How AI Is Changing the Way We Work
How is AI reshaping the future of business leadership and innovation?
In this episode of the Las Vegas IT Podcast, Hunter Jensen, founder and CEO of Barefoot Solutions, shares how his company evolved from a small development firm into a strategic partner for enterprise businesses navigating the digital age. He dives into the transformative role of AI in driving business growth and the importance of building ethical, scalable AI platforms.
What to Expect in This Episode:
🧠 AI’s Business Potential: Explore how AI is accelerating innovation and the real-world impact of enterprise-level AI solutions.
🤝 Relationships Matter: Why building strong client relationships is still the cornerstone of business success—even in a tech-first world.
🔐 AI Governance & Ethics: Understand the need for ethical frameworks and governance as AI tools become more powerful.
🛠️ Private AI Platforms: Learn why businesses should prioritize developing their own secure, proprietary AI tools to gain a competitive edge.
🌍 The Future of Intelligence: Insights on the evolution toward Artificial General Intelligence (AGI) and what that could mean for the future of work.
Hunter offers grounded, forward-thinking advice for tech leaders who want to harness AI while staying rooted in values and strategic vision.
Let’s Connect with K&B Communications!
If you enjoyed this episode, let’s keep the conversation going:
📱 Follow us on social media for insights, tech updates, and behind-the-scenes content:
Facebook- https://www.facebook.com/kandbcom
Instagram – @kandbcommunications
LinkedIn – K&B Communications
▶️ Subscribe for exclusive interviews and IT business strategy:
https://www.youtube.com/@kbcommunications
💼 Need data cabling or low-voltage solutions?
K&B Communications specializes in network infrastructure, security systems, and fiber optics.
📞 Schedule your consultation today:
https://kandbcom.com/schedule-las-vegas-commercial-lowvoltage-consultation/
Your vision. Our expertise. The future—built together.
Transcript
when Apple launched their app store.
2
:We had one of the first 100 apps in the app store.
3
:You know, I'm an early adopter for all the things that are tech, and I'm just constantly
evaluating new stuff and trying to think a year ahead or two or three years ahead.
4
:And when I have something that I believe in, that I think has a future, it is not about
getting rid of people.
5
:It's about making your people more productive.
6
:Welcome to the Las Vegas IT podcast.
7
:And today I have the pleasure of speaking to Hunter Jensen with Barefoot Solutions.
8
:How are you doing today, Hunter?
9
:I'm doing great, how are you?
10
:I'm doing amazing.
11
:I am super excited to get to know you a little bit better.
12
:Could you just tell us a little bit about Barefoot Solutions?
13
:Yeah, sure.
14
:So I'm the founder CEO of Barefoot.
15
:We've been in business over 20 years at this point.
16
:Primarily we are a custom software development shop.
17
:You know, we got started as kind of an extension of my freelance web development work when
I was in college and just picking up gigs on Craigslist and I grew the business
18
:organically from there.
19
:And so, you know, we started off as a uh
20
:a web development shop making WordPress websites back in 2004.
21
:Moved into mobile apps very, very early on, which fueled our growth for a number of years.
22
:Then we got into IoT, connected devices, medical devices, that kind of thing.
23
:After that came blockchain, which was a fun little ride we had for a couple of years.
24
:then most recently, everything's AI.
25
:project that we see has some sort of AI component to it.
26
:And so we've worked with enterprise like Microsoft and Salesforce and Samsung have been
clients of ours.
27
:We work with venture funded startups, which is a lot of fun, but not necessarily like
reliable.
28
:And mostly we work with medium sized companies, either building products for their clients
or business process stuff, internal stuff, you name it, we've done it.
29
:Got it.
30
:Now that's so awesome.
31
:So you've been in business now for about 20 years.
32
:I mean, you said you got started with creating WordPress websites and I'm sure that's a
huge difference from what WordPress is today.
33
:It was, it was very, very early days.
34
:PHP three was the programming language at the time.
35
:And it has evolved immensely.
36
:Like we've, you know, we've been able to be in business as long as we have a lot of shops
like, like mine don't make it this long.
37
:And the reason we've been able to stay relevant is by staying ahead of technology, right?
38
:Like we were, we had one of the first 100 apps in the app store for when, when Apple
launched.
