Techstrong Gang – March 27, 2025

What's AI going to do to bi or is it DYI? You're watching Text Drug Day. Hi everyone, it's Happy Thursday.

Alan Shimel. Happy, happy Thursday, and happy to be here today. Welcome to Textron Gang.

You know, it's not often we get visitors down here in Boca Raton. Well, we get a lot of visitors down here in Boca Raton. Who am I kidding?

This time of year I can't wait for them to go home. But it's not often. We get visitors here to our tech strong studio, and I'm really tickled to have Guy Courier who's a gang member.

You've seen him on here before, but usually he's in Austin, Texas, or gallivanting somewhere in the world today. He's gallivant down to Boca and he's here with us. So, first of all, guy, welcome.

Thank you. It's great to be here. In person, In person on the show Techstrong Studios.

Ah-huh Mm-hmm. So we're having, we're having some fun down here, but as usual, we have our, well, our gang's kind of concentrated today. We've got two in Boker and two in Colorado.

So let's go to our Colorado contingent. I don't know which one of you is at a higher elevation. I don't even want to guess Kimberly.

'cause she's near Boulder, right? Boulder's higher than Mitch. Near the closer to the airport.

So Kimberly Bates, future analyst, always a, a bonafide gang member here. Hey, Kimberly. How are you?

Uh, good morning. From, um, we're 70 degrees just like you guys there. Booker Raton today.

Good. It's The same temperature all over the United States. Yeah.

It, it's being controlled. Did you ever see that show Paradise? Mm-hmm.

Yeah. They control the temperature. Yeah.

Inside the, the mountain. It's a great show. If you've ever, Mitch, have you seen that one?

I've started watching it. That's a great series. Anyway, speaking of Mitch, also in Colorado, but he's, was it northeast Denver?

I'm, I'm coming to you from the low lands of the front range of Colorado. How's that? All right.

The low ends of the front range, we're the buffalo roam. They still roam. I thought the buffs are all at uc, Boulder now.

It's pretty much just cattle or C Boulder. Uh, that's just cattle. Anyway, fu VP Analyst Mitchell.

Ashley. Hey, Mitch. How are you?

Good. Good, good. And then he's all by himself in the Empire State still camp outside of Yankee Stadium, waiting for opening day.

He's our chief content officer, Mike Ard. Hey, Mike, how are you? Good.

Five hours from now, I will be sitting in the upper decks behind home plate at Yankee Stadium, assuming my usual position. And now's a good time for me to tell you that I won't be making that three o'clock meeting with you today. Yeah, Probably not.

We, uh, well, you know what, when we looked at the baseball schedule this year, we, we kind of deleted you from any meeting. So we, we knew where you were gonna go. Or B Max Reed's opening, right?

Or no? I don't know who's opening. I gotta go look.

I thought it was Redan. It might be Redan. You're right.

It's Carlos Redan. It's not Max Street. Well, So, so little maybe little known fact.

We don't open here in Denver until April. Mm-hmm. Because it's snow.

Well, That's prudent. But who knew it was gonna be 70 degrees? But again, that's the weather in, in Colorado.

Right. Never know what to expect. Anyway, hey, let's jump in.

I, I, I mentioned it before, AI transforming bi. Do you DYI, what, what are we doing with all these eyes, Mike? What do we got here?

All right, well stay with me on this one. But AWS has this BI application service called Q Insight, and they added some AI capabilities to that late last year. And now, yesterday or the day before, they, um, added the ability to create what if scenarios.

So rather than necessarily you have to upload all this stuff and find some business analyst to go and code all that stuff, you could do it yourself. Of course, business analysts might use this to lighten their own load. But the interesting part of it to me is you can interactively do analysis by just changing a couple of parameters in a spreadsheet or whatever data you uploaded.

And you don't necessarily need a business analyst to go all do this. And you can think about how I can model different things, or I could change the, uh, you know, the discounts that we're gonna set and see what the impact that might have on the profit and loss statement. And, and it just gives people a lot more control over planning.

But Guy, I know you've been in this space a long time, and I know also we've had what if capabilities and other BI apps. I think we're just kind of getting to the point now where maybe we can democratize this. Uh, we definitely are democratizing it.

Um, it always amuses me to hear how you can cut the professional out. In this case the business analyst and everything's gonna go just fine. Um, I do think that it allows for the business analyst to be a lot more productive.

Uh, they're a lot of, we, you might not call them rote exactly, but let's say repetitive, but highly skilled, um, functions that business analysts, uh, uh, have to do. Um, it's not just, you know, maintaining, um, you know, the analytics and the dashboards and, and you know, the pipeline of information and reports that they're constantly producing. Um, while handling the host of custom requests, there's a lot of elements of that that are repe are repetitive.

