Riding the Wave of Technology to Elevate Innovation and Individual Distinction

WEBVTT

1
00:00:00.080 --> 00:00:03.319
The topics and opinions expressed in the
following show are solely those of the hosts

2
00:00:03.319 --> 00:00:06.000
and their guests, and not those
of W FORCY Radio, it's employees their

3
00:00:06.000 --> 00:00:10.039
affiliates. We make no recommendations or
endorsements for radio show, program, services,

4
00:00:10.080 --> 00:00:13.439
or products mentioned on air or on
our web. No liability, explicit

5
00:00:13.519 --> 00:00:17.000
or implied shall be extended to W
Forcy Radio or its employees or affiliates.

6
00:00:17.160 --> 00:00:20.480
Any questions or comments should be directed
to those show hosts. Thank you for

7
00:00:20.559 --> 00:00:31.000
choosing W FORCY Radio. What's working
on Purpose? Anyway? Each week we

8
00:00:31.120 --> 00:00:35.200
ponder the answer to this question.
People ache for meaning and purpose at work,

9
00:00:35.399 --> 00:00:40.159
to contribute their talents passionately and know
their lives really matter. They crave

10
00:00:40.240 --> 00:00:44.399
being part of an organization that inspires
them and helps them grow into realizing their

11
00:00:44.479 --> 00:00:48.039
highest potential. Business can be such
a force for good in the world,

12
00:00:48.200 --> 00:00:53.079
elevating humanity. In our program,
we provide guidance and inspiration to help usher

13
00:00:53.200 --> 00:00:58.039
in this world we all want working
on Purpose. Now, here's your host,

14
00:00:58.240 --> 00:01:06.519
Doctor Release Cortez. Welcome back to
the Working Purpose Program. They optunity

15
00:01:06.519 --> 00:01:08.200
again this week. Great to have
you. I'm your host, Doctorleis Cortez

16
00:01:08.319 --> 00:01:11.120
Join you live from Dallas, which
is home based for me. If we've

17
00:01:11.159 --> 00:01:15.439
not met yet and you don't know
me, I'm a management consultant, organizational

18
00:01:15.439 --> 00:01:19.319
logo therapist, speaker and authors.
My team and I italyst Cortez and Associates

19
00:01:19.319 --> 00:01:23.400
help companies to enliven their operations by
building inspirational leaders and cultures activated by meeting

20
00:01:23.400 --> 00:01:27.920
and purpose to turn those companies from
a flatline EKG to a vironrant workplace.

21
00:01:29.239 --> 00:01:33.599
There people are intrinsically motivated to perform
with their best, grow into their highest

22
00:01:33.599 --> 00:01:36.879
potential, and are committed to stay
and help deliver on the company's mission.

23
00:01:37.079 --> 00:01:38.879
You can learn more about on how
we can work together at a last Cortez

24
00:01:38.959 --> 00:01:44.560
dot com. Now let's getting Today's
program with us is the chief technology officer

25
00:01:44.560 --> 00:01:47.560
of a cutting edge tech company who
is a big believer in the power of

26
00:01:47.680 --> 00:01:52.400
language models like chat GPT to revolutionize
the way we work and learn. Travis

27
00:01:52.480 --> 00:01:57.280
James is a fractional CTO of Tripe
Health Solutions based in Texas and Texas Frisco.

28
00:01:57.319 --> 00:02:00.640
Specifically, he has more than twenty
years experience in software development, machine

29
00:02:00.719 --> 00:02:07.200
learning, architecture, natural language processing, and decentralized distributed systems. Today we'll

30
00:02:07.239 --> 00:02:09.960
be talking about what he is currently
pouring himself into at Tripealth Solutions, and

31
00:02:10.000 --> 00:02:15.680
here his perspective on how the AI
tool chat GBT can be leveraged to elevator

32
00:02:15.680 --> 00:02:21.919
productivity and creativity and distinguish our uniqueness
as humans journs today from Frisco, Texas.

33
00:02:22.039 --> 00:02:25.000
Travis, Welcome to Working on Purpose. Oh, thank you so much

34
00:02:25.000 --> 00:02:30.719
for having me absolutely, And you
know, this idea came because of a

35
00:02:30.759 --> 00:02:35.479
conversation that we had about me sharing
something about how what I had read about

36
00:02:35.520 --> 00:02:38.919
how universe professors were going to have
a hard time actually dealing with this new

37
00:02:38.960 --> 00:02:44.520
tool that's come out here and students
in deciding who had actually written the paper,

38
00:02:44.560 --> 00:02:46.919
And you gave such a great perspective
on that, which I want to

39
00:02:46.919 --> 00:02:49.639
get into later. That that is
what made me say, you got to

40
00:02:49.639 --> 00:02:52.800
come on Working Purpose and talk about
this and help us understand. But most

41
00:02:52.800 --> 00:02:55.319
of us don't get about this new
tool, so we'll get there eventually.

42
00:02:55.360 --> 00:03:00.159
But first, Travis, I want
to start by just talking about your background.

43
00:03:00.560 --> 00:03:04.520
Um, it's pretty unique. I
think that you started your technology background.

44
00:03:04.560 --> 00:03:07.680
I think you told me that you
started writing code at age seven?

45
00:03:07.919 --> 00:03:12.479
Is that right? Age seven?
About? Really about age eight? Okay,

46
00:03:13.719 --> 00:03:16.400
So well you have to figure out
first for why what was it about?