39
:their app store.
40
:you know, decisions like that to kind of get in early on new tech has enabled us to
continue to evolve.
41
:If we were building WordPress websites right now, don't know that we'd still be in
business.
42
:Yeah, well, especially now that everybody wants to be a WordPress developer.
43
:So I'm sure that's the, you know, a lot of people want it.
44
:That's the direction that they want to go.
45
:So, but that's awesome.
46
:And one of the things that you did mention is, you know, staying ahead of technology and
how is it that you do that personally?
47
:You know, it's my favorite part of this job actually, which is I get to learn and tinker.
48
:So I, you know, I read voraciously every day.
49
:I'm reading something for an hour or two.
50
:I, you know, am an early adopter for all the things that are tech.
51
:And I'm just constantly evaluating new stuff and trying to think a year ahead or two or
three years ahead.
52
:And when I have something that I believe in, that I think has a future, then I get my team
ramped up on it.
53
:And we do like little demo projects and we tinker with stuff or we'll have like a lunch
and learn event so that we're ready.
54
:we reposition our marketing and everything and get that one flagship client.
55
:in with a new technology and then use that case study to drive new business in that area.
56
:uh It's the same playbook.
57
:We've done it.
58
:We've kind of done it five times now.
59
:And every time is a little bit different, of course.
60
:yeah, just getting to tinker with...
61
:I was tinkering with Mid Journey a year and half ago when it was like a brand new thing.
62
:And I was like, wow, this is going to be important, I think.
63
:And I was
64
:beta user for a chat GPT and all of that stuff.
65
:And so that's how we kind of stay ahead is, you know, it's a big part of my job to look
into the future and make sure we're well positioned to deliver what our clients need.
66
:Understood.
67
:And so with what you've said, for one, I mean, some people may not know what mid journey
is, can you just kind of share exactly what that is?
68
:Yeah, Mid Journey is its generative AI for images primarily.
69
:And this was kind of early days before, I mean, not before Dolly, but before Dolly really
was the huge leader in image generation.
70
:You had to be pretty tech savvy to use it.
71
:You had to use the Discord app to generate images, but I just thought it was so neat that,
you you could just...
72
:write out prompt and get a set of images.
73
:Back then, people had three arms and missing teeth and their thumbs were weird.
74
:There was all sorts of problems with it that have since a lot of them have been ironed
out.
75
:Not all of them, but a lot of them have been ironed out at this point.
76
:But that was just very, very early days of generative AI.
77
:got it and then when you say discord are you talking about the application could you
explain exactly what that is was I have an idea what you're talking about for discord but
78
:I may be wrong
79
:Yeah, so Discord is like a community chat platform and it's like, it's really big in like
in gaming and like certain communities, blockchain.
80
:was like web three communities were on Discord.
81
:And it was, it was a place where you kind of create channels, like you create a Slack
channel, but they're more open and free and mid journey made the choice to not develop
82
:their own user interface for a long while.
83
:And so you jumped into Mid Journey and you installed the Mid, sorry, into Discord, you
installed the Mid Journey bot, and then you wrote your prompt in to that bot and it would
84
:respond back with an image.
85
:And so that was an interesting choice by them, but it did help them develop like a pretty
strong community that way, because people could share their images and chat about stuff.
86
:And here's a cool prompt I tried and all sorts of things like that.
87
:So it's more of a community for people.
88
:It is, but it can be used as a platform for, they're called bots or for apps that you can
install into your Discord.
89
:But it's very much a chat-based interface where you're kind of chatting with the bot and
it's responding.
90
:And with that, I'm guessing, is it very similar to chat GBT?
91
:At this point, well, no, not really because it's really focused on images.
92
:ah Whereas chat GPT is does both they it does regular natural language as well as images.
93
:You know, if people are in the early days, they were not integrated that way.
94
:There was Dolly and there was chat GPT and both of them owned by open AI.
95
:But I forget.
96
:how long ago it was, probably a year ago, the OpenAI integrated Dali into ChatGPT such
that you could then ask it to generate images for you.
97
:under the hood, those are different models, right?
98
:ChatGPT is a large language model in LLM, whereas Dali is a diffusion model.
99
:And they work, the math and the algorithms and the training and all of that work very
differently between those two types of models.
