DevOps, AI-Generated Code Quality and Platform Engineering – DevOps Chats Podcast EP2

Welcome to DevOps Chats, the industry leading podcast for DevOps, digital transformation, cloud native, and cybersecurity. If it's happening, it's happening on DevOps chats. Hey everyone, this is Alan Shimmel and Mitchell Ashley, and you're listening to DevOps Chats.

Yeah, baby. Yep. So Mitchell is the, uh, second episode for 2024 in this relaunch of DevOps Chat that we're returning to after all these years.

Mm-Hmm. I hope the audience, if you haven't heard our, uh, previous episode, it is available on your favorite, uh, podcast, uh, platforms, whether it's Apple, Spotify, or, or what have you. Best, by the way, I titled the Origin Story of Security Podcasting.

Yeah. We got started. It was sort of an origin story.

You're right. At least that was the front end of it. Yeah.

You know, it's a 20 year, uh, new thing, and we're still working to get it right. Yeah. We're figuring it out.

Hopefully this, this week's better. We hope you'll enjoy. Sure.

com, which of course is a flagship site run by us here at Techstrong. Um, I'm Alan Shimmel, CEO of Techstrong, and Mitchell, CEO. Yeah.

Mitchell, CTO and, and Head of research. Head of research, GM Research Principle Analyst. So we feel like we have a good handle on the market and the space, and hopefully you, you get something out of this, this, um, for this week's show, we're gonna focus on a couple of things.

Mitchell and I are in person together. We're actually up in freezing Boston, where it's like 20 degrees, balmy, balmy, 20 degrees for me, not for me. And, um, we're up here, you know, meeting, meeting with a bunch of clients, uh, sponsors, friends in the industry and, and doing a little fact finding.

And, you know, it's interesting, Mitch, just in a couple days, you, you start seeing sort of trends emerge, right? Yeah, definitely. And, you know, for me it's, look, the biggest trend I think that's most relevant to our audiences.

We we're seeing layoffs. Yeah, definitely. And budget cuts, you know, everything we've been hearing, reorganizations, reorgs, uh, you know, with the start of the new year, people are implementing kind of these new plans.

And, um, while the economy in general seems to be strong, all kinds of new jobs being done, and unemployment remains low, it, it's tough times. And, and for our friends in the tech space and us, all of people interest rates still being high. And Yeah, of course that's hit mostly the, for sure the startup world.

Well, it Right, it's hitting the VCs. Yeah. And that stuff flows downhill.

So it goes from the VCs down to the startups, and so they're cutting. But, you know, an interesting thing, I was talking to a, a friend of ours, Mitchell versus a friend of mine Okay. Um, to get all mobster on you, but, um, and, and she was recently laid off, her whole team was laid off.

Mm-Hmm. And I said, how, how did they lay off the whole marketing team? And, and the kind of the skinny was, well, you know, they, the marketing team did a great job and they over delivered leads into the funnel, and now they're gonna take some time just like kind of harvesting those leads before they turn the spigots back on from marketing.

So they're saving money by laying off the marketing team. And I, I'll tell you the truth, I, I don't claim to be Warren Buffet or, or, or, you know, some management business whiz, but everything I know tells me you can't time stuff like that. You can't wait till you're out of leads to turn, to go higher up and train up and start up your marketing team again, that has to be kind of a, a balanced approach with a steady stream in there, because otherwise you, you know, you get the, the, the snake who swallowed the rat kind of syndrome, where once that lump is through the snake, now what?

So I, I think it's a mistake. Right? Well, well, I've learned this lesson a couple times, not in marketing, but the first time was early in my career, uh, it was with actually with EDS and they decided to cut, stop doing recruiting 'cause they had enough people.

Mm-Hmm. And this goes into their whole education program and what they, they stopped it, and about month and a half, two months into it, they said, oh, you know what? Yeah.

Um, we're in trouble. Because they realized once they stopped it to start it up isn't turn it back on. No, exactly.

It was a year and a half to get it back up to what it was to meet the people requirements. So you think about it, all right, so you're using those leads by the way, you're using those leads by the way the markets keep moving and shifting and what they want and where they're going. So you're losing track of that plus, you know, you just don't go out and buy leads, right?

