Techstrong Gang – March 20, 2025

Hey, everyone. Nvidia has ambitions to rule the world, will they? You're watching Textron Gang.

Hey, everyone. Happy Thursday. Wow.

It's Thursday already. This week's flying by. It seems like.

I, I don't know. You know what they say As you get older, the, the, the, the days grow longer, but the years grow shorter. My days are growing shorter and long and as well, so I'm not quite sure.

Anyway, we've got a lot to go over on, on Textron Gang today, and we've got some great people to discuss it. I feel like I, I want to introduce her first because we don't get a chance to see her as often on the gang these days as she's doing her thing. But let me introduce you to Futur analysts all around.

Good person. Kimberly Bates. Hey Kimberly.

How are you? Good morning. From Colorado.

Doing pretty well. Oh, you're home in Colorado. That's great.

Home in Colorado. I'm taking tomorrow off to go ski. Hopefully.

So we'll see. Good for you. Enjoy it.

Enjoy it. We might as well stay in Colorado then if we're already there. Just go into, well, you are northeast of Denver, right?

Mitch? Yep. I'm Northeast.

He's a, he's a stones throw from Me. A stones throw a Colorado Stones throw Becausecause the air slider up there. A rocky stone throw.

Right. But, but they travel further. Right.

I remember playing golf up, up there hitting it like ti in Aspen. I was hitting it. Tiger Drop a club.

Yep. Yep. Got Drop Club Bank.

Mitch Ashley, uh, future vp DevOps analyst. Hey, Mitch, how are you? I'm doing really well.

That's the end of my golf advice that I have. But other than that, you're on your own. No, you, well, you know this, Mitch, when I was there, it's still secure with you.

I used to go play golf a lot. I used to play with Larry Middle, who, who was our CFO. Mm-hmm.

And he was a good golfer. Mm-hmm. Man.

And I used to see him hit the ball up there and I'd be like, holy Mac girl. These, these people, they must grow like crazy people in Colorado. They hit that ball so far could hit it.

Could hit eight iron 160 yards. Yeah. No, nowhere else.

You could. Yeah, you could. And that's the why we don't have a decent baseball team.

Well, yes, those balls fly outta that Rockies Park. But, you know, I used to enjoy, well, we gotta talk about the news. We'll, we could do Colorado reminiscing over John Denver music and cocktails another time.

Yeah. Let's move, let's move over though, to the West Coast where we have our eye on Silicon. The thumb on Silicon Valley.

He's our editor. John Schwartz. Hey, John.

Sporting a Georgia sweatshirt nonetheless. I am. I know.

Isn't that weird? Incongruous. Yeah.

California and Georgia. It just makes sense. It's like oil and water.

Nothing in common. Um, but I went to school there. But it's good to be here.

It's good to have you on here. And, uh, you know, in spite of the Bulldog, we're, we're a Gator family, so, but we'll underst we'll let you stay on. And then moving from there up to Harrison, he's the Dean of Techstrong.

I dunno if it's the boys dean, the girls dean. Remember you used to have that in school? You like, you didn't have a dean for both sexes.

I don't know what that was about. But anyway, our Chief content officer, Mike Ard. Hey, Mike.

How are you? I'm well. I'm here in New York where, Hey, everything's turning red man.

And I'm not talking Maga Red. It's St. John's Red.

Red Storm. The Red Storm. So I'm, I'm old enough that it was the Red Men when we were there when I was there.

It was the St. John's University Red Men, little Lou Conker. He was very far.

Well, I've, I've, I've reminisced enough. I'm, I am getting old. Let's stick to the, stick to the plot here, Shimo.

All right. Um, our first, our first block today is, uh, a report on what's happening at the Nvidia Conference where Jensen Wang and Crew are laying out their plans for world domination. Do a little Doctor Evil, um, Mike, what's it, what's going on?

All right. Well, it is their annual GTC AI show. And, um, they got new chips, they got better reasoning models and a robot named Groot, which I guess is a nod to, I am Groot from, uh, guardians of the Galaxy, but it's lions and tigers and bears.

Oma. John, you were there. Walk us through what happened.

It was, as you said, they, they threw everything with the kitchen sink and give them some time, and they'll probably come up with some AI version of that. So, GTC as Mike mentioned, is in a sense, I think of it now as like the new Apple event. It's this annual event that is larger life.

It's almost like a Martin Scorsese, like production. Before Jensen went out to give his 90 minute keynote, there was a 90 minute pregame show with the acquired co-hosts of the podcast. The guests were Jensen who came on and, and, and did his shtick about working at Denny's, Michael Dell, pat Gelsinger, bill McDermott.

This is just the, the lead up. So the, this is presented as a Super Bowl of ai, the center of the AI universe. And it was a very slick, polished event.

