Walmart AI

Hey, everyone. Looks like Walmart's playing small ball with small ai. You're watching Tech Drunk Gang.

Hi everyone, it's Alan Hummel and I'm back. Happy Tuesday. I've been away for two weeks traveling, traveling the the countryside of Italy and really enjoying myself.

Highly recommended to many of you ever get a chance to go to the mountains of Tuscany and just really enjoy the lifestyle. But I am psyched and happy to be back here on Textron Gang and back with my gang members talking about great things. As I mentioned, we're gonna talk a little bit about Walmart playing small ball or small ai, talk about that.

A few other AI things. We've got an update from Bonnie Schneider at a recent event she was on, and we've got our Steven Foskett to give us an update from a recent cloud field day. Let me introduce you to our gang members talking about all this today, though.

First of all, joining us from the data center down in Florida. It looks like resident expert on many things is a good friend of mine though, too. JP Morgenthal.

Hey, jp, it's good to see you, man. Same here. Glad to be here as usual.

Fantastic. Then jumping from there out to Boulder, Colorado where she, I hear she was on a little biking adventure herself. Yeah, South Dakota Black Hills and the, and the Badlands.

You oughta Go the Badlands, the bad girl from the Badlands. She's analyst Kimberly Bay. Hey Cam, it's great to see you.

Good to be here. Oh, and then joining us in, I guess he's home in Hudson, Ohio, the Hudson Flash, uh, tech Field Day founder President Steven Foskett. Hey, Steven.

How are you? Uh, good. It's, uh, good to be home for a little while.

Uh, a lot of traveling going on over here too, but, uh, not, not quite the bad Badlands though. Sounds like a, I Guess that's Akron, right? Akron can be the Badlands.

Well, if you're in Hudson it is, I guess, but, you know, I think of the Bruce Springsteen song, but I digress. Speaking of Bruce Springsteen from the other side of the river, that's another Bruce Springsteen reference there for you if you didn't catch it. He's our chief content officer, Mike Ard.

Hey, Mike, how are you? I'm good. I'm gonna be close to New Jersey later today.

I'm going to the Datadog conference over Javit so I can wave at it. Nice. All right.

We're always, always close to Jersey. Can see you from the Jersey side. Anyway, let's jump into our first, oh, You forgot somebody.

Oh, Did I forget? Yeah. Hello.

Um, joining me here in studio, she is our Echo Insight analyst and expert editor Bonnie Schneider. Hey, Bonnie. Hey.

Good to here. Good to see you, Bonnie. I, I apologize.

No problem. Alrighty, let's jump into it. Mike.

Walmart's playing small ball in AI with a kind of different strategy. Tee it up. Yeah, I'm not so sure how different that is.

I was just curious to me that A, they actually bothered to articulate it. And b um, it does seem that this is the route forward for a lot of organizations and maybe there isn't gonna be this kind of handful of magical AI agents, and there's gonna be a small army of them that you have to figure out how to orchestrate. But jp what's your take on what's going on here?

Yeah, I, uh, I think I had the same impression as you, Mike. When I read the story, I was like, okay, what's the news here? Um, maybe I, I do think the, the author, uh, you know, was trying to differentiate from the fact that these large mo large language models that are answering everything from health questions to, uh, you know, supply chain questions, to travel and transportation, you know, it'll book you a trip and also find you're a cure for cancer.

You know, it sound, it sounds a little too good to be true. I think people are starting to get lost, and I think what Walmart is trying to call out is, is a, an an architectural approach that says, you know, right now, um, and let me stop for a second. Let me over the weekend, apple, which has been kind of bashed about for their lack of involvement in this AI community, came out with a, a slam paper from left field that was very interesting, that's getting widely distributed that kind of basically put a kibosh on the reasoning, uh, theory and thinking that's going on.

And, and the reality is that reasoning isn't really reasoning and here's why. And, you know, putting, putting the mathematical proofs behind it to show that, yeah, if you put the right data in, you can basically make it look like it's doing the right stuff. And as soon as you're outside the scope of, you know, what it's been trained on, reasoning falls apart, uh, which is getting apples some kudos for, you know, in certain circles and in others, you know, the, the fanboy hype factor or, you know, coming out like antibodies and trying to destroy the virus that is apple for puck cooing on their parade.

But in many regards, it does play into what Walmart is doing and saying and recognizing, and they probably, you know, Walmart's always had great researches when, when they've, they amazingly have always had these incredible teams of people on the forefront of technology helping to drive their business. And, um, and they probably do a lot of research that doesn't get documented that is on par with stuff that you would find in Google or, or you know, Amazon. And so, you know, they recognize, I assume themselves that, hey, this reasoning stuff really isn't, you know, we can't count on this.

This isn't ready for prime time. It's not what the, you know, what the, uh, philanthropics and the, uh, you know, and the Googles and you know, uh, you know, the other players are, are saying it is open ai. So let's look at it for what it does well, and let's look at it for what we are.

We're a retail company. We, you know, uh, clearly AI can play a, you know, a big area for us in our main area, which is supply chain, right? That's what Walmart is known for in, in driving costs out of the supply chain to keep costs low at home, uh, in the store.

