Platform Engineering Scales DevOps and the AI Dev Bubble – DevOps Chats EP11

Hey everyone, I'm Alan Shimel. This is Mitch Ashley, and you're listening to DevOps Chats. Yeah, man, DevOps chats.

We are back. So, Mitch, I don't know, we got off track here for a couple months. As you know, you this fu gig that you've undertaken where you're now VP of research for DevOps and related activities.

And of course I think it was really kind of the whole fu uh, merger acquisition that mm-hmm. I know maybe that had a lot to do with it, but we, we, we, we dropped our cadence there and we gotta get back into it, but it's good to have ya. It's good to be all with you.

It's good to be doing it. I mean, we've been doing this in, what, 2004 or so, somewhere around, yeah, we've been at the podcasting thing. So yeah, it wa it's been an interesting, we're only nine, 12 months or so.

Nine months I guess, has been since the future deal kind of popped up and we're, you know, it's all closed. All good. And we're all, yeah, I'm in this, uh, speed modified role now and so, you know, interesting times for all of us excited about the future.

Yep. I, as am I. Um, but we could talk about that on the FU and podcast.

Oh, always. Well, RUM podcast, but maybe we'll do one, but let, let's stick to DevOps for this one, if you don't mind. Absolutely.

So, Mitch, I, I can't help but feel not complacent, but there's sort of a lull in the DevOps world. Mm-hmm. Mm-hmm.

Right. And I think it's not necessarily a bad thing, it's a sign of the maturity of DevOps in that it just is. Right.

You're not, you're not fighting religious or political wars about DevOps anyway. Um, you know, people I think understand what it is. I think that the relationship between DevOps and, and let's say platform engineering, cloud native, SRE, you know, all of these, not competing necessarily, but all of these adjacent frameworks, disciplines, they're all, you know, they, there was a lot of jostling for, jostling for, for space and priority and what have you, but they seemed to have settled in a little bit waiting for, to see how AI disrupts it all.

And I think we've reached, I dunno if it's a plateau or just a point of convergence where enough enough of these things have come together. And DevOps, you know, we, for long, for a long time, we felt like DevOps is on the leading edge, you know, in adoption. And there's more people gonna kinda like cloud native was, and even cloud natives, Kubernetes, platform engineering, DevOps, even SREI think our, our mainstream, I mean, it's not like everyone everywhere is doing this, but if you look at the data, you know, I've done enough surveys, multiple surveys, now you're looking at, two thirds of organizations are in the, either they are kind of accelerating their adoption, they're in the standardization phase, or they're in what they call a mastery phase.

They feel like they're doing DevOps really well. And it's used on a majority of project, maybe all projects in some cases. Same is becoming true with Cloud native.

Same, you know, Kubernetes is underlying everything now. So a lot of stuff we've worked on in just in the past two years seemed like it was evolving and now it's evolved. We've reached the next kind of, you know, maybe we're homo sapiens now and, you know, you know, prior areas, areas of, of humanity or whatever in these technologies.

But I don't, I'm not saying we're done. You know, we've reached, we've reached Nirvana and this is it. But it is, it's a different place than when you're climbing the hill.

Yes. And I'm, I don't think this plateau we've reached is necessarily nirvana. I I think there is something to what I said earlier, which is I think people in, in some respects waiting for the other shoe to drop regarding AI and how this is gonna change everything.

Well, that we've, we've built these years. I think we're gonna talk about some of the anticipation, whether it's warranted yet or not. Um, but, but, but clearly, you know, as AI projects are trying to get them crawl out of, in some cases, get out of the pilot and prototyping stages and get into production, Matt's starting to ask the question of, of like, okay, how does this fit with our workflows are pipelines and is, are they the same?

Are they different? It's got this data thing that goes with it, you know, with the model or with, with the algorithms. Um, and they update on their own cycle.

It's not the same as just kind of the software we've been developing that's, you know, in whatever rust or Python or whatever language that we're using. Sure. There's, that's kind of the next hill to, to really take is understanding better, better how those pipelines converge or align or where do they intersect.

And getting more sophisticated of that. Because my concern, if we don't, then DevOps become the hold back for we've got, now we've got the AI assistance to production, but your software is not tying to it. You know, how to update it in a way that can work with our release cycles of models and data, et cetera.

So it's an important thing to work on. I think we're see a lot of vendors stepping up to start to address this. Yeah.

I also see things Mitch starting to evolve into what I call the app dev. And I hope it's not rebuilding the silo. Mm-hmm.