100
:got it and I personally have seen it where you you ask Chachi we teach it create you an
image and they give you something like that's really off but I do believe that they're
101
:getting better.
102
:uh
103
:It is not nowhere close to perfect, but improving all the time.
104
:Yeah, very true.
105
:And then with you getting started about 20 years ago and got started within with
WordPress, what took you to this direction of business with where you're currently at?
106
:Yeah.
107
:So, you know, like I mentioned, was so kind of origin story here.
108
:I was in the computer science department undergrad in Virginia and dropped out not of
college, but of that department because they weren't teaching new stuff.
109
:They were, you know, they were behind by a few years as, as traditional.
110
:Academia tends to be and so I got a liberal arts degree in philosophy and economics, but I
checked books out on the library from the library and taught myself how to code and I was
111
:just getting gigs on Craigslist and you know nights and weekends and outside of classes
picking up jobs and when I graduated it was very natural for me to want to continue that
112
:work.
113
:I loved it and you know, I was my own boss and making my own hours and all the rest of it.
114
:But then you get.
115
:You get to a point where you realize that, you know, when you're billing your own time,
there's very much a ceiling on the money that you can make and the value that you can
116
:drive for your clients.
117
:And so that's when I started hiring either for people that for jobs that I wasn't any good
at or that I just didn't like and grew it organically over the years.
118
:One big milestone though was back in 2019.
119
:I actually sold a controlling interest of Barefoot to a very large global software
consultancy based in Poland.
120
:And so we're like kind of a boutique shop, you know, it's very white glove and hands-on
long-term engagements, that kind of thing.
121
:But at my parent company, we have 1500 engineers.
122
:So we've got the horsepower of like a major software shop.
123
:Like we can spin up a team of tenor
124
:engineers in a matter of a week or two on new projects.
125
:And so that has put us in a very unique position.
126
:There are not a lot of software agencies that can say that, that are kind of have the
benefits of being small and the benefits of being very big at the same time.
127
:That's kind of how we've differentiated from our competitors.
128
:And that, and, you know, honestly, I would say the number one driver of business for us.
129
:It's just been relationships, just building relationships with people over many, years is
the way that we've maintained and grown this business.
130
:Understood.
131
:No, I totally agree with you.
132
:Just the importance of building a relationship with people.
133
:What are some of your most recent examples of a project that went well due to a
relationship that you've had created possibly, you know, recently or in the past?
134
:Yeah.
135
:Let's see.
136
:So I won't name names, but I was in a fraternity at the University of Virginia when I was
in college.
137
:And there was a guy in the same house that was younger than me.
138
:And we became great friends.
139
:But then you leave college and you kind of drift apart.
140
:And many years later, so this was like three or four years ago.
141
:like 15 years later, I haven't talked to this guy in many years and he pops up out of the
woodwork and he's running an auto dealership group and they have been our best client for
142
:the last two years.
143
:And so that's just like a good example of just if you, you know, if you're good to people,
it may not like immediately or directly benefit you.
144
:But in the long run, if you just build relationships and trust with people, eventually it
actually is very much in your own self-interest to do so because that stuff, the universe
145
:brings it back to you in some capacity.
146
:No, it's very true.
147
:And I just love that.
148
:I have relationship that started back in college.
149
:oh some time ago, which is awesome, when it comes to your current tools, because one of
the things that we did talk about is AI, and that's been a huge topic on this podcast.
150
:What does it look like on your guys' side for implementing that into a business?
151
:Yeah.
152
:So we've seen a trend.
153
:We've seen a trend.
154
:In 2024, it was a lot of like prototypes or like, let's dip our toes in the water with
this, right?
155
:Or let's address this like single use case at our business and see how it works.
156
:2025, it's all in.
157
:Like if you don't have an AI budget line item, like you're missing the boat at this point
in:
158
:And so what we're kind of preaching is that we need to stop talking about use cases and
start talking about platforms.
159
:And what do I mean by that?
160
:eh You know, companies need to...
161
:boost their employees' productivity.
162
:And they need to keep growing, but they don't necessarily want to hire all that much in
this environment, right?
163
:So how can we grow without hiring?
164
:Well, if our employees are 30 % more productive every day, then that can drive revenue and
profit.
165
:And so what we've been building is kind of on-prem.
