Subscribe
Share
Share
Embed
Join the the Artificial Intelligence Student Organization, RAISE, at the Crummer Graduate School of Business and Rollins College, to discuss the reality and implications of AI in the business and academic world.
EP 6 Part I - Bradford Neal_mixdown
===
Pristine: [00:00:00] This is the AI across the campus and beyond podcast. Today we'll be talking with Brad, the founder of Tech Armada, a custom software development and AI solutions firm, partnered with VEFI Solutions, a cybersecurity and risk assessment firm to develop AI advancements in cybersecurity and information assurance.
I'm pristine, your co-host and co-president of the AI Student Organization race. I'm self designing a major at in AI at Rollins.
Casey: And I'm Casey, your other co-host and the graduate side co-president of Raves. I also work at the AI Edge Center here on campus as a graduate assistant.
Pristine: Uh, Brad, could you tell us a little bit about yourself and your background as an IT professional?[00:01:00]
Uh, what were your first impressions of AI and how did you first become involved with VEFI?
Brad: Absolutely. So, um, I am a type of person that just kind of curious about things and I guess it started back whenever I was, you know, a small, small child. And back in the eighties you were just learning what a computer was, you know, they were just, PCs were suddenly popping up in homes and, uh, we got one and, uh.
It was an IBM PC with an Intel 80 88 processor. You know, it had the, uh, floppy disc. It didn't have a hard drive, it was just nothing. And, uh, I just started poking around and it was interesting. Uh, there was this book that came with it and was, it was a, it was a basic book that said, [00:02:00] um. How to program using GW Basic.
I'm like,
Brad: I don't know what this is. I'm sitting here, I'm this 11-year-old boy, I'm sitting here at this, this terminal, and I just open up the book and I just start outta curiosity, just start typing everything in there. I'm like, this is actually kind of fun. I can actually write basic procedural programs and you know, just by following this book.
And then it kind of went from there to. To other things, but I kind of set that aside. And then, um, from there I went to, um, college and didn't know what I wanted to be. Um, but uh, I stumbled on computer science and I remembered how much fun I had as a kid doing that kind of thing. And so I picked it up and I started learning about AI really in college.
My first experience with, uh, uh, AI [00:03:00] development was on a risk six, uh, risk risk 6,000 supercomputer. Um, and the programming language, uh, that I was, uh, developing in was scheme. I was just learning it. I, I didn't have a class in that, uh, language, but I was a, a, a technical, uh, lab person. I was a lab tech, you know, managing different rooms.
And there was a risk system in that room. And so I just started getting on it and playing around. I was like, oh, what's this scheme thing? And so, um, I started learning, uh, a little bit of scheme. And then, uh, a year or so later after I went to UCF, um, I, uh, was able to, uh, learn, uh, common lisp. Which is a recursive, uh, AI language and interesting enough, uh, lisp, you know, it was like the first Yeah.
You know, AI language, you know, and it, you know, today it's still used. I, I, most, most of the languages out there are Python. Mm-hmm. [00:04:00] Um, uh, that are most of the AI out there, uh, AI development is, is developed in Python. Mm-hmm. But, um. But Lisp was the original, you know, and, and so Lisp is still being used for, um, things associated with, um, uh, planning, uh, symbol, uh, type a ai.
And um, I'll give you an example of one company that's actually using it full force today. Still. It's Grammarly. We know Grammarly, right? Yeah. Yeah. Well, they use it for their engine. Interesting. It's common lisp. It's still around today, and I think every,
Casey: every student here at Rollins has probably met Grammarly at, at some point in their experience.
Pristine: You seem to like Lisp a lot more than Pie Torch or anything. Yeah.
Brad: Well, um, I'm old. I'm, I'm long in the tooth, so you gravitate to the things that, you know, um, but, you know, I'm not, I'm not a against Python at all. I've developed in Python. I, I like it. But [00:05:00] um, again, I started off with the primary language.
That I focused on in college and at the start of my career was Java and c and c plus plus. And so I still gravitate to those. But it doesn't mean that as a software developer, I don't know when to use Recursion. You know, there's specific times when I need to use Recursion, um, and I know when to use it and when not to use it.
Um, I know, um, uh, how to make a, a program I. As intelligent as possible, you know, and but, uh, when this AI movement really started taking force and, uh, the, the, uh, language models, um, had a large enough dataset and it was able to, um, actually provide some value as opposed to, you basically have to do a lot of work to get a whole lot of little AI out of your program.