But does the app dev aspects of things, so platform engineering makes it easier for app dev, cybersecurity for the app dev cloud native for the app dev process. And then there's the, not to use the word ops, but you know, the, the deployment aspect of it and that, you know, the observability pieces of it and, and testing sort, I think is still in, in that app dev. But where in in DevOps we were really good at viewing it.

It is one ball. I think you are starting to see lines emerge here, right. Uh, and I think part of it is a reaction to the, to the shift left or a, uh, you know, I think first we've shifted left and the, and now the pendulum's swinging back a little bit from, hey, you can't, you can't just shift everything left.

There's some stuff that still belongs here on the right. Um, or shift everywhere. You know, it's interesting, I did an interview with our friend Jeff Williams, the CTO of contrast security last week.

And his, his new thing, their new thing at contrast, I think they call it a VM, uh, it, it's, it's, uh, it's vulnerability management for production. Mm. Yep.

Vulnerability management for production. Well, gee, that sounds familiar. We been, we've been to this concert.

We saw this movie. I think we helped write that movie. I think we're in the, you know, the, uh, the black and white version of the movie.

And wait a second, you are going to tell me which vulnerabilities are critical and which ones are reachable and which man I wish I would've come up with these ideas. And, and my boss Colonel can see this too. Oh my God.

Yeah. And, um, but you know, in talking to Jeff, so what they're doing is they're really taking a very abec view of vulnerability. Not maybe the vulnerability scanning that, you know, you and I were doing, it's still secure 20 years ago on infrastructure and operating system.

Yeah, yeah. It's more AppSec scanning, but in, in any event, again, you know, shifting Right. Shifting right.

So, you know, I, I, like I said, I think this is a sign of the maturity of DevOps and people sort of waiting for this next shoe to drop. Um, I know you had mentioned you, there was an article out in the Times, I believe about, you know, looking at unemployment figures that for the first time in as long as I can remember, unemployment is actually up for software developers. Yeah.

7, it went up like two points close to two point percentage points for it. Um, you know, unemployment rate, which is like, said, when did that happen? Right.

What are, what about all these commercials of you want a recession proof job go into it? Maybe not. And it's, uh, here, here's my opinion about it, is if, if you remember, we saw sort of the efficiency cuts after covid, right?

We've gotta get more efficient. So we're gonna like lop all these people off that we didn't hire that we don't need anymore. 'cause we overhired in covid times.

And that was, you know, a bit of a reaction, maybe a big overreaction In some cases. We kind of we're going through this cycle again of, oh, AI is here. We won't need the mid-tier developers, you know, according to Zuck, which I don't trust his analysis of developer skills, but he had this a whole nother topic and we won't need those people anymore.

So we're gonna lay all those people off and not hire anymore. Um, but it's more than just, of course, Facebook, it, it is this view of we're gonna need fewer developers so we can not hire as many as we had planned. So it's less hiring as opposed to always being layoffs.

And, you know, I follow this base really closely. We're not there yet. We're not there where AI's taking over developer jobs in any whole scale manner.

Maybe it's, you know, a year away maybe, but it's, you know, people are still experimenting with, with ai. I mean, I talked with a major compute platform, um, hardware manufacturer and, you know, they've gone through the, uh, ideation and experimentation to, okay, now what are the, you know, 500 things we could do with AI and development across the SDLC? What are the handful of things we're really gonna center around and were we gonna go pick a product to, you know, build around with a co-pilot or a Udacity or, or whichever circle?

Um, you know, and there we're at this early mat maturation phase. So I think we're over rotating on not needing all these IT people. At least, at least not yet.

I, I don't disagree. I, but I, I I don't know if it's solely AI that's responsible for it. Mm.

Okay. Well say more then. I think, I think part of it also is the current political, political economic climate interest rates, business climate, if you will.

You know, Mitch, I was talking to a good friend of mine, you know, him, I, I don't want to mention names here, but he is the, he was the C-T-O-C-P-O for one of the, uh, larger, uh, gaming companies. Uhhuh Uhhuh in the world. And he, I mean, he led the team that built their online gaming.

He left or he is leaving because look, they, for the last whatever it is, four or five, six years, they paid him handsomely to build out this platform. And not only paid him, but gave him virtually carte blanche to hire the software developers and team to build and operate this platform. And it's been successful by any way.

You want to measure wildly. Yeah. Yeah.

But the bottom line is that this year, the push is not necessarily to keep growing that platform or keep upgrading it with new features or, you know, to keep innovating necessarily. It's more about maintaining, it's more about, we're not gonna grow the top line as much, but we're gonna grow the bottom line instead. And so we're not gonna pay people exor, we're not gonna pay, you know, developers $300,000 a year.