166
:I say on-prem, but normally that just means on-client cloud infrastructure.
167
:AI's assistant and agent that can perform all sorts of different tasks for all sorts of
different employees because it's connected to their document management.
168
:It's connected to their single sign-in and authentication.
169
:Maybe their CRM or their accounting software.
170
:You know, it's, connected to all these, all these kind of data sources and it can be an
assistant and a tool for employees to do their jobs better, more efficiently, faster.
171
:produce more with the same amount of time.
172
:And so that is where we have been focused in this year specifically is building out these
platforms.
173
:Like all the big major companies, they have these already.
174
:Goldman Sachs has something called GSAI, BAE Systems rolled one of these out.
175
:JP Morgan has one, but they have the resources to build these internally.
176
:When you start looking at small to medium sized businesses, which I believe a lot of your
audiences is in that segment, they don't have the resources to build that out internally,
177
:typically, right?
178
:And so that's where we come in at Barefoot.
179
:We've actually built a product around this and that we're calling Compass, which is very
customizable and it's not SaaS, right?
180
:Because SaaS doesn't work for these businesses.
181
:They might handle use cases.
182
:But in terms of a platform, like an operating system for their business where they can do
a lot of different things, they need something that's for them, that's built for them,
183
:that's connected to their weird legacy database or their accounting system that nobody
uses anymore and like all sorts of things like that.
184
:That's what we've been building and selling in the last like six months, we'll say, and is
a real core focus of ours right now.
185
:Got it.
186
:So with this, I'm guessing with what you're able to offer to businesses, what I'm hearing
is that it helps with not having to use multiple AI systems.
187
:Is that kind of what I'm hearing?
188
:It's that, but also, you know, there's a lot of companies that are dealing in confidential
information all day, right?
189
:Think financial services, insurance, IT, uh banking, you name it.
190
:They can't just go to ChatGPT and upload customer information and ask questions about it,
right?
191
:That's a violation of the confidentiality.
192
:And so this is about providing tools for your team.
193
:So they're not leaking confidential information into these third party large language
model platforms, right?
194
:And so it is about boosting efficiency and not having to use a million different tools and
having a single place.
195
:But it's also about being able to connect to your proprietary data sources, right?
196
:When you're using third party tools, you often can't do that, but you need to.
197
:You need to be able to like...
198
:query data, it's not enough to just ask a large language model when you're trying to
operate a business and you're trying to do your job.
199
:And so it solves a few different pain points there.
200
:Got it.
201
:So you're pretty much saying that with the tool that you guys can help with, and I'm just
trying to make sure I have a full understanding, is that it's more of like a private
202
:sector so their information is protected.
203
:That's right.
204
:Yeah.
205
:One of the main reasons that we built this is that, you know, people whose jobs involve
confidential information are feeling left behind because if you're in sales and marketing,
206
:you're on chat GPT all day and you're doing all sorts of cool stuff and you're writing,
you know, AI emails and generating images and fun.
207
:But if you're an accountant, no, you're not.
208
:You're not using it at all.
209
:In fact, because
210
:you're doing work for a client and it can't help you with that unless it knows certain
things.
211
:And so that's where companies that are dealing in that kind of information on a regular
basis, maybe it's a code base that you're dealing with like for us, any number of things,
212
:like you can't just share these things.
213
:And so it's about having an internal system to help your employees be more productive.
214
:And this is actually I have a friend who's a real estate agent and she was telling me that
she couldn't use really chat BT when it comes to like certain things because there's
215
:policies that they have to follow when it comes to making sure that they don't share
certain things.
216
:I'm guessing that that's one reason why it would be private for a real estate agent so you
can customize it for them.
217
:Is that kind of what I'm hearing?
218
:Yep, that's a great use case.
219
:Maybe you want to use it to help you review a purchase contract.
220
:That's confidential.
221
:You cannot upload that into ChatGPT.
222
:But if you had your own internal private system, you can't.
223
:so there's many, many, most jobs, in fact, have some form of that that Compass is trying
to solve.
224
:Got it, now that's so awesome, because I do know many, you know, if it's real estate
agents, doctor offices, like all the things that you mentioned as well, and accountants.
225
:So that's a great tool for these people.