47
00:03:16.400 --> 00:03:19.919
What was the impetus? Why did
you start coding at such a young

48
00:03:19.960 --> 00:03:23.120
age. Well, it's really about, UM wanted to spend time with my

49
00:03:23.159 --> 00:03:29.639
father. Actually, my father was
a research chemist at the time, and

50
00:03:30.080 --> 00:03:35.479
he was getting an eight degree in
computer science a message a green computer science

51
00:03:36.039 --> 00:03:38.680
m you know, at night on
the weekends. And so it was right

52
00:03:38.680 --> 00:03:42.840
around at the time that we were
living in Virginia at the time. Him

53
00:03:42.919 --> 00:03:47.599
he got his degree from VCU in
Richmond, Virginia, And if I wanted

54
00:03:47.599 --> 00:03:51.800
to spend time with him, I
spent time with him doing what he did

55
00:03:51.800 --> 00:03:54.759
on the weekends, and so I
would learn everything about, uh, you

56
00:03:54.800 --> 00:03:59.560
know, the kind of programs.
Who was writing. Um I even used

57
00:03:59.560 --> 00:04:02.039
a couple and everything to gave up
and write my own little programs and things

58
00:04:02.080 --> 00:04:05.960
like that. The power of love. That's still beautiful, Travis, I

59
00:04:06.000 --> 00:04:10.639
love that. Okay. Well,
so then from there you have a mass

60
00:04:10.719 --> 00:04:13.400
quite a career. I know,
you how to stint at Microsoft and later

61
00:04:13.479 --> 00:04:15.839
year we in American Airlines as an
enterprise architect. And you have found your

62
00:04:15.879 --> 00:04:20.319
way along some pretty interesting entrepreneurial places
as well. So can you say a

63
00:04:20.360 --> 00:04:23.319
little bit of it, just kind
of give us a bit of a tour

64
00:04:23.399 --> 00:04:27.240
of your background. Yes, So
it was always a dream of mine to

65
00:04:27.399 --> 00:04:32.160
work for Microsoft. When I was
in my mid twenties, I was doing

66
00:04:32.160 --> 00:04:38.759
some work for McAfee, which on
the virus scan product and some network solutions

67
00:04:38.800 --> 00:04:44.040
products there, and it just happened
that a former manager of mine there joined

68
00:04:44.120 --> 00:04:47.680
Microsoft and then recruited me to go
to go with them, and so I

69
00:04:47.720 --> 00:04:55.360
took the opportunity and started with Microsoft, and it was one of the best

70
00:04:55.360 --> 00:04:59.000
experiences I've had in my life,
you know, even to this moment.

71
00:05:00.279 --> 00:05:06.040
And why did you leave? Well, ironically, I left because I kind

72
00:05:06.079 --> 00:05:13.040
of rose to a point within Microsoft
while I was working in the consulting services

73
00:05:13.120 --> 00:05:18.079
division, to a point where I
either had to take long periods of time

74
00:05:18.199 --> 00:05:25.360
on the road to places like South
Africa, or I had I was married

75
00:05:25.360 --> 00:05:29.480
and had young children at the time, and um, and I kind of

76
00:05:29.480 --> 00:05:35.360
made the choice of at the time
of family over career. Understood understood,

77
00:05:36.399 --> 00:05:41.199
Okay, Well, so how did
you find your way into the entrepreneurial world.

78
00:05:42.160 --> 00:05:47.000
That's very interesting, Um, so
after I left Microsoft, Microsoft was

79
00:05:47.040 --> 00:05:53.399
still the way I made money.
I started immediately a Microsoft Solutions provider,

80
00:05:53.800 --> 00:05:58.759
which at the time whether a route
for people who had my expertise. Microsoft

81
00:05:58.839 --> 00:06:03.639
was trying to limit their risk by
having external entities writing code on their behalf

82
00:06:03.680 --> 00:06:08.600
after having worked at Microsoft, and
so I did that for about ten years

83
00:06:08.600 --> 00:06:14.600
from nineteen ninety eight to two thousand
and eight or and during that time I

84
00:06:14.720 --> 00:06:19.600
did some work for Mark Cuban on
Broadcast dot Com. One of my specialties

85
00:06:20.160 --> 00:06:26.480
is video streaming technologies as well,
and they were using Microsoft Media Encoder to

86
00:06:26.800 --> 00:06:32.040
encode video that was sent over the
Internet from Indiana and other places for those

87
00:06:32.319 --> 00:06:36.480
basketball games you want to transmit.
And I wrote some tools that allowed you

88
00:06:36.839 --> 00:06:43.160
to insert local commercial content in the
middle of those streams without breaking them up.

89
00:06:43.839 --> 00:06:48.600
And it was that technology that he
sold to Yahoo for four billion dollars.

90
00:06:49.639 --> 00:06:55.120
So I made I don't know,
maybe a million and a half in

91
00:06:55.240 --> 00:06:58.560
nineteen ninety nine doing that work.
He made four billion. That was all

92
00:06:58.639 --> 00:07:02.920
I needed to sell me on Entrepreneur. What a great story brows. That's

93
00:07:02.920 --> 00:07:09.040
amazing, totally amazing. M Yeah, that's kind of an interesting return on

94
00:07:09.160 --> 00:07:15.000
investment there. So what's also interesting
to me is you specialize in building innovative

95
00:07:15.040 --> 00:07:18.639
solutions that leverage artificial intelligence, natural
language processing, real time communications, and

96
00:07:18.800 --> 00:07:24.439
IoT. That's just a quite an
interesting combination there. How did you end

97
00:07:24.519 --> 00:07:28.759
up heading down that particular path,
Well, you know, it was my

98
00:07:28.959 --> 00:07:32.439
experience of being a consultant during those
two years that Microsoft sent me to their

99
00:07:32.519 --> 00:07:40.160
most difficult clients. And there's a
component of the technologies that we have at

100
00:07:40.199 --> 00:07:45.360
Tribe Health today that are directly descendant
from my work during those ten years.

