Techstrong Gang – August 12, 2024

Hey, everyone. Happy Monday. We've got a great show for you.

We we're going to talk about, is there such a thing as an AI developer? Do they need an IPC? You know, Microsoft seems to think so, but we'll find out.

We've got NIST weighing in Onis, and then our friends at CloudBees have acquired some old friends of theirs. We're gonna talk about all that and more. You are watching Textron Gang.

Hi everyone. Alan Shival here for Techstrong Tech, strong gang. We've got a Power Pack Monday show full of some great news and, and opinion and analysis.

Couldn't do it, of course, though, without some of our fantastic gang members. And let me introduce you to this edition's version of the Textron Gang. First of all, joining us out west.

He's fresh off his return from Las Vegas blackout engagement. He didn't play the sphere or anything, I, but he, he was in Las Vegas for, uh, that was, That was the dead last week, the last few engagements. That was it.

We missed it too. We came home too early. But it's our editor, roping editor extraordinaire, John Schwartz.

Hey John. Welcome. Hey, Ellen, it was great seeing you last week.

Great. Seeing you. Was pitch my friend pitch.

Yeah. We'll talk more about Black Hat in a moment, but let's introduce the rest of our gang joining us from her perch in the, uh, I, I assume you're home. It's, I am Home.

Our open source developer testing expert board member Tracy Reagan. Hey, Tracy. Oh, I should mention CEO and founder of Deploy Hub.

Hey, Tracy. Hey Ellen. Hope you guys had a great time at Black Hat last week.

We did, we did. It was, it was good to see a lot of the old gang. Um, and then now joining us from Alabama, I assume he's home.

He's an expert on cloud. Yeah. You know him from DevOps and AI now as well.

I think he has a new book that just came out too. I saw. It's my good friend John Willis.

Hey, John. Hey, Alan. What's going on?

There was a new book, a follow on a Deming though. Yeah, my, the original Deming book, um, I had to cut out about 15 stories, you know, going through all the editors and all, and I decided to take those stories and put 'em in a, a supplemental book or compendium. So profound stories.

So, yeah, we just, uh, that just came out like about a week ago, so, yeah. Very cool. I think we call out the director's cut, is that right?

No, no, this is a full on, second on, this is multi CD case. Blu-Ray. That's right.

Then. Anyway, chiming in here on my left. Our editor, not editor in chief, our Chief Content Officer, Mike Ard.

Hey, Mike. Welcome, welcome, welcome home. It's good to be home.

I, I, I got home yet last night and I'm, I'm happy to be home at least for a couple days. I'm on the road again this week. Um, so we'll do what we have to do.

But anyway, before we jump into the, the blocks that we had in there, I mentioned, uh, John and I spoke a little bit. I'm fresh home from Las Vegas and Black at, and just a couple of observations there as I, me, it, it's good to see, you know, I'm in security 28 years and doing the interviews or walking the floors one thing, but going to some of the parties and seeing folks you haven't seen in a while or catching up, I, I saw so many good friends and it was so good to see them. Um, you know, it was almost worth hanging out in 115 degree heat and getting slowly roasted.

If I was on a spit or something, it would've been better, but it, it was hot out there. I don't know how people live there full time. So you were cooked unevenly.

Is that the problem? Well, I was baking from the inside out, it seemed. Um, but that being said, a a couple of observations I heard, tell, and John, I know you and I spoke about it, but then I checked with, uh, one of the people I had interviewed who had insight, you know, from the Black Hat folks themselves.

There was somewhere between 17 and 19,000 people there. Right. Which, right.

I, mm-Hmm. Right. John, is that what you heard as well?

Yeah, So I, so when I checked in, I was talking to an organizer who was kind of walking me around, and she said it was about 17,000. Yeah. Now, I thought last year was like 40,000, and maybe I'm, I wouldn't be the only one who, who's getting older and exaggerates numbers, but I'm not running for anything.

Um, but that being said, I thought it was as big as RSA now, RSA clearly is over 40,000 people. So it would seem to me that RSA is still twice the size of Black Hat. Yeah, I think it, I think it's significantly larger.

It's interesting though, because those two c, those two shows are al almost, um, talked about interchangeably, right? Yeah. And one, almost black hat's almost become an extension or, or vice versa.