And so, uh, you know, I think these, when we say small, I think it's really what they're looking at is applying these technologies in a very directed sense to accomplish that mission and not get taken out by the, let's use it for everything under the sun right now. Mm-hmm. I think your point is spot on in the sense that it seems like they're also trying to limit their risk, right?

I mean, if I have an AI agent or whatever it is and it goes haywire, I'm only gonna have x amount of scope of possible damage to worry about, whether if I have some bigger AI thing, you know, the risk level goes up. 'cause the reasoning model that I'm counting on has to be more accurate. And well, that, to your point, a lot of the claims right now seem a little aspirational, if not downright, fanciful.

But Alan, you've been away for a while. What's your take? So, I got a lot to say on this.

Number one, this to me, is Walmart dipping their toe in the water. Look, I mean, if that's what you wanna do, you're never gonna be the leader. You're never gonna shake things up, dipping your toe in the water.

But it's a cautious conservative approach. And if you are, if you are Walmart, you could get away with that, right? Walmart, even though they actually are a great technology company, and all of us probably knew or know a lot of people who work at Walmart Labs and Walmart technologies, though they've laid off a lot of people recently as well there.

But they, they do some great things technology wise. I'm a little, I'll be honest, I'm a little surprised that they're not putting a bigger bed on the table here. But, you know, they've got, they've got football teams to worry about and super yachts and planes and, you know, they've got other priorities as far as Apple goes.

You know, did I ever tell you the story of Dr. Henry Pucci? My, my, yeah, my professor Poly-sci Professor St.

Johns senior year, his claim to fame. He was the conservative party candidate for Senator for New York in 1964. He ran against a guy named Bobby Kennedy.

He didn't do so good, he got killed. As a matter of fact, I don't even think the Republicans ran anyone against Bobby Kennedy that year. Anyway, Dr.

Pucci and I coming home from Italy, I like to say that now, Pucci, Dr. Pucci always talked, you know, go all the word Dr. Pe Dr.

Pucci always talked about the haves and the have-nots. And the have nots always throw stones at the haves because they really want to be the haves in today's AI world. Apple's a have not.

And so they're throwing stones. Of course, if they were leading ai, they would say the reasoning was fantastic. Is the reasoning perfect?

No. Will he get better? Yes.

But for Apple, and we'll see what comes out of, you know, they have their worldwide developer conference, I think today or something like that. We'll see what they come up with. I'm sure there'll be AI mentioned there.

But for the laggard here to be throwing stones, it, it's, it's unbecoming. It's a have not trying to become a have, I think, or, or throwing stones at the halves. So I Was thinking, I, I really, I'm sorry, Kimberly.

Go ahead, cam. No, go ahead. Jp.

No, no. Kimberly, JP went. We gotta give you a chance.

Go ahead. So, I'm not gonna comment on the Apple stuff 'cause I haven't read it or anything, but looking at the Walmart piece, I think it makes, to me it makes a whole lot of sense. Um, they innovate by taking what, you know, systems and implementing systems out in the world, not necessarily as in being an innovator in new technology and in the processes.

So there's a couple things here that are going on. One of them is, as I read that, I think a couple, first of all, there's prompt engineering. So we talk about how, how good you are at prompting.

Essentially what the AI agents are doing is breaking up those prompts into smaller pieces and putting those together. So if I break up the prompts into smaller pieces, now I'm thinking about Henry Ford and how he put together the factory. So, or microservices developing under like a microservices where I can reuse pieces of the AI agents together to go out there and implement it.

So the first thing they're taking on is maybe this some piece of customer service. And when you think about Walmart customer service is massive. You know, customer service is everything from somebody buying something to somebody returning something, um, to questions about, you know, techno questions about their products, et cetera.

So then you're talking about how do I reuse some of these agents that are consistently being able to operate? So again, I'm looking at that Henry Ford meets microservices meets systems of operational systems that Walmart super excels at. And taking the techno new technologies and applying it to what they know best, which is retail operations.

So I think it's advanced. I I would not think of them as being behind. I think they're gonna let other people create the GPUs, the CPUs or whatever.

They don't need to go out create an LLN, they do new, do need to create their foundational analysis for their customer base, et cetera, but not compete against all some of the other big things that are going on. Fair. Jp, do you wanna, I I Just wanted to, you know, address your point of the haves and haves nots.

It's, it's an interesting philosophical point. I do think that it's so important, uh, especially with a technology like this, that there is a recognized that, uh, uh, established pragmatism. And I think right now the height factors off off the charts.

And it would be good to establish some well-respected leadership that is coming out with a pragmatic voice around this at this time that, you know, I think it's important. I, if ABO wants to establish themselves in that camp, I, I think we should support them and not say they're, they're a wannabe that is, you know, outta the game. What I'd like to see actually is some sort of, for lack of a better word, council of customers who get together and kind of provide a mitigating voice against all that hype factor noise that's being generated.

And somebody who's been playing with this stuff, and actually try and Hop, you're asking Tigers to lose their stripes. That's not the way this game is played. There's a hype cycle.