Well, I guess you can with, with, we have an option, right? Well, well, that's, yeah. We could announce that we have lead gen express, so you could come in and buy leads.

There you go. That wasn't a setup from our, but yeah, you can, but normally can't just like, they're not falling off trees. You gotta go cultivate and develop that and get them into the queue and Yeah.

I mean, but even, you know, buying leads, let's say whether you come through us or you go to an agency or whatever, that's a marketing function, right? Marketing has to make sure what they're getting and why, and, and the, you know, all of the, the background on it, it's not, it's not as simple as going your faucet and turning on the water. Hmm.

That's the bottom line. Mm-Hmm. Exactly.

Yeah. It's like waiting for the hot water, but it takes a long time to come. Yep.

And I'll tell you something else, we're gonna talk about it a little later, and it's what went on at Dell this week. But, um, you know, I, I think in a lot of these tech layoffs we're seeing, especially in the startup world in venture backed companies, the marketing team, marketing teams are taking a, the brunt Mm-Hmm. Of, of these budget cuts.

Yeah. Which is pretty common, you know, tough times marketing and hr Yeah. Get hit first.

But, you know, I think an interesting thing here, Mitchell, is it's not just about cutting the budget. Yes, it's about cutting the budget, but it's not just about cutting the budget. I think there's also sort of this underlying kind of feeling of, you know, with ai Mm-Hmm.

We're gonna be able to accomplish a lot of these marketing, uh, functions without the people Mm-Hmm. Automate it with ai, we're gonna automate it with ai and that we need less people to do more with it. And, and, you know, hey, I'm all for ai, and, and that's making a bet on it.

I'm not quite sure if, if it's there for this particular use case right now. Mm-Hmm. Well, it still is an aid, right?

It's a, an aid to doing those functions. I was, I was talking to someone last night at dinner, I said, you know, the problem is AI doesn't know whether it did a good job or not. Right.

You know, it, it writes some stuff and it might be relevant, it might be accurate, it might not, might even be what you want, but that's, at least how I use it is mm-Hmm. You know, help kind of organize some thoughts, get a straw man, straw person thing. You know what, it could help polish something, it could help you start something Yeah.

When you ask it to start and polish Mm-Hmm. It, you know, to totally replace the human in the equation. I, I just don't know if we're there yet.

I I definitely got, I, and so, you know, I think there's a little bit of wishful thinking there and, and rose tinted glasses Mm-Hmm. If you are cutting back on staff because you think AI at this point is going to replace them. Yeah.

Um, I think you should be looking at AI to make your staff more productive, not to have less of them. Right. But We'll, you know, we'll, we'll see what happens there.

You know, also on the road trip Mitchell, though, it was good to see, you know, some old friends spoke about upcoming events, CubeCon in Paris, for instance, RSAC, a bunch of stuff, man. Yep. And, uh, and again, you know, common themes there, reorg and layoffs and, and, and stuff like that.

Um, it'll be interesting. We're gonna be continuing this road trip. We're gonna be going to New York and the valley, San Francisco, you know, bay area, uh, Austin.

And, and we'll see if what we're hearing here kind of is, you know, consistent across the, at least across the us we did, we did hear some optimism of things seem to be coming back. Yes. And people are getting budgets and they're starting to spend more.

So it wasn't all, we're in the middle of both spending more, we're kind of journey through this. Yeah. I, I think what it is, Mitchell, it's the old story about economic uncertainty.

Mm-hmm. Right. If we're in an up economy, great.

And I've got, you know, big fat budget, great. If I've got budget cuts, but I know what my budget is, I'll prioritize based upon what my budget is. Mm-Hmm.

But when you're in limbo and you don't know what your budget's going to be, you don't know what the mandate and, and you know, goals are, then you kind of paralysis by analysis. Right. And, and you can't, you can't move you until you know what you're doing.

So now we're starting to see whatever plans were being made roll out. Mm-Hmm. And, and so we, you know, well, people can function.