And before, as we talked about this in the last time we were, when we were in Vegas, it was, um, more of the, the black hat dev con con combination. Now, this is clearly like a marketing show, right? Interestingly, John, I I was reading an article on Politico this morning.

They had some fundraiser out at Defcon, and, but Defcon is at the Las Vegas Convention Center, the big, the big house. Wow. So could it be that Def Con is actually bigger than Black Hat now?

And maybe that's why that, maybe that's why 17,000 were a black hat and the number's not as large as you thought it would be, because there was kind of a, there is like a bifurcation between the two shows, it seems. There, there you, before we had folks going for, to both events. Now perhaps maybe we have another set of people, or, or, or blue comment Just with Go.

Yeah. So that was exactly, I, I will tell you the conversation in my house goes something like this, honey, wanna go to Vegas in August? No, no.

Next subject. Yeah, it is. I gotta tell you, it is unbearably hot, but it's a dry heat.

Anyway, um, John, I don't know, we, I don't wanna do a whole deep dive here, but, um, you know, You know, one thing I was gonna mention to you all is right before I left, I met with Dave DeWalt with at Night Dragon. And one of the things I wanted to ask him about, and again, we talked about this already, was this kind of debate or this obsession with AI for the second straight year, and whether it's a friend or a foe of the folks, especially in security. And he was of the opinion that, um, gen ai, which is obviously it's revolutionary versus the kind of evolutionary path of, of security has it, gen AI has like a two year head start on defense.

So that was a major theme that kept coming up over and over again. But again, that, that, uh, gap is, seems to be shortening. So there, there we will talk about AI development and the proliferation of it, which will be coming soon, I think.

But anyway, that was something I kept coming across. I, I think that was a, a theme. I think another theme I heard from numerous people was, geez, there's a lot of companies here I never heard of.

But in terms of what they sell, what they've developed, what they're advocating, it wasn't any revolutionary things. It, it was, you know, very evolutionary and some may say not very exciting at all. Interest.

Interest. You know What, one thing I was gonna also mention that was, that was interest that, that I think was actually very encouraging is I ran across a number of, of companies or startups that were co-founded by women. And so one of them in particular, I think Mitch had talked to her, she was inspired by science fiction and science fiction writing and literature.

So that was her inspiration for, for her company. But there were, there were quite a few. It was, it wasn't as predominantly male as I had feared or dreaded before.

So it was kind of a more of a, uh, kind of like a diverse No look in that regard. We, you know, the, the industry's changed. It, it, it is a lot better.

Anyway, though, it was a great black hat. Def Def Con was also fantastic from what I hear from people who were there. Um, and now it's onto the next conferences, you know, the next, in the conference season though, August, it gets slower.

Anyway, let's jump into our first, uh, topic today. And that I, I hinted at it in the opening the rise of the AI developer. First of all, is this just a marketing term?

Is there such a thing? Do the, does the world need an AI developer? Does the AI developer have his trusty IPC and that when he presses the button, he turns into Heman or Sherah or something?

What, what, what is this about John? What do you think? Yeah, no, I mean, it, you know, it's a reality, right?

And, you know, and I think, um, you know, Microsoft is trying to drive this, and I, I think m you know, if you look at the, what Microsoft has been doing for, you know, starting with the Azure DevOps, you know, uh, visual Studio, uh, code space, like all these different tooling, pie and GitHub, right? Like, they've been like preparing for this and then the investment in open ai. So, um, they've been eating their own dark food and, you know, that's how they've been developing software.

You know, I, you know, um, Sasha Rosenbaum, I'm like, um, she, she did a presentation a few years ago when she was on Azure DevOps team. And, you know, it's not fake they're building their software, but the tools they're promoting. So, you know, I don't know currently how they're using sort of copilot and copilot workspace and all these different tools to develop, but I assume they're following the same path.

So, uh, yeah, I, I think, I think it's reality. And again, I think, um, you know what, what's what's really interesting is that, you know, Google and, and, uh, Facebook or Meta have been doing this for forever. They just, they don't have the sort of the, the tools in a commercial or sort of a way that Microsoft can promote 'em, right?

So in a lot of ways, I think Google, if you go back to how Google built the original Kubernetes is the Borg, and, you know, talking to people at Meta, I mean, it, it's all push button at, at at meta, you know, including ads, which are, you know, they, they've been doing AI extensively for many years, right? And so all that stuff's push button, you know, how you get a model, how do you create a model? How do you, you know, how do you run PyTorch?