Maybe this hype cycle is off the charts compared to any Torical. It's, and you know what? Hype it's up.

It's up to the people who keep their hand over their wallets to decide how much of it is hype and how much of it is real and how early they want to get in. I gotta share something. I'm gonna share something, which is you folks in here.

So I was away on vacation. I had a dream. I had a dream.

This is a real God, God strike me dead. I'm telling you the truth. I had a dream.

It's the entrepreneurial, the entrepreneurial blood was, was pumping, I guess. 'cause I wasn't busy at work. I had an idea to start a company, and it was, it was, the name was Ask j Eves.

But we already haven't ask j Eves remember Ask Jevs. That's an, but this was an Ask j Eves, but it was, it was an AI agent Butler, who you want to use the Walmart AI agents, you want to use the targe ones. You want to use whatever Apple's coming out with.

This was a consumer, this was gonna be, this was a consumer, uh, driven, not a B2B a B2C play for consumers to have their own Butler, AI butler who handles all of their agent tasks for them. And you weren't locked into any other, you weren't locked into Google, you weren't locked into Apple. You weren't locked into meta.

It was generic, not quite open source. And I was going up and down the, the venture capitalist meeting route for any of you guys who have ever raised money, right? You go, you go, like you schedule these meetings and you do your dog and pony show, you put on a suit and tie even sometimes, or at least a sports jacket and a tie and, and you go to raise your money.

Um, and I was pitching, I was pitching hard about don't get locked in AI agents from any one vendor. You need something that's generic. And I didn't know about this Walmart thing at the time, but I'm rea I'm, I read this and I said to myself, was I precent with this dream?

Maybe I should go do this. I have to come up with a different name than NA Chiefs, but mark my words. All of us are gonna have a butler that handles our AI agents for us.

And I'm not calling it an orchestrator, you know, isn't That my husband? You've got your husband. Well, you don't dress them up in the English Butler uniform, do you?

The I don't want to go there. I don't want to go there. I think, I think we're not only gonna have that butler, but that Butler's gonna do battle with AI agents from various vendors to try to come to some sort of negotiated mill that we all agree on.

Because ultimately, my, my agent's gonna be optimized for a specific outcome. And it may not be their agent's outcome. Well, right?

You've gotta have an agent that watches out for you, but he has to deal. And he is, that's, that's wrong. It will have to negotiate that with all of these agents.

But we're all going to need one. We're all going to need one. And I mean, this dream was, it was, I had my PowerPoints all set up.

I was, I was really pitch, and you know how it is. You get a lot of sounds good, big enough market. I I, it was like the whole VC thing I was playing out.

Right? It's a big market, big opportunity. What makes you think you can do it?

And because I'm not gonna lock us in it, it, so obviously subconsciously, this has been playing on my brain, I guess, for some time now, and it came out while I was away. But, um, I, I think this plays right into what Walmart and, and the rest of these folks will be doing. Steven hasn't said anything yet, so let's let him have the last word on this one.

Uh, I've been biting my tongue the whole time. Yes. Well, first off, jp, thank you for bringing up that incredible Apple, uh, report, uh, to Alan's cynicism.

I do not think it was a mistake or, or a surprise that that report, that that, uh, academic paper came out right before WDC because of course Apple is probably, certainly, well, let's just put our money down and say they're not gonna announce anything revolutionary regulated to AI at WDC. Uh, they clearly, um, was smartly evaluated their AI and said, this ain't good enough and, uh, pulled it. And I don't think that there's a good chance that they're going to have a, a new story here.

And frankly, I think they're right. Uh, I have read the paper. Um, it's very, very good.

It academically shows that these supposed reasoning models with their chain of thought are actually nothing more than the LLMs that they're composed of. And they're not reasoning, essentially, they do pretty well, uh, up until a point when you ask them to actually do real cognition and reasoning, and then they just fall apart and they can't do anything. And that's really a reflection of what this technology is good for.

I do, like jp, I see a reflection here in what Walmart is doing. Essentially, they're saying, wait a second, rather than having, uh, or wishing that we had actually intelligent artificial intelligence, let's use this technology for what it's good for, which is what I suspect Apple is gonna do. And what I suspect, frankly, a lot of enterprises with, uh, decent, uh, management are gonna be doing.

I I heard another similar story about the, the Campbell Soup Company of all people, uh, who are also using AI the way Walmart is, which is to say they're building small agents to accomplish limited but useful tasks in various areas. They're, um, having the LLMs handle, you know, data processing, ingestion, transformation, translation, that sort of thing. Ai, you know, these LLMs are great for that.

In fact, they're probably better for that than humans. But, um, to ask them to be sort of this all seeing, all knowing God emperor from science fiction, it's just not gonna happen. And I'm really glad to see companies like Apple and Walmart, uh, pulling back on that instead of continually putting more and more and more and more money down on a dream that just isn't real.

So, so you're saying my AI agents are unreasonable. I'm saying your AI agents are limited and useful. Well, well, maybe they won't replace all of us, then.