Um, there were, as Mike mentioned there, there's a Blackwell Ultra, uh, flagship AI chip for building and deploying models that will help apps reason and act on users' behalf. There was the, uh, Vera Reuben next generation GPU, which is twice the performance of the grace chips. There was, uh, I mean, it's hard to keep track.

There was an announcement with gm. There was a factory related announcement in terms of AI productivity. There was Groot N one, the first open, fully customizable foundation for general wise humanoid reasoning and skills.

There's cosmos and opening customizable reasoning model for physical ai. In a sense. They did throw down the gauntlet, and in a sense, what they did is they announced what Jensen alluded to and teased it, the CES keynote in January.

It, it was, it was pretty interesting. You know, the funny thing to me, and, and I, I kinda use this as a kicker in the story, was that everything was so polished and scripted. There was one miscue, and this was kind of eyeopening to me.

There was a small glitch during the robotics announcement at what, at which point Jensen was somewhat annoyed in the essence frustration. Is there a human who can help me? Right?

'cause he couldn't hear in his earpiece. Nobody was saying anything. And he's surrounded by robots.

There's, there's laughter in the audience. And then he kind of smiles and says, it's okay. These things happen, but really, is there a human?

You can talk to me and help me here? And that kind of underscored a lot of what they were doing. Um, you know, it's still the human touch.

There's still that element that AI has to work with humans. And, and in one of the things that he also mentioned, where there will be 10 billion digital laborers and next several years, and he's saying right now a hundred percent of NVIDIA's engineers are as assisted in some way by ai. It's, it, it was pretty impressive presentation.

It was a bit overwhelming. It was a fire hose of news, maybe too much news. But then again, this company's up to, to a lot, they're partnering with virtually everyone.

And they made a very strong in emphatic statement that they are still the company at the middle of the AI revolution, so to speak. Are they gonna make those robots themselves, or are they, is that like a demo unit that other people are gonna make? It's a demo unit.

Other people are gonna make, I mean, it's something that it, there're elements of it. I remember a couple months ago, I went to Carnegie Mellon and I was told that these starter robot kits are becoming less and less expensive, and that we're gonna see kind of a brush on in that area, so you don't have to build it from scratch anymore. So that was, that was an interesting, an interesting angle.

Um, they're basically covering all bases. I mean, it's not just this announcement. There's the healthcare announcement they made a couple of weeks ago.

Uh, seemingly every vertical market cer certainly factories, the automotive market. Uh, it's, it, it's pretty incredible the tentacles that they have in the partnerships they've struck with virtually every big name company. You know, I was, I sat through the, virtually the, the, the hour hour two, was it two and a half hours or something like that?

It went over about, Oh my God. It was, yeah. And then I came away from it and understanding that we were gonna be talking about this this morning, and I was trying to figure out rank order, what was really important, what was the most important announcement they made, or what was just mm-hmm.

The next continuation. And I kind of came away. And I'm not, and so much of this is way over my, my pay grade.

I, I just, it just like goes woo right over the head. And I try to, you know, kind of parse it out. But I'm thinking what he talked about from the Cuda libraries mm-hmm.

Me, me, stood out as the one that most advances as also creates biggest moat around his company. Because that is where, and from an Nvidia chip situation Yeah. Intel and a MD and the other guys can develop it.

And it's that software that they've got that's enabling this development thing to me that said, okay, as long as you stay on that edge, this is what you've done. You nailed it, Kimberly. You nailed it.

That's exactly their competitive advantage. So you got the, the big thing out of it. You know, another just sort of side of this, you're talking about John, it being such the next kind of new Apple event, right?

And, uh, Nvidia being at the top of the pyramid for, um, AI hardware. I also gotta believe it's not, maybe not the main motivator of how they're doing this, but you also gotta believe they've gotta rebuild their image in the market after losing so much value in their stock. I mean, that one day is not gonna put 'em back at one 40 a share, but they've gotta rebuild that momentum back up.

It's just not automatically gonna swing, uh, back there. 'cause it's such a competitive market. So the more they can define the higher value, people are gonna see both wanting to work with Nvidia, use their software to Kimberly's extremely insightful point there.

I'm just giving you props camera and, uh, that Colorado thing, you, we gonna stick together as Colorado folks. Anyway, that was my point. I think there's a financial motive to both how this is done and what they're also driving to bring that share price back up.

The, well, Daniel was on, uh, Daniel Newman, our Daniel Newman was on one of the Fox business, CNBC or something yesterday. Um, and, um, he, I think he's on vacation and he came on just to talk about the Nvidia piece of it. And you know, what they're probing about was what they're talking about is, you know, how big is this?

You know, does this change? The ticker number? Does this change where, you know, their, their, their standpoint is from the stock price.