So I think it's a reality. I just think Microsoft's, you know, in the right position at the right time to drive this narrative, and, and they've got the, the available tools, you know, that, you know, again, yeah, I, I use some of the Google tools, but I've switched over to, uh, to basically all the, um, the Microsoft tools, you know, the code spaces and studio. I mean, you just can't get around AI development without really sort of using those tools.

So That's, that's powerful, Johnny, hear you say that. So, Tracy, let me ask you this question though. Will not every developer become an AI developer?

And is this kind of, you know, maybe an exercise in Microsoft marketing who, you know, to John's point is a house that developers built, and I get all that, but at what point are you no longer a developer if you're not using ai? Well, we're still gonna have API developers. We're gonna still have, um, you know, sequel with, uh, developers.

We're gonna still have backend developers and we'll have people who are specializing in ai because that's just kind of how we have always, uh, siloed, uh, our skills. So we will have AI developers. And I don't really see ai though, as you know, I always see AI as a tool.

It's not a product. And we've had these conversations before and I think it, we, it, it deserves more conversation. Is AI a product or is it a feature, or it is, is it just a tool?

But I wanna say this on this topic of where we see Apple going, where we see, um, you know, Microsoft going Google, I almost feel like they're chasing open ai. You know, they wanna, they, if Alan always tells us, if you're not in the top three, then get out of the, you know, get out of the barn. And I feel like they're all chasing that top three.

And what's going to get them there is developer adoption. So the, the sexier you could make this, the more fun it is, the more you are going to bring developers to your party. So if we're at the top three, those top three have to fight.

And anybody playing in this space right now, they have to fight for the hearts and minds of those, those potential new AI developers, AI PCs, as you kind of mentioned before we started the show, having a PC with NPUs on it all that sounds really fun and sexy, and it's going to get developers interested in playing in ai. So John Willis, if there is such a thing as an AI developer, I mean, we see new drops of models almost all the time now. I think Open AI just dropped one last week.

And developer really think about which model to use. What are you tracking and, you know, what leaps outta you. Yeah.

Um, I mean, I think, I want to step back a little bit because there is this idea of AI developer, it, it's sort of a little bit of misnomer because we go back to things we've talked in past through, so the inference and training or training in the inference, if you wanna put 'em in order, right? So when we talk about an AI developer mm-hmm. Um, in the pure sense is people who build models, there's gonna be less of them.

Uh, but the, but I think when we say AI developer, we're actually, and you know, I think what Microsoft's saying is the AI developer is gonna use all the AI tools available to them. And so what you're gonna have, you know, uh, you know, all the different types of development paradigms are going to be, uh, enhanced by AI tools. So I don't know that they're sort of focusing on, okay, now we're gonna have people that build, 'cause very few people build AI models, you know, especially in a general purpose use scenario, right?

More people. Exactly. Yeah.

In the same way, not everybody builds a database. That's right. But, but like everybody uses a database.

So I think when they say AI developer, what they're really saying, there's a new way to develop everything. And, and it is that you're gonna use things like copilots, you're going to use these tools. In fact, the whole, you know, if you've been following the whole, um, open Devon Devon scenario about automated dev tools or sort of agentic, uh, system software engineering tooling, right?

Like that's the next wave. Or not just writing help you write code, but help you close issues on GitHub. So I think, you know, all of this is what Microsoft is driving as far as what you choose for your tools.

You know, it's funny that, um, for Open OpenAI just dropped another snapshot, right? It was, uh, what was it sort of, uh, it was eight six, so it was like, it was basically, uh, last week. And, um, and it didn't seem like much, 'cause it was, it still was October, you know, the sort of snapshot of the date.

It was still October, so people kind of missed it. But what happened was it was like twice the size of sort of tokens. It was basically half the price.

Um, you know, and, and then it scored on some of the, uh, benchmarks incredibly well, right? Like, it, it, um, you know, um, there's sort of a zero val benchmark, and if you look at it, it's just like overall two per two points higher than the, the sort of the next on the list, which, um, I think was anthropics. Uh, and, uh, but what's interesting, it forced me to go in and do something I've been wanting to do for quite a while, is look at a little deeper to benchmarks and see which models are leading in which categories.