Time will tell, or, Or, or have the, or Skynet take over the entire world Soon enough. All right, let's take a break. We're gonna come back.

We'll, we'll keep talking about AI though, but we'll talk about testing and ai. You're watching Text on Gang. Hi everyone.

Welcome back to the Techstrong Gang. Recently I was at Code remix a summit in Miami where the fo focus actually was AI and also testing. And I had the opportunity to speak with Andy Piper, who came all the way in from Oxford, England.

Um, and he spoke about his company Diff Flu. He is the vice President of engineering there. And not only about their testing that they're doing, but also how it's, uh, making their processes more efficient and the environmental mental impacts of it.

Hi everyone. We're at Code remix in Miami and joining me now is Andy Piper, who is the Vice President of Engineering at Diff Blue. He came all the way from England for this conference.

So Andy, it's great to speak with you today. Thank you. Very pleased to be here and enjoying some of the Miami Sunshine.

Yeah, Exactly. Exactly. So tell me about Diff Blue and your role there.

So, diff Blue is a spin out from Oxford University. So it was started by a couple of, uh, ex academics and we use AI to write tests, Java tests automatically for Java code. And, uh, we've been going about eight years now, I think.

And, uh, there's 45 of us mostly based in Oxford. And So how has that testing field changed? Well, uh, we used to be the only product in the space until Gen AI came along.

So there used to be no AI products at all. Uh, and originally people were basically generating tests in a very standard way. So just sort of template type tests, not using ai, but we actually, because we're able to run the code, we're able to analyze what the code does.

Developers are a crowd that are hard to please. And so they like things like, does the test look good? Does it, is it understandable?

Is it named? Well? So we are able to do all of those things.

Um, but the other thing that we do is we test what the code actually does. So Gen AI tends to test what it looks like. So if it'll read the comments and base the test based on the comment rather than based on what the code actually does.

But we, because we're able to execute the code, write the test based on the code execution, we can write tests that are a hundred percent accurate. Well, where, where does AI come in? Because there's obviously a lot of controversy with that, with people using it and trusting it too much.

It needs to have a human overseeing it. Yeah. So how do you balance that?

Well, So one of our value adds is that you can trust us. So the tests that you get are always correct. We tend to, one of the things we do is discard tests that don't compile run past all the rest of it.

But also because we're, we're not using Gen AI at all. We're not using lms, we're using reinforcement learning. We're able to take a, uh, very deterministic approach.

So you always get the same test, same test that is, right. So with testing, you want to be able to trust it. You don't want something that you've gotta keep checking, otherwise you're testing the test effectively.

So What have you noticed in recent years that developers are asking for that you've kind of had to maybe shift or, or adjust to keep them happy? Sure. So, um, developers very much want idiomatic tests.

So they want, so for instance, we, we have a lot of support for Spring. They want tests that look like spring tests in the way that a developer would write a spring test. They don't want just kind of a bog standard Java test that, uh, doesn't understand the context, doesn't understand the domain.

So we've had to respond to those sorts of developer needs. Developers always want to, they kind of want to fire and forget when it comes for testing. They want to focus on actually doing their, their day job, which is writing code.

They want the testing part to be handled for them. And often if they use these sort of gen AI assistance, they have to stay in the loop checking things. Whereas what we do, we, we, you know, we're able to do that a hundred percent correctly.

And that's one of the ways that we sort of responded to developer needs. So What do you see going forward in terms of testing? Where do you see the space going?

Sure. So, um, I mean, one of the trends is that the one we've announced here, so we've announced a, a sort of, uh, hookup with modern where they're using customers can use modern to do automated upgrades, and then they can also use us to validate the outputs of the automated upgrade. And that's a very powerful thing.

A lot of customers are dealing with issues where they've just gotta upgrade all the time. How do they know that that upgrade is gonna be safe? And testing is the way that you can do that.

And if you can automate that part of the process, then that's, that's very powerful. Suddenly you've released a lot of, uh, developer productivity to, to spend on other things. Right.

And also, do you feel that it's also helping with efficiency in their productivity? Uh, it definitely helps with efficiency, but, you know, developers will fill the time available, right? It was also about unleashing their creativity to some degree.

So there's a lot of rules and regulations imposed on developers today and really wanna be doing that stuff. They don't wanna be doing the security checks and balances all that. If that can be taken away by tools and ai, then they're quite happy about that allows them to focus on their creative passion.

Absolutely. And since you're based in Europe, the movement of green software and being more sustainable is bigger there than here. I was just wondering what you've seen in that space, people's interest in that.

Good question. I, I, I mean, I've seen it in conferences. Um, I think, uh, in terms of what we do, one of the values of unit test is they're supposed to be fast and light.

And, you know, that's one way of saving energy For particular, you know, a lot of people, their testing infrastructure is very sort of end to end oriented, and that can be very, very expensive computationally. Whereas if you have a unit test approach, then you can save on some of that output. But often people don't write the number of tests that are required in that space.

'cause it's just too hard, too, too time consuming. That's True. Well, have a great conference.

Thank you so much for Joining me. Thank you. No problem.

