AI is Getting Real – Techstrong Research Review EP 29

Hi everybody. Mike Rothman here, general Manager of Textile Research with another episode of the Drum Roll, please, tech Strong Research Review. I mitch, your, your drum ball.

I'm joined as always by Mitch, my evidently silent drumming partner, uh, over there. Mitch, how are you doing? Good week.

You know, I know you were in Boca this week, so you got to, uh, visit the mothership. I did. I did.

It was a, it was a great trip, and I think we're gonna talk about some things we're working on there as part of our conversation there. So it is a purpose-filled drip. We are, and, and, you know, so it's interesting when we kind of get in the virtual green room, right?

That's the two of us, you know, shooting the breeze before we hit the recorder, uh, on that. But like, have we talked about AI in a while? It's like, well, we talked about AI a lot and then not a lot, and now a lot, again, It's a little sprinkle here and there, but we haven't focused on it.

No Sprinkle, but we haven't really focused on it in a couple of weeks and everybody else has Right. Simple as that everybody else has been talking about, uh, ai. And interestingly enough, I did a little, uh, side project, a little, uh, independent learning over this week.

I, I'm, I'm a subscriber to a service called Section, uh, and a section is kind of like a business school online business school type thing. But they had a workshop that was about generative AI and how to use generative AI in your, you know, business functions. And, you know, really mapped out a framework to try to evaluate the different options and, you know, how ready your organization was to embrace generative ai and, and, you know, just got me moving, got me thinking about how, yeah, if they're starting to talk about these in these, you know, kind online business class type scenarios, every company everywhere is having the same discussion.

And it gets back to two things, right? One is how does that start to really impact the digital transformation efforts that organizations have underway, right? You know, they were moving anyway.

Oh, we gotta get closer to customers. We gotta embrace these systems. We gotta overhaul these processes.

So that's been underway for a long time, as evidenced by the fact that we launched digital cxo uh, dot com, I guess two years ago now. Mm-hmm. It's been maybe two and a half years ago.

Um, you know, to, to really document the decision processes around that. So obviously we are doing a digital C X O conference. I figure, what's it called, Mitch?

I can't, I can never remember. Um, I'll pull it up here. Just kinda, I wanna make sure I got the date right here.

Digital C X O summit. And it's on September 13, Summit. Summit.

I couldn't remember summit because, uh, the brain just doesn't work. Like, You know, executive leadership conferences are summits. I think that's the Yeah.

Digital CX o summit. So again, obviously AI is gonna be a huge topic there because it really is taking the oxygen, uh, out of every room. But I, I want to, you know, we, we need to dig in on that.

We need to talk about, you know, kind of what's happening with digital C X O summit. But we also can't neglect the role of the tech team in these initiatives, right? Because what I found out during this workshop is you're gonna have a whole bunch of business planners.

You're gonna have business leaders, you're gonna have task force, you're gonna have red teams, you're gonna have all sorts of different folks that are tasked with trying to figure out what the hell do they do with this AI and generative AI thing in order to solidify and improve relationships with customers. But at some point, somebody has to implement this stuff, right? At some point somebody has to build the systems, train the LLMs, right?

Protect the intellectual property, make sure that it works, and it can be operationalized to deliver value to customers. And that's really where the tech team comes in, right? So I wanna hit on both of those because we're getting, and, and it's really incredible in terms of how quickly this stuff is moving.

But we went from, what the hell is generative AI to you? You know, are you a three do five or are you a four? Oh, you, you know, kind of G P T and why haven't you started yet?

G P t uh, person, uh, on that front too, you know, we've overhauled our business strategy and we're focusing all our efforts on this stuff within the period of eight months, right? Eight months. It, it's, it's unbelievable from that perspective.

So we really do have to, you know, think about how do we get out ahead of it, both from the planning standpoint, how does it dovetail with our digital transformation efforts, as well as the tech reality that somebody has to make this stuff work. Mm-hmm. So Vince, why you tell us a little bit about the little initiatives that we have undergoing even at Textron, we've had Exactly.