'cause the, the benchmarks have these different, like there's a coding category, right? And GPT OpenAI wins there. Um, multi, uh, multitasking things.

GPT wins, again, based on sort of a, a smaller set of, of vows, but this one seems to be the one most people go to. Um, but then things like, um, a deductive logic, um, tropic wins basically. Um, you know, so it's, it's funny, you can start looking at some of the, and then like on math, like if that's something you like need to do outside of a calculator, um, you know, llama went.

So it's, there's, I think the benchmarks are telling us some information. So to, you know, I think, to answer your question, I think, I think we're gonna learn more and more as these models drop and, and some are gonna be better at other things. And, and I've, you know, I've instinctively found that certain things I do go to, you know, I go to Claude for, for this, and it's sort of like, I, I have a feel, but I didn't know why.

Like, certain things that I wanna do, I'll use Claude, otherwise I'll use, you know, uh, GPT-4 oh or Omni. And so I think as the benchmarks and we, that this race that seems like it's sort of crazy is actually gonna give us more information. And as a developer, I think we're gonna have more, um, ability to choose.

And then, and I think like the other thing you have to be aware is the small language models, right? The high three, the, um, the Misra, right? So, so I don't think the game is over in terms of anybody.

I, I think what we're seeing is AI Darwinism in a Petri dish, right? Totally. And as time goes on here, each of these models is going to digress from, let's say a common base and have sp specialization.

We see it in the cloud providers too, right? There are things Google does well, there are things m Azure does well, and there's a lot of things AWS codes does. Well, you're gonna see that, but it, it's still early in that game.

To me, the question is, is it the AI developer versus the AI as developer? Maybe, I guess, you know, what I'm trying to figure out and, and try to keep the answer real short, John, if you don't mind, but sorry, are people, um, switching out LLMs from day to day, week to week, month to month? Or do I commit for a year?

And so I'm kinda locked into that one Project. I don't think, excuse me, John, I don't think it's even the LLMI think it's the whole AI Either way. But on this example, we're talking when things are being updated every other week, and it's hard to keep track.

No, I mean, I, I think there's companies that, probably the biggest thing you're seeing in enterprises is they're trying to lock in with the big three, right? And so, like, we're a, I was at a internal hackathon of a company last week, right? They're all Google now, whether that's the right decision or not, that's a whole nother, but, but they built their cloud infrastructure on it, now they're building their AI infrastructure, so they go with Google and Vertex, right?

Other ones, you know, majority of the large enterprises already bank banking on Amazon. And, uh, but the, to your question about the model, um, I, I think they're interchangeable. So I, the, the point is, if I go with Vertex, I can use any model and I can swap 'em out just by putting a parameter change, right?

If I go any, you know, Amazon, they all support all the, all the models. They're the sort of, the model selection is pretty much interchangeable via a parameter. I mean, there's some other sort of things you have to do in terms of orchestration and all, but in general, um, you know, the models, uh, themselves are, um, reasonably compatible for the code you write.

Uh, if you can serve, if I'm a, if I'm a visual studio developer and you can serve the tools to be, at that point, I'm gonna adopt a lot faster and I'm gonna be more in love with you. And so I believe that Microsoft, while that list that you gave us, showed us who was on top. Now, if Microsoft could make the adoption easier and serve it right to the developer, they're gonna win.

Well, well, I gotta add it there because I can use, um, BS studio and all that work with, um, really any tool I can use, Google, vertex I can use. So again, don't conflate the tools with the, uh, the, the engines, if you will, because they're gonna win the sort of the, I agree with you on, they're gonna win the desktop, they're gonna win your environment, they're gonna win. But I don't know that that necessarily equates to like open AI wins or the Azure Azure ai.

Fair enough. Wins. Fair enough.

Guys, we're gonna take a break. We have to, someone's gotta pay for this. We're watching text and gang.

We're gonna be, we're gonna be right back. All right, folks, we're back. 0.

John Willis has taken a look at it. John, what is the driver behind this? Why is NIST weighing in now on this subject?

And what is the state of AI model testing as you know it? Yeah, I think it, it's a really good question, right? I, I've, um, I've sort of publicly, you know, um, ragged on nist, if you will, uh, in, uh, in the sense that, um, you know, I think Oasp has done, I I've said this many times, oasp seems to have been up until now, seem to have done a better job on ai, you know, some of the earlyness stuff early, like in, within the last year, there's been more like, you know, what are the harm thing, harmful things you can do with Google?