101
00:07:46.160 --> 00:07:53.839
Ironically, while those sound like vastly
you know, different technical areas, they

102
00:07:53.920 --> 00:07:59.480
all work with the same dynamics,
They have the same kind of structure and

103
00:08:00.040 --> 00:08:03.560
to them. And as a result, I built engines because my background was

104
00:08:05.480 --> 00:08:09.680
in building compilers. I worked on
the compiler for CPUs, PASS and other

105
00:08:09.759 --> 00:08:15.079
languages including dot net just before I
left. And I build things as an

106
00:08:15.120 --> 00:08:18.319
empty container that you populate with the
knowledge of the things that you want to

107
00:08:18.399 --> 00:08:24.199
manage, and then a way to
execute on those things. And as a

108
00:08:24.319 --> 00:08:28.759
result, just about anything that you
could write can be modeled that way,

109
00:08:28.160 --> 00:08:35.000
regardless of the dynamic, and then
they can be combined to create an ecosystem

110
00:08:35.480 --> 00:08:39.759
of things that communicate with each other, and that is a structure essentially of

111
00:08:39.200 --> 00:08:45.879
what we've built with Tribe Core at
Tripell Solutions. I wanted to hear about

112
00:08:45.879 --> 00:08:48.120
that as well, but before we
do, I think it's interesting. I

113
00:08:48.159 --> 00:08:50.480
don't know if you remember this and
some of our other conversations, but I

114
00:08:50.639 --> 00:08:54.679
started my human capital career in recruiting
and I was specifically an IT recruiter for

115
00:08:54.840 --> 00:08:58.639
those first five years. And what's
interesting to me about you working in the

116
00:08:58.720 --> 00:09:01.759
space is that you don't have an
electical engineering degree. You have a mechanical

117
00:09:01.799 --> 00:09:05.759
engineering degree. Correct, That is
correct? And how does that You've said

118
00:09:05.799 --> 00:09:09.320
something before about this, but how
does that perspective, How do you think

119
00:09:09.360 --> 00:09:13.440
that changes your perspective the way that
you approach coding. Well, there are

120
00:09:13.480 --> 00:09:18.120
two things that that work together to
change that make my perspective and coding a

121
00:09:18.159 --> 00:09:22.919
little different. One is simply the
fact that I physically can't remember a time

122
00:09:22.919 --> 00:09:26.960
that I didn't write some kind of
because I started so early. As a

123
00:09:28.080 --> 00:09:33.559
result that then it's something that more
of like what I am than you know

124
00:09:33.679 --> 00:09:37.080
what I do. And because of
that, I have a different perspective on

125
00:09:37.240 --> 00:09:43.120
how to learn how to do it
and how to execute the steps. So

126
00:09:43.279 --> 00:09:46.559
that's, you know, that's the
first thing that comes to mind when it

127
00:09:46.639 --> 00:09:50.679
comes, you know, to doing
that. The second thing is that when

128
00:09:50.720 --> 00:09:54.240
you are taking engineering at a university. I was at University of Texas,

129
00:09:56.080 --> 00:10:00.799
isn't possible to memorize everything they need
to know about fluid animates, about you

130
00:10:00.840 --> 00:10:03.559
know, leadens, their law,
about all these things. There are a

131
00:10:03.600 --> 00:10:07.600
few core formulas that you remember and
you derive everything from those, right,

132
00:10:09.600 --> 00:10:16.080
And it is that thought process,
that pattern that is different from the way,

133
00:10:16.639 --> 00:10:22.320
at least from what I've understood so
far my career hiring computer science students

134
00:10:22.600 --> 00:10:26.799
is different from their perspective that they
memorize a lot of syntax linguistics for a

135
00:10:26.960 --> 00:10:33.879
specific language and they get kind of
tied to that technology. UM and as

136
00:10:33.919 --> 00:10:37.360
a result, UM, it's less
it's more of a memorization exercise than engineering

137
00:10:37.440 --> 00:10:43.600
exercise. And as software has evolved, it has been become a lot more

138
00:10:43.639 --> 00:10:50.720
important to be more engineering focused structure
and UH and understanding that the behavior and

139
00:10:50.840 --> 00:10:58.080
dynamics, and and that is what
I believe puts UM not just those who

140
00:10:58.159 --> 00:11:01.720
have an engineering background ahead, but
also those who are have degrees in the

141
00:11:01.799 --> 00:11:07.440
languages, who speak multiple languages,
multiple grammars. They end up with a

142
00:11:07.559 --> 00:11:16.759
lot easier transition into understanding multiple types
of languages and technologies and a stack fascinating

143
00:11:16.879 --> 00:11:20.240
love that. Well, now to
learn one other bit here that I think

144
00:11:20.360 --> 00:11:22.879
is really interesting about you in the
way that you approach your work, because

145
00:11:22.879 --> 00:11:26.679
I know you have a passion for
beauty in the design of things. I'm

146
00:11:26.759 --> 00:11:31.000
interested in how you express your this
kind of passion in your work. Yes,

147
00:11:31.159 --> 00:11:39.720
so it has to do with the
personality that is attached to the software

148
00:11:39.759 --> 00:11:43.519
you build in how it behaves.
For me, software is a living thing,

149
00:11:45.759 --> 00:11:50.039
is a thing that never as long
as it's being used, it never

150
00:11:50.159 --> 00:11:52.559
dies. It's not a thing that
you just complete and then walk away from.

