are investing, people are still planning to invest going forward.

They are upping their AI spend and decreasing their hiring plans, but it isn't cut everything out. We don't need developers or DevOps or platform engineers anymore. We just don't know how many we need.

All right. Alan, does this feel like a DevOps renaissance to you? What's going on in your head?

I thought DevOps was dead. Wasn't it? Yeah, that was like six years ago, right?

Didn't it already die? What about all those T-shirts and stickers? But, just wanted to say that.

org folks. But here's the deal. Of course, you're going to need more of this, right?

The real question is, are you going to need people? How many new people are you going to need, right? Because people process technology.

" Right? But Mitch, this is the AI data center riddle. I'm spending $200 billion to build the data center, and I'm going to give you 37 jobs, right?

Yep. It's also because- Are we going to spend that much money on DevOps? Is it result in more people working?

Well, think about it. This is also who moved my constraint, right? Who moved my cheese?

The constraints moved because now we have, of course, the verification debt is real, right? We don't have enough people to verify everything that AI is doing, and that's why so many things leak and drift into production. But it's not just at the verification of doing a code scan or something the development team does with maybe somebody in testing.

It's all through the pipeline. So, as we apply more and more AI to the front end, it may help alleviate some of that gap. I don't need so many people.

I'm not spending more time instead of less verifying stuff. But then it pushes farther down to the pipeline, right? It's hitting the production team.

It's hitting platform engineers or testing or ops for that matter. So, I think we'll see this bubble, kind of like you squeeze the balloon on one end, and it creates a bigger bubble at the other. And the people needs are going to shift through that whole process while the jobs are starting to evolve and change.

It seems to me, everybody I talk to, on the one hand, they may be cutting developers, and that may or may not be a good idea, but from what I read, everybody seems to be looking for more software engineers. So, that suggests that there isn't enough people to manage the tools, even if we do have AI running through the DevOps workflows. Or, Alan, are you suggesting that the AI agents are going to get smart enough that we don't need any software engineers either?

No, I'm not suggesting that. Okay. I think this is really a question of scale.

If you're going to keep a human in the loop in the CI/CD DevOps process, and you are going to 2x, 5x, 10x, pick your number, right, the amount of code that we're processing through that pipeline, at some point, thinking that you're going to keep the human in the loop at 10x the volume we are now is unrealistic, to say the least. It's insane. You're f*****g- It's not going to happen.

Excuse me. You're crazy, right? So, now you've got to move the humans up a level from human in the loop to human at the helm, which is a different level of- Focus on what's running through the factory, so to speak.

And- Can we? Oh, I'm sorry, Al. No, no.

And I don't know how well that's going to scale either. I don't know if we've even tried that. Well, I got good news for you.

We have. Okay. And we did a really good job of that about 25 years ago.

We said that if pair programming was a good thing, let's do it all the time. If testing unit tests are good, let's do it all the time. If deploying and integrating all of our code streams together is a good thing, let's do it all the time.

And we had extreme programming, and the X was capitalized because we were doing that all the time in the early 2000s, because everything was extreme. But I would like to see companies that are adopting more and more AI to start focusing on the lessons learned from 25 years ago and do test-driven development. So, let's have gates that AI must go through before we release it to production.

Let's start validating all these things. Let's treat AI like a sophomore undergrad intern. Let's give them access, let them look at the stuff, and propose things, but no way they're pushing it out.

And then give clear entrance and exit criteria for our code. These problems will fix themselves, but we're treating AI like a person, and it's not. Yeah.

It's like- That there is like a- ... you've got a manager. Think about it this way, Robert, is apply the goal-driven stuff that people are doing on the coding end.

I'm not telling AI how to do this. I'm telling what you need to do and here's your conditions of success. Put that all through the pipeline.

This is exactly what you're talking... You don't go into production until all of these things are met. Right?

These are my criteria. And I'm being simplistic to it, but essentially looping on those things until it figures out how it can meet your criteria, and then you're doing it at a higher scale. You're not doing it- Yeah ...

with a human trying to solve the problem every time. That's what it takes to press the bubble back down of where the constraint moves. And that sort of scale comes from- That was the one thing, Robert, touching what you said, that sparked me.

That 75% incident rate, that's like the predictable cost of treating AI agents like they're conventional developer tools. Yes. They're humans.