These discussions. Exactly. These Discussions.

Yeah, we have, well, you know, I think with AI there's some little bit of fomo, right? Everybody wants to be, what do we do? What do we do, right?

Or maybe Fear of missing out for all you Luddites out there that don't hang out with teenagers, so Go, or maybe, uh, fear of losing your job, right. Of even, you know, that kind of thing. F O L G F.

Yeah, I guess so. Yeah, exactly. Um, so we, we know we've been talking, 'cause I think one of the first thing a lot of people are doing is like, let's take our data and what can we put it into a large language model and put use generative AI to, to query it?

What could we use that for? Right. com, security Boulevard, uh, cloud native, now digital, C X O, all of our videos, all of our 400 webinars a year that we do, which transcripts that have fantastic discussion content perspective, 7,000 videos.

You know, it, it's, it's, it's massive amount of content. So there, there's obviously, so you only just like, I'll throw that in one big bucket and okay, let me query and see what I get. Because they call it, um, sometimes you get insanity responses from, you know, Chachi pt.

It's like no, oranges and houses are not related with each other. Might have an orange house, but not oranges and houses. Um, so, um, actually John Willis has been, uh, working on this for a while.

And so we're, we're doing a couple things. One is we're setting up a hackathon. We're bringing in a lot of early DevOps leaders who helped get it off the ground and, and set it into direction.

'cause a lot of 'em are interested in working on AI and using some of the things we're our business scenarios, at least as a start for some use cases and see where the hackathon takes us. Because that's one of the things with this technology, you know, you mentioned how fast it's changing. Um, a friend of mine just said, well, yeah, if you have a bug with that l l m, just wait two days 'cause it'll be updated.

That's how fast people are changing this stuff, right? Right. And, uh, I think, you know, I think he's right most of the time.

So we're, um, you know, we're in the process of our setting up our own, um, lang chain, uh, with, uh, deciding what we're gonna do for a vector database. You know, there's the pine cone and chroma. We're also, you may maybe use something like Mongo.

We'll see, we'll see what we do. Um, and kinda learning about this. We're in this heavy learning mode.

This isn't like, go take three developers and build yourself one. Well, they gotta learn how to build it, right? Unless they've already been working with this stuff.

So I feel, feel, I feel like we're in this. Okay, we need one of those, you know, the airline magazine, I read this, so we must need to do it, um, syndrome. And then also the, but yeah.

But how do we do it and what do we use it for? And then let's get smart about how we create a strategy around it. I'm curious about your, your course or whatever it was you took of, you know, putting a strategy behind it before you really kind of know what you can do with it.

Or did that get in the way? You don't really to know too much about it. It, It, it was, it was really about how do you start to frame up the, just really the discussion internally and it's business focused, right?

It's not tech focused at all. And you know, basically what you do is they have a series of questions where you kind of analyze your organization in terms of your ability to embrace generative ai. Right?

You know, are you gonna be resistant? How is your culture? You know, you know, stuff like that, that is obviously self-selecting, but gives you a sense of where on that train you, you know, where on the continuum you are.

And they, and they identify basically three different types of use cases, right? One is optimize, and that's really to make the stuff that you already do work better, right? That's pretty straightforward.

You know, kind of how do you optimize your customer support, you know, database? How do you, you know, make your reps more effective? Those kind of things optimize, right?

Mm-hmm. Then there's accelerate, which is really how do you take the things you already have going on and make them work better because you are integrating better content, more effective search, what have you, and then there's transform, right? And the three, I mean, you know, almost anybody can do an optimized type of thing, right?

You have existing processes. How do they work better? You know, using gen ai, right?

Accelerate and then transform is, you know, we're really gonna just blow something up and, and start it over again with this different thing in mind. And that obviously is a much more disruptive, expensive right. Time consuming initiative where you have to start prototyping things out and moving in really like a product development, you know, type of environment.