Diff Blue also just entered a partnership with Modern who was hosting the Code remix Summit in Miami. So they were, um, very prominently focused there at this summit. And, um, it was fascinating to see, as I'd mentioned, um, there was a lot of interactive activity between these companies, and Andy was just one of the, the global people that I spoke to that had come in overseas for the conference.

You know, it, it's interesting, the, the, you know, the whole reinforcement as a learning model kind of thing, of course, came out with deep seek, really came to the forefront of, you know, maybe there's a better mouse trap to be had here. Um, but it, it'll be, you know, to, to do previous to our previous, uh, segment about how real is this and, and all of that. Look, I don't think it's the answer to every question, but I, I think the AI you're seeing today isn't gonna be the AI you see three years from now, or five years from now.

And I do think we're making a lot of progress on how to better train them, reinforcing and everything else. Now I look, I think there's an argument to be made if you took a human child and didn't stuff it full of learning as, as that per, as that child grew, what would their reasoning be? What would their, you know, cogni cognition of things be?

I, I think this is the way you learn. You, you learn based upon what you've been taught, what your exposure is. And I think as we get better at reinforcement, we will get better at, I don't know if it'll ever reach human reasoning, perhaps, perhaps it'll surpass human reasoning, but I do know that, you know, it's thesis and synthesis, right?

You, that that's how reasoning there's a lot of that working that you need a basis to, to build on. And to me, that's what the reinforcement part of it is. It's the learning aspect of it that allows you to build on from there.

So very interested to see. Now, modern is a very interesting company, right? They, they do code optimization, but for the huge, huge, huge enterprises where you have hundreds of thousands, if not millions of lines of code, right?

And, and it does, um, modernizing and, and correcting and finding vulnerabilities and, and remediation. So this is a great, that diff blue mo modern, uh, partnership is sounds like it'll be a really, uh, yeah. And the CEO who I interviewed there as well told me they just received series B funding From modern From modern, yeah.

Yes. I I actually interviewed him mm-hmm. On text Trump tv.

Okay. On that. So, so I am reminded of the woman who said, um, I don't want an AI that does art.

I want an AI that cleans up the kitchen so I can do art. And testing to me is kinda like cleaning up the kitchen. Wait, wait, wait a second.

What woman, what woman said that Maybe there's a quote in an article where she basically said she didn't want the AI to do the art. She wanted to do the art herself. She wanted AI to clean up the kitchen so she had more time to do that herself.

The reasonable assumption, this is another use case of that, um, testing is maybe the kitchen and something that a lot of people don't wanna do or don't have time to do. So maybe this is one of the better use cases for ai. It, it, it, it, so I really like, and I think the reason we're seeing so much, uh, positive news driving the hype, uh, curve out of the coding environment is because programming languages are so finite, which gives us a really great link to neuro symbolic ai.

You, you basically are, you have a limited number of patterns that the, you know, this transformer technology has to be aware of when it comes to language, when it's feeling around a programming language. And because of that, you get very high quality results, right? And so it's also very easy to add neuro or attach the neuros symbolic AI into the transformers when you're doing coding.

Not so easy when you're doing spoken language, right? Because the vast differences in understanding syntax and semantics of NLP, you know, are, are va you know, massively different, exponentially different with regard to the data sets, right? And so you will see that.

And the reason testing is so great is because, you know, the one thing that people are really awful lot programmers are awful at, is trying to break their own code. It was something that was instilled in me when I was a young developer, because I had a manager who, that was his thing, you know, he, he was like, you need to try to break your own code. And, um, so many people and developers I met after that don't know how to break their own code, nor will they try.

And so, um, you know, qual QA environments are, were, that was their goal was to try to see where vulnerabilities were, where opportunities to break the code existed. And so the good thing that this thing does is it has this limited set of, you know, semantics that it has to worry about, and it will devise for you what we call code coverage. And it can do a hundred percent code coverage.

That's a, that in the past we've discussed, no one in their right mind would've expected a hundred percent code coverage from a human. In fact, the averages were like, you're good if you're 80% code coverage. And what that means is how much of your code are you checking?

How much are you checking the input? How much are you checking the output? Are you validating that the code does what it's supposed to do?

I've used it for building some tests, and I gotta say, it's awesome. It's perhaps one of the best uses of this technology in, in the coding and development world. And if I can jump in here too, on diff blue, the one, one of the things that's interesting here is that this company predates the current LLM mania.

This is a company that, uh, was founded, um, I I guess almost a decade ago or based on research, almost a decade ago. They were out there, uh, before chat GPT, before the LLMs. They had, um, you know, they, they, their, their basis is in the core technology that, that, that forms all of these LLMs.

But this is not a company that said, Hey, let's just take off the shelf LLM Tech and pretend that it knows something about coding. This is a company that focused on coding and, and as, as jps, uh, pointing out on a specifically limited and useful problem set, which is let's develop unit tests around Java code. Great idea.

And, um, you know, they, they released a freeware version, uh, five years ago. They've gotten a lot of adherence in the DevOps space and uh, you know, I really love to see this. Another thing I really love to see, they may not love to see this as much.