151
00:11:52.960 --> 00:11:58.360
It's something that forever lives through multiple
iterations and cycles. And when you

152
00:11:58.519 --> 00:12:05.039
start a project, you kind of
pick a personality for that software that governs

153
00:12:05.240 --> 00:12:07.919
how it's built, what the rules
are for extending it, and um and

154
00:12:09.000 --> 00:12:13.679
in how you ensure that can bring
bug free and then always does the job

155
00:12:13.799 --> 00:12:16.720
that is expective of it and UH. It is that kind of type of

156
00:12:16.799 --> 00:12:22.399
thought process that UM, I believe
makes things a little different in in in

157
00:12:22.519 --> 00:12:28.720
how I build and UH and it's
the essence of what I mean about art

158
00:12:28.960 --> 00:12:35.279
and technology, because not everything can
be uh you know, programmed in to

159
00:12:35.759 --> 00:12:39.799
your your software. There's unexpected conditions, there's there's a way that software UM

160
00:12:39.879 --> 00:12:46.519
can be designed to um to respond
to conditions it didn't expect. And a

161
00:12:46.639 --> 00:12:50.720
lot of that can be aggressive in
a certain way. It could be very

162
00:12:50.799 --> 00:12:54.320
defensive, it could be proactive.
It's just like a it's just pretty much

163
00:12:54.360 --> 00:13:00.320
like a human being. Really,
m that's fascinating again. A thing,

164
00:13:00.399 --> 00:13:03.240
None of these things that you've been
saying so far are things that I think

165
00:13:03.279 --> 00:13:05.799
that I really have heard others talk
about. So I think that's your very

166
00:13:05.840 --> 00:13:07.679
fresh perspective, or maybe have been
out of the game way too long.

167
00:13:07.720 --> 00:13:11.639
I don't know. But so then
now we have to hear about what you're

168
00:13:11.679 --> 00:13:16.000
now doing at tri Health Solutions.
What's that all about. Well, Tripalth

169
00:13:16.080 --> 00:13:22.399
Solutions is all about empowering the individual
and to be the constodian in management of

170
00:13:22.480 --> 00:13:28.600
their own medical data. That is
probably the most broken thing when it comes

171
00:13:28.639 --> 00:13:33.200
to our healthcare system today is the
fact that the individual is not the one

172
00:13:33.320 --> 00:13:37.480
the steward of their own data.
It's trapped in some silo that is actually

173
00:13:37.519 --> 00:13:41.080
owned by someone else who's the one
making money off the fact that they have

174
00:13:41.200 --> 00:13:48.559
your data. This is what restricts
the movement that you can have within the

175
00:13:48.159 --> 00:13:54.360
healthcare ecosystem. It makes it difficult
for you to you know, change it,

176
00:13:54.519 --> 00:14:00.200
change providers, or or share your
data with specialists that you might need,

177
00:14:00.799 --> 00:14:05.240
get multiple opinions from multiple sources,
um and and create an ecosystem.

178
00:14:05.320 --> 00:14:09.440
You can't create your own ecosystem around
your own data. These are the things

179
00:14:09.679 --> 00:14:15.360
that we are wanting to address at
Tribe Health because it's the key to the

180
00:14:15.440 --> 00:14:20.720
future, especially as more disruptive ways
of providing medicine have appeared, like wellness

181
00:14:20.799 --> 00:14:24.720
practice practices, um. You know, more people going to chiropractors, more

182
00:14:24.720 --> 00:14:28.759
people going to aesthetics practices for skin
and other things that have a little bit

183
00:14:28.840 --> 00:14:33.120
of a medical component to it.
And then a lot of people are into

184
00:14:33.240 --> 00:14:39.039
optimization, whether it be you know, mental health optimization, counseling, meditation,

185
00:14:39.559 --> 00:14:45.080
UM and UH and all of these
things require when you do them to

186
00:14:46.159 --> 00:14:50.000
TOM at a high level, then
require your data like an ordering like what

187
00:14:50.120 --> 00:14:54.679
I'm wearing here keeps track of my
sleep patterns, It keeps track of like

188
00:14:54.759 --> 00:14:58.440
heart right variability and other things that
can be points to other types of warnings

189
00:14:58.559 --> 00:15:05.000
for medical conditions or simply tell me
that I'm over training and I probably shouldn't

190
00:15:05.000 --> 00:15:09.759
work out today. So who is
your ultimate customer who buys your is it

191
00:15:09.840 --> 00:15:16.320
a device? Who buys well?
Our software is provided as a service and

192
00:15:16.519 --> 00:15:20.480
is sold in I would say three
different ways. One, it's software as

193
00:15:20.480 --> 00:15:24.279
a service and we simply we have
a module that does a specific job for

194
00:15:24.360 --> 00:15:28.600
you that will integrate well with your
EMR system or what other system that you're

195
00:15:28.639 --> 00:15:33.320
looking for looking for, and you
can just buy that and pay a subscription

196
00:15:33.399 --> 00:15:41.759
fee and then you're off and running. Right then there our software that is

197
00:15:41.759 --> 00:15:46.399
a platform that you can build your
own software on top of. So our

198
00:15:46.440 --> 00:15:50.679
customers for that would be other startups
that are trying to do something disruptive in

199
00:15:50.399 --> 00:15:56.360
the field that require that they have
a way to build an ecosystem around their

200
00:15:56.559 --> 00:16:00.919
particular part of the services they provide. Um. You know, those types

201
00:16:00.960 --> 00:16:04.559
of things can be sleep centers,
like like what we're doing with a company

202
00:16:04.600 --> 00:16:11.279
called meta Bio in in Australia where
where they can basically allow you to be

203
00:16:11.360 --> 00:16:15.360
able to take control of and share
your sleep study data with you know with

204
00:16:15.559 --> 00:16:19.840
any of your other practitioners, UM
it can be uh you know, any

205
00:16:19.879 --> 00:16:23.639
of these software companies that we have
with two software companies that are using our

206
00:16:23.679 --> 00:16:30.279
platform UM with in conjection with the
generative AI services to do things like fix