We're not gonna pay exorbitant amounts, you know, to lock in the leadership with, you know, all kinds of incentivized deals that are top line or code related. You know, how much code you put out. And I, I think that's something we're seeing in corporate America, which is, you know, what we can't develop just 'cause we, everything has to be tied to the bottom line.

Everything has to be tied to revenue. And if, if that revenue doesn't make sense, we can't do it. Well, it's, it's, we, we see this over and over again, right.

It's the, the time of I'll spend whatever it takes to be, to gain my market share and be competitive in the ga in gaming industry for one. You know, let's pick that. Okay.

Well that's, you know, now it's commonplace as I'm every other ad on sports TV and every for ad on every other channel. You, you know, it, it's reached that point when there's some who argue, I think it was Paul Kuman, is that his name? The Economist?

Yeah. Yeah. From many times.

Yeah. He, uh, he thinks we're gonna hit an AI bubble, right? We're kind of headed into the next cycle of this again.

And, you know, we, we of course always overrate rotate. That's my favorite new word about it. Or rotate, um, on, on AI investment.

We saw that with deep seek, you know, pulling back a little bit. Like, wait a minute, maybe it isn't always the next biggest hammer. Maybe there's more than one way to innovate, and whether we're gonna see a bubble or not, you know, eventually some kind of bubble's gonna happen with ai.

I don't necessarily know if it's gonna be a big one or not, but that's, you know, today it's, I don't need developers tomorrow and when, where can I find them? Um, at the same time of, I don't have enough security, people still can't find the people that I need to, uh, secure all this AI stuff. Right?

So there's all these kinda successive waves of Yeah, we saw that bubble too. And the next one will happen, just like the next one will grow. Yep.

What, and you know, Mitch, I, as we were talking offline, and you mentioned an interesting twist is that while unemployment went up, like in IT general or software developers, the, the plan is that cybersecurity needs even more people than we always say. Mm-hmm. And that's gone down.

Mm-hmm. So, you know, if AI is one of the driving forces, or if government, you know, regulations is one of the, another force in this, clearly the, the message is, Hey, but we need to lock down what we have. Well, it's, it's the, um, theory of constraints all over again, right?

We, we pushed the constraint to the next thing, you know, the, now it's security, right? How are we gonna secure all this? And we already didn't have enough security people, and it'll be ops next or whatever the end next kind of constraint is.

And, you know, we're following that pattern. It, it's, it makes a lot of sense if you think about it in a systemic way like that. Yeah.

No, I think you got it. Holistic, right? Um, mm-hmm.

But it, it's gonna be interesting to see how it plays out, because the good news is the market has a way of evening these things out, right? It's like Lou, it's finding its le lowest level there. Mm-hmm.

Mm-hmm. You don't have enough software being developed. They'll get more software developers.

I, I don't think, as you said, that we're anywhere near to the point where we could just turn AI loose and develop the software someday. We'll still be saying, where are we gonna find those Java developers now to go support all that Java code that was written way back when, you know, just exactly like COBOL today. We'll go through that phase too.

I, I think we will, Mitch, you were, you were out in Vegas, uh, last week, or actually two weeks ago now, and, uh, for the, uh, Dynatrace show speaking or maybe some legacy stuff, right? A lot of people don't know what to make a Dynatrace. They've been around a long time.

They've gone through some changes and evolved. I mean, there's a mainframe aspect of things, but observability and give us a report from, from the field in Vegas. What, what'd you see there?

Yeah, it was that Dynatrace perform. They had a number of, uh, of announcements and, uh, it was a very good conference about, I think it was about 2000 people's, what they said were attending. Um, the standouts for me was observability for developers.

We've talked about using observability and development and test, but it's sort of like, yeah, same tool. Somebody else using it. The number one thing developers can sniff out is a tool that wasn't developed for them, right?

That's been being repurposed for them, and that tends to get rejected. But now we're, now we're talking about, um, being able to do kind of real-time data log information, call entry information tied back into, eventually it isn't in the IDE yet. They've got their own, their own observability platform, but they will put it in the IDE to be able to say, here's, okay, got this issue from ops, I think it's in this area, pursuing it.

I can pull that data together through my lens of the world, which is my IDE as a developer, and start to really kind of analyze and do some real, real time assessment. And then you have their Davis ai, which is, you know, they've invested early on in ai some one of the companies that actually did, as opposed to claims to, and they're maturing that. At the same time.