And it's like, I don't think it does because I think they've been whatever. But anyway, we don't get, you Know, Kimberly, it's, it's funny you mentioned, you, you, your your point about the moat was, was in the, in the pre-game show, there was an analyst on who's very well respected by Nvidia. That's why they put him on this show.

And he was with, uh, Michael Dell. And that's exactly what they talked about. Software was the Moes.

Um, and also Mitch, you're spot on in a sense. What Apple would do was they would make a ton of announcements. They're all pretty much incremental announcements.

And then they would have maybe one big announcement. They don't do that as often. That's kind of what Nvidia was doing.

They're reasserting themselves. And in a sense, a lot of this stuff didn't jump out or pop out at you. But in a sense, what they're trying to do is kind of, is reclaim a little bit of a lost grasp.

But they have on the market, they still do. But they were, you, you're right, there were lingering doubts after the stock plunge and the, the deep seek moment. John, what's the timeline on the processors though, that are availability?

Is that a this year thing or a next year thing? It's, uh, it's next year. So John, did those processors plug directly into the sun for energy?

Or how are you supposed to run 'em? Very, the Quantum Two Underneath each one. That's a very, and their second half of 2026 before they, they show up and Yeah, energy was very rarely mentioned yesterday.

Well, thinking that 2026 State has something to do with that stock ticker not moving too far. Well, so, so here's my take. And I, I didn't watch the virtual, I I read the Cliff notes.

An important thing is they acknowledged, right, that that Blackwell ultra chip, when it runs in a, in a stack, it's gotta be liquid. Cool. Right?

The days of of air cooled, uh, processors and, and servers are done, done. You gotta have liquid cooling. Think about what that does to the cost and complexity of every data center that wants to run this.

Number two, Look, we've heard this not just from, from, uh, Nvidia, but that, you know, physical AI robotics is a big, a big kind of, I I, I don't know if it's just a PR thing or, or they really, you know, think it's that close that we're gonna have these sort of robots that do, I mean, it's not that we don't have robots now, right? You go to any car factory, it's robots that are building the cars, right? They're single use robots.

They may not look humanoid, but they do one job and they do it well, they make a weld. We've got robots that do surgery. That's, that's very delicate, right?

So we already have some very advanced robotics putting AI into them. What does that give us? Well, you know, it gives us the ability for them to act autonomously.

Okay? Right. Is that gonna, you know, how big is that gonna be?

Well, until we actually have, you know, a Rosie the robot, like on the Jetsons or something, um, uh, I'm not quite sure, but Cam, Kimberly, you're a hundred percent correct. You know, we discussed this, I think it was on yesterday or the day before. Everybody and their mothers making chips today.

The, the special, you know, the recipe for designing advanced chips is, is out in the open, I guess because everybody is advi designing advanced chips. There may not be the Blackwell Ultra, but they're designing advanced chips. So what does Nvidia have, right?

Because this is a lesson, tell the Wintel monopoly taught us, right? Gotta have the software. If you control the software that only runs on your chips, well then you've built that moat around your chip, around your castle, which is your chip.

And so I think that's the real thing there. And, and though they mentioned it a lot in several different ways about the Nvidia, you know, software and, and SDCs and soft and stuff like that, that's going to be the true test of, of how long Nvidia retains this preeminence in the market now. And so you're saying software is the spice Alan, the spice flow, I think he's saying, I think he's saying CO is a set of golden handcuffs.

That's what he's saying. Well, it's a lock in. And when people realize it, they may or may not want to do it in terms of the stock price.

Look. Yeah, the deep seek stuff had a, it's momentary blip. The bigger issue is the, the, the hype around Nvidia is so high.

And this happens with these kinds of stocks around earnings, right? It, it, it leads up to earnings and then no matter how good the earnings are, and they had a solid, solid earnings, it takes a hit. And then you have macro conditions on top of that.

Are we gonna be able to build these chips in the us? Are we gonna pay a 25 to 50% tariff on them? Is Taiwan gonna be in existence anymore?

And by the time these chips are done, where are these chips coming from? Who are we gonna be able to trade 'em with? Where are they gonna be able to run?

How many you gonna go to Singapore and backdoor their way into China? There's so much uncertainty and chaos in the world today. How can you go make a bet?

How could you go pay 45 times revenue for a security company for $32 billion? You know, this. That's thinking in a past, there's this part of this is thinking, this pie is getting so big.

It's not like we're dividing this small little pie up between whoever's, whatever's doing. But what's happened is that we've expanded where the TAM is, the opportunity is in this market. Um, you know, I picked up something this morning out of, um, one of the reads that I have is Foxcon is talking about saying that by the end of this year, about the end of this year or next year or whatever, their server revenue will be higher than their iPhone revenue for factory.