226
:When it comes to, do you have any real world examples that you'd like to share with us of
it, you know, helping a business, save them time, resources when it came to using your
227
:guys' services?
228
:I do, I have a couple examples, it's still early days, but here are some of the use cases.
229
:I had an attorney tell me that it saved six hours per week of his time.
230
:When you multiply that times an absurd hourly rate and the number of attorneys at a
particular firm, the number gets in the nine figures.
231
:I had a defense contractor say that they could probably write.
232
:25 % more proposals per year for government projects, which equates to a 25 % increase in
revenue over time.
233
:We have a pharmaceutical company that estimated they can design, I think it was 20 % more
clinical trials in a particular year, which is how they get paid.
234
:So these are like,
235
:These are not just like incremental, you know, like these are game changing numbers for
these businesses.
236
:You know, it's like hiring a whole new team to get more done without the cost of that.
237
:And it, you know, and I want to be clear what it's not.
238
:It is not about getting rid of people.
239
:It's about making your people more productive.
240
:Right.
241
:We're not trying, we're not going in and saying, we can, you can cut 20 % of your customer
support staff with this new tool.
242
:We're saying your customer, your existing customer support staff can now handle way more
tickets way faster than they used to.
243
:We're going to increase the velocity of it.
244
:That's the point of what we're trying to do.
245
:Right.
246
:It's not, it's not replacing AI, replacing workers.
247
:It's AI enhancing them and boosting them and raising them up.
248
:No, it's so, true.
249
:You know, just the importance of understanding that you still need a human being behind
the tools.
250
:Yeah, in data science we call that the human in the loop.
251
:Yes, I do think that's super, super important when it comes to, you know, I guess who,
guess who are you currently looking to?
252
:Are you looking more to reach out to the corporate?
253
:Like, you know, I mentioned a real estate agent that I know I'm guessing you would be
interested in reaching out to her person that she works under or over.
254
:Yeah.
255
:Yeah, this is not a tool for individual personal use.
256
:It's focused on businesses.
257
:you know, talking to CEOs, owners, CTOs, sometimes CMOs, sometimes at, you know, small to
mid sized companies is really who we're targeting with this.
258
:We're not trying to be an enterprise tool because they don't need us.
259
:They can do it themselves and they are doing it.
260
:themselves, but going after a law firm that has a couple hundred attorneys as an example,
or a defense contractor that has 30 employees.
261
:Both of these are real world examples.
262
:Those are the types of people that we're delivering this for.
263
:Yeah, totally.
264
:totally understand that, which is so awesome.
265
:And then when it comes to, you know, with AI, where are you seeing that it's going to be
out in the next few years?
266
:Yeah
267
:How much time we got?
268
:I mean, the pace of change is like never before.
269
:We've never seen this.
270
:And it's accelerating.
271
:It's actually getting faster.
272
:And you know,
273
:Within three years, you know, we're going to have AGI, artificial general intelligence,
that's going to be better than humans at most things, which is the definition of AGI that
274
:I like.
275
:And if you combine that with a few other things, so, I mean, one of the most exciting
things about that for me is in health and medicine, drug discovery specifically, I think
276
:is going to be the thing that
277
:blows this worldwide open.
278
:um With enough compute and sophisticated enough models, discovering new drugs that even
for, like right now the way it works is a disease has to be a big enough problem for these
279
:pharmaceutical companies to pour millions and billions into research to find new drugs for
them.
280
:And so if you have a rare disease, you don't get medicine.
281
:because the pharmaceutical companies, they can't get an ROI on that research.
282
:All of a sudden, the economics of that are completely different, right?
283
:And you can't, you can solve these long tail medical issues.
284
:And so I see a massive increase in health and the lifespans of people and crazy worldwide
changes, I think.
285
:It's like, what a time to be alive.
286
:We're going to see that in our generation.
287
:We're going to be on the front lines for that kind of change.
288
:Now this gets pretty out there, but I'll say it anyway.
289
:When you combine AGI with some couple technologies that are not quite as far along, but I
think are coming in within 10 years being nuclear fusion and quantum computing.
290
:Those three things.
291
:It's a whole new world, once we have, we basically have limitless compute and PhD level
AI, mean, beyond PhD level, genius level AI, right?
292
:Like, it's hard for me to imagine that the world isn't almost completely different than it
is today.