Look, humans... And okay, remember, laziness is a virtue of a software engineer. It is horrible for a leader.

And so these leaders are like, "Oh, well, this is great. " And yeah, head in the sand. Yeah.

We've got to have entrance and exit criteria. Let's take the lessons from DevOps, CI/CD, that we've done over the past 26 years, 36, all these years. " Great.

Humans are really good at that. AI, not so much. So we need to give it coding conventions.

We need to give it test-driven development. We need to start playing AI off of each other and do some code pair programming. Why don't we do that?

I want to see some calculations that go something like this, though. If there's an incident or some software needs to get rolled back, can we keep score on that relative to what we consider the ROI investment in DevOps tools in the age of AI? Because it seems to me everybody I talk to is so obsessed with measuring how fast they do something, and if there's some sort of inconvenient issue that happens in the back end, we just ignore that because, well, that'll take the numbers down the wrong way.

Mitch, what's going on? Well, you're right. That's why I think productivity is a false flag.

It's the wrong thing to be pursuing because, A, nobody knows how to measure it, and if they think they do, nobody else agrees with them. So productivity isn't the output. It's either capacity of how much work we can take on and successfully deliver it.

It's also to the degree that we are now performing more functions, whether it's through the pipeline or growing or doing more products or doing more services across the business. That, I think, is the real end measure of whether AI is making a difference. Because figuring out whether Robert or I or anybody here is more productive using AI, we can tell the stories about, "Hey, I wrote that document in half the time.

" Yeah, sure. But that's a hard thing to measure when everybody's productivity is different. So I think you have to focus on what the outcomes of what AI is producing.

That's the ultimate irony, though, is the productivity was the overwhelming AI messaging from these executives when they were pushing this, right? That's absolutely right. Yeah.

I mean, it's like 30%. You mean they picked the wrong thing to promote, like let go of all your people? How could that happen?

Well, our cloud providers also back in the day marketed cost. Mm-hmm. It's going to be cheaper in the cloud.

Ask any CFO if it is. All right. You know what?

They weren't lying. They were working with the facts they had at the time. Yeah.

I know. They were mistaken. You're such an apologist.

They would never mislead, Alan. I'm going to take the Fifth. Every road is just as uneven.

That's the problem. All right. All right, Mike, we're ready to go to C block?

All right, let's go to C block because this is one that has me scratching my head for a while now, and Alan has a post up on Cloud Native now Talking about how cloud-native technologies are the foundation for all these AI workloads. And we've heard this story before where especially the CNCF is touting the fact that Kubernetes is going to be the ideal platform for running AI inference engines. And I look at that, and I nod my head because I'm like, I get it.

IT people don't want to have an extra platform. They already have Kubernetes, so they just want to throw in a new workload on that platform. However, Robert, I have to also ask myself, Kubernetes itself is not yet optimized for those AI workloads, and just because I have a platform doesn't make it the right answer.

So, and I also frankly don't feel like the Kubernetes community's moving too fast here on making that platform optimized for AI workloads either. So, are they going to take too long and miss the opportunity, and something else is going to show up? If it does, if something else shows up, great.

Let's celebrate it. No, seriously. Look, the idea that we have to, as a community, do something with Kubernetes to help them out.

They'll figure it out. If they're not moving fast enough, companies can join. If a company, like a major bank, is all in on Kubernetes for their development and they're trying to figure out where to run AI workloads, Kubernetes is good enough.

It's not optimized. I guarantee you, the top four banks in the US, if they're using Kubernetes for AI workloads, they're going to work on optimization. Whether they keep it for themselves and give it back to the community, that's another issue.

So look, yes, personally, I just had to wait how many years? Five years, six years, to get a consensus around persistent storage in Kubernetes. It does take time, Mike.

It does. And if individuals, companies that are relying on this want it to move faster, they can help out. But if this was behind in proprietary closed source, they couldn't even do that.

They would just have to negotiate with their vendor. Here, at least they have a way out. Let's never underestimate the power of good enough, and inherency, the fact that it's already there.

Yeah. " Just like life finds a way, so does software. Software will find a way.

Oh, I thought you were going to go with the other thing. I thought you were going to say you were so worried about if you could do it. There's that, too.

Just be really simple about it. You're in a situation, you're saying, "Hey, we've got this AI. We need to be able to run it.