Yes. And it was like, okay, that, that's a statement for the AI market of what's happening here and what we're doing to throw the money in there. So then the question comes is that I had this discussion yesterday with somebody, is what, in order to do all this, we have to be able to see an ROI for these implementations, and are they there?

So that's the, that's the, to me, that's the big lens to look at saying, is this going to continue? Is that, do the businesses, the governments, the institutions find the ROI and are they able to deploy these applications there? And I think the answer is yes.

You know, we, we discussed yesterday's is, is one of the analysts was saying the AI is in the child disillusionment or whatever. Um, you know, yes, AI is helping engineers, AI is helping marketing people. AI is helping salespeople.

Is AI replacing people? And I, I think, I don't know if today it's replacing people, right? What job is it replacing?

Well, statements get made, Alan, you know, like, um, meta Zuck saying we're not hiring more software engineers. Um, I I think one of the things we, that, that contributes to the, you know, overinflated expectations around AI is statements. Like everybody in the company now is using ai.

Okay, but what does that mean? Right? I'm not, we're not gonna hire more developers.

Really. I doubt that's true. True.

Usually those, that's maybe generally the direction, but often it doesn't pan out that way. So, and, and you know, it, uh, it AI could solve this problem in 30 seconds versus, you know, three weeks that it normally takes. Well, that all depends on the conditions that those happen in.

So I think we're, we're so interested in, in trying to validate the ROI or the benefits of, of AI without doing a real ROI analysis, um, that those, those statements just kind of are meaningless to me. It, it's maybe true, but what, so everybody has co-pilot and Microsoft co-pilot. What that's nice.

What difference is it making? What, what's the impact of the benefit of it? You know, those kind of things.

Yeah, that, that's right. I mean, they, they, they start, it's Zuckerberg and even Jensen, Jensen plays into this all the time. So does Benioff, they oversell this concept.

I mean, even when we go back to the physical ai, uh, concept that, that Nvidia threw out in January, it's almost kind, kind of conflating with genic ai. I mean, they're kind of getting ahead of themselves, but I think they're just trying to stake their name to that phrase. Uh, it's, it's, it's, it's all kind of marketing BS and, and a sense in the real world, this very incremental use or minimal savings of time so far.

I mean, practically most companies, To that point though, you have the Wall Street Journal and the, and the Washington Post echoing those comments and various contributed articles that have published in LA either two days ago or in three days ago. But Kimberly, I wanted to ask you a question, and maybe I'm wrong about this, but as far as I can tell, the utilization rates on these GPUs are horrible. So can we get better at increase the utilization rates so we don't need as many of them?

Maybe we should create the Department of AI efficiency. What do you think? Um, have no idea.

I am not necessarily, I have heard that the GPU utilization rates are much lower. They, they, we don't run them at a hundred percent. That is for sure.

I don't think we run 'em at, I've heard numbers like 60% or 50%, but that is like what we run our servers at, right? We don't run them beyond that. Um, No, but I've seen numbers As low as 10%.

Mm-hmm. Me too. I, I heard something.

I don't, I, and this, I don't know about this, but I heard something that in these models and in the training, um, you line up how many GPUs are gonna work on that. But for each step, they all work on their part of it, but it doesn't move to the next step until all of them are finished. So there could be a lot of idle time in CPUs.

That's sort of the architecture of, doesn't mean they aren't dedicated to that task, maybe. And that's probably where there's a lot of room to say, well, wow, they're not busy. Could other things multitask with that GPU?

Again, I'm not a hardware expert on that side of it, but I was told that at a conference, one of the, uh, tech Field day events, And one of the pieces that I, I mean, Jensen did, uh, talk about Edge, and I kind of went, I think that's gonna be the domain of the CPU. You know, maybe there are some use cases that are gonna need a full blown GPU out there because of video analysis or that kind of thing. But I think we've got powerful enough CPUs that are out there that don't suck up all the energy.

Um, you know, this, this much more efficiency that's out there. So, you know, that's the other thing when him claiming, you know, the entire landfall, um, And, you know, that's The name of the talk To people about that. And it comes down to what the level of paralyzation required at the edge is gonna be.

And if it's a, if it's a high amount, people are leaning towards the GPU, if it's a low amount, they're going with a CPU as the engine. I almost think the edge part of this is that's the, you know, this device that you can use to put a supercomputer on your, on your laptop or whatever. That's kind of the raspberry pie strategy of let's get that out at the edge and not people build things with it.

Uh, not that it's as low cost as a raspberry pie, but it seems to me that's part of that strategy of building up the edge part of it. Excellent. Guys, I gotta cut us off here because we are way over on this block and we got too many good things to talk about.