293
:Hmm, I'm sure.
294
:uh
295
:And I really believe that.
296
:know it sounds like science fiction, but go back and read science fiction from 30 years
ago.
297
:We're living it right now.
298
:Star Trek had mobile phones.
299
:That was their little try quarter.
300
:We have those now.
301
:Science fiction is actually just prediction of the future.
302
:We have robots.
303
:They're getting better.
304
:So I do see, I've drank the Kool-Aid on this one.
305
:And I'm not, I'm actually a pretty conservative person.
306
:I'm always saying, eh, pump the brakes.
307
:Like, let's not get over our skis here.
308
:But on this one, I don't know.
309
:I think we're, I think it's all coming and we, and we get to be a part of it.
310
:And that's so exciting.
311
:I'm, I am the last generation that remembers the world before the internet.
312
:And I get to see AGI.
313
:Like, man, nailed it.
314
:on timing.
315
:What an exciting, you know, lifetime to have, right?
316
:No, very true.
317
:And then when it comes to AGI, just in case our audience doesn't know what that is, can
you just share with us real quick?
318
:Yeah, AGI stands for artificial general intelligence and it's kind of like a major
milestone in the development of AI and different people define it in different ways.
319
:But the way that I like to talk about it is it's an AI that is better than humans at most
things, not necessarily all things, but at most things.
320
:And when that happens, the whole way that
321
:you know, work happens and productivity happens and energy production, like, like
everything changes.
322
:And then there's another one above AGI, which is like super intelligence, which is
basically like Skynet from the Terminator movies, I'm dating myself, but you know, one
323
:that is many, many levels above the smartest humans that we have.
324
:I don't know that we'll see that in our lifetime, but I think we'll get to a point where
the AI is better at most things than most people.
325
:understood.
326
:And that is so awesome, the directions that we're going and just, you know, I guess seeing
the internet come around.
327
:I think that's so awesome.
328
:I had an encyclopedia set in my home when I was a child.
329
:That's when we did reports.
330
:We looked up in the encyclopedia.
331
:Got it.
332
:feel like I, when I started school, I want to say we did have internet, so I did not get
that option, but that is still cool.
333
:When it comes to a company, do you feel like there is any pitfalls that comes with the
transformation of getting a private sector for AI?
334
:Yeah.
335
:So to me, I'm like raising the red flag.
336
:I'm telling this to anybody that will listen to me.
337
:You need an AI governance policy.
338
:You need rules for your employees about how, what they can and cannot do with AI.
339
:Right now, people are leaking confidential information all day.
340
:onto these third party platforms and it is putting your business at risk.
341
:And, you know, just like you have data governance and other HR policies, you need an AI
policy.
342
:I've like, I've, I've, all the companies I've talked to, I've talked to maybe two that
actually have a strong policy in place.
343
:I've asked HR people about this and they said, no, we don't even offer that as a service.
344
:We don't know how to do that.
345
:And so putting that, so that's step one.
346
:or one important thing and that will restrict your employees.
347
:And then you need to provide them with tools to unlock the potential of these things,
right?
348
:Lock it down, but then provide them what they need to be able to leverage this technology.
349
:It's too powerful to just say, there are major companies right now that say you are not
allowed to use AI with your job because it's too dangerous, but it is too powerful to
350
:ignore.
351
:because your competitors are going to figure out a way to use it.
352
:If you don't, then you're going to get left behind.
353
:So to me, that's the major pitfall.
354
:The other, would say, I mentioned this earlier, but stop chasing use cases individually
and start thinking about a platform for your business.
355
:Think about it as an operating system, right?
356
:Rather than having to log in to nine different things to figure out the answer to one
question,
357
:having a platform that's connected to all those things where you can just ask and it can
find the right information for you.
358
:Or you need to update a record in your CRM and you can just say, hey, can you update?
359
:I just got off the phone with so-and-so.
360
:Can you move that from phase two to phase three in our funnel rather than logging into
HubSpot?
361
:Like that's how I think business owners and executives need to be thinking about deploying
AI and stop chasing use cases.
362
:got it and so is that something that your current system can do?
363
:It is.
364
:It is.
365
:Well, we're on our journey there.
366
:It can do some of that and it's on its way to do all of that.