Matter of fact, we want to be able to scale it. " Right? We've got this mindset of Kubernetes as the container platform.

No, it's not the container platform anymore. It's the workload platform across tons of different workloads today, not just microservices and service mesh and the great things from the original cloud-native model. It is the workload.

It's the Linux layer above Linux, right? It's the workload layer that we run on everything. You're going to run AI on it until something better comes along.

In his basement or in his garage, Robert's right now working on the next replacement for Kubernetes, I'm sure. Because I haven't seen him for a long time. No, I'm too lazy.

So he's been working on something. Now, I made my first trip to Rochester not too long ago, and I was introduced to this delicacy called the garbage plate. Oh.

" Well, is Kubernetes- Ooh, that's going to sound special ... going to become a garbage plate? Because it's starting to sound that way.

No, it's not a garbage plate, but it's versatile. And let's be clear, it's not just Kubernetes. Mm-hmm.

When we talk about the cloud-native stack, you're talking about observability, which is another key piece of this AI- Absolutely ... infrastructure application, AI-powered applications. And to really, really understand, Mike, I think you're playing glib with this.

Right? That this is not your grandma's Kubernetes and Docker. This is the platform of choice for building scalable applications.

And whether those applications are AI applications, IA applications, or any applications, right now, it's the most scalable platform we have for applications. And so this is more than a cloud-native issue. This is also platform engineering.

AI, this is the platform for DevOps, for SRE, for data, and for AI. That's what this has become. And Mitch, you are 100% right.

There's all of these pieces. What it is, is we've spent this Kubernetes... I don't even want to call it the Kubernetes platform.

The cloud-native platform has been under development now for damn near 15 years. And over that 15 years, it's not as clunky as it was when I saw it in Austin at DockerCon all those years ago. Right?

Which was a great party. It was a great party, but it was a s****y Kubernetes. But it's come a long way.

Give it its due. Here's what I would argue for, and maybe I'm wrong, but I'm going to toss it out there anyway. " Because right now it feels like we're just waiting on, there's three updates a year, and we'll eventually get there, and we'll have the right platform, and I'll set my watch by 2029, and things will be awesome.

Next one will suck. Yeah. But no, but I think you're not giving open source its due.

Robert said that it's very hard to get the official Kubernetes apparatus to update things. But the beautiful thing of- Everybody had different opinions. Right And there needed to be, they needed to coalesce.

Consensus. Yeah. And everybody doesn't get everything they want.

But the fact of the matter is, running underneath that are a million Petri dishes with a lot of people experimenting and sciencing the s**t out of this. To quote a movie. And that's what's going on underneath the surface here.

You are seeing... And I can't wait to get to KubeCon in Salt Lake City. Right?

And see some of the experiments that are taking place in real-time. Mitch, sounds like someone wants to go with me to Salt Lake. He got excited when you said Salt Lake.

It is a dog-friendly city. I don't know what it was. Yeah.

Or he's a Matt Damon fan. One or the other, right? Somebody's listening to what you's saying and wants to go.

That's what I got. Right. Because I think- He's got science the s**t out of it.

Isn't that exciting? when we get to Salt Lake, we're going to start seeing these million Petri dishes of what people are experimenting- Oh ... with because everyone's using AI.

Everyone's pushing, and that's the beauty of open. Everyone gets to experiment. It's one of the best show and tell experiences ever.

Yeah. I love it. I love seeing weird stuff there.

This is the age of AI, when we're supposed to be accelerating the crap out of innovation, and I still keep hearing you guys say, "Yep, we'll use that same old process that we've been using for the last decade to keep building out platforms and other-" It works, man. "... " You are free to come up with something else.

Yeah, Mike. And if you do, I'm all with you, buddy. But in the absence- Yeah, let's go, Mike.

Let's Pepsi challenge this. You go- Yeah ... and we'll race.

And you do it your way, I'll do it my way. Let's go, man. Come on.

All right. Well, I will- It'll be awesome. We will get together a smaller number of people in a committee and accelerate that versus waiting for everybody to vote for every little stupid thing.

God, I get it's a democracy, but time's of the essence here. Oh, no. It is a meritocracy.

It is based on your contributions. Your voice in open source is based on how awesome the stuff you're providing is. Uh-huh.

Whether that's in documentation or code or- Or cash ... or cash. All right.