The other standout thing for me was, um, enhancing Davis AI with Kubernetes information. Really leaning hard into Kubernetes observability, you know, is Kubernetes goes everywhere. Okay.

It's gotta be easier to run. We talk about how complex it is, and it is getting easier. Um, but it is not only because of tools, but also AI support.

And as that gets built more into tools, uh, that's also interesting. And I, I guess the third thing was a, a, a bit of a look forward on security, uh, cloud security, posture management coming out with something a little bit later this year, midyear or so in that space. So, you know, they've been around a long, a long time, or in, relatively speaking in, you know, in DevOps years.

They've been around a long time. Um, but they're continuing to innovate, I'd say. Good, good, good to 'em.

And they're, yeah, I know. I, they're taking it seriously. They're not just, just hanging around.

I agree. They, they have, they've reinvented themselves Yeah. In many ways.

And look, they've got a great solution set. Mm-hmm. I, I think they're, they're definitely a, a, a player to be reckoned with there.

Um, Mitch, I, I wanted to close things off though, with a, a heartfelt thank you to a friend of you and I, our friend Sasha Lare. Yeah. Sasha is the co-founder of CloudBees.

Of course, before that, Asha was one of the key people at JBoss. Right. But, you know, CloudBees always had a special place in my heart.

com mm-hmm. CloudBees was the second sponsor that signed on and helped us get off the ground. And you know, we've been working with CloudBees ever since.

Not as much recently. They've undergone, you know, their own set of changes. But certainly for most of those years, Sasha's been a mainstay.

Him and KK and K, of course, KK left, then he's come back, but, um, Sasha and Nas last week that he'll be stepping back from an active role. com, but for the DevOps community in general in the open source community. And, you know, I don't think he's done innovating, creating Businessing, you know, I'm looking forward to what's next for Sasha.

But wherever it is, Godspeed and good luck and we'll be watching. Yeah. Big thanks to you, Sasha.

And, you know, that was part of my entry into, into DevOps, you know, Jenkins we're early Hudson, I mean late Hudson and then early Jenkins, um, getting involved in that. So, you know, he's done a great deal for this community and I would, like you said, I bet there's more communities that will benefit from his, uh, DNA and innovation and entrepreneurship and technology. So look forward to Absolutely.

So, thanks Sasha. Alright. So Mitch, you and I of course spoke, let's, let's share it with our audience.

I think going forward for DevOps Chats, we're gonna focus on bringing in some key folks every two weeks when we do this. Um, key people in the DevOps community, whether practitioners, executives, vendors, what have you, will be lining them up. Of course, if you're listening to this and you think you'd be a good guest or you have a suggestion for a good guest for DevOps Chats, let us know.

We'll probably be doing an interview and talk as the new format going forward. Mm-hmm. Right.

So two parts every episode, and I'm looking forward to it, man. I think it'll be, it'll be great. We used to do more guests when you and I do.

Yeah. Early podcasting on Still secure after all these years and be shimmy after all these years. Um, and there'll be some opportunities if people, you know, we're doing more and more, um, kind of full episode of things where companies on a sponsored basis come in and, and get a really chance to dive in a little bit deeper in conversation with us.

So there'll be those opportunities too. So, we'll, we'll have some spotlight time where we take a little bit of time with companies, uh, or individuals coming in, and as well as being able to offer something a little more, um, you know, thorough to folks that, or interested in sponsor kind of thing. com.

Is that the best way to do that? That'll work. com.

There you go. com. And we, and we're, we'll work from there.

Where can they hear about Shimmy sits? Where's that at? Tell us about that.

Thursdays, Thursdays two 30 on LinkedIn live. You know, we took a little excerpt from that Mitch Uhhuh and put it on YouTube shorts, uh, from two weeks ago, or three weeks ago. And it's got like 15,000 views already on YouTube shorts.

It's pretty cool. Um, but yeah, I'm doing that every Thursday, always looking, if you want to join in on LinkedIn live, you could ask questions, comments. We try to put it in there.

We're just experimenting with different formats and it's fine. It keeps, I, you know, me, I, I don't have a shortage of things that I like to talk about. Yep.

I keep with the times, or suddenly you aren't with the times anymore. Well, you know, I, I'm shooting for the next Super Bowl halftime show. Is that?

Yeah. Well, you know, Hey, I, I can just see you up there now. Rapping DevOps.

You bet, Mitch. This is a great way to get back on the wagon here with DevOps chat. Looking forward to doing more.

But until next time, I'm Alan Shimmel. And this is Mitch Ashley, and you've just listened to DevOps Chats.