367
:But yes, that is what I believe businesses need right now and that's what we're building.
368
:Well, and I've used like the AI assistants and a lot of it's just providing you a chat but
not actually like they call themselves an AI assistant, but they're not actually like
369
:doing the work if that makes sense.
370
:And so, right.
371
:And so the distinction there, which is an important one, is the, it's the difference
between an assistant and an agent.
372
:You hear agentic AI sometimes and talking about AI agents, AI assistants kind of provide
information.
373
:AI agents take actions.
374
:They have tools that they can use.
375
:And what that means in a very practical sense is that
376
:AI assistants have knowledge that they can search and provide you, whereas AI agents,
they're connected via APIs to other systems and they can read and write, right?
377
:And create new records and update and all the rest of it.
378
:So right now, Compass is mostly an assistant on our way to becoming an army of agents.
379
:got it.
380
:and that's so awesome to know just because I've personally have used softwares where they
tell you they're the assistant.
381
:And then you find out are the agent and then you find out no, that's not exactly or you
maybe I just didn't understand the actual term for it, which is good.
382
:It's actually like, I've been writing about this because people are misusing this all the
time.
383
:Another one that people misuse all the time is, know, pre-training models, training
models, almost nobody's training models.
384
:The big, companies are training models.
385
:The kind of application layer are using these foundational models and applying use cases
to them, but in most cases, they're not actually training their own.
386
:But people say it all the time.
387
:You look on websites, it it everywhere, but it's technically not true.
388
:Got it.
389
:And when you say training a model, are you saying more of like the memory base of the chat
GBT or the AI tool?
390
:I'm talking about machine learning, training machine learning models, training deep neural
networks, which is how large language models work.
391
:And that's like what open AI is doing and what Mistral is doing and all these folks,
right?
392
:They're downloading, they're aggregating every piece of information they can find.
393
:And they're feeding it into these neural network algorithms to train a model to then, it's
a predictive model.
394
:You ask it a question.
395
:And it predicts what it thinks you want to Right.
396
:And so machine learning is all about prediction and these large language models are doing
that natural language processing, but they're prediction models.
397
:Really?
398
:What's the next word?
399
:And so when people say they're training models and they're just like using chat GPT
through an API and maybe give it some special instructions, that's not training, right?
400
:That's instructing that's prompting, but it's not training.
401
:So.
402
:Sorry to get on my soapbox, I'm getting tired of all the marketing materials that are just
like completely untrue.
403
:No, I'm sure and as a consumer or you know, I wouldn't know that.
404
:So thank you for sharing that with us.
405
:of the things that you did mention earlier is the way that you stay ahead of trends is
reading books.
406
:What are some of the books that you found very useful to stay ahead of trends?
407
:I got one right here.
408
:Hehehe
409
:The AI playbook, this is a really good one.
410
:It opens with this like staggering fact that I think it's like 90 % of the models that
businesses build never make it into production.
411
:Because there's a disconnect between the scientists and the business people.
412
:And there is often not someone in the middle that can coordinate and understand both
sides.
413
:Because data scientists, where they find joy typically, is in building models that can
predict things well.
414
:What they care less about, and I'm generalizing, of course, not all data scientists, but
what they care less about is predicting the right thing.
415
:And when you're deploying AI, you need to be thinking about the return on investment as
you're in the R &D phase.
416
:Are you trying to solve the right?
417
:And that is why 85 % of models never make it into production because you did solve a
problem and it's neat, but neat doesn't cut it.
418
:Money.
419
:Money is what drives businesses.
420
:Right?
421
:And so you need to make sure that your models are predicting the right thing and providing
the information that's going to move the needle.
422
:You have to cut costs.
423
:You have to drive revenue.
424
:You have to improve profit.
425
:You have to do something with a return for it to make sense.
426
:to deploy it across an entire organization, because it's a costly thing.
427
:So anyway, that's what this book is about, is like a kind of a playbook for making sure
that you're aligned across your different teams when designing and deploying AI at your
428
:organization.
429
:It's a really good one, I recommend it.
430
:Another thing I'll mention though, real fast, is a newsletter that I love, comes out every
day, it's called The Rundown AI.
431
:I read it every morning.
432
:It's like one of the first things I do when I wake up because while I love books, books
have lead times.