You mentioned a funding round. Uh, they haven't raised much funding. This is not a company that's gotten a billion dollars from, you know, the, the AI frenzy.

This is a company that is trying to build a product, thank goodness, you know, we need more companies out there that are trying to build a useful product instead of boiling the ocean with ai. I don't know if the VCs agree with you there, right? They want 10 x on their money.

Well, I, I think Steven has a point though. There's multiple types of AI models and engines out there, and they're not all L-L-L-L-M, sorry. And they're just as equally useful, if not valid.

And we'll be part of the monopoly of things that you're gonna orchestrate. So I wonder, maybe, yeah, we're not paying enough attention to these things. 'cause you know, investors are looking for the words LLM and that's it.

Well, not all investors are VCs. True. This is also true.

Mm-hmm. Alright, let's take a break right here. We're gonna come back and we've got Sea Block an update from a recent cloud Field Day review.

Who's leading that one? Stay tuned. You'll find out.

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. Welcome back everyone. I'm Steven FoST.

As, as Alan mentioned at the top of the show. I lead the tech field day business unit, uh, with futurum. And, uh, we actually wrapped up our cloud Field Day event last week in California.

Now, I wasn't the leader of it, uh, Mr. Alistair Cook, uh, was the leader out there on site, but we had a couple of days of great presentations from companies. And, uh, most of those, well, I think all of those were live streamed, uh, right here on Textron tv.

So let, let's talk about some of the key takeaways. So as always happens with Cloud Field Day, it's sort of a reality check. Um, I like to say that we don't ever have really a theme for Field day, but a theme evolves.

And this time, uh, was no exception. There was really a couple of big themes. Um, and number one, um, this is a recognition of the fact that cloud strategies are evolving.

Uh, we're seeing a new generation of, uh, cloud technologies that reflect the, the hybrid nature of cloud that reflect the risks that companies are, are, are, are seeing, uh, whether, uh, it was Commvault, which, uh, focused on cyber resilience or scalability, which talked a lot about data sovereignty or Qumulo and HPE talking about cloud architectures. The industry is really trying to figure out how to do this at scale in the second half of this decade. So first off, uh, among the key points was cybersecurity and data governance.

Um, these have moved from, IT concerns to real business imperatives. Companies aren't just worried about ransomware, they're actively developing systems for operational continuity under attack. And that was a really interesting aspect of the Commvault presentation.

A as we've heard from that company for the last couple of years. At the same time, companies are trying to prepare for the rise of ai, as we discussed in the first two segments here. We're trying to figure out how companies can deploy this technology in a reasonable way and what that means for infrastructure teams.

So we saw companies like MIN io with their AI store, um, and of course we heard from our own Signal six five talking about how AI infrastructure is being built and the fact that these workloads aren't just experimental, they're being actually deployed in production. So let's talk through the three big takeaways from Cloud Field Day. And I'm, I'm gonna invite the folks here on the panel who know quite a lot about this, whether you watch the whole thing or not, to talk about these three big topics.

Number one, the rise of cyber resiliency and sovereignty. Uh, number two, the shift toward hybrid and private cloud as sort of a default mode of operation for enterprises. And number three, how AI is reshaping everything from storage to hardware acceleration.

So I guess, uh, dive in, uh, cyber resiliency, uh, data sovereignty and, and continuous operation. What, what are your thoughts? Yeah, I, I wanna combine two of your things.

Um, the AI and the, uh, in, in, you know, as part of the, in the hybrid, because, uh, one of the things I've written about recently and, and in compare the evolution is in hybrid, right? I, I do think that, uh, as an example, we're gonna see AI follow the model of cloud hybrid, uh, capability. But a good representation that shows the, the architecture and the need for it is that I'm going to have local models that I train on my business running in my facilities, right?

They're smaller, they take less memory. They don't have that heavy weight overhead from an infrastructure perspective that, you know, open AI has to run chat, GPT, they can run on a simple server and provide tremendous value because it's trained on my business and it really only knows about me. But that's okay.

I only want to ask you questions about me, but occasionally I need to augment me with data that's outside of me. And at that point, it's going to go to the larger LLMs, the ones that have up-to-date information, the one that have tools that can go out and scan and bring in external information and summarize it and make it part of me when I need to. What's the weather?

You know? Oh, what's happening geopolitically, right? I, I don't wanna keep that data, right?

So I'm gonna go out and it's a great model to look at for why do you need to do hybrid cloud? Exactly the same thing. I want to train on me, I wanna run efficiently on me, and occasionally I need to go bigger than me.

I need to think bigger than me. And when I need that, that's when I want to leverage cloud. That's when I want to expand out and take advantage.

And that mechanism for recognizing that I think is still kind of missing from a maturity perspective. I think it has to be created by an enterprise architecture group right now. And kind of, it kind of made it into a fixed model where it's like, we're around this, we're doing hybrid, and I'm setting up the AWS connect and I'm setting up the VPNs and I'm setting up the VPCs and, and it's too much fixed.

And, you know, architecture to make it happen, we need to get more dynamic. It needs to decide, hey, this is where AI may be a great fit for helping. I need this side.