Mm. Um. That kinda reached
Casey: critical mass.
Brad: [00:06:00] Yeah. Reached critical mass. Um, uh, it was time, it was time for me to just delve into it. Mm-hmm. Um, and so, um, I, I, I, and again, I I, I'm the type of person that likes to evolve, you know, I don't wanna be the dinosaur. Um, and, uh, and so this is the ne next greatest thing, and I want to.
Be right at the forefront of it as as much as possible and ride the wave. Yeah. This is a, this is a.com type scenario. Or even greater. Yeah. And, uh, we'll get into, I'm sure some other, um, questions related to the future of ai, but mm-hmm. Um, the, the history has been that it was slow coming. Yeah. But it's here and it's not going away.
Casey: Absolutely. I remember I was shocked, uh, when I first got the, the job at the AI Edge Center. I had to do some research into the history of ai. And, you know, the research academically goes back to the 1950s, which, you [00:07:00] know, you think about chess programs and, and basic AI applications. I guess you can kind of see it, but it still was, was shocking.
I mean, it, I guess, has a bit of a snowball effect or critical mass, however you wanna think about it, but. Now that the data's there and we have things like distillation where AI models are able to train using other models, it uh, it definitely seems like it's the next big thing. How, how you're putting it.
Pristine: Yeah. I wonder how much more advanced it would've been by now if we didn't give up on it in the past.
Brad: Mm-hmm. Yeah, exactly. Um, and there wasn't enough. Um, well first of all, there's there technology was advancing so rapidly in so many other areas. Hmm. Everybody jumped in and put all of their processing power, their brain power into, um, into the internet when it came around in the nineties, you know, everybody was jumping into that.
And, and when, um, when the cloud came along, everybody jumped into cloud, you know, and in the two thousands. [00:08:00] And so things are evolving and, uh. That's, those, those systems is per those, those, those, um, it, uh, services and things like that have already reached almost like a, a maturity peak. Mm-hmm. And a plateau.
So what's the next greatest thing? And that's what people do. People innovate.
Mm-hmm. They're
Brad: looking down the list of all the potential AI solutions out there, all the potential. And they go find areas in that, that aren't matured yet. Let's, let's, let's grow this. Um, and that's just the way of, uh. Of mankind, I think is, is just to find the weaknesses and to make them better and stronger.
And, um, and this is the next greatest movement.
Pristine: I agree with that one.
Casey: Well, I'm really glad we have, um, somebody like yourself with experience and development and kind of entrepreneurship and business. 'cause I mean, that's a common theme is looking for sectors of the industry that have growth potential. I, you know, just started my MBA and it's a far cry from history, but I keep seeing [00:09:00] that, you know, you don't.
You want to be in industries that have plateaued, you wanna be in industries where there are high margins, potential for growth. 'cause that's where you really have opportunity. Yes. But, uh,
Brad: that's true. You do have opportunity. Um, and you have, uh, also the most room for, uh, uh, of interest and excitement. Yeah.
There's, there's energy, uh, when you're doing something new and mm-hmm. And you're, you're part of, of building, building something better. Um, there's just a lot of energy associated with a lot of positive energy, yet there's also a lot of risks. Mm-hmm. And, and you have to weigh those things out. Um, there was risk with the.com, you know, there was the.com bubble and things like that.
So there's always a risk with ai. Mm-hmm. But the upside is so great. It's like with cry crypto, there is a risk with crypto, but the, the upside is so great. It's, it's, and there's so much momentum behind it. It's not going away. Yeah. AI's not going away. Um, and so you, you ride that bubble [00:10:00] and, and one, one phrase that I live by, uh, especially during these interesting times because, uh, we we're living in a very dynamic world and things are changing constantly with the government, with, um, the, um, uh, just with, with the private sector, with mm-hmm.
Just everything, um, with just, just society in general.
Yeah.
Brad: Focus on. How you can improve the world in your world. You don't have a big world. Yeah, you have a small world, but impact that world and change it as much as you can in a positive way, and thrive in chaos. So right now it's chaotic. Yeah. You know, let's learn to thrive in chaos.
Look for opportunities and chaos. And right now. AI is the wild, wild west.
Yeah.