Like, it, it, it didn't see a whole lot of difference between what they were expressing. But, um, I, you know, there was the executive order that, uh, Biden order on ai, uh, last year. And I think what we found, what I, you know, found is this has just been working really hard at getting it right.

And, and I think this, uh, dioptre I think it's called, I don't know, I I'm not good at pronouncing, but, uh, we can put a link to it, but it is a testing framework. And, um, and I think, um, it's, it's pretty interesting. The only thing I'll see or say right now is you crawl, walk, run this, this crawling walk run, which again, is a good sign.

The crawl is, they're really doing model, you know, sort of model building testing. And as I said earlier, I don't think the, I think the real risks are in the inference side, not the, not the sort of the, the, the sort of the, the modeling and, and the training. And so, but I, that just tells me that they're on, they're on the path, right?

You know, I think they're, you know, this is a good first start. But, um, and then, but the other thing that I noticed when I was doing the research on it that, um, they dropped a whole bunch of new papers that are really interesting and that puts them, that those aren't tools, um, that, that are available. And this, this diopter is a complete open tool.

It, it expresses that you get help from the government, you free help. And, you know, I mean, it, I, I haven't installed it, but I went through the code, I looked at it, and it, it looks pretty real. Uh, but the papers are more interesting 'cause that sort of puts 'em in line with, um, these sort of AI papers, uh, literally just dropped in July.

Um, it was like five of them that dropped. And, uh, and I can give you a link to those. But, but that, that seems to put 'em on par with, you know, from my perspective, which is, you know, sort of just me.

But, um, it puts 'em on par with o aas in terms of what, what Owas has been saying for about the last year or so. So to me, this I was su Oh, I'm sorry. Go ahead, chase.

I was surprised to see NIST have a product as opposed just a document on standards. That's what surprised me about it. It's not good.

Yeah. Interesting. Yeah.

A lot of work. Yeah, I think I, you're right, it, but I, maybe that was like, I was starting to say they weren't doing their homework, or they were just sort of spitting out stuff that just seemed, you know, not relevant, but maybe they were, you know, buckling down and putting something together that, that, you know, may maybe one of those other three letter agencies slipped something under the door for 'em. Like, Hey, you gotta take a look.

Well, I, I will tell you though, to me, this bleeds into our first block. It's a continuous of, you know, for people who you ai I'm Sorry. Oh, I'm sorry, Alan.

Can I, Can I, Yeah. It's kind of an extension of the, of what we were just talking about in AI development. Can I just kind of like broaden it or, or step back ahead.

Wall Street point of view, and I think it's Daniel Cher Group has talked about this and written about it, is like this whole kind of concept about the speed of AI's proliferation, especially when it comes in terms of development, is gonna take months and years, not days and weeks. We have this kind of ultra speed viewpoint of ai, but there's also a claim in, in a Goldman Sachs note recently that suggests there are zero killer gen AI applications, which, which seems wrong because we already have these use cases for advertising search, content creation, video development, rendering, private models that, that helps streamline financial services or personalized healthcare. They're moving fairly quickly.

I just wanna interject that idea or notion that there it is actually expanding and moving in in some sort of, some sort of way, and it's not just emphasis shouldn't just be on a handful of companies. I just wanna throw that out. I, I agree.

And I, I think, you know, putting these two blocks together, what we're witnessing is, is history in real time, right? We're looking at a day by day, day-to-day view of something that's gonna stretch out over the next five to 10 years. And it can't happen soon enough for, for some of us, right?

John lives this every, you know, significant portion of every hour of every day. And so when you look at it day to day, you know, the paint's not dry yet. But when we look at it, and, and the, and, and that may get some frustrated, especially people who are making big bets money-wise on this, right?

Going to the Wall Street thing. But you can't have that day-to-day view of this. You've gotta look at this as just a, a reweaving of our whole platform.

A a new paradigm. Dare I say, there's Like A three that doesn't happen in a day, Right? There's like a three step network process or effecting the first wave or servers memory, network storage, second wave, probably independent software vendors, hyperscalers in the third wave is then we get applying AI to new and improved experiences from retail to financial.