I need to go hybrid now. And by the way, let me tie up and go quickly, you know, and generate the, uh, the DevOps scripts to go provision that architecture up in the cloud, do what I need to do, and then tear it down. So, you know, it's terraform run it same way it does with a Python script, right?

Or it could do it through the API Amazon or Azure API, I think that's kind of where this is all gonna blend together. Yep. It, it kind of reminds me a little bit of Alan's dream from earlier here, right?

That, that we would have our own, uh, uh, ai, you know, agents, whether they're running, you know, locally or out there somewhere. Uh, and they would take on different tasks and, and we would be able to kind of mix and mend and blend based on, on what we want. Um, you know, Alan, uh, does this resonate with you?

Well, yeah, I've got a few thoughts. First of all, in terms of my dream ask chiefs, the beauty of it is I only had the one agent, I only had the one agent, but it intersected with everyone else's agents. So when everyone and their mother's making an agent, why do I need to make so many agents?

I just need an agent to talk to those agents. That was part of my pitch to the VCs is I'm not gonna create all of these agents. Everyone else is, I'm gonna harness them, but let me return 'cause I digress.

In terms of resiliency, look, cyber resiliency is all the rage in, in, in this cyber space these days. And it really is something that's been coming on now for a while, which was instead of putting all our eggs into the prevention basket, we're never gonna be able to prevent all these bad things happening. We've gotta be resilient.

What happens when, and part of that resiliency is kind of spreading your eggs into multiple baskets, which goes to this hybrid thing. But it also is having a, a real plan in place for when stuff happens. And, and it's not just cyber incidents that happen.

It could be geopolitical incidents. You're no longer allowed to host your stuff in some country because of some nationalistic kind of programs or tariffs are in place or something else happens. And you've gotta, you've gotta be, you've gotta have not just cyber resiliency, but it resiliency and right.

And that's one of the great things about cloud is it gives you that expandability, burstability, mobility, all the, I ities I guess to, to, to handle that kind of stuff. And that goes hand in hand with this sovereignty issue, which is in an increasing, in a world where increasingly the internet is being balkanized, right? It's no more Yugoslavia.

Um, we need the ability to really segregate, and I hate to use that word, but to segregate our data and infrastructure into the jurisdictions that political, socioeconomic, political dictate where they need to be. And so that's part of the world we live in today, right? You've gotta have a sovereignty.

And it's not just data, it's it's infrastructure. It's, it's everything. You've gotta have a sovereignty, uh, strategy that plays in, in a balkanized world that, that we're, you know, multipolar world that we're living in.

Yeah. And those were topics. Yeah, those were topics that came up in the discussions at at, at cloud field.

And some of the more interesting parts, you know, they're talking to Commvault who's talking about data protection and, and, and business continuity. And, um, and the delegates are starting to say, wait, could this be used for migrations? Could this be used to move data from?

And, and absolutely it could. The same thing came up with scalability. Uh, you know, certainly that's one of the things that, uh, you know, HPE is really keen on with a lot of the things that they've been trying to work on for customers.

It's about where you, where you run is no longer a simply a question of cost. It's a, it has a lot of different factors. Yeah.

Kimberly, Kimberly, I'd love to get your opinion on the third topic though, that Steven brought up. 'cause we talked about this on yesterday's show. If we have AI agents that are constantly using infrastructure, well, we already know that infrastructure kind of tends to keel over when utilization rates get too high.

So do we need a whole different set of infrastructure to accommodate all these AI agents that are gonna be constantly processing data? I don't think the AI agents are taking that much p processing power to execute. Um, they are, you know, there are iOS and that kind of piece of it, but I, I would, I'm gonna take it at l slightly different play.

I do wanna remind folks that 70% of the enterprise data is still on on-prem, right? And so while we're looking at hybrid, I mean, the hybrid piece of it is that part of what's going on with the CIOs are making some decisions about where to best place their applications. And so that piece is going on in terms of where that piece, just as JP is saying, if I need to more, if I need more power, if I need more bursting or I need more capabilities, yes, I have the cloud that's up there.

Um, so I think all of that piece of it, how it all fits together, and those, some of those decisions are not straightforward. And that's why I think you're seeing, you know, the vendors that are coming in is, they're not just talking cloud. They're not just talking on-prem.

They're talking at, how do I look at this in a unified way? How do I look at data? I, I did look at a little bit of the, um, cumula piece and like, how do I look at data that, so that wherever I'm at, wherever it's being processed, it, the speed and the application in response time is significant.

Is is at the level that I expected that goes as well for the cybersecurity cyber resiliency. How do I manage wherever the data is, wherever the application is to make sure that I am resilient and I can recover? So that complexity has grown as they're examining these pieces of it, but what they're looking at is I've got all these different tools that I can do and I can put those pieces together.

So I, Micah, I know I didn't directly address your, the question that you had there on the AI APIs, but I don't see those as taking significantly more transactional effort, um, to process, um, than we have in any other transactional kind of system that we're using. JP you wanted to say something? I I wanted to ask Steven a question about, um, the event itself, you know, you have all these great ideas being put forth, uh, oh, uh, cyber resiliency and addressing business continuity.