Brad: Get [00:11:00] out there with your, your, your gun slingers and learn how to shoot. You know,
Casey: I think we definitely saw that with, uh, deep seek and just the disruption that that caused. I mean, coming outta nowhere, and it was right on the heels of Salesforce's agent force announcements.
So, um, it, it is kind of hard to keep up. Um, but let me ask you like, uh, from a developer side of things, I mean, how do you, how do you see this say, you know, I. You have students that are majoring in computer science. I mean, how does this change the process of writing code from five years ago before the explosion of this generative AI bubble?
Brad: Absolutely. So never, never, never take shortcuts. Mm-hmm. Okay. As a and I, I encourage anyone getting a computer science degree today as an a I it degree. Mm-hmm. Anything related to technology. Learn it. Yeah. Learn what's under the hood, understand it. Well be able to develop [00:12:00] yourself. And then whenever you start using the AI tools and you start leveraging those tools, you'll understand how they work and you'll, how you'll be able to integrate them better.
Yeah. And that's one thing that I've found I've, I've developed, you know, from assembly language all the way up to high level languages. Mm-hmm. Over the years, I mean, all of them, all the stacks. Yeah. And when it comes to AI now and it's doing the work,
yeah,
Brad: I, I know how to tailor, uh, the requests that I make to the AI engine properly in order to get the results that I want.
Whereas if I had just been a fly by night computer science person who really didn't have a deep, uh, understanding of it. Mm-hmm. Um. The results that I get back will have holes, will have vulnerabilities, it won't be as quality. And, and I just give you an example. Just recently, uh, uh, before I retired from [00:13:00] nasa, I, um, it wasn't of course my responsibility as a program manager to, to do this, but there was a great need for a, a, uh, a tool that.
Uh, uh, grab a, a bunch of, uh, API points. Mm-hmm. And then, uh, and then from any, from any system and then. Take those, um, a, um, the, those endpoints. Mm-hmm. And then create a retrieval from those, um, in a CSV and uh, PDF format. So basically, yeah, we had a lot of need with all this data that we needed to consolidate and, and provide reports on.
We didn't have a tool to do that in the SBIR program at nasa. Um, be because we were actually transitioning from our old tool to our new one. And that's another long story. I don't wanna get into that.
Casey: No, I, um, I, I ran into a similar use case where we were doing some social network analysis mm-hmm. Of, uh, sort of a cross section of [00:14:00] AI and medical research for the, the AI Edge center.
Mm-hmm.
Casey: And, uh, you know, I didn't get into it, but I, I put under the, you know, possibilities for future research. If you can train an AI model to take this data and format it as a CSV. Any researcher, you know, a older senior researcher that only knows history or doesn't know anything about computer science could create a social network model of their genre, their subject.
Um, but formatting and, and I guess data transformation is a, is a term I heard actually, um, from a colleague of yours, I guess at nasa. Uh, Dr. Michael Bell, who was the Chief Knowledge Officer, and they were looking at taking their, their lessons learned and just transforming them. But, um, you know, it's like, like you say, you don't wanna trust it too much and you need to have a human centered, human in the loop, at least process.
But being able to take data that you already have and transform it seems like a very attainable task.
Brad: Yes, absolutely. And, and so, so AI can do that [00:15:00] for you. Mm-hmm. Um, and also it can do the development for you on the gooey side and the backend side. Wow. Uh, and so, uh. I used to develop by hand the backend and front end of applications all the time, and so mm-hmm.
But I didn't have time to do it as a, as a program manager. Mm-hmm. Um, and we needed to have this system very quickly. I, we needed it within a month, and I needed to be custom. So I went in and found a, a, an ai, um, uh, uh, that would work well for me. I used Gemini, Google's Gemini. Mm-hmm. I architected the application as a, as a software architect.
I, so I knew each of the classes that I needed. I knew all the mm-hmm. Um, capabilities and functions that I needed. So what I did was I piecemealed each of the requests and tasks that I needed the Gemini to create for me. Mm-hmm. And then I validated it and make sure it had all the parts, and then I fit it into my [00:16:00] application piece by piece until I had an application pretty.
Pretty hardened and it worked. Um, okay. And had it done pretty, pretty quickly.
Casey: So you just kind of segmented it for the segmented, AIed it in,
Brad: um, and by well-defined requirements. Mm-hmm. For a specific task, I was able to build this application.