I, I think it's a misnomer though, to say, I think what's happening is, I, I've been going to a lot of internal hackathons of companies and they're building stuff. They're just not talking about it. Yeah.

Right? Um, so there's real stuff being built. In fact, JP Morgan is talking about their stuff and they're doing some real stuff.

So this misnomer that AI is sort of stalled, or, I, I think what we're seeing is the sort of commercial aspect. What we're not seeing is a lot of companies that are doing incredible stuff are not basically publicly disclosing that they're doing it, but I, it's also the nature, John, these things day to day to day, there's progress made every day. Now, unless you're doing time lapse photography of ai, you may not see that progress.

'cause you're just, you know, the, the, the watch p never boils, but of course there is. And, and then, you know, at some point you're gonna look at it retrospectively and say, wow, look where we were a year or two ago to compare today. It's night.

Literally night and day. So I think that's what we're dealing with. Let Me come back to that first question, the top 'cause, uh, and maybe provide a little more context behind it.

But John Willis, one of the things I hear all the time from folks is they're like, well, I don't know how well this model's gonna perform. I don't know how well it's gonna cost. And there isn't a lot of testing tools upfront, they say, so they're kind of all in on a proposition, and then they get, you know, concerned that the next thing's gonna come along and outdo the first thing.

So what do we need for testing in the AI space? We have it, it it, it's called, it's another flavor of observability, right? And fortunately, they use an overloaded term.

And I, you know, I've said this a couple times. It, most people think observability is like, you know, CPU performance, um, the large, you know, the sort of the Dynatrace and honeycombs, but like, it, there is a form of observability that's basically been around for a while called evaluations. And that that space is like lightning speed right now, and it's using sort of l LMS to sort of look for these things like guidance and, and relevance and QA correctness and hallucinations.

And so I think a lot of people don't understand that part of your stack of AI has to, you know, I, I've said this before, I so loosely turned something called Alarmer Stack, right? I'm not trying to become famous or anything, but you've said what, you know, it's language model, um, orchestration. You have observability, which is these evaluation tools.

You have RAG and you have model management, now you have Gentech processing, right? And that's sort of evolving to be the stack. So you, you, if you're not aware of the tooling that is available and how fast it's moving, and then the, the big people like, you know, uh, vertex has a whole lot of guides model guidance and bottles sort of guardrails.

Uh, Azures had it for a while, so it's there, it's just a lot of people are moving so fast. They're like, I'll get to that later. And by the way, that's the way DevOps went, right?

Yeah. Like we, we didn't really start playing with DevSecOps in anger until about five years into DevOps, right? So, Absolutely.

Hey, we're gonna take a break on the AI out. Let's talk about something else. You're watching Textron Gang, Discover Textron Group, the epicenter of tech innovation.

We are your go-to, for reaching IT leaders and practitioners worldwide. Our secret impactful content that sparks awareness, engagement, and top quality leads with us. You'll access editorial websites, streaming videos, virtual events, custom content analyst research, and more.

Join our satisfied clients. Let's revolutionize your tech journey. Contact us today and tell your story to the world in the most powerful way with Textron Group.

All right. Last week in all the noise that was happening around Black Hat and ai, something may have missed your attention. CloudBees acquired Launchable, and I know you've talked to these guys all the time, but, um, what is your sense of what is going on with CloudBees these days?

What are, when does Launchable fit in your mind? And, and kind of, you know, what's, what's, what's the backstory here? Okay, well, so keep in mind, I have not spoken to Sasha or KK or, or Har Pete KK Harp being the co-founders of Launchable, both former CloudBees people and Sasha Lare being one of the co-founders over at CloudBees.

I haven't spoken to them since this news went public. But first of all, congratulations, congratulations to CloudBees Congrat, congratulations to the folks at Launchable. For those who are familiar, launchable was a, a company started by Koa k Ka SI always mispronounce it, but I call him kk.

kk for those who may not know, is the founder of Jenkins. And he was the CTO and then chief scientist at CloudBees, when Cloudy was primarily a Jenkins company for many, many years. Har Pete Singh was also a CloudBees.

He ran product there for a long time. Both, both bright, bright guys. They left CloudBees, you know, as CloudBees was kind of moving away from being the Jenkins company and started launchable because they felt that testing and intelligent testing and efficient testing was the biggest bottleneck that we face in this CICD world.