But the issue has been the business itself has not taken care of its business continuity over the years, right? In fact, there are many organizations that don't practice and test that BC plans on an annual basis. I remember sitting at a conference and having preface with someone from an insurance company who, I don't know how we got on the topic, but he is like, he, he said, yeah, if, if we had an incident, we wouldn't be back up for two months.

I mean, that is incredible to think about. Two months the company would be down if they had a major incident because they don't, they haven't tested the DR plans, the VC plans, they don't know if they work, they don't re, they wouldn't necessarily be able to restore in a timely fashion, you know, based upon what they had. And now you're gonna, so is is, did anybody talk about the other piece that goes with this, which is we have to rethink the organization, we have to think about who owns this and who's running this?

Yeah. And, and certainly that is a topic that people are aware of. Um, as somebody who just moved houses, it occurs to me that, um, you know, e every time you move, whether it's moving houses or moving your business in the event of a, of an outage, it's extremely disruptive.

And there's always something you need, something you forgot, right? Oh my goodness, I don't have any milk. Oh my goodness, I don't have any, I don't know, towels, you know, whatever it is.

When you're moving, you know, you, you find these things out. The only way to not have it be disruptive is essentially to live in two places all the time. Essentially to maintain two residences and to move between them, which I think some folks tend to do, uh, once they get a little bit older, businesses are doing the same thing.

Uh, I had a wise, uh, business associate, uh, talk to me about this about 10 years ago where they said, you know, we're, we are ready for any kind of disruption. And I said, oh, you're, you're fooling yourself as a a veteran like me, you and me, jp. You know, and I, I know that Kimberly, you know, you've seen this too over the years.

You know, you talk to these companies, you know about their DR plan and the dirty secret is their DR plan is basically panic and find a new job because there just ain't gonna be no way they're gonna recover. But this guy was completely confident and the reason he was gonna completely confident is because their company actively ran in two locations all the time and switched their nexus from one to the other on a weekly basis as part of their regular business practices. That's sort of the Commvault strategy here with their continuous business approach.

And frankly, that's the strategy I think of a lot of modern cloud infrastructure. You know, you look at how these things work. Um, I've been actively switching the location of all of the tech Field day websites on a weekly basis for two years now, and nobody has noticed, I don't think I've even mentioned it, it mo moves from the cloud to on-prem, uh, every week.

And, um, and it runs in both places. And so I'm a hundred percent confident that if the, if there's an outage, it will run smoothly and seamlessly wherever it happens to run, wherever it needs to run. And I think that that's what businesses are looking at too, which is why products, you know, you talked about Qumulo, you talked about S Scale, you talked about Min io, you talk about HPE, all these companies, I think their, their idea now is that it has to be continuous operation in multiple locations, not disaster recovery, which is elf is a disaster.

So, but I, go ahead Cam. So to that end, I think where we've come and migrated as we've gone through the cybersecurity five, six years ago, actually more than that, when we were interviewing people about their cybersecurity, um, capabilities, what they were doing, how they're architecting, we were pretty horrified on those interviews about how people thought, as you were saying Steven, they thought they were perfectly fine. And we, you know, we weren't, we were interviewing, we weren't advising.

And so we get off the phone and go, they're screwed. Uh, you know, all it takes is gonna be hit. So now we've come full, you know, and we've seen more and more companies getting hit or small companies or whatever.

So whether it's the state and local or, or the educational system or a bank or a credit union as, as I, you know, big credit union that, you know, went down, um, last summer, those pe that, that has a trickle effect into your other peers within that industry. And so then they start looking a little bit deeper. And then once you get hit, then is when unfortunately is when they start hardening things a hardening and then realizing I have to be recoverable and what am I doing to be recoverable?

So I think right to be resilient, so over a period of time. So it makes sense that we've shifted from protect to recoverability, um, that they finally have gotten the notice that it's gonna happen and, um, that it, you know, trickles up and down the executive staff to say, can we really actually do this? So the investment is being made and it still is at the number one investment area for the CIOs is cybersecurity in cyber resilience.

Absolutely. Alright. Hey, we gotta wrap things up here.

As Steven mentioned, tech Field Day was streamed live on Techstrong tv, but you could also, if you didn't catch it there, you can go check it out on the Tech Field Day YouTube site as well as on our Techstrong tv, TV site, our YouTube tech Strong tv, YouTube site, and our tech strong tv t app on Roku, Amazon, and Apple tv, where we have all the tech field day content up there now as well on an ongoing basis, as well as our tech strong tv, as well as some Signal six five content. So check it out there. Um, but for now we're gonna pull the plug on this episode of Textron Gang.

Can't wait to see what I dream about tonight. We can talk about tomorrow. Um, but Mike, Steven, JP and Kimberly, as well as Bonnie yeah, right here.

Thank you for joining us. Thank you for watching this. As usual, we have a full text, strong TV lineup following up, so stay tuned for that.

Until next time though, it is Alan Shimmel for Text Strong. We're out.