They create templates, they have notes, they have process sheets. There's lots of ways that they have had to accomplish this in the past. And like a lot of promising areas, the most promising and the most encouraging areas of ai, um, where you can turn to AI to do this.

It never gets tired. It's always quick. The more it does it, the better it gets at it.

You know, the more productive those business analysts will be. Um, can the developers, um, and the data scientists create interactive systems where, um, just like, let's say in the platform engineering world, um, the, instead of the developers, it's end users, business users are interacting with the platform and they don't have to deal with pesky people as much who are still behind the scenes keeping things going. Yeah, I think that's where it's going.

But talking about cutting out the business analyst, I think misses the point of democratization, which is, um, so q uh, came out I think in 2020 or something like that. And it was a sort of a precursor in that that was common in business analytics and business intelligence of use, being able to do natural language queries where you could, you know, enter in plain language what it was you wanted to look at. And the system knew how to produce, uh, charts close to or at what you wanted to do.

Now, AI helps take that to the next level and add additional functionality and capabilities that the business analysts can take advantage of as well as the developers. And that is the what is gonna create this, this sort of democratization I think that that, that you refer to. But no, we're not cutting out the business analyst.

I worked in research years ago. You, you, you, that's actually when you and I met and it always amused me how, uh, uh, folks, we were doing surveys for, thought they could write the surveys without the use of a professional. You can't really do this properly without the use of the professionals still.

Mm-hmm. I think the key word there was properly. Properly, yes.

Thank you. Should have just said properly and we would've all saved a lot Of time. So of the things, one of the things I do like about this though is that they are letting you watch the steps that are used to create the outcome and Observability.

Yeah. Of observability and auditability. Yeah.

Go on. Yeah. One Of the issues I've always had with business analysts is, you know, they're, they're like human black boxes.

I don't really know what went into their output that they created. So I don't know whether I'm gonna trust it or not. And so there's always this question that people have.

It's like, so here comes the report from the business analyst and we don't know what went into, uh, waiting that. And so we always kind of like, when the outcome doesn't kind of align with what our hopes were, we always get a little suspicious. But, but there's the other piece, the other issue that goes along with this piece of it in terms of like, let's talk about a financial analyst that's doing that.

Either business analytics. And so often we have systems that don't match. I mean, I've been, we've all been in companies, big companies especially, somebody walks in with the truth piece and somebody else walks in with the truth piece and they don't match up.

And so you're asking, well, why? If he's asking the same question and getting different answers, because different systems have different numbers, then can I ask the AI to kind of resolve about why these numbers are different that might be beneficial to me, or kind of guess. Mm-hmm.

Or maybe I'd have to ask a third AI agent to go do that, to check the work of the other two. Right. Well, and for a penny and for a pound, I guess.

And, and just, yeah, it just some, So, so Alan, the point of this whole exercise is allegedly, so business executives will make more quote unquote fact-based decisions versus, uh, intuitive decisions that might be based on their, uh, experience or what they had for lunch. 'cause they have a case of indigestion. So, um, do you think that we will get more back based decisions that are better?

Or is there always gonna be some level of intuition involved? So, I probably have an unorthodox view on this. I, I think, I think quite frankly, a lot of executives use these types of business intelligence, uh, research to reinforce their gut feeling, right?

And they tend to discount things that don't reinforce. And, and it's quite frankly, it's a, a lot of what goes on in our country today, right? We, we, we, we tend to read news that reinforces our tribal beliefs and, and tend to shun news sources that don't reinforce those beliefs.

I, and I think that happens with analysts too, right? I, I want, and, and to be fair, I think that when the analysts, some, and I'm not impinging the integrity of analysts guy, so, and Mitch and so forth and camp, but I think the analyst sometimes gets their cues from the customer about what the end result we'd like to see or what, what our beliefs are, prove me wrong or reinforce it. And, and so I, I, you know, I, I always cast a suspicious eye at these things because, you know, is the outcome baked beforehand, right?

Well, so maybe AI will stop. Yeah. Does this change corporate culture or institutional culture where politics and lines of power and that sort of thing, like you, you referred to two different elements of this, and I didn't do like, you know, whatever they call social engineering or whatever it is in, in B school.

But, um, there's two different elements here. One is the researcher or analyst, um, understanding that the mission is to help support a decision that's already been made or a point of view and for political or whatever reasons, you know, helping out with that effort as opposed to exposing the, you know, quote unquote truth. Um, and then the other one is how decisions were made at all in the first place, which seem often to be based on last week's meeting that was really important, where something happened and the executive or person in power had an insight and brought that out.

And then they're picking and choosing among a lot of facts, the ones that support their case. Those could be affected in some way that, uh, is a little hard to see by the advent of ai. Uh, but they'll still be there.