Um, and they, not to bring up this word again, but they brought, built a lot of AI into their, uh, automated and intelligent testing. You cannot escape it. No, you can't.

Um, well now, you know, let's be real. com okay? And we follow this pretty closely.

CloudBees is a kind of a dog fight along with some of the other, you know, major DevOps platform players like Jfr and GitLab and GitHub and Harness and others. And I'm glad to see them reaching out and bringing Harpe and KK back into the fold, as well as bringing Launchable and its AI powered testing, intelligent testing to the platform. Here's what I think we need to see.

Is Launchable gonna be integrated? Uh, CloudBees came out with a new platform that was independent of Jenkins, uh, maybe a year ago now, less than a year. Are they going to integrate Launchable into this?

Is it a, is it become part, you know, a, a real part of it? Or does it stay as a standalone? Right?

And then more importantly, sometimes you do acquisitions like this for the people. You've got two bright people coming back into cloud base. I, you know, is k is KKD, you know, on his way back to being CTO, maybe is Harpe on his way to, you know, being a, a broader vision of prog of product?

'cause he has a good view of the CICD product space. Maybe if they do that, do other people drop out? Maybe we'll have to see.

But I, I am thrilled. I'm, I'm so happy for my CloudBees, all the bs my friends there and, and for KK and Harpe and the Launchable folks, We had a video. I agree.

Alan getting, getting KK back, getting co k back into CloudBees is he was the heart of CloudBees. He is, uh, an an amazing, uh, engineer. He was the heart of CloudBees jc.

He is the heart. Yeah. Mike, what, say he needed to be Back?

Um, let's take that again from the top and we'll cut it. Okay. Um, Well, we actually have an interview with Sasha on Techron.

Yes, we do. Last week. And, and one of the things that I found intriguing that he was pointing to was the convergence of DevSecOps and finally quality assurance that all these things were starting to move in the same direction because of, I know I'm gonna use the phrase ai, but um, John Willis, do you think, you know, as you look at the way DevSecOps and quality assurance is evolving, is that these, there's a convergence finally gonna happen?

Yeah, I think, uh, as Alan said, it's, it's gonna be sort of a little messy getting through to get there. You know, like I, I look at, um, you look at the, the pre gen AI tools, the harnesses and all those, you know, and then, you know, there seemed to be a lot of noise and nobody was really sold that it actually did what it said it did, because, uh, you know, you shouldn't ask me a question. I was gonna stay silent a second.

Uh, but, but the, uh, you know, the, the dependency map, like the, the, you know, it is so complicated, right? And, and Tracy knows way more about this than I do, but, but like, like I just owe, before we got to Gen, I, I was like, I, I just don't, I don't buy it. I'm not buying, you know, sort of pre gen ai ai to solve these complexities.

Now I'm like interested in hearing about it. And, but the problem right now is, you know, every vendor that's classically been doing sort of nothing with AI or doing sort of ML ops version AI are now doing Gen I versions. I dunno where you start.

I think that's, it's, it seems a little messy right now. I think it's an opportunity. Chase, should we try your internet connection again?

Maybe. I don't know. We probably will.

We probably should try it. Um, so my, my question for the whole thing is, you know, bringing testing into CloudBees is going to be an interesting kind of, uh, uh, journey for them. Yeah.

They, it launchable it. While, you know, I've, I've looked at that, that solution, and I say, you know, I say it with, with True Heart, I love kk and I think he is a brilliant engineer, but they weren't sales reps. And bringing it into CloudBees may change that dynamic for launchable.

And that's what I'm looking, I that's what I'm, I'm eager to see if that happens for them. Um, because they're not sales reps, they're engineers and bringing it into an organization who already has an enterprise sales team will be interesting. Will the enterprise sales team, sales team push that product that's, that's yet to be seen?

Uh, and I, I think because it's que they will Fantastic. Are we gonna see more of these kinds of deals? I mean, we talked about the market.

Absolutely. Yes. You had me at Hello there.

I I honestly think, look between what happened on Wall Street last week, and, you know, and I'm glad the market came back a bunch though. It didn't come back all the way looking at my accounts. Um, but all that being said, we're in a jittery phrase.

We're in the middle of an election cycle. Uh, there's, you know, what's the Fed gonna do? I think for even public companies in this space, and in tech in general, it's, it's shaky times out there.