207
00:16:30.360 --> 00:16:34.799
the fix the notes that that a
doctor takes to assure that when they show

208
00:16:34.840 --> 00:16:38.399
up in a medical record there they
show up in a way that is likely

209
00:16:38.480 --> 00:16:41.720
for them to be able to get
paid on time. And then also UM

210
00:16:42.399 --> 00:16:48.879
likewise other generative AI technology that that
can fix uh medical records that provide UH

211
00:16:49.080 --> 00:16:55.840
notes that that defend a particular charge
when a charge or a preorthization or something

212
00:16:55.879 --> 00:16:59.639
like. It's like that is denied. Not as fascinating, Travis, That

213
00:16:59.720 --> 00:17:03.320
just cool. Now if that weren't
enough already what you're doing. I also

214
00:17:03.360 --> 00:17:07.279
noticed on your LinkedIn profile that you're
up to something called think King. What

215
00:17:07.519 --> 00:17:15.640
is that? Yes, So thinking
is really kind of my incubator of ideas

216
00:17:17.599 --> 00:17:22.240
entity where where we actually started with
an earlier version of what is now Chat,

217
00:17:22.319 --> 00:17:26.839
GPT and all those types of technologies
I just work with in natural language

218
00:17:26.880 --> 00:17:36.160
processing on a project with Stanford University
perhaps eight nine years ago, where we

219
00:17:36.480 --> 00:17:38.960
work with their computer science department to
build a project that was called Almond,

220
00:17:40.119 --> 00:17:44.839
which was basically an open source version
of Alexa and allows you to give you

221
00:17:44.880 --> 00:17:48.720
a model for doing skills and things
like that same thing as Alexa does.

222
00:17:48.799 --> 00:17:52.720
The difference is all the models will
will execute locally so that you can preserve

223
00:17:53.240 --> 00:17:59.839
privacy and all of that, and
it varies similar to how Siri works in

224
00:18:00.279 --> 00:18:07.240
principle, where everything gets processed locally, but then giving you access to well

225
00:18:07.559 --> 00:18:11.839
an ecosystem of other providers that can
that can provide you skills that then you

226
00:18:11.960 --> 00:18:17.240
can download and run locally. And
this is this is a lot better than

227
00:18:17.759 --> 00:18:21.039
you know, uh, than a
model like Alexa, where everything goes up

228
00:18:21.039 --> 00:18:22.519
to the POWD you are listening in
on all your conversations and you get it

229
00:18:22.599 --> 00:18:30.119
all. So so in Google's very
similar, uh. This basically guarantees your

230
00:18:30.160 --> 00:18:34.559
privacy will be preserved while at the
same time ensuring that that you can have

231
00:18:34.720 --> 00:18:41.200
models that are that are consistently updated
and trained at the same time so that

232
00:18:41.240 --> 00:18:45.240
you get the same level of performance. Troubles. That's amazing. You know,

233
00:18:45.319 --> 00:18:47.480
I really have a trumous respect for
what you do, because, as

234
00:18:47.519 --> 00:18:49.680
you know, it's way out of
the line of what I do. So

235
00:18:51.160 --> 00:18:55.119
really appreciate you sharing getting some foundation
here on what where you come from and

236
00:18:55.240 --> 00:18:57.720
really what goes into your work.
Let's grab our first break on your host,

237
00:18:57.799 --> 00:19:02.599
doctor Earleis Quartez on Aero Travis James, who is the fractional CTO of

238
00:19:02.759 --> 00:19:04.920
Tripe Health Solutions. We've been hearing
about where he's come from, how he's

239
00:19:04.960 --> 00:19:07.960
built us expertise. After the break, we're going to get into how we

240
00:19:08.000 --> 00:19:11.680
can actually use chat GBT to become
more efficient, effective, and distinguished.

241
00:19:11.960 --> 00:19:30.160
Stay with us, We'll be right
back. Doctor Elise Cortez is a management

242
00:19:30.200 --> 00:19:36.279
consultant specializing in meaning and purpose.
An inspirational speaker and author, she helps

243
00:19:36.319 --> 00:19:41.680
companies visioneer for a greater purpose among
stakeholders and develop purpose inspired leadership and meaning

244
00:19:41.720 --> 00:19:45.640
infused cultures that elevate fulfillment, performance, and commitment within the workforce. To

245
00:19:45.839 --> 00:19:49.559
learn more or to invite a lease
to speak to your organization, please visit

246
00:19:49.640 --> 00:19:53.839
her at Elise Cortez dot com.
Let's talk about how to get your employees

247
00:19:53.920 --> 00:20:03.759
working on purpose. This is working
on purpose with doctor release Cortez. To

248
00:20:03.880 --> 00:20:07.519
reach our program today or to open
a conversation with a lease, send an

249
00:20:07.559 --> 00:20:14.480
email to a lease a l se
at Elise Cortez dot com. Now back

250
00:20:14.559 --> 00:20:22.319
to working on Purpose. Thankteresting with
us, and welcome back to working on

251
00:20:22.440 --> 00:20:26.440
Purpose if you're just joining the program
now. My guest is Travis James.

252
00:20:26.519 --> 00:20:30.759
He is the founder of Tripe Health
Solutions. He specializes in building innovative solutions

253
00:20:30.799 --> 00:20:34.400
that leverage artificial intelligence, natural language
processing, real time communications, and IoT.