Mm-hmm. Mm-hmm. So building the strategy was really about defining where you were on that continuum, and then brainstorming in terms of how can these things impact our business?

And then really kind of talking about, you know, how quickly can you get it done and how much r o i there is. So you basically have a quadrant chart and you can evaluate, Hey, let's try this one because it provides, you know, high r o i and it, you know, kind of low, uh, amount of time or short amount of time, uh, to deploy a prototype. So it really went after it that way.

Not, you know, kind of the, oh my God, we decided to do this. What the hell do we do? Because it seemed that most folks were kind of at the point of, well, you know, once we do that, we decide to green light it and, and allocate some funds behind it, you know, then we'll figure out what third party can, you know, help us do that.

Right? And I think that's gonna create a real opportunity for those service providers that have probably been doing cloud implementations since I guess cloud platforms are really gonna be the main, you know, kind of environment that, that folks start playing around with, with AI given the, you know, compute and, and resource Easiest way to do it. Yeah.

This one easiest and, and quickest. So you're gonna start seeing each of these guys developing very quickly and they're already doing it, right. You know, kind of AI practices within, you know, their specific thing.

We're doing some work with, uh, a W S and Accenture, right? They're, they're actually, it's called the Accenture a w s Business group. Um, and they did that, right?

They, they did a, a webinar recently on generative ai and they talked about, you know, kind of the use cases and the accelerators that they have, which are basically, you know, kind of packaged up things that they've done for a couple of other clients that they can kickstart and work on, uh, with you guys. So, so you, you know, again, it's one of these things where you're gonna need specialty, uh, services. You're gonna have a whole mess of folks that are getting in there.

There's gonna be a bunch of charlatans, uh, a handful of shysters and, and a bunch of comb poops, right? So, you know, kind of trying to figure out how to determine somebody can actually execute on this thing is gonna be a big, uh, success criteria. Those must be business terms.

'cause I haven't seen that in lens anywhere. No. Tum poop command.

Yeah, absolutely. Mount, Mount this. Yeah.

Okay. You know, it, it's interesting. So for me personally, when I go through these kind of, oh, you know, new thing, what could we do with it?

There is sort of like thinking about it at a vision level and you know, what does this mean? Where might we go with it? And kind of exploring there.

I al but I also at some point kind of hit the, hit the, you know, hit the top of the mountain. I'm like, okay, now I really need to kind of do it kinetic learning, even if it's not something's ever gonna go into production, but just get some hands on and like, okay, I understand how you kind of put it together and what it takes to do a little bit of that so we can at least start to figure out what questions. 'cause when we're thinking about doing like disruptive things, especially, or, or even if you wanna do an optimize, it's what does it take to do that?

I, I don't know even what to commit to yet, right. On what we might do for a generative AI till I get a little bit of grease under the fingernails, right? And, okay, I now, I think, I think we've got enough understanding, you know, helps the team feel like we're engaged in new stuff and staying current and not stuck in maintenance mode on whatever, you know, but that's, that's, that's kinda learner.

I am, it just works out that way. Yeah. And, and, and that's part of it, right?

The other part of it is that every business is gonna be impacted by this, and I could make the case that media may be the most impacted, right? Yeah. You know, entertainment is gonna be up there, but I mean, how media is developed, how it's, you know, kind of curated what it's, and, and how it's delivered.

I mean, all of that's fundamentally going to change. Mm-hmm. Right?

And, and a lot of this came out of, again, just kinda hallway conversations, right? Alan was talking to John Willis, and John Willis was, you know, talking about how, you know, he's such a devotee of, uh, professor Deming and he just wrote a book. Check it out.

I'm sure it's wonderful. Uh, S it's coming out, it's not out yet, but it'll be, yeah. Oh yeah.

