Technology’s Impact on Infrastructure with Ant Weiss – KubeCon Europe 2025

This is Textron tv. Hey, everyone back here at CubeCon. Introduce my friend Ant Vice.

Now you've seen an before probably in a green suit. Yeah, yeah, probably, probably. Now he's in scales.

We got a cleft scale. A bass scale, Exactly. Some musical notes, something to do with scales Looking for the Yeah, yeah.

Basically it all starts with this. Do you know what this is? A tuning fork?

Exactly. So it goes like this. I can't hear with these things.

Yeah. Ho hoping the, the mic can catch this. I don't know if he caught it.

Did he catch it? Do it again. Okay.

Let's see. Now we got it. Okay.

Wonderful. So, once upon a time, musicians used to use these to tune their instruments to perfect scale. Now, nowadays, well first of all, uh, I do now work with a company called Perfect Scale, which does Kubernetes optimization.

But what I'm trying to say is that nowadays musicians don't use these anymore so much. The rye, Well, you know what they say The rye electronic tuners. Right.

But does It? Well, no. You can tune a fish, but you can't tune a piano or something like that.

But no, Using, using modern tools doesn't make them worse as musicians. No, no, it doesn't. Right.

They in some ways makes them better. Exactly. Exactly.

And that, that's exactly what I'm saying. You know, we have all of these new technological tools. There's a lot of scare around AI now.

Like, is it gonna put us out of our jobs? Are we gonna gen AI ourselves out of our, out of our jobs? I'm just saying we have better tools.

Mm-hmm. It's making us better engineers, not worse engineers. So, you know what I say to this?

Show me one person who lost their job to ai, Who wants, Who lost a job to ai, Not yet, not happening yet. Exactly. If anything, it creates more complexity And more jobs.

Yeah. And, and not only that, every technological innovation that I've seen in my life, and they say, oh, it's going to, you know, it's going to kill your job, take away your job. It always makes more jobs.

Yeah. We, we, we do all this technology with the hope that we are gonna work less, but, right. It more work more.

So there's a lesson to be learned there. Um, just saying hi to our friends here, but, so as I was saying, it makes us work more. Yep.

So an but tell us the clear scale story now. Yeah. It's Perfect scale.

Perfect scale, Perfect scale. Excuse me. So basically, uh, what Perfect Scale does is, uh, optimizing Kubernetes.

So Kubernetes is a great tool for workload orchestration. It allows you to manage all this complexity of, you know, running hundreds of containers on, uh, dozens of, uh, compute nodes. But then you need to allocate, uh, memory and CPU to those containers.

And, you know, in this new world, we also have GPUs and NPUs and all kinds of new, new types of processors. And, uh, you know, nobody knows exactly how much resources to give to these containers because developers, they don't have any idea what's going on in production. It's been like that 20 years ago.

It's still this, this way today. And, you know, the ops people don't know what the workloads are actually doing. So it's, it's, it's always kind of a guessing game.

What Perfect Sale does is monitor the utilization in the runtime and provide you data-driven recommendations on how to actually give those containers just enough for them to run reliably, but not too much so they don't waste money. Because at the end of the day, you know, CPN memory costs you money, especially if you're on public cloud. So, uh, that, that's basically what Perfect Cloud is about, and it's that that's there.

So if you are a Kubernetes administrator, you want to play your clusters like you are, uh, an actually excellent musician. So you want it fine tuned in general, you know, the more technology we're using, the more optimization and fine tuning starts playing a role today. They're, they're talking a whole lot about optimizing the large language models.

Yeah. Because they're not efficient. They're burning a lot of money, a lot of resources, and you want to make sure nobody, you know, burns money just because they want to generate another senseless AI slot.

So two things. Number one, a big theme I heard this year at CubeCon here in London is making Kubernetes easier. Whether it's, uh, Microsoft putting in a, a new, um, like a Windows like interface to Kubernetes.

I don't know if you saw this. No, no, haven't seen it yet. They announced, it's called, uh, I forget what it's called right now.

We did an interview with it. Other people, I don't wanna say sating, uh, Kubernetes, but putting a, a friendlier face on Kubernetes comes, It's all about user experience, basically. Yes.

And then, you know, when it comes to ai, it's how can we use AI to make Kubernetes work better? Uh, I'm not sure better, but Yeah. Easier.

You can make it easier. Yes. Basically, you know, the, the new concept is talk to your clusters.

Right. And To be there, but that's huge. You know, to me, this is like when Scotty and Star Trek says hello computer.

Exactly. It would be nice to be able to do that. Yeah.

Basically what the, what what we are looking at is hopefully, you know, Democrat democratizing computer infrastructure. Yes. So, you know, once upon a time, you need to learn a language in order to talk to machines.

Now they're slowly, it looks like they're, they're finally learning our language so that we can talk to them, you know, like we talk to a person. Absolutely. Um, where can people get more information on perfect scale?

io. Io. Yep.

Uh, you can sign up. There is a free community edition, uh, free of charge 30 days. Connect up to five clusters, start monitoring, getting recommendations, and uh, talk to us if you really want to get the best out of it.

Alright. What else is doing at the show here, my friend? Uh, well, uh, I think in general, uh, we're starting to realize that, uh, Kubernetes is, uh, becoming the foundational layer so the injuries is happening on top.

And, uh, that's been a story for a while now. Yes, it has. Yeah.

I agree. But there's nothing on the horizon to replace it either. You know, most, most of the bus in the last couple, and that's funny, like most of the booth here, there's very little gen AI in the, because you know, those boots were prepared in probably half a year ago Right.

Ahead of time. And The, but yet they all have an AI story, But it's, it's, It may not be a good story, but they got a story. Yeah.

It's very half-baked one. Yeah. But in general, you know, it's, it's learning.

But whoever is watching the, the industry, like the last couple of months saw the buzz getting to un unforeseen dimensions ab Over hype, but that's a whole nother story anyway. But that, that's the fun in our, in our Absolutely. It is.

We've been doing this, we're chasing the shiny trinkets since as long as I'm in here And more hype, the merrier and, uh, You know, and so it gets us going. Yeah. We're still, we work more.

Yeah. We're, we're at Cube Con. We'll be back in just a minute.

Ed Weiss. Perfect scale. Take care.

Bye-Bye.