And when times get shaky, there's, there's people who wanna cash out and there's people who wanna bottom feed, and there's people who want to look at this as a great time to build. And so I think we're gonna be entering a hyper consolidation phase over the next year where you're gonna see a lot of this kind of stuff, as well as some bigger companies that'll surprise you that, you know, this is tech, right? The little fish gets swallowed by the medium fish, the bigger fish gets swallowed, swallow the medium fish, the gigantic fish swallow the big fish.

You're gonna see a lot of it. I think we have seen John Schwartz, um, He, he, he came up there, right? There must be, uh, I think maybe in the last two weeks, I think I've seen stories from the financial press that basically says every company on the it's for sale.

Yeah, that's right. You know, given the uncertainty, it seems like it, given the uncertainty of, um, the election, let's just put that out there. Um, probably it's gonna be very close.

A lot of the comp given the state of the market and given the state of the uncertainty, whether are we in the recession or are we not debate over inflation, et cetera. The Fed, I think everyone's playing out off multiple options. They all have plans A, B, and C.

And I think anything could happen. I think there'll be a lot of consolidation. And I also think there's this uncertainty, even going back to our first block, um, in terms of things like development, in terms of regulation.

I, I would not, um, diminish what happened between the DOJ and Google, because that's gonna have an effect. Oh yeah. It already is having an effect on Apple and how anyone else looks at AI development, whether, not just in Europe, but in the US.

Well, well, I, I, I think if you wanted these tech giants, John, you gotta be real careful about doing any acquisitions. Absolutely. The FTC might be in bolded, probably is emboldened by what happened with DOJ and the d OJ is going after Apple as well.

But You know, you have, you have the Chevron decision as a counterweight there, right? Say, are these agencies, do they really wanna push it because of these companies taken to the Supreme Court? Will the court tell them that the agency overstepped their bounces so that that Wouldn't even, I wouldn't even overlook FEC and the influence they're gonna have in terms of, um, but They're also an agency and subject to the Chevron Yeah.

Decision. So, And given the courts. Exactly.

So there are all these from every executive, every branch of government. Um, these companies are slowly, I mean, they're, they're political animals now. They have been for years, and they are hedging the bets, You know, like that old Jewish folks saying, may you live in interesting times.

You know, I, I don't, I won't I The Chinese jersey? No. You know, the Chinese jersey.

Exactly. Let me just ask this question though, and, and I'll ask it to you and maybe Tracy can jump in, but do we not overlook all the integration that has to happen all, every time these companies get acquired? Everybody goes, yeah, this is great, rah rah.

And then like, you know, six months later they're still stitching this stuff together. Yeah. Six months, six years later, sometimes they're still stitching.

Look, you know, I, I've done a lot of acquisitions in my day, you know, for companies I helped. And, um, you, you gotta break a couple of eggs to make omelets. Mm-Hmm.

And not every integration works. And, you know, the, the road to perdition is aligned with good intention. So yeah, sometimes they don't work Well if Launchable did it right, they already did that integration.

Yeah, probably. And, and, and KK and Harpe knowing what they know about the innards of CloudBees, I would think, yeah. But that's just what I think you're gonna see a lot more.

Perhaps they built it to be acquired in the first place, could very well be. Mm-Hmm. Um, a lot of ex Cisco people used to, they leave Cisco, start a company and get brought back in two and three times.

And that's, anyway. Hey, we're about outta time. I hope we got your Monday off to a great start here.

It's a, a busy, busy week. I'm gonna be on the road this week, but Mike, Mitch, and the rest of John, the rest of our gang members will be on and we'll, you know, we're here to give you that day to day. So until tomorrow, check out the rest of Tech Trunk TV theory.

We have a great lineup. Also, do check out, I believe our Black Hat interviews are now that we streamed live while there are also now available, and they'll be as part of the show or available on Techstrong tv. Um, but until tomorrow, thanks for joining us here on Techron Gang.

This is Alan Hummel. We're out. I'm Bonnie Schneider, sustainability contributor to the Techstrong Group.

I'm excited to introduce you to a groundbreaking new initiative from Techstrong Research, the sustainability pulse meter. The pulse meter offers valuable insights into how environmental responsibility factors into tech purchasing decisions for key players in the industry. Position your company as a leader in the industry and differentiate from your competitors with a sustainability pulse meter offered exclusively from Techstrong research.