Cash can be awesome. Yeah. But remember, there's all those folks that are treated like gods at CNCF that are doing great contributions.

People know this. It is a meritocracy. It is not a democracy.

It- And all of these meetings where they decide, guess what? They're open. We can all go to them.

Yeah. And we can all speak. Look, if something better were out there, we'd see it.

Software finds a way. Yep. Oh, nice callback.

Good job. Good job, Mitch. It's the tie-in.

Way to put a bow on it. Yeah. The other question I would have, though, Mitch, and you're closer to this than I am, but to what degree do developers care about Kubernetes anymore?

Because I might just be writing to the, let's take the NVIDIA NIM framework, and I just write to that. And if there's a new platform that comes underneath NIM, great. Let's face it, Kubernetes is baked into the infrastructure.

The developers don't really care about it because they're containerizing their workloads already. That's part of the packaging and shipping through the pipeline. I think we forget that Kubernetes isn't a standalone unique thing.

It's built into so many different products. It's built into OpenShift. It's a major part of what now OpenShift supports, just to pick one.

You say Tanzu from VMware, you could argue about, I think Rancher- Rancher ... Rancher. Tanzu, so.

All these platforms have Kubernetes baked in, that may or not be at the surface level where you actually see it. So I think that's another thing that's been pushed down in the stack, not because of irrelevance, just because it's fundamentally running the workloads. It's done things like scaling and adding security and the things that you need to be able to run it at an enterprise level, which isn't easy still, by the way.

So I think it's just a non-factor. Use it. It's there.

It's sort of like, are you going to replace the muffler system in your guitar? In your guitar. In your car because you're thinking you might find a better one?

No, you're going to leave it alone till you really need to do something about it. Unless I go and buy a new car, right? There you go.

Unless I want to have a hot rod, then I got a reason. All right. Guys, we're about out of time.

Mike, you got to get busy. You got a new platform to build. There you go.

I got- See you at the race, Mike ... it's going to be called the Giddyup Committee. The Giddyup Committee.

In the meantime, we'll all be working on something that works now. I thought it was going to be called Are We There Yet? Are We There Yet?

Yeah. That's nice. Normal.

Can you hear me? Can you hear me now? Anyway, look, what a great show.

Robert, always, we got you excited today. I think I got to get you a T-shirt with a big X on it. I didn't realize you were an extreme programming dude.

Dude, I'm a- But you lit up. You lit up with that ... laziness, impatience, and hubris.

Whatever works and gets me done with the job so I can do what I really want to do A-OK. Right. It's great seeing you on here, Mitch, John- Thank you ...

Mike. Thank you. Hey, as always, we're here Monday to Friday, noon every day Eastern Time, live on the Techstrong network.

tv, Techstrong TV YouTube channel, Techstrong TV OTT channel for any screen you like to watch on. We've got Techstrong TV running on our sites almost all day and night. You can check out some more great content there because we do a lot of video programming.

We will be back tomorrow, though, with more gang. I won't be here, unfortunately, but Mike and the gang will be hamming away. Unless, if Mike's not here, you know we've had a breakthrough.

We got a podcast coming up, too. Right. We got a podcast coming up preview on Black Hat, so we'll promote that.

Yes. We're going to have a Black Hat podcast preview, hopefully up by tomorrow on Still Cyber, and Shimmy Says coming at 2:30 today. And John and I will be talking about Anthropic tomorrow, so stay tuned.

There you go. We'll be channeling you, Alan. I'll channel some of your comments.

Dario. So Shimmy Says is As the AI World Turns. It's a soap opera.

What a cast of characters. Yep. Right?

You got slippery Sam. Dario, I'm in it for you. Right?

Elon Musk, if there was ever a more lovable character, having my baby. You got the chameleon who's Zuckerberg, who will change his position at the- Zuckerberg, another- ... at the drop of a hat.

Yeah. Yeah. Let's not forget the present administration.

Blake Carrington himself over there. " What, Mitch? " It basically is.

Yeah, absolutely. That's a great comparison. It's crazy.

Which is really a soap opera when you watch it. Exactly. All right.

I think this one, though, might be darker than Dark Shadows. Anyway- Wow. You like the way I went there, huh?

But- Barnabas Collins. Jesus. Barnabas Col-- exactly, John.

All right. We're out. We'll see everyone tomorrow.

Bye-bye.