There's, as long as corporations are made of people and not of AI agents, and who knows, maybe when artificial general intelligence comes along, it'll be replaced by a different set of politics and forces. So, Well, I think this means that, uh, natural language is, is the new sequel. This is, we don't need every person to have to know sql, just like we did.

Don't need every person to know how Very good point, including the analysts Yeah. To Get. Yeah.

Uh, but this is, this is the outer layer of what data analysts do, right? Getting the information out of the systems is one part of it. But there's building models, there's aggregating data, there's doing analysis, so you can give answers.

Not e executives are not gonna go in out there and build complicated, sophisticated models. That's what data engineers are about. Maybe the business analysts are now more data engineers and, and require that requirements have kind of shifted of what they need to do.

So they can produce it in whatever form. Here's our traditional reports, here's the things that feed into the GL general ledger. Here's things that feed into business analytics for executives or product managers, or whatever it might be.

It's the accessibility to that information that's now through natural language, and it'll go deeper and deeper, but it's kind of like, um, Mike going to the Yankees game does not make a Yankees game. There still has to be a Yankees game that happens that he can watch. Well, Not if you listen to Mike, but Mitch, let me ask you a question.

Where do the data scientists fit into this one? Well, da data scientists, data engineers, a lot of people are arguing data engineer. Uh, job is now an ai AI data engineer.

They're crossing the lines into it. And this is a perfect example of that, where it isn't just about building the models, um, you know, whether it's for machine learning requires more curation of data to be able to get the kind of reliable answers that you want out of it. Generative ai, less so because now you're asking questions of it producing results.

You know, I, I can, I remember doing banking reports, right? Whenever somebody would ask for new reports, uh, in, in one of the departments, inevitably I'd get a call and say, why does my number not equals Alan's number over in the credit department? What's, what's up with this report?

Well, they both say balance, but they're two different things, right? So there's, there's, there's the meaning and the structure and the context of the data. That is what really has to happen to not only build the models, but also the IA that's gonna serve it.

You know, I wondering if, as we go down this path here, the, the principle of five why's becomes really, really, really more critical than it has been in terms of why are we seeing this? Why are we seeing this? Why are we seeing this?

To bring that critical thinking back into that the data that we have presented to us, um, and making sure that we're not, you know, looking at it in the wrong way. There's transformation of data, right? Account means something else in than it does in this system, right?

So you have, um, you have to do all that translation, transformation of data, data management. There's still a ton of work to do. So business analysts may not have to program SQL anymore, but they still need to make sure that whether it's an executive writing a simple query, or someone doing a more sophisticated query, extracting that data out of the, out of the interface and putting into their spreadsheet that they're getting the right data.

You know, this, this is an, so I know you've been banging on this drum for as long as I can remember, but are you hopeful that as we kinda look at all this AI capability and we're not dependent upon sql, that maybe just maybe we might have increased data literacy among our executives? No, Definitively, that was a very guy answer to the question. No.

Uh, what would, what would cause data literacy? I mean, in what way? Would, would, would you tell us, Alan, hey, what elevates someone to the executive mantle vp or whatever, based on data literacy?

Uh, I think talking the talk, understanding, showing results, and, um, showing how, ooh, we knew how to leverage AI and use it to get these amazing results. Sure. I think that will help.

But by my gesticulations, you can see that it's, it's, it's a lot about what it was always about politics, sales, sales, personship, um, ability to motivate people, that sort of thing. So I, I do think, I do think, uh, this is a really interesting, um, area to look at the effective ai because it is so pervasive. I mean, think about it.

Um, a lot of, uh, bibas, uh, uh, work, um, has been based on predictive analytics, which is being really changed by a lot of different AI models accelerated. Um, you, you get a lot more, uh, possible models to look at to begin with, a lot more lines of inquiry thanks to ai. And so you are a lot more likely to come up with, through human intervention to the intervention of the analyst with what are promising paths or what are good and correct, like truths, possible truths to, to Kimberly's point, there can be more than one.

Um, but that also, uh, uh, its pervasiveness means that our attention needs to be a lot higher all along the way. It seems to me, um, one of the great things that business analysts do, and this this touches on what Mitch was talking about, um, they, they kind of smell, they, they can, they can kind of tell when there's something funny in the results that they're getting, and they can go back into, make demands back on the data side to say, Hey, you know, do we have a complete set? Do we have a clean set?

There's something weird here. And I, I think it's a really interesting potential problem that with the intervention of ai, which is trained to look true, whether it is or not, that might be more difficult. And that might raise the, the, the literacy level of the whole enterprise.

But of executives, no. Well, Or I don't think so. Anyway, Let, let me wrap a bo a bow on this.

And Mike, note to self here, you, you don't pull up, you know, the, the death of analysts when we have three analysts, people on the panel. I dunno what you were expecting. Um, we know people in New York too, Mike, But I just wanted to see who was gonna jump on that grenade.