Let's take a break here on, uh, text Gang. And we're gonna come back and talk about, you know, DevSecOps. Is the glass half full or half empty?

You're watching Textron Game. 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. Hey folks, we're back and we're talking about, well, DevSecOps, which of course is one of our favorite topics on the gang, but it seems like there's been just a wave of attacks against software supply chains.

A lot of it revolving around GitHub, and some folks are writing opinion articles that suggest that at least, well, our first wave of attempts at DevSecOps might have failed. Uh, Mitch, I know that you've been following this space pretty closely over the years. Um, I can't figure the following out.

Is it just that maybe, you know, we're paying more attention, so we're discovering more attacks, but there aren't actually more attacks, it's just that we're figuring out that they're being made? Or are there actually more attacks? Well, there's always more attacks.

I think it's kinda like, you know, what size of a TV do you need? Whatever the biggest one they've got, there's more attacks, right? Um, the, what, what's happening is the attack surface is changing and expanding.

And attackers realize now that I can try to compromise an enterprise by these exploits or these vulnerabilities. But a better way to get a wider spread, um, impact is get into the supply chain of where software is created and distribute my code, my nefarious code into lots of production environments. Kind of back to the solar wind example, which really brought attention, I think the most prominently into the supply chain attack.

There was a, um, examples of this where, uh, GitHub recently with their actions, um, capability there, there are some code that you can use to monitor changes to files in the, in the repo. And someone built up enough trust or kind of slipped under the, the, the tent and got some code in there. The, the code only did a call out to go get more code.

It was an executable, so you didn't see it. Um, clearly if, if you weren't paying attention, that got distributed. Now all of a sudden, people who were running that, they've been running it forever running that same code or infected that happened with the, um, the, the xz or XY xz, um, utilities in, uh, Linux where someone broke into, got into the development stream and introduced code into the repo.

So it's, it's bringing a lot of tension to say, well, it is the, did, did DevOp DevSecOps fail? It's, we can't take a narrow view of, it's just about developers working earlier on better software security. It's the whole process, both, not just your code, but everything goes into making up the code and the platform that you, you build software, test software delivered on that tool chain also has to be secure.

'cause that's the attack vector to get in those endpoints end underneath the tool chain. Well, I, you know, Mitch, everything you said, I don't disagree with. I I think the thing we need to remember here though is security Security is hard.

If it was easy, it wouldn't be discouraged than it is. DevSecOps is hard. It's even harder, I think, than regular security because it's not just the purview of the security person.

You've gotta deal with the platform engineer and the DevOps person and the developer and the tester and all these people. They don't, they're not all big fans of the way Security pros go about doing security. They're all fans of quality.

No one says, I wanna create inferior software, right? I want to be obsolete when I'm 30. But it, it's, it's, this is such a hard thing.

And so when I look at some of the, uh, sightings of the, of the, uh, articles and so forth for this one, right? Um, yeah, stuff happens. This GitHub action compromise, it's not the first time a GitHub action has had been compromised, and it won't be the last time.

What, what's our realistic expectation? What does success look like in DevSecOps? Is it that we never have a security incident again?

You know, I don't wanna go all Hoover in animal house on you and stand up and say, this is un-American, right? But When the Germans attack fraud harbor, We didn't stop him then, right? But no, but the fact of the matter is, no matter how good we are, no matter how hard we try, we're never gonna eliminate these kinds of things.

It's about being vigilant. It's about being resilient. That's good DevSecOps.

That's good security. And I, I just feel that we don't need people piling on about, oh, another DevSecOps debacle. We are failing, failing, failing.

No, we're doing a hell of a lot better than we were five, 10 years ago. I'll tell you that. And we continue to make progress, but this is a very hard issue.

We're learning how to create better tools for developers to develop more secure code. We're learning. Not everybody wants to be a security pro or have that level of understanding, but to say that DevSecOps has failed, no it hasn't.

No, it hasn't. You made several good points. I just wanna jump in a little bit on that, Mike.

O one is we need to learn the lessons that have already been learned in general security, right? We know today that is, is equally important, if not 1%, 51% more important than prevention. Because it isn't just that attacks are gonna happen.

There are gonna be successful attacks, whether it's in the software supply chain or against our network or a cloud or whatever it might be. So response is, is actually extremely important. Maybe more important than prevention.

'cause without, without adequate response, then you really have a problem, right? This thing can, things can spread very quickly and get outta control. So I just wanna make that point of there, there zero trust, things like that.

We're kinda learning these lessons again in software that we've already learned in network security. And maybe that's what we can benefit from. The network security folks who understand that have been doing this for a while.

So, so in the LA we're supposed to anyway, build out, uh, more software in the next couple of years thanks to AI than we did in the last decade. At least. That's the theory.