So, so I, I got a thing from, you know, gene Kim's, uh, email address, so I know it's, it's maybe a pre-order or something like that real quick. He's become, but obviously he follows everything that happens with Professor Demming. So he set up this Google alert that then, you know, scrapes it and pops it into an l l m that gives him a daily summary.

And then, you know, he can click on whatever it is and it posts it to a different place. And, and, you know, Alan's just like, yeah, we need to do some of that. So that was really the genesis of the hacking.

Um, then we've gotta figure out what we can do and we gotta do it fairly quickly because the world is literally changing underneath our feet, right? The foundation is shifting, and if you're not in front of it, you're gonna be underneath it. Right?

So, so you really do need, again, depending on the business, you are in our business for sure, right? It is going to upend, disrupt, and transform, right. How people get and consume media.

And if you're not leading the trail on that, uh, again, you're just, I think it already is disrupting it, right? Of course. Right?

But you, you know, you're, you become an endangered species if you're not, you know, kind of on top of it. Mm-hmm. Agreed.

A agreed. It was interesting because you mentioned Deming's book. One of the things I learned from John, uh, this week and doing some work together is his strategy is, is take, you know, take something that you know really well, like some documents or PDFs in case it's his dimming book.

And then he uses that as part of his training of the l l m 'cause he knows the content. So when it says something stupid or insane, like that's not right. Okay.

How do we, how do we adjust the prompts? How do we, you know, create some things about different characteristics of, you know, what might be in the, in content in the book to help kinda guide some of those, uh, questions. So it was really interesting.

He's like, pick something, you know really well, and, but, but don't make it too big, right? com as your first project. So it, it's a high learning environment.

Um, and, but it's a lot of fun. It's very fascinating. It It, it is.

And back to you, you know, kind of this idea of the services and the platforms and stuff that can get you fairly quickly to get your hands dirty. Uh, right. You know, a w s has their SageMaker, you know, service and, um, I forget they have a, they have like basically a, a packaged up thing that you can use to start it.

I should know this off the top of my head. Yeah, No, my head too, You know, but they have a couple of different offerings. You know, you can use commercially available, uh, l l m models, uh, in order to drive it.

You can customize those, you can do all that. And the answer is yes. Yeah.

Right? Because you don't know which of these things are gonna work. But you have to start playing with this stuff now, because if you don't, you are going to be behind the curve.

So one of the other interesting things about this course is, um, you know, they kind of say, all right, you know, this is what we wanna do. This is how we wanna frame it out. And if you're looking at optimizing type things, why don't you try something like this prompt?

So they actually gave you a chat GT four prompt that said, you know, I'm in this kind of bureau analyst and I'm in this kind of business and, and you know, here are my different, you know, kind of, uh, value propositions and y you know, here are the, the, the, the things I want deliver to customers. How can I use generative AI to make that better? And generative AI figures out your generative AI strategy.

If this is not inception, I don't know what it is, right? But all the same. Use the tools that you have available to you to at least get that brainstorming going, right?

And to start throwing some ideas out there. Then you say, oh, that's a crap idea. Oh, let's dig in on this one.

Or, you know, dig in. And that's another thing. Well, what would that look like?

If you can ask, you know, chat G p G and have a conversation along those lines. And again, it's not gonna be right, but it's gonna give you some ideas, right? It's gonna give you some Ideas.

Take it the next step further. Ask chat. G B t, Hmm, what l l m might I, should I check out?

Uh, can you write me some Python code to load data into L M M using this as a data source? That's, you know, so it's, it's actually, you could use it, not, it's not gonna solve all the problems for you, not yet anyway, but yeah, you could sort of use this as it train the trainer, train the human, you know? That's right.

Using Chad G P t as at least as one source, of course, building the code editor As the plugins start to, you know, kind of get baked to a much greater degree. You can start, hey, pointed at this. And, you know, kind of look at that.