Alright. Um, but, but that being said, look, I I, I do think this is a, a use case where AI can add real value. I don't think it's going to be the end of analyst or business or bi, you know, an analyst.

I, I think the better analysts will learn to leverage ai. Mm-hmm. Just like better programmers will learn to leverage ai, better marketers will learn to leverage ai.

Right? Let, let's not, the sky's not falling in the end of the analyst space. But let's take a break here on Textron gang.

We're gonna come back and we're going to talk about AI meets platform engineering. Is this the, is platform engineering dead? No.

All right. You're watching Textron Gang. Discover Textron Group, the epicenter of tech innovation.

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

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

com that I encourage you to check out. And it talks about how AI agents will be applied to platform engineering. I'm not entirely clear, this means the end of platform engineering as much.

Maybe it just means that everybody can do it. But Mitch, you've been following AI platform engineering, DevOps, GI ops, and every other form of ops that I can possibly remember. Um, what is your assessment here of how AI agents might be applied to platform engineer?

Well, it's interesting, Mike. We just launched, uh, the new dataset for DevOps and, and application development practice and future and analyst business. And it was, it was very close actually.

Platform engineering teams said that they use, uh, AI in their work. Actually, a couple percentage points more than developers, people doing DevOps and developers did, which I was pretty surprised about. Now, that can be anywhere on the spectrum.

This article is a great example of that, where they were essentially kind of doing some user testing. Developers saying, okay, show us how you use this. Now that we've got some generative AI to do searching and explanations, uh, some automation to build environments.

And they actually went through a whole scenario and said, wow, this is really cool. It's amazing the developers are doing more with this than we even thought they might do. Shocking, that's what developers do.

But I think what it, what it goes to is kind of what we were talking about in our last segment, the approachability and the access to this is changing. It's not a wiki anymore or, or you know, the tools on the rack on the tool shelf. And you have to go kind of pick what I do and go searching for documentation.

They actually use some metrics, uh, to look at like, how much time did it take them to complete the task? How long did they have to go looking up documentation, reading how to do it, how long did it take to set up? All that kind of thing, compared it with and without ai.

In some areas they say a lot of improvement, you know, it was about a 30% improvement in just the overall task and a couple of things. But documentation was substantial, you know, cutting that. They have to do a lot of reading.

They could query it. So again, it's, it's making the tools, the platform engineering is putting together the, the platforms, the configurations, containers, environments, all of that stuff more accessible as opposed to every time we do something, the wiki gets harder and harder to find things. 'cause it gets more complex and filled with stuff.

So, did I hear Mitchell say platform engineers is somewhere on the spectrum That explains a lot. Aren't aren't we all Alan? Yeah.

Speak for yourself. I'll raise my hand. No, not at all.

Okay. So Mitch, how hard is it to set this up in your mind at this point? Because, you know, when I was looking at the article, there were a lot of coding examples and it still didn't feel to me like it was turnkey, but they were talking more about there'd be self-service capabilities and the IDP that I can invoke through the AI agent.

But, um, where are we on the journey? I think we're in this constant thread of improving and automating. And because you're, you're end users, here are people who develop software, right?

Who will develop their own tools, you know, if they don't get what they want from somewhere else. So I think that the platform engineers can look at this as an opportunity to not only create things for developers that use AI to make it make it easier to get those kind of tasks done, but also see what developers do with it, right? Incorporate stuff that they may build something with agents, uh, to help their individual work or something on their team that can be rolled back into, uh, what platform engineers provide to the rest of the organization.

So I think we'll see this, you know, ongoing, pretty rapidly moving incremental process of more and more AI driving or facilitating the tasks that developers are performing, test engineers are performing, uh, DevOps people operations coming off of, you know, the work of the platform engineers, but maybe more largely the whole community contributing that back. So, Mitch, question I have for you is, how do you, how are you separating out platform engineering for IT operations and AIOps? Because, you know, there's quite a bit of engineering going in by the vendors into developing AI operations that have to do with the infrastructure pieces of it.

And I'm not as, um, up to speed on the DevOps thing. That's not my, my area of expertise, but kind of where do you see one of those and ending in the other beginning and how does this reflect into the kind of things that you were just talking about? Well, your, your question is perfectly positioned because the line is very blurry and is getting blurrier all the time.

AI ops is, GE general definition is using AI in operations for whatever process, data ops, same kind of thing. Um, and as we move, you know, in a good GI op sense we were talking about earlier that Mike mentioned, that's things flowing all the way from development to test into production environments and supporting an in, in a production environment. So I, I can very easily see, you know, the tools that we want to use in, in production like observability, we're pushing back up and we want to use more of that and development vice versa.