Pristine: Oh, I was just gonna say, I think the requests like knowing what to ask AI is definitely going to be the most important part of computer science in the future.
Maybe. Because I had a project like in my programming with AI class, and basically it was to make a search engine. And I didn't know anything like about making search engines. But, um, once I actually looked into the structures I needed and like tried to understand what I was trying to build, then I could just have AI do the coding part for me.
So I feel like knowing the request is gonna be really important
Brad: and that's great. And, and that's the power of ai. I, I just so powerful. And you don't have to, I. [00:17:00] You know, be a, an in-depth, you know, developer to, to get a lot of work done. You just have to understand what you want and to get it out of that ai.
You know, you might be a CPA, you might be a doctor, uh, of medicine, you mm-hmm. You might be, uh, somebody, you know, sitting, sitting at the, at the end of a, a, a, a assembly line and you have a certain function that you need. Created.
Yeah.
Brad: You know, and you can tell it that. And uh, but I do feel that for the more advanced type functionality that you wanna build, you really do need to have some support in that area, um, on hands on.
So it just depends on the, on the development, uh, something that's repeatable multiple times. Mm-hmm. You can, you, I mean, that's AI's great for that, uh, but you're always gonna have to work in, um. [00:18:00] Uh, your own, uh, uh, um, understanding of the system to, to get what you're wanting out of it, I guess is what I'm trying to say.
Pristine: So you're saying like, you'll always have to work with the ai so you're not thinking about like being replaced or anything?
Brad: Yeah. I, I don't see, if you evolve, you won't be replaced. If you evolve, you won't be, uh, replaced. And it's the same way with, um, uh, the assembly line when Henry Ford. Yeah. You know, created the assembly line, you know, people, jobs were, you know, changed.
And the same with automated, uh, manufacturing. Well, people find it, you know,
Casey: we, we did actually mention that same thing during a previous podcast episode. Um, I think with, uh, Dr. Myers, uh, from the computer science department. And what we were getting at was that. Obviously there's still a need for people, but a big change with AI that we've heard is that it closes the performance gap.
So the difference between top [00:19:00] performers and mid-level or lower level performers is smaller. So kind of just like the assembly line, where in the past, you know, say you have a guy who's in his garage making custom, you know, classic cars and selling them for half a million dollars. Once he's done, he's uh.
A master craftsman and reducing the value of the skillset of, say, programming, which was an, is an expert level skillset to a, um, more easily commodified and replicable skillset that that has a tougher job market. Do, do you see any of that? I
Brad: see, I, I see. I see. I see that yes, it does enable
mm-hmm.
Brad: Uh, under performers or people that really don't want to be a developer because it takes a special person to want to be a developer.
Mm-hmm. You're not, developers aren't the smartest people in the world. They're just people. Mm-hmm. Um, you know, some people have talents in one area and they don't have a talent in another area. So when you come to a, a developer, you know, they just, [00:20:00] they just fixate on the code. Mm-hmm. Okay. Whereas somebody else might.
It might be focused solely on, on law or whatever. Well, that lawyer with AI now can develop something related to their expertise that that developer couldn't develop.
Casey: Interesting.
Brad: Um, but on the other hand, the people that are using this every day as a software developer that are just aren't major performers in that area, in the genre.
Mm-hmm. They're gonna build, they're gonna build the Hugo. And the people that are experts in the area that know what they're doing, they're gonna build the Mercedes-Benz, they're gonna build the Lexus. It's, it's going to, it. Knowledge is always impactful. They, they, they may build something, uh, somebody that with lesser knowledge may build something.
Uh, that's just as quality. Mm-hmm. They may do that, but um, that, that person that has the expertise, I [00:21:00] guarantee you they'll do it faster. Yeah. And so they'll be more productive and so they'll be the first person called on for the next job. I. Okay.
Pristine: And I also had this conversation with someone about how, like between different models of AI that we have right now, even if they all perform at the same level, like the branding and the marketing, like the name behind it might have more value than like the actual performance.
So I think it could be like that too.
Brad: That's true. That's very true.
Casey: This has been our conversation with Brad Neal, the founder of Tech Armada. Uh, stay tuned for a special bonus episode coming out next week. This was a great conversation and we have about an, uh, an additional 30 minutes of content that we'll be posting again within about a week. See [00:22:00] them.