254
00:20:34.880 --> 00:20:38.519
I'm your host, doctor Elise Cortez. Okay, so now we have

255
00:20:38.599 --> 00:20:41.319
to get into how we can actually
use this stuffic before we do. For

256
00:20:41.400 --> 00:20:45.759
our listeners and viewers who don't really
know anything about chat GBT, let me

257
00:20:45.839 --> 00:20:48.599
just say a couple of things about
it that I've learned from my research and

258
00:20:48.839 --> 00:20:52.799
what you've taught me by sharing some
of these resources. So I understand that

259
00:20:52.200 --> 00:20:56.640
chat GBT is. Well, we
can already say it isn't a psychic oracle

260
00:20:56.720 --> 00:21:00.559
or something out of sci fi well
like some people might be aching. It's

261
00:21:00.559 --> 00:21:03.920
a language model using huge sums of
data. Um, so I think that's

262
00:21:04.000 --> 00:21:07.200
really interesting just to be able to
start there. So can you say,

263
00:21:07.480 --> 00:21:11.519
help us understand a bit more about
how how it actually works. Sure,

264
00:21:11.680 --> 00:21:19.079
so you know what GPT stands for
is a generative pre trained transformer. And

265
00:21:19.799 --> 00:21:23.960
the third yet you to hear talk
of GPT two, three, and four.

266
00:21:25.359 --> 00:21:30.839
You know, those are generations of
the model. In each generation they

267
00:21:30.960 --> 00:21:38.000
is trained on an increasingly larger set
of of of seed data and data points

268
00:21:38.480 --> 00:21:44.559
you know, so um uh like. For instance, GPT three was trained

269
00:21:44.599 --> 00:21:48.839
on one hundred and seventy five billion
parameters. GPT four was trained one close

270
00:21:48.880 --> 00:21:57.119
to a trillion parameters. Now what's
unique about GPT technology is in the t

271
00:21:57.319 --> 00:22:03.839
part. The transformer part is important
because it is a it's a new approach

272
00:22:04.079 --> 00:22:08.920
to machine learning and how to train
a model and UM and how to have

273
00:22:10.119 --> 00:22:15.279
that model execute um at runtime based
on changes that it might have but you

274
00:22:15.400 --> 00:22:19.920
might have in the environment UM and
and that is what makes GPT magical.

275
00:22:21.359 --> 00:22:26.079
There have been models for natural lanage
processing and and and natural landage understanding and

276
00:22:26.200 --> 00:22:32.559
things like that that could perform a
lot of these functions UM based on a

277
00:22:32.640 --> 00:22:36.079
similar method. You train it on
a huge amount of data and and then

278
00:22:36.160 --> 00:22:40.960
you get these models are able to
predict things or classify things. The difference

279
00:22:41.119 --> 00:22:44.960
is that in order to make changes
to those older models, you had to

280
00:22:45.279 --> 00:22:49.720
train them again, right, which
on a data set um uh and that

281
00:22:49.920 --> 00:22:55.839
that training is extreme expense, it
takes a long time to do and uh.

282
00:22:55.920 --> 00:23:00.359
And it's also fixed. It's fixed
with regard to the types of ambers

283
00:23:00.400 --> 00:23:04.279
that it can understand. And then
there are limits on its on its function.

284
00:23:04.720 --> 00:23:11.519
Transformer technologies allow you to take a
base model and expand it in its

285
00:23:11.960 --> 00:23:18.240
features and functions based on data you
provided. Right. And what's interesting about

286
00:23:18.359 --> 00:23:25.559
chat GPT is that you can make
it remember your h the history of your

287
00:23:25.640 --> 00:23:30.799
conversations by providing what are called embeddings, where basically you take the other the

288
00:23:30.880 --> 00:23:34.640
previous questions and answers and things like
that, and then you can feed those

289
00:23:34.759 --> 00:23:41.160
back into the next request so that
the engine has a context for understanding things.

290
00:23:41.799 --> 00:23:47.960
Now, the base chat GPT model
can really only remember about the last

291
00:23:48.039 --> 00:23:51.839
four or five things you said.
Much have you ever played with it where

292
00:23:51.880 --> 00:23:53.720
you want to try to do a
long conversation, but about four or five

293
00:23:55.839 --> 00:23:59.839
iterations in, it's going to forget
all about It's what you're talking about right

294
00:24:00.200 --> 00:24:03.839
in the base model. And that
has to do with the fact um that

295
00:24:03.400 --> 00:24:07.000
with each request you're sending a larger
and larger amount of data. The tokens

296
00:24:07.200 --> 00:24:12.279
the amount. A token is a
personally uh three force of a word that

297
00:24:12.720 --> 00:24:17.960
um that is all the words that
you have sent before it into the engine.

298
00:24:18.319 --> 00:24:21.279
And then you approach a limit,
like in the base model there's like

299
00:24:21.319 --> 00:24:23.960
a two thousand to token limit.
If you're using the paid version, there's

300
00:24:23.960 --> 00:24:30.240
a four thousand token limit. UM. If you're using upt uh four with

301
00:24:30.799 --> 00:24:36.000
open ai services like through um uh, you can get up like five hundred

302
00:24:36.559 --> 00:24:40.240
is a huge limit and you get
up the five hundred pages of information back

303
00:24:40.279 --> 00:24:44.480
and forth. Well you know so
UM, but if you're paid by the

304
00:24:44.519 --> 00:24:48.319
token um when it comes to the
paid models, and so you want to

305
00:24:48.319 --> 00:24:52.519
be efficient with that and and to
be efficient that is the reason why you

306
00:24:52.640 --> 00:25:00.400
have word embeddings, which basically build
an equation that represents a catalog of all

307
00:25:00.440 --> 00:25:06.039
the similar stuff similar to the to
the inputs that you have given, so

308
00:25:06.200 --> 00:25:10.480
that you give the engine a very
rich context with a minimal amount of data.

309
00:25:11.440 --> 00:25:15.599
And that's how you optimize those engines
so that they can perform better.