Won't most of that software hopefully be better from a security perspective? Because we'll be able to use things like AI to say, explain this vulnerability to me while I'm writing this thing. And can we not maybe replace all kinds of legacy applications with something that is a little more secure, a little more modern?

And maybe we should make that a goal. Well, I'll, I'll chime in on this. Um, 'cause I, I did the, I did an article or a, uh, research note around who's gonna win the ent ai race in software development.

And the, the trend is most certainly moving, not just doing things in the developer's IDE tools, but also doing them externally. I mean, Google, uh, Google Gemini just announced, uh, a new, basically coding inside of the, the Gemini interface. That's also happened with, um, anthropic at the command line.

So things are moving out of that. I think where we're headed is people generating code who don't know code, even though there are other tools that are, you know, no code tools. You can generate a full application in these tools.

Now the problem is not everyone's gonna have the skill to look at it and say, well, that's not right. You know, you're, yep. Are they gonna ask the, the tool to say no?

Is this secure and it's gonna come back and say yes or no? And they just won't know. So I think we've gotta get to a point where we know we're, we're generating more secure software, not just software.

And we aren't there yet. That isn't the focus at this time. I, I think what we may find is that AI is better at fixing code than generating code, or at least the current, the current iterations of ai, where we are in the evolution of ai, it, it might be a better fit for that, which goes along back to the last block of is AI replacing people or enhancing people?

And, and there's, and by the way, guys, if AI enhances people, makes developers better developers, makes marketers better marketers, makes salespeople better salespeople, that's a fine thing. That's a great thing. It's a great thing.

You wanna be disappointed that we're not cutting heads as a result of it. We're just being better at what we do. All right?

You know, shame on you. As far as I'm concerned. I, I think, you know, me having AI make us better at our jobs is a worthy, worthy goal.

Having AI eliminate jobs, you know, you want that if that's what you want. I don't know if we're there yet. I think the ai, well, that takes us to a different topic, which is saying that AI is helping us build more critical thinking will help us with critical thinking skills if we deploy them.

Which goes back to the DevOps, what you were saying, Mitch, about what, what we're doing now, what, what you're asking the DevOps people to do, that you have these senior people that can perceive, can understand what's, do what's doing well. And those skills are skills will start to change in terms of what we're, what they're doing, um, for the day-to-day job. Yep.

If I can make an analogy, I think we're at the, using a gaming analogy. We're at the mods stage where a AI is the mods to the game that enhance change. You know, help it make it better, do things easier, better interface, whatever it might be.

And well, at the same time we're working on, well, what are the next generation of games look like? What's the next Fortnite look like's a new, a new format, no new form of gaming, which is the, how do I genically create software and not, not have to worry about code. But right now, it's really in the mod stage, I think in a big way, in a very helpful way.

And that's appropriate 'cause we're learning a lot. And that way we all can sort of mod software development the way we want or whatever role that we're doing. Love it.

DevSecOps, folks. Keep doing what you're doing, right? We're making a difference.

Don't believe the hype. We're gonna take a break here on Textron. We're gonna come back to our C block today, which is a, a look inside from Citigroup.

com is the leading resource for news analysis and education on challenges facing the cybersecurity industry. com covers all aspects of cybersecurity, including data security, DevSecOps, cloud security, application security, network security, security threats, and more. com has the largest selection of security content featuring breaking news, blog posts, podcasts, and more.

com to learn more. com. Home of security bloggers network.

Hey folks, we're back in. Yes, as Alan alluded to, we're gonna be talking about an announcement the C group made where they said that they're gonna move more and more of the work that they do in it, do an internal team, and they're gonna start eliminating more and more contractors. Now, if you've been around it for a while, you can walk into organizations and there's more contractors than there are employees.

So, you know, if this becomes a significant trend over time, it will change the employment algorithm, I guess, or equation for a lot of IT professionals. So Kimberly, do you think that this is gonna be a, a major shift in general, or is it just unique to Citigroup in the banking industry? Actually, I have no clue.

And I read this piece and I'm like going, okay, what does this really mean? I mean, especially since we've got all the noise coming out of Washington, DC and all that kind of stuff that's going on with Doge. But so there's two pieces that came out of that article.

One is reducing this, reducing contractors and going internal. And that to me is something that companies do all the time. That they go this way, they go that way, they go this way, they go, we that way in a way to improve something.

And the second piece of that article was, why? What are they doing? And they've been fined, they've had some data governance issues and controls issues and they're having to bring it in.

Okay, so what is that all about? Well, I kind of did some digging on it. And what I'm understanding, because it's saying data management, of course my, my little, my my antennas are out out there.