They may not have enough money to do everything they want. They may not have enough budget man head, you know, head count and so forth, but they know what they need to do and what they're funded to do. Yeah, exactly.

Important. Well, let's turn our attention to, uh, some articles Sure. com.

One of 'em, I I wanted to highlight was this, does using AI assistance lead to lower code quality? There've been a lot of questions about what is, you know, sure. Code generator, generator code or assistance from AI by Bill Orfield.

And what kind of prompted his article is about GitHub's stats that 92% of developers should, they use AI tools and developers complete tasks 55% quicker using co-pilot, little bit self-serving there, but whatever. Um, and, and he had some quotes from people saying they anticipate that, you know, you're introducing more code and if you kind of get into this, um, just passive, I'll just let it do what it's doing, and you don't kind of do that same thing, right. That we're just talking about with writing, marketing content.

Just accept it blindly and move on. Well, for a co small code change, that's not a big deal. When you're asking it to write sections of code or doing things like that, you gotta make sure it's good stuff.

So again, I think you use AI either to check the code you wrote Mm-Hmm. Or to generate code that gets checked by a human. Mm-Hmm.

You don't use AI to do both. Right. You need that balance.

Yeah. Autopilot in there, copilot not autopilot. Right.

And that, that's a good way of putting it. And, but you know, let's be clear, co-pilot's been a tremendous success. Absolutely.

It's a real help. Yeah. Um, I, I think one way of looking at it is, you know, you look at like the low-code, no-code market.

It's, it's an exploding market. Mm-Hmm. People using a lot.

Uh, AI is a great tool in that low-code, no-code kind of vein. Um, AI's a great tool to teach you to code to help look at your code. But again, 92% of developers said they're using AI tools.

They're not letting AI just write the code. That's not the percentage of work they're doing. Right.

Right. Exactly. Yeah.

That, that's it. And that's an important thing. And I think we have to remember that, that that's what AI is.

You know, it will get better, it gets better every day. But at this point, you know, I, I don't think it's autonomous driving. I think it's assisted driving to your point.

Mm. It's lane keep you in the lanes, it's parking assist, you know, those kind of things to use the car analogy. Go ahead.

Some direct good recommendations about Yeah. Why don't we go through some of those. So if you are using AI to generate your code, these are, these are, you know, using to make a copilot kind of, uh, assistant.

These are some of the dos for that. So, you know, use it thoughtfully, don't blindly, you know, no. Yeah.

Be intentional about how you're using it and, and it's not for every task and every code you're gonna generate. Right. Use it in the, you know, pick the, pick your spots Mm-Hmm.

For that. And I would augment that and say, still learning. We're still learning how to use it too, so Yep.

Just don't throw it at every problem. Mm-Hmm. You know, do code reviews, security reviews, both with tools and with people.

That's still super important. Maybe we need to start to tweak how we do some of those things. Yeah.

With, you know, with, uh, gen AI created automated, uh, assisted development. Now, there, I hadn't thought about this one. Avoid redundancy.

Right. If you're having it right code for you, you may already have a function that does that. You may, that may be the same routine.

You could turn into a function now that four different times you've created the same code. Mm-Hmm. Um, so I imagine there's lots of opportunities.

Yeah, sure. And maybe double down or, or emphasize more measuring and monitoring the quality of what you're doing as you, 'cause you're changing, right? You, you're bringing in a new technology that could have a big impact.

Yeah. I, I don't positive or negative. Yeah.

I don't just set it free and, you know, set it and forget it. It's, I think you put it out there, and even though it may pass the sniff test initially, you gotta watch what happens in the wild. Yep.

Absolutely. Life will find a way. Yep.

Life will find a way it to survive. Yeah. Mm-Hmm.

Exactly. Okay, great. You know, uh, next article we're gonna talk about is this, how to get platform engineering just right.

Big topic, right? Mm-Hmm. That was DevOps dead platform.

Engineering's the new thing. Agile's gone. Oh, no.

Those things are gone, right? Yeah. They're all, it's all still part of our ecosystem, but look, platform engineering is a real function.