310
00:25:15.960 --> 00:25:21.440
And another trait of GPT technologies is
that you can fine tune them, which

311
00:25:21.559 --> 00:25:25.599
is a mini training method that doesn't
require to retrain the whole thing, which

312
00:25:25.720 --> 00:25:30.920
you can train on data sets that
you provided that that prevent what are called

313
00:25:30.079 --> 00:25:34.000
hallucinations. If you ever have heard
this, yes, yeah, if you're

314
00:25:34.039 --> 00:25:40.920
heard of hallucinations, chat GPT will
will depending on the parameters you give it,

315
00:25:41.680 --> 00:25:44.400
it will attempt to give you an
answer, which means sometimes it'll just

316
00:25:44.559 --> 00:25:48.720
make it up and the thing you'll
find it sometimes the things that tells you

317
00:25:48.799 --> 00:25:52.920
don't actually exist, and things like
this, but they logically will fit,

318
00:25:52.640 --> 00:25:57.440
but you don't really exist. And
so to prevent hallucinations, you're certain parameters

319
00:25:57.559 --> 00:26:03.799
that you need to set in order
to ensure that the engine will only answer

320
00:26:03.880 --> 00:26:08.680
back things that are that actually exist, and then also provide a boundary around

321
00:26:10.000 --> 00:26:15.880
the logic that it can use in
order to construct an answer. Fascinating.

322
00:26:15.920 --> 00:26:18.039
I'm so glad I had you want. This is just I'm learning all kinds

323
00:26:18.039 --> 00:26:21.680
of things, and I did actually
read several articles about this before I even

324
00:26:21.720 --> 00:26:23.559
brought you on, so that you're
taking it a whole little level. So

325
00:26:25.039 --> 00:26:29.079
one thing I have learned is that
there's the criticality of prompts, and the

326
00:26:29.480 --> 00:26:33.759
quality of your prompt is really what
delivers the quality of a response. Talk

327
00:26:33.799 --> 00:26:37.240
to us about prompts. Yeah,
So prompts. Prop engineering is what is

328
00:26:37.279 --> 00:26:42.200
now called is extremely important and is
the thing that can ensure that you can

329
00:26:42.440 --> 00:26:52.000
use GPT type technologies and large language
models to expand the number of tasks that

330
00:26:52.079 --> 00:26:57.519
it can perform. So to give
an example, currently in the base model,

331
00:26:57.559 --> 00:27:02.279
you put a prompt in there,
um um, it will provide you

332
00:27:02.440 --> 00:27:07.880
back um, you know, the
the best answer that it can um.

333
00:27:07.039 --> 00:27:11.440
But that is without context, there's
an error. Is a language reeken where

334
00:27:11.680 --> 00:27:17.400
you can provide as Microsoft supports,
where you can provide a context for the

335
00:27:17.480 --> 00:27:22.200
system itself to say system you are
a an expert in accounting, right,

336
00:27:22.720 --> 00:27:26.240
right, And then you can also
set context to who the user is.

337
00:27:26.559 --> 00:27:30.839
The user is um a person that
really needs accounting help. Right, You

338
00:27:30.880 --> 00:27:37.200
set that context, and then you
can provide other context to uh in the

339
00:27:37.319 --> 00:27:40.559
you know, in a form of
lists and all and things like that,

340
00:27:40.799 --> 00:27:45.519
and embeddings about the situation that you're
wanting to ask about. And then you

341
00:27:45.640 --> 00:27:51.759
ask the question right now, with
that formed prompt structure, you're going to

342
00:27:51.799 --> 00:27:56.000
get back very accurate responses, right, that are more accurate than if you

343
00:27:56.119 --> 00:28:00.400
just blast away, you know,
without any structure to your prompt. And

344
00:28:00.680 --> 00:28:03.920
I've even played with us because I
have I did, I did, I

345
00:28:04.000 --> 00:28:07.920
do have the free version. I've
actually done done a few chats now and

346
00:28:07.000 --> 00:28:11.079
it's been interesting. But even in
what I discovered is that when you tell

347
00:28:11.119 --> 00:28:15.519
it to do something like give me
a satirical response, something that specific,

348
00:28:15.240 --> 00:28:18.759
M was pretty interesting that you could
actually really define or something that's cutting edge

349
00:28:18.839 --> 00:28:23.079
or something that would be considered controversial. I thought that that was really pretty

350
00:28:23.160 --> 00:28:27.200
pretty interesting that you could be that
specific and it would actually deliver a pretty

351
00:28:27.440 --> 00:28:33.880
pretty finite response to that. Oh
absolutely, um chat GPT. GPT three

352
00:28:36.119 --> 00:28:41.240
was the version that was starting to
become aware to that extent to understand nuances

353
00:28:41.279 --> 00:28:49.079
and meaning and understanding. Now GPT
four goes way beyond that. It's able

354
00:28:49.119 --> 00:28:56.880
to actually construct logic around those areas
of subject matter. Um it is.

355
00:28:56.880 --> 00:29:00.559
It is more than a little leaps
and bound the head of GPT three,

356
00:29:02.119 --> 00:29:06.519
and it's in its potential to provide
guidance, answers and things like that.

357
00:29:07.920 --> 00:29:11.720
How long has this technology been around? Transformer technologies have been around for about

358
00:29:11.839 --> 00:29:19.160
four to five years, and it
started with chat with what not Chat with

359
00:29:19.319 --> 00:29:26.680
the GPT two, which was the
first model that really started getting good results.

360
00:29:26.839 --> 00:29:30.880
With regard to asking random questions in
a particular subject matter area in getting

361
00:29:30.920 --> 00:29:37.960
back really good responses, GPT three
went another level and can actually construct lists

362
00:29:38.039 --> 00:29:42.720
of things, of suggestions for things
to do right when you give a certain

363
00:29:42.759 --> 00:29:48.599
condition, and JP four goes even
beyond that to be able to draw conclusions.