What some of the issues are with C Corp or Citigroup is they've been buying so many companies and they have a mishmash of data and silos and their ability to report to the federal government in a way that the federal government has confidence in their regulatory reporting evidently is what's getting them in trouble. So if you go back down to the underneath, part of what their problem is, is that they have silos or systems out there. I mean, I read one of the articles I read, you're talking about, you know, 40 or 50 different, 70 different systems, you know, computer environments that don't talk to each other.

And they were also talking about some of the problems they were having in terms of each actually even, you know, doing work that they have. So they, one of the examples they gave on wealth management people is saying that it takes them, you know, a week to set up that system for a new person and wealth management. Whereas their competition can do this in a day and a half or two days.

Okay? So when you un when you look at that, you kind of go, okay, is bringing in your staff into your own staff going to improve this problem that they have for data governance? I have no clue because really don't understand what their problem is other than they're saying we have a problem with people being external.

It's interesting that you, you mentioned that, um, Kimberly, so Citigroup had been fine with $136 million. We're not making enough progress on these quote unquote longstanding data management issues. And in, in a sense, it's probably something that other companies are gonna be concerned about.

But there was, there was a report, I don't know if it's substantiated, that those problems that Citigroup had in particular were tied to some outside consulting, regardless of how you look at it. They, they, they pretty much brought in a new head of technology, A guy from PWC named Tim Ryan who said, we've gotta, we gotta address this because we're spending, I mean, we're, we're spending more to address our IT issues and, and, and these shortcomings. And that's having an effect on their bottom line and in the way Wall Street look, wall Street looks at them.

So in a sense it's a self-protective move, but it's also, it comes in a climate of we don't know what's gonna happen in terms of regulatory penalties. We're getting mixed signals from DC maybe the banking industry. I don't know how it stacks up against the tech industry in terms of this regulatory oversight, but, um, it's, it's something they're trying to get ahead of.

And I don't know if, if they necessarily are gonna solve the issue, if they continue to add on or craft on other companies or, or acquiring of their properties. Well, their day, this, this problem is not just a recent one. This is going back decades in terms of the, the ones that I've saw the sightings of.

Here's where they've been cited multiple times they've been cited onto. So when they're saying they're, they're dragging their feet or not getting this done, somewhere along the line, there's a management problem. So the fact that they've changed over the management, the management's gotta come in and say, okay, so what's the, what is the, you know, the base problem here?

And there's a lot of speculation of that. You know, since I'm not sitting in that person's seat and I haven't done the analysis, I have no clue what it's, but when they're talking about, you know, they acquired so many companies, and we've already seen this with the AI problem, you know, that silos of data, being able to bring that data in, doing data management, ha is the one thing that's stopping us proceeding forward in an easier fashion. Now, one of the things that was said, could AI help them fix their data management and regulation problems?

And I think it could, you know, if they wanna take all that data and dump it into a vector database of some sort, you know, spend that money, drop it in, you know, create this entire thing and then say, okay, so give the report, you know, what does the, the federal government need? The problem with that is I need to show the logic about how I got there as well. I need to prove what that data is all calling me about.

And that's what the regulators are asking for is like, I need that transparency. It's reporting back to me. That is accurate.

Does anybody find it ironic that the person leading the charge to get rid of the contractors used to work for one of the largest contractors in the world? And does anybody think that maybe he knows something about contractors and how maybe inherently He knows all the flaw, he knows all the flaws, right? They brought somebody in who did troubleshoot.

He knows all the problems, right? Well, I go back to Alan was talking about security is hard. Data is hard in large, in a large organization.

I mean, we're talking about behemoth sized organizations, to your point, Kimberly, about all the acquisitions. It isn't just data, it's data in motion that's changing. And also they have generations of data, right?

They have systems that those databases are still running on something designed 15, 20 years ago and things that are brand new. And you can put 'em into data lakes and different tools to try to get that, that, I don't know if this is true in this case, but my, my experience with audits and regulations is, um, it's, it's a, it's sort of a layered onion, right? You provide data, then there's more questions, then you gotta go back and find out how to get that data.

Then you've gotta respond to the next layer of it until answers are are satisfied. And when you can't give them reasonable answers or don't know how to give them those answers, okay, is that a problem or is it a something's going on here that, uh, we is suspicious about and you know, we've gotta investigate further. You know, you know, Salesforce just had this, uh, developers conference, TDX and, and one of the things that cropped up, and I don't know if this is related, but this whole issue of data readiness, and there were a lot of unhappy customers of Salesforce, and I heard this through somebody who works for Salesforce as a consultants who said that a lot of these projects involving AI in particular, the data's not ready.

There's a lot of, uh, teeth gnashing over this. And that adoption of things like genic AI are not gonna happen until probably even as late as early next year, not this year as a re as a result. So I mean, it's just part of this kind of whole equation of weighing cost and benefit and then fines you have to pay Well, and there's also just the, uh, the meaning of what the data is, meaning what is an account.