You know, I heard someone, we were at dinner last night and someone said it in a, I think it was Brad said it in a really succinct way, you know, if we think of DevOps as culture and DevOps tools, right? As, as DevOps tools, platform engineering has become the DevOps teams, right? We used to have DevOps engineers, but they ruin everyone always said there's no such thing as like DevOps engineer.

Right? But I think the platform engineer has become, and the platform engineering team has become the, the new DevOps engineer, if you will. Mm-Hmm.

Doing that kind of, of task. So, you know, it's not competitive or replacing, it's again, part of this whole continuum, uh, with Agile and, and DevOps and SRE and, and platform engineering. Agreed.

You know, and there's so many parts to, uh, you know, what, what does platform engineering mean? Right? Is it, uh, developer productivity and portals for, for that?

Is it standardizing environments for dev test and production? Is it optimization? Things like that.

It's all of them. I think it's all of those. And I think that's one of the reasons why it's been successful is 'cause it can adapt, you can adapt it to what your issues are, right?

If you're really focused on, Hey, I can't hire more developers, how do we get 'em more productive or point it to whatever you want? Um, but it, it of itself isn't a panacea, just like none of these things are. But, you know, so as the report or as the article points out, the Google state of DevOps report says that on average, it, it takes three years for platform engineering to start showing tangible results.

Mm-Hmm. Look, that might be a bridge too far for a lot of companies who that's are, that's a long time. Yeah.

Who are, you know, they're doing six months sprints here. Yeah. In terms of, of projects and so forth.

Most projects, if I was ever on a project that took three years to show value, I would, I think it'd be canceled about a year into it. Yeah. Or 18 months, something like that.

Yeah. But in reality, even the, even the three years might be, uh, optimistic. Best case scenario.

Yeah. And, and so again, this isn't something you just wake up and say, Hey, let's go do platform engineering. You, it's a long term commitment on your part of your organization.

Yep. Um, Benjamin Brows the author of this. Yes.

And he has some recommendations, some guidance around setting up your tool buffet. He, he emphasizes the importance of observability and that role role as well. How to plug into the CICD pipeline and focus on finops and green ops and, you know, there's probably some other things to do as well.

Yeah. We should mention Benjamin, I think is with cycloid. And, uh, he's also a platform.

Do Eng nng. com, which we'll have in our notes here, and maybe a QR code on you if you're watching this on video. Uh, you'll, you can go to Platform Eng and get more there.

Yep. Absolutely. We kind of talked about the next topic already around the, or reorganization.

Dell just announced this week. Well, it's been kind of announced coming Mm-Hmm. Um, around doing some layoffs around 5% of their workforce, which I guess works out to be 6,600 jobs, or at least.

Yeah. I mean, and again, from what I was told, it was heavy in the marketing space. Mm-Hmm.

I, I think Mitchell, the question is, is that just the first shoe dropping? Mm-Hmm. Right.

I read an article last night, I think I might have sent it to you. Yeah. Was it in the register or something?

You know, Dell has always been, I, I did deals with Dell back in the late nineties, the dot coms. Mm-Hmm. Uh, Dell host, and even back then at the upper echelons of Dell, almost everyone came from Bain.

Mm-Hmm. Bain Capital, right. Mm-Hmm.

Bain was, the Bain took, you know, Michael Dell out of his college dorm in Austin and made them a world power. Right. And, and ever since then, Michael Dell, I think there's still Bain people there running it with him.

And they, they have this, you know, Bain has this philosophy that you, you basically build an organization by saying each employee represents about a million dollars in revenue. Mm-Hmm. So to have a hundred, what was it, 133,000 employees, right?

A hundred. Right. But to have a hundred thousand employees Mm.

At a million dollars in revenue, was that, I forget what the, you're adding terrible with math. You're adding five zeros to the end of that, I think Right? Is what you're adding.

Five zero is the end of a hundred thousand. So it becomes a million, 10 million, a hundred million. A thousand million.

I think it comes to 10 billion. Yeah. 10 billion.

I think that's what the number was. And so if you look at Dell's revenue now and where they're, you know, planning, according to the Bain formula, they should be at about a hundred thousand employees, not a hundred mm-Hmm. Well, they're down to about 125 now.