The things that we want to be able to do, observability in production, we can facilitate better by what we incorporate into our workflows. And, uh, AI to do, you know, triage and analysis of, uh, telemetry data. Just to pick one example up and down the chain.

I'm not saying it's all gonna be one continuous fluid thing and it, and you can't tell the difference between anything. But I think there's a lot of dots that can be connected across that full lifecycle. Yeah.

And from my perspective, the infrastructure side of the house, one of the things we saw attempting to happen, I think maybe it is a little bit, is pushing down things like data protection, um, you know, for this application or security protocols for this application to the DevOps people. And then we started sliding it back into saying the DevOps people don't really wanna do anything with this. Um, it's, it's, it's another task to do that they don't wanna be responsible for.

They wanna just, you know, develop. And so then you have this concept of moving from a dev operational environment to the platform engineering. So that's kind of what I, you know, kind of looking at is that you're talking about once you deploy, now you really have to address the security and data protection, the governance issues.

So, Well, we talk about, you know, developers being the customer for platform engineering, it's really a lot more the right data engineers is operations 'cause they use the platforms, right? That, that platform engineers creating. So they're kind of in, like DevOps was in the middle of the development flow, right?

All the way to deployment. Platform engineering is in the middle of the flow upstream, downstream. And then that's right.

I, I think you gotta go back to what gave rise to this platform engineering movement that we, we see today. com and we have a thing or two on there about this and including the article that gave rise to this. We also have a podcast called the Platform Engineering show that I co-host with Luca Ante.

org site, which is two, 300,000 people strong. What, where, where the platform engineering came from. Well, as part of that DevOps movement, you know, this whole concept of shift left came into being, and we're just gonna shift left.

We're gonna shift security left, we're gonna shift testing left, we are going to shift deployment left, we're gonna shift integration left. We are going to shift building the very platform that developers develop on left. And so it's okay if you're are a small startup with tenor a dozen developers, but when you get to a large enterprise where you've got thousands of people involved here, it, it gets messy quickly.

And, and I think a lesson we've learned in the DevOps movement over the last 10, 12 years is that we can't just shift everything left onto the developer. The developer's probably the highest paid person in that food chain. And you know, what developers like to do develop.

And so when we ask them to be security pros, when we ask them to be testers, when we ask them to be platform people, when we ask them to be ops people, you're taking away from what they do in Dev. And, and so platform engineering is an outgrowth of this. And the idea here is let's give the developer the platform that they need to go as fast as they can in a secure manner, guardrails, all of that good stuff, right?

And, and so that's the mission of platform engineering. Now how can AI help with that? I don't think it's the Yes, Mitch, you're right, everyone will use AI up and down this food chain, right?

But there's no reason that the platform engineers aren't using AI to develop better platforms. And part of that, I envision and I I am gonna explore this on a future platform engineering show plug, we'll be at Q Con live recording some of these. Mitch, you'll be there.

Hopefully you jump in on this. Is these platform engineers should be creating agents for their platforms that the developers can use to make their life easier for the testers and for the security people. Right?

Let's, you know, these platform engineers are building the rollercoaster. Let's go as fast as we can on, well, maybe not a rollercoaster, I don't want to go, but they're building the highway, right? They're building the highway and you know, we wanna make it an auto box.

Yeah. Mitch, I'd love to get your input on this. 'cause you know, I have this conversation and I make that description that Alan just gave maybe a little shorter, but, um, Was that Dick Mike I see today it's Yankee's baseball.

Yeah. I'm sorry. You're gonna be late for the first pitch.

But when I talk to IT people, I get this vibe all the time, right? It's like, yeah, have you heard of platform engineering? No.

What is it? Describe it. I describe it to them and then they look at me and they just nod their head and said, yeah, we do that.

Yeah. It reminds me the day, the original days of cloud. So, so I was, I was, uh, I was a, uh, marketer, um, marketing, uh, cloud solutions, uh, or cloud capabilities, what have you at a, at, at, at a big vendor.

And um, this was in the earlier days of cloud. And, uh, you go in the field and customer after customer after customer would tell you, yeah, we have a cloud here. We're doing a cloud there, we're doing a cloud there.

And, and at the time, I, you know, I was like, I'm, I've never really been a purist, but I was like, look, is it on demand? Is it self services? Is it this that, right.

We had the NIST definition and all that other kind of fun stuff. And it turned out what they had was a, a virtualized therapy. Virtualized.

Yeah. That's The, they they called it a collapse. But We, we, we went through this, Mitch, we went through this in DevOps.

Mm-hmm. Nothing new in DevOps. DevSecOps of course, it's always got security.

You don't need that. The second there, it's, it's, well, Before we had DevOps, we had a CIS admin who did the build server. Yeah, yeah.