An an account in one system may mean one thing and very different than another system. So you've got the semantics of the data and how do you, how do you transform or bridge those gaps? Um, because those are apps that were built by another company at another time, maybe even internally, just generations of different products, what Salesforce calls an account, just not what our accounting system calls an account, right?

And, and that's just one tiny of, you know, literally thousands and thousands of data elements. So that transformation, that sort of mental math that you have to do to bring it together to be able to answer questions can be daunting for a large organization. And that's part of it.

The other piece to that, the data readiness, John, to your point is, uh, governance issues. So I have, if I'm going to train or do anything with training and analysis, I have to make sure the privacy information is protected. And so that means I'm gonna be masking some of that data and I need to make sure that as it goes through that pipeline that masking all the way goes through.

So that's another piece of it that, you know, when I get to another end, it doesn't reveal itself for one reason or another. So that's part of the other data readiness that we have in terms of the, the being able to train ai. And you're absolutely right.

That is a, I mean, we've had these master data management capabilities for years on data warehouses. Um, and you have these data managers that that's what they do. Um, that skill base is a unique skill.

Um, and it takes a, I mean, I had a friend of mine that was CIO for the state of Colorado, and this is like 10 years ago. And what she was trying to do at the time, they were trying to bring in data from all the different divisions to be able to analyze information. You know, one of the pieces is matching police record information of juveniles with schooly school information to be able to track and see is there, is there a trending here?

Information to do that, though I have two incredibly different organizations that I need to bring the data together and the privacy. Um, that's tough, tough to do. Yeah, Guys, I, I think there's a a another aspect and, and cam you hit it, Kimberly, you hit it right off in the beginning of this, and that's the pendulum that swings between insourcing and outsourcing and, and there's something else tied in here.

It's remote work versus work and work in an office, right? This, we're not talking about a couple people we're talking about right now. 50% of the people working on it at Citi, I don't think they're called Citi Group anymore.

I thought they were just Citi now. But 50% of the people at Citi working on it are contractors. They've, they've outsourced 50% of their IT, and they're getting hit with fines and there's regulatory pressure.

And most of those outsources, I guarantee you may probably were hired during covid or the contracts we're giving out during CVID, they're remote workers and they don't feel like they've got control over things. They wanna bring it in-house. They wanna bring people back to the, to the offices, to the data center where we can control these things.

And it also may be that give and today's hire, it was much harder to hire people. IT, people three years ago today, we got a lot of unemployed IT people, we could, we could hire them maybe cheaper than we were. Maybe it'll be cheaper now to keep 'em in-house versus doing contractors.

It's an excellent point, Alan. Yes. I I think that is it.

That plays into this man. It has to. I I think if you're paying contractors based on outcomes, you'll generally have a better experience If you're paying them on time and labor, you're gonna have a bad experience.

You know, I, they're smarter people than me who do this stuff, so I'll leave it to them Anyway, speaking of smarter people than me, thank you all for showing up today. Um, I appreciate it, Kimberly. It's always great to have you on.

Anytime we can get you, you know, you're welcome here on the gang. John, Mitch, Mike, thank you, thank you out there for watching. I think we have day two of, uh, tech Field Day on tech drunk TV today, as well as some more of our seus subcon coverage from last week and a whole bunch of other great stuff on Tech Field Day.

So stay tuned for that. Um, I wanted to also give a plug, I don't know if we've had a chance, Mitchell, the, uh, I know you guys did a, uh, a, uh, a research note on the Google Wiz deal that we spoke about yesterday. It's up on the FU and research site, you and Fernando and Krista.

Yes, we did. Yeah. Kinda looking at all dimensions and what, where that might head and everything from little pass regulatory, you know, muster and is this a multi-cloud deal?

Is this more than that? What is the, what is the purpose? Um, is it, is it so, um, you know, go, can take on Cisco to try to vie for the, the big, big gorilla in the security market?

So really some really good thought and then really kind of bringing all the heads together and put into a research note. So thanks for mentioning That. You're welcome.

Also wanted to mention, for those of you who have never heard of Futurum Intelligence, no, it's not an oxymoron. It's actually a really great portal of information that our friends of futur have been building and they just released an update to the DevOps, uh, information in the portal that, again, Mitchell was in integral, in integral piece of this will be the topic on Monday show. So stay tuned.

Absolutely. We're gonna talk about it Monday. If you wanna sneak peek though, go check out Future of Intelligence.

With that, we're gonna wrap up today's text Drunk Gang, thanks for joining us. Stay tuned for Text Drunk tv. Have a great day everyone.

We're out.