433
:Somebody's got to write that thing.
434
:Somebody's got to edit that thing.
435
:Someone's got to publish it.
436
:It gets printed and shipped to a bookstore.
437
:And like, it's obsolete by the time it hits my desk, you know?
438
:And so I've actually shifted my reading away from books for the same reason that I dropped
out of computer science when I was in college, which is that it's not current enough.
439
:It's all, you know, like unless it's something that has like long lasting value.
440
:But when I'm trying to learn about brand new technology that's coming out of books, don't
do it.
441
:But these like daily newsletters actually are really, really helpful for me.
442
:got it and I know a couple maybe weeks ago months ago we had somebody come in saying that
the that there's AI books that were written like 10 20 years ago are actually better than
443
:the ones currently and I know we're talking about trends but would you say that's true or
444
:think to a certain extent, yes.
445
:I think there's a bit of a gold rush thing happening right now where everybody's just
trying to get their book out.
446
:Whereas there wasn't the same feeling of urgency back then.
447
:And you could really think profoundly, you know, and think decades ahead and write about
that versus a lot of these books are writing about right now.
448
:And right now,
449
:is gone, just like that.
450
:And it's over and over every day, every week, every month, every year.
451
:So to some extent, I would agree with that, that some of the older stuff, like lands
better than that.
452
:Like, I'll give you an example.
453
:The Singularity is Near.
454
:That's a good one.
455
:That's a good one.
456
:And he just put out a new book that I haven't read yet, which is called The Singularity is
Nearer.
457
:Mm.
458
:But that's a good older AI book that ah I really like.
459
:Awesome.
460
:And these books, would you recommend them for professionals like you or a small business
owner or someone that runs a corporation?
461
:Are these books good for them as well?
462
:So the singularity is near is not practical.
463
:It's philosophical.
464
:So it's only for somebody that's interested in this stuff and wants to have kind of a
framework of thinking about it.
465
:But like, it's not going to help your business all that much or help you do your job any
better.
466
:Something like the AI playbook.
467
:It is that's a practical business book about the business of AI.
468
:And so, you you have to just keep in mind what you're consuming.
469
:Is this for
470
:just my own personal desire to learn or is this like something practical that I can, you
know, help me run my business?
471
:So.
472
:Which makes sense.
473
:Which makes sense.
474
:And before we go Hunter, do you have anything that you'd like to share with us?
475
:Feel like I've shared a lot, maybe overshared even.
476
:now I here I'll leave you I'll leave me with a thought, which is that like.
477
:There are people that are very afraid of this technology, a lot of people.
478
:And I've had kind of a core principle or value or I don't know what you belief we'll call
it.
479
:And when it comes to things such as this, and a lot of people like to make a connection
between nuclear weapons and AI if it falls into the wrong hands and all the rest of it.
480
:But I don't, I'm not, I'm not in that camp.
481
:While I think...
482
:We need to be safe and careful about how we build this stuff.
483
:do, because it is extremely powerful.
484
:I believe that people in general are good and that there are way more good people than bad
people.
485
:And so overall, the good will prevail and this technology will be put to amazing uses and
save millions of lives.
486
:and improve the life of everyone.
487
:And so I'll leave you with that on the positive note that this is going to be the most
powerful technology that any of us have ever seen.
488
:And while it can be scary, I think very good, very smart people are going to make it
improve everyone's life.
489
:Very true.
490
:No, and I agree with that.
491
:And so thank you for sharing that with us.
492
:And I do believe that if the people that do decide to use it for negative, they're going
to do it either way.
493
:So it's important that you have an understanding of it.
494
:Hunter, if people are looking to get in contact with you, how do they do that?
495
:I encourage them to, I love talking about this stuff.
496
:Please reach out to me.
497
:I will respond.
498
:Just say, Hey, I heard you on Shae Toye's podcast and wanted to chat about something like
I'll go back to you.
499
:Email me hunter at barefoot solutions.com.
500
:I'm also pretty active on LinkedIn or you can go to our website at barefoot solutions.com,
fill out a contact form there and we'll get in touch with you.
501
:But I want to be clear, like you don't have to be a potential client for me to
502
:talk with you or so hit me up.
503
:Thank you so much Hunter and you guys thank you for watching