Or we had CICD, right? I mean, it's an evolution of that role just as more automation is done. And it's, it's a common Story.

Okay. But to, to Mike's point, I did a recent kind of tour through the enterprise talking to, uh, not so much IT as, as enterprise, like business analyst types, um, uh, IT business analyst types. Those are the folks who are delivering, um, the IT in the way each of the business units need, you know, whatever it might be.

And, uh, um, and I gotta tell you, there's just a ton of development and compute and platforming going on out there that would not resemble, that does not resemble what we call platform engineering right now. So, so in some ways this is, this is, it's, it's, I I take the parallel with cloud. They were calling it cloud back when it was virtualization, still calling it cloud only.

Now it is cloud, it's multi-cloud. It's, it's, it's hybrid cloud. It's all of these things that is the paradigm.

And so I think platform engineering, I used to say way back in the day, what we call cloud computing. Now we're just gonna call computing. That didn't really turn out that way.

But in the same way, what we call platform engineering right now, we're just gonna call the infrastructure later on op No, no, I don't. I, I, I mean ops. So what are platform, what are platform engineers?

They're coders, right? Right. Tell me I'm wrong.

They're coders, they're coding necessarily coders, the security developers, the networking developers, the, you know, the integration developers, all of those folks, they're sort of coming together to take that. It's like, it's, it's still shift left, so to speak, but what they're producing is a platform for the rest of the pipeline. No.

So am I, it's Interesting that you're calling them a coder because I see them as systems people. Me too. I see them as ops Is coding.

Now I don't, I, I don't know. Well, infrastructure is code that, okay. That's, But it's not because Infras infrastructure is code, but you still gotta sit on something and that something has to do with the hardware piece of it.

So you Mean eyes on glass, right? Like the people. But, But you know, going back to our previous block of AI transforming bi, you know, and we said, you know, executives sort of have their predefined motions.

It's the same thing here. org has two to 300,000 members. Platform Con, they're virtual and in-person event, they're expecting 40,000 people signing up for IT.

Platform engineering in CubeCon next week on, uh, Tuesday zero day is the biggest single satellite conference going on. org through the roof, the market is responding. It, it's, it's clearly aimed at enterprise, right?

You, you may not need this if you're not at that large team level, but I'm gonna, I'm gonna back up what support, what guy was saying is, you know, coding is, is a spectrum, right? There's, there's developing software applications, et cetera. But is is a bash shell is is a shell script code.

Yeah. Um, is writing some Python, pearl, whatever to automate something that's coding. So yeah, I mean, CIS admins, people in operations, people in cloud environments, you know, running the cloud, some people don't do that, but very much there isn't as much a separation of of that's not coding.

This is writing. You have to write in c plus plus to be coding. No, I mean, there's a lot of scripting is what we would consider coding today.

I don't know if I would exclude, you know, somebody with a graphical tool managing VMware installation. That's a platform. And that's engineering, right?

Developing an agent. Is that coding? Well, is it if it's an operations or sys admin or or platform engineer to do It's What about, what about if I have the AI do the, the scripting or coding And it's a developer?

Yeah. Well you, hopefully you're taking a look at it. And, uh, But you know what?

Gene Kim something, gene Kim Phoenix project, you know, celebrity and the DevOps thing. Hi. I don't know if you guys follow him or if Gene's a great guy, you should follow him.

But, you know, he's on this mission these days of he fancies himself a coder. He's never coded. Gene was never a coder.

Gene's a more of a business guy, but he's using AI and he's generating code. He does it to your point. He doesn't know if it's good, bad, or indifferent.

Okay. Okay. So hold on.

So are you saying Vibe coding is coming to platform engineering? Is that what you're saying? Well, I Think vibe Coding's coming.

That sounds great. Good vibrations. We, we don't have anyone from California.

We're finally talking about good vibrations and we don't have one of our California people lot Pull out that guitar. Mitch. Yeah.

Pull it up. Let me hear you say start doing the Beach Boys. Yeah.

A little beach Boy. Help me, Rhonda. Um, anyway, um, let, let's take a break on this one at this point 'cause we ran a little over.

We're gonna be back with our next block. And, uh, is the cloud guy you've already touched on. This is the struggle real in cloud.

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 and we're talking about cloud computing, the, kind of complete the trifecta here a little bit. Um, one of the issues that we're starting to see out there is that the number of workloads in the cloud has finally kind of moved above that 50% mark as a percentage of the total, at least according to our latest report from flexera that you can find on Techstrong itts m soon to be techstrong. It.

And it kind of suggests that we are now struggling with managing all those workloads, optimizing, and people are very concerned about the cost of the cloud. And, uh, we're also still struggling with how to provision this stuff. There's a separate report from Space Lift talking about how it might take people in some instances a week still the provision cloud infrastructure by the time they get all the processes done and everything nailed down.