And, um, so again, it's just, it's one of these things that you're using the substrate to figure out how the foundation should look as you move forward, which again, is extremely cool, but all the same, right? This leads specifically into your digital transformation efforts. Once that gets nailed down, then the tech folks have to come in and figure out how this is going to work.

So if you're on the tech folks side, right? Learning how these different things playing around, as you said, Mitch, getting some of that kinetic experience, again, you can't wait on this. You have to do this now because this, an initiative is gonna get spun up in your organization.

You're either gonna be prepared to do it or you're not. If you're not, they're gonna give it to somebody else who's been spending nights and weekends playing around with this stuff. I guarantee you that.

So that's just a little bit of a, a word to the whys. Get your hands dirty. Get 'em dirty now, and start to figure out how this stuff works.

Well, 51% of planning is doing. 'cause every plan is wrong. The minute you hit safe, it's, I mean, it just is, is this, you know, you, it's, you have to learn as you evolve your, your strategy in your plan.

So it's that, uh, it's that kinetic learning and applying it. So very cool stuff. So we could go on for hours, but we can't because we both have other stuff.

Um, so, uh, the, the key takeaways here. One checkout, um, the digital C X O summit, that is in September. Uh, this will be a major topic.

There will be new stuff discussed because this is moving faster than I think any of us have, have really seen before. Mm-hmm. Um, start playing around now, you know, get involved to the degree that you can understand where the strategy people are thinking about how to integrate these things in.

Then you can get a little bit of a head start if you are somebody looking at it from a, a tech background standpoint, um, and just start playing around with it again. It's, it's very, what we were talking before is, right, it feels like the early days of the cloud, right? Where, you know, yep, it was this thing and you knew it was coming, you weren't exactly sure how you were gonna use it.

But the folks that prospered were the ones that, you know, opened up their a w s account before they even had a free account. Started playing around with it, started understanding how it worked, started building things. And again, they don't have to be big things.

They don't have to be complicated things. They have to be things right. And similar type of of environment here.

Just start playing around with it. The cloud platforms have capabilities in order, uh, to, you know, allow you to kickstart your efforts, uh, and get going. That's, that's the best thing.

Just get going. Excellent. Hey, you know what?

I wanna make sure we get a plugin. We launched, uh, something new collaborating with a w s called the, uh, builder Community Hub. com.

And it's working with some of the a w s partners that are focused on app modernization, that particular kind of slice of projects and work. And I think to talk about digital transformation. So we just launched, launched this.

It's about a dozen or so partners that are on this. And it's all content focused on that topic at different levels, and we're continuing to add more and more to it, but it's been extremely well received. com, it's slash builder dash community dash hub.

I know it's kind of long, but we'll figure out a way to make it shorter. Um, and there's some great stuff there already. So congrats to the a w s and, and, uh, Techron team for getting that off the ground.

And, And, and the other thing we should, we should pimp a little bit out is, um, tech strong ai, right? Yeah. Get the newsletter, start looking at the site because we're gonna be documenting all of this stuff as we continue to move forward.

So yes. We're, We wax poetically and well, I think we're gonna have a, an AI conference later in the year too, because mm-hmm. We are have, Right, because it's, it's one of those things where we have to cover it and has to be something that becomes a major area focus.

Not just in delivering our value, but in documenting how other companies are delivering that value to their customers. Mm-hmm. Yep.

And I'm glad you mentioned that. We would be, we'd Be in trouble if we didn't. Yeah.

Yeah. We would've to re re career record this. Well, good.

It's been a great week. Good talking with you, Mike. And You bet.

Alright, so we will be back. I'm not sure if we'll be back next week 'cause we're both heading out the black hat at some point, so we'll be back at some point. Maybe we'll record a black hat and that'll be Yeah, yeah.

Let's do next. Yeah, we're gonna record a black hat, so Yeah, we'll do that. So, all right.

Very nice. All right, good. Thanks everybody for tuning in.

Great chatting with Mike and, and, uh, with you, our audience. So take care.