Yep. So we might be seeing as much as 20 to 25,000 more layoffs at Dell, which, I mean, that's a lot of people. It's worldwide, but that's a lot of people.

But it's also in line. Look, I mean, Microsoft, Google, Amazon, uh, they all, all these big tech giants who by the way, reported decent Yeah. No numbers on the street.

Good. Right. And their stocks are doing well, but they're all taking this opportunity to skinny up.

Mm-Hmm. And, and it's a cycle Mitchell, right. You, it's that, you know, it's like dieting.

You, you get down to your weight and then you add the pounds back, and then you do it again and again and again. And it's probably not healthy, but you're hiring while you're laying off people. Yeah.

Because you, whatever. Well, so we'll, we'll keep our eye on that. But a along with that though, from what we know, there's, there's a bit of a reorg, huge reorg going on there.

Yeah. Uh, this fellow John Burn has, uh, taken over global responsibility for all regional and Dell technology, select sales. So they're, you know, retooling, marketing, sales, internal teams, and, you know, it always creates for a little uncertainty.

We'll see how it comes out. Mm-Hmm. Well, we'll see where they go.

You know, I, I gotta believe they're also gonna adapt to where the market is now and change some focus internally on products or channel or whatever it might be. So it's more than just losing people usually. Yeah.

Well, I think that some of the AI may take some of this functionality over going on. That's good too. Maybe some over optimistic, who knows.

Yeah. Well, let's talk about your interview that you did with Kurt, uh, with Kit. Yeah.

So our good friend, ki we just talked to him at Reinvent. Yes, we did. Yes, we did.

So, you know, I originally met Kit, he, he was still at Google, I think, uh, when Kubernetes was first, you know, going public. And, um, kit then went over to Jfr and he was VP Biz Dev at Jfr. And Adam interacted a lot with Kit during those years.

He left jfr and, and helped start up Noble Nine with, you know, s SLOs, SRE company. And he was, uh, I forget his title, chief Commercial Officer, or Chief Revenue Officer Marketing. I mean, kit was responsible marketing and, and sales there.

And you know, he, over the last, I don't know, two to three to four years, he's really helped make Noble Nine a leader in their space. But he's now taken a job as CEO, first time CEO. So congratulations to Kid.

That's awesome. With a company called Plain Sight. I, I did an interview with him.

It's on Textron tv. So Plain sight's been around a while. You know, when we think about ai, we tend to think of chatbots.

Mm-Hmm. But there's other AI types of ai, uh, one of them is, is computer vision. So in other words, you, you feed in a camera feed to a computer and it uses AI to figure out what it's seeing and, you know, act on that.

Mm-Hmm. Um, interesting technology, especially when it starts to get applying to it and stuff. Um, but the, the company company's well financed, but it wasn't, uh, what's the word I'm looking for?

It wasn't well known in the market. Oh, okay. You know, it was kind of still a best kept secret.

Best kept secret. Yeah. Yeah.

And it's Kid j it's KIPP's J Kid's job, nato, uh, making used a great evangelist. Yeah, he's good. He's, he's a great front guy.

Um, and he has a real idea of what he wants to do there and, and how to do it, and some interesting functionality that this technology can bring to it. Um, so I would, I would definitely check it out. It's called Plain Sight.

Plain Sight, uh, technologies. And, you know, we wish Kit the best of luck with it, Mitch. Mm-Hmm.

Absolutely. Absolutely. Well, another DevOps chat in DevOps.

Yeah. Another week of busy DevOps should mention, we we're gearing up to head towards, uh, well, about another month and a couple weeks we'll be in Paris for Q con. Yeah, absolutely.

Uh, but until then, lots going on. com for the latest and greatest DevOps related news and, and information as well as Textron tv. Right.

tv as well. Absolutely. Send us your feedback too.

com. You'd love to hear from your ideas, your reactions, what's good, what you don't like, what you'd like to hear more of, all that kind of thing. Absolutely.

So until next time, this is Alan Shimel and Mitchell Ashley. Keep Dev, keep DevOps. There you go.

We'll see you soon.