Kimberly, we've been playing around now with Cloud for a decade or more and, and it feels like we're kind of still wrapping our arms around this thing. So where are we on this particular adventure and why does it seem like it's still more difficult than it should be? Well, about two years ago, as I think is when we started the trending back from taking a look at saying I'm cloud first with a lot of comp with companies to the point that I'm doing a very, very, um, thoughtful process of where do I place a particular workload.

And that has continued, um, this year, especially as they're looking at what the budgets look like, et cetera. What we've seen all along is that once you get to a very stable application, that stable application is probably best served and operated on prem. Um, because I don't have the, I don't need to take the advantage of some of the things that the cloud brings brings me, which is a lot of flexibility, et cetera.

So there's some cost equations there. And then if you add in the pressures that it is feeling today, the CIO is feeling today where they, the budgets are continue to be pinched, and they're even pinched more now because of the AI initiatives. So I've taken budgets that used to go into whatever I am doing, and I'm moving it into AI because we've gotta go out that way.

So the other areas have got to be lookeded at, you know, very, with scrutiny to say, where can I save some money? Um, in order to maintain. So, so some of those transitions and changes that are going on, um, in terms of where I place my application when I'm using those applications are, are becoming part of all those pressures.

And, um, and then you, then you have some of the, there's some, you know, big case studies that, that people are out there that have talked about how they've saved millions and millions, hundreds of millions of dollars by bringing their, you know, a applications back on-prem. Um, and so that's gonna cause people to say, wait a second. You know, should I be doing this as well?

And examining what, what decisions they're making. Um, so I think that's part of it. Um, the, the last thing I'll mention is that the second half the how long it takes you to provision up in the cloud, um, if you're provisioning, if you're setting up a well architected environment that has got high availability, it's not dependent on one zone of whatever, because we've all seen the stuff go down and you know, you're out for whatever, you know, you, you have to set up a very well orchestrated or architected environment that takes time to do.

And it's not just a pure cookie cutter kind of process, maybe that some people think it is. Um, so that takes time to set up. And if you're all you're go doing is DevOps, sorry Mitch, all you're doing is DevOps and writing code and I don't need to worry about high availability, et cetera, or data protection or some of those other things.

Yeah, I'm, I'm gonna be up and going, but if I'm, if I'm delivering a transactional system, different game, I, I have some thoughts on this one. So Kimberly, I think you're a hundred percent right talking about how quickly you could stand up an instance really depends on what you're standing up, your experience level of standing up what your requirements are and all of that. And that's gonna vary.

Your mileage varies when it comes to cloud costs. I think there's two different kind of scenarios. One is, am I being wasteful with my cloud usage?

Am I being a resource pick? Am I not shutting down an instance when I'm done with it? Am I adding more capacity than I really need and not bringing it back down?

Am I just blowing the balloon up and never shrinking it back? That's the kind of thing that I think finops can really help with and, and try to show you where you're just, you're throwing money out out the window. The other issue though is when are you too big for the cloud?

You know, I remember about 12 years ago I was at the Boulder Dorado Hotel, downtown Boulder. I know both of you know that Michael probably, Mike probably knows it too. Um, it was a, uh, it was a Foundry portfolio and Techstrong, not Techstrong, we're Tech Trunk, Techstars, uh, portfolio, uh, conference.

com. But Mitch, you remember Rally software? Kimberly, you probably do too, from Boulder Rally.

I knew the executives there. Yeah. Yeah.

So the t he was the original CTO at rally. Mm-hmm. He got up and this was 12 years ago, he got up and gave a, a presentation on exactly when the dollar point was and compute Point was where it didn't pay to use the cloud, that you were better off private data, centering it.

You're laughing, Cameron, do you think he was wrong or, or what? No, I, I mean I, I I, we, we can see that, I mean, when you, it depends upon, you're talking about size of company, I don't know about that. No, No.

It wasn't size of company was size of cloud spend and how big your infrastructure for, For, well, yeah, for a workload or For right, what your footprint, what your cloud footprint was and when it would pay to move back in, into a a, a private data center situation. And I would assume that there are some certain characteristics, not just a size situation that goes along with that. Um, it, it may be kind of like, so one of the areas that they haven't found a whole lot of, um, financial efficiency is in the data storage side of the house.

Um, that's one of the, that that's the big buck, the bucket that keeps on growing that never shrinks. And that, you know, Amazon keeps pounding over the head telling us, here are all the things you can do to make yourself more efficient. Nobody does.

Uh, so I mean, I'm so, I'm so glad you brought, I'm so glad you brought that up. Uh, that's been on my mind, Kimberly. So, so that is a really decisive factor is where the data is.

It, it is enormously expensive, um, to, uh, move data from one place to another now because of, because of not just how big it is, but because it's in mixed formats. Usually there, there's every given application that, not everyone but many applications, especially enterprise applications, use a lot of different data source data sources. There's an architecture and so on.

Um, for years, uh, I think, you know, the, the, the cloud service providers, uh, made it easy to bring the data in and then they've made it hard to take the data out. And so that is a really important cost factor. The kinds of charts you're talking about.

I used to see them too. I think that you're talking about, I used to see them too, I think refer to a large degree on how much it costs to run power, you know, uh, host manage and all those, you know, total cost of running this infrastructure. But it does not take account of data gravity or of the, you know, the need to right now.

There's just this big, really big movement to bring the processing next to the data instead of trying to bring the data next to the processing. The good news is these on-prem environments, wherever it is now, as we were just saying in the last segment, is cloud, almost always cloud in one way or another. And the, the cloud service providers are pri providing, you know, uh, on-prem extensions, um, of cloud of their cloud environments and vice versa.

I mean, Oracle's been doing this really interesting work lately. 'cause that's like a major data, you know, repository in most enterprises as well as enterprise applications and so forth. So they are putting infrastructure into all three of the big clouds, um, so that you can access your oracle from those clouds and then vice versa, you can access those cloud services and back.

So that is almost the window out if you ask me. But, you know, as far as cost and usage ballooning these I that, you know, I, I confess to being a little uncertain. I feel like we're just back in the old days of, uh, of, uh, runaway costs and bloat.

I think finops is, is very much a way out of it. It's just poorly adopted. Yeah.

It's, it's cloud hygiene is what, you know, that's a cloud hygiene problem to me, which Is different. I think you're trying to apply rational thought to irrational behavior. The irrational behavior is a bunch of CIOs thought it would be cool to say cloud only, and then they got rid of all these data centers and, and colo facilities because, you know, that was fashionable.

And then, then they were surprised when developers showed up, much like drunks at an open bar and said, look at all this compute that we can use. And they didn't really think about much about exactly how much that cost, because to them it was free. So I think, you know, we're just kind of paying the, the, the bill finally for behavior.

Then we incentivized in reasons that have nothing to do with logic. It's also difficult to turn it off. Mm-hmm.

Um, so I mean, my, my experience, little itty bitty company, you know, my, my company here and that we were using the cloud for doing a lot of things and turning things off was default. And so when you come to the finops, they're doing this analysis now I gotta go and figure out what that is. So maybe getting back to the first conversation we had, can AI do the business analytics on the Finops?

I did not wanna be the one to say ai, but Yes. No, no, but, and can it also make it worse? Agents going off the AI agent that shut shuts down instances or Sets up, Sets up, you know, ways I look, I, but I think that's the future.

I, I do think that's, that's a way out of that. Uh, all I gotta say is my work is done here and I'm happy when a theme comes together. I wanna take what you said, Mike, kind of a step further in, uh, you know, I appreciate you, you must have been in one of our hackathons talking about all the drunken developers.

But anyway, that's another, No, it wasn't the developer showing up drunks. It was the drunks showing up. It was the platform engineers.

They're on the spectrum anyway, Those platform engineers. No, it's, it's, you know, if a resources, they're all use it, right? Oh, spin up another server, spin up another whatever.

If I can do that, you know, it doesn't cost me anything. There's also, you know, just like, you know, the, the resource itself, the compute. Do I build an application that's high performance?

Whether you take, for example, should I build it in Python or maybe I should build it in Rust 'cause I can build a high performance application. It's not as kind of flexible and easy to use as, as Python, but maybe I should do things a bit differently for so high performance applications to get the most efficient use out of what we're creating. So there's the resource itself, and then there's the way we use it, you Campbell's point about data, right?

99 terabytes. Mm. Same thing for CPU and then we'll upgrade the drive.

Same thing with my clothes closet. There you go. I, I you Said it.

Fill up, meet the space, right? You said it. All right guys.

It's been a great, I like I, I hear somebody calling beer here. I gotta go. You gotta go get hot dog.

Hot dog here. Um, anyway, Mike, enjoy the game. Good luck to our Yankees, b***h Ley, enjoy the beautiful weather out in Colorado.

Guy, it's a pleasure having you here. Pleasure being Here. Thank you.

And thank you for watching. As a reminder, as always, Textron Gang is followed by the rest of our great Textron TV lineup for today. So stay tuned to this channel.

If you're watching it's streaming, uh, today. If not, you can find Textron Gang as well as 8,000 odd other videos on Textron tv or our Textron TV YouTube channel, and soon coming to a TV near you on our Textron tv o TT channel. But until, I guess tomorrow, this is Alan Shimel for Textron Gang.

Have a great day, everyone. We're out.