364
00:29:49.680 --> 00:29:55.359
Wow, that's cool. Yeah,
that's amazing, absolutely amazing. Okay,

365
00:29:55.400 --> 00:29:56.759
So now I want to bring this
into the application here. So,

366
00:29:57.599 --> 00:30:00.160
as I said in the very beginning, what made me want you to come

367
00:30:00.200 --> 00:30:03.680
on the show was we had a
conversation about I was saying something about gosh,

368
00:30:03.839 --> 00:30:07.960
universities, I'd hate to be a
professor today, and trying to decide

369
00:30:07.000 --> 00:30:11.119
who actually wrote these papers, and
you said something like, well, it's

370
00:30:11.160 --> 00:30:14.039
going to make it's going to require
them, professor to be more on their

371
00:30:14.079 --> 00:30:17.599
toes and know their area of their
area. Which all that was really interesting.

372
00:30:17.680 --> 00:30:19.480
So now what I want to pull
up for you is I want you

373
00:30:19.519 --> 00:30:23.200
to talk to us about how chat
GBT can be used to challenge traditional models

374
00:30:23.200 --> 00:30:29.960
of education and can it replace human
teachers? Yes? Actually, okay,

375
00:30:30.079 --> 00:30:37.200
so the answer to both those is
yes. I've long been had as a

376
00:30:37.440 --> 00:30:41.599
side interest, you know, disrupting
education And to give you an example of

377
00:30:41.680 --> 00:30:47.920
where I believe education needs to be
disruptive is that educational systems, especially public

378
00:30:48.359 --> 00:30:52.039
education, they are forced to orient
their model around the low common denominator.

379
00:30:52.319 --> 00:30:59.559
The average student you know who is
who is either a visual or auditory learner

380
00:31:00.160 --> 00:31:03.359
UM, not so much the hands
on type learner and UH. And as

381
00:31:03.400 --> 00:31:07.319
a result, there are kids who
are really smart that suffer from kind of

382
00:31:07.359 --> 00:31:10.480
being a square peg in a you
know, in a in a round hole

383
00:31:10.880 --> 00:31:17.720
going through standard education because of GPGPT
four special specifically, which is what we

384
00:31:17.839 --> 00:31:22.559
call multimodal, meaning that it can
process more than just text and things like

385
00:31:22.640 --> 00:31:27.279
that. It can process sound,
video, UM images and provide those back.

386
00:31:27.880 --> 00:31:33.400
So as a result, UM as
a training tool it is is particularly

387
00:31:33.519 --> 00:31:37.880
powerful. UM. There's there's not
anything, any subject matter that you couldn't

388
00:31:37.960 --> 00:31:44.920
explore using with a proper, proper
set of prompts and and learn about UM

389
00:31:45.200 --> 00:31:49.759
using GPT four because you can set
your level of context as as UM you

390
00:31:49.799 --> 00:31:53.240
know, the user at any level. If you ever have tried to say,

391
00:31:53.279 --> 00:31:57.599
you know, explain you know you
know Meeton's third law of thermodynamics,

392
00:31:59.559 --> 00:32:04.119
UM know to as if I was
a five year old it will give you

393
00:32:04.519 --> 00:32:10.319
an excellent response that that will take
an into account. Take five, that's

394
00:32:10.359 --> 00:32:14.359
amazing, and you can move up
from there to say, you know,

395
00:32:14.599 --> 00:32:17.880
whatever background you have, you want
to learn about you know mathematics, and

396
00:32:19.079 --> 00:32:24.480
mathematics, I believe is one of
the places that UM GPT technologies or uh

397
00:32:24.839 --> 00:32:30.359
you know, general of AI technologies
can be of particular use because it's hard

398
00:32:30.400 --> 00:32:37.319
to find truly good math teachers understand
math at a level that a. You

399
00:32:37.359 --> 00:32:44.119
know, it's notoriously the UM the
hated subject for most kids in school is

400
00:32:44.200 --> 00:32:49.599
math, and it's all because it's
not taught properly. And T three and

401
00:32:49.759 --> 00:32:54.279
four can provide you a math teacher
that will understand you and how to explain

402
00:32:55.000 --> 00:33:00.880
certain nuances to you, whether it
be there about the equations you know,

403
00:33:00.039 --> 00:33:07.200
different equations, calculus, algebra,
trigonometry, all those can be put in

404
00:33:07.319 --> 00:33:13.000
context in your context, and it
really helps when it comes especially since you

405
00:33:13.039 --> 00:33:21.400
can also eventually train GPT to take
into account cultural background type of things.

406
00:33:21.880 --> 00:33:25.119
So when if you come from a
particular cultural background, you have a greater

407
00:33:25.279 --> 00:33:30.519
chance of understanding the content that is
in textbooks and things like that than someone

408
00:33:30.599 --> 00:33:37.400
from a different background and these things
can be accounted for and in the use

409
00:33:37.759 --> 00:33:45.640
of a generative AI technologies to assist
in training people to understand anything of any

410
00:33:45.720 --> 00:33:50.160
complexity from where they are. It's
amazing, Travis, this is so cool.

411
00:33:50.559 --> 00:33:52.359
Let's scrub our last break here.
I'm your host, Dctriley's Quartez.

412
00:33:52.440 --> 00:33:55.559
We've in the air with PRAVS James, who is the founder of Tripe Health

413
00:33:55.640 --> 00:33:59.480
Solutions. We've been talking a bit
about how we can use CHAGBT. Now

414
00:33:59.599 --> 00:34:01.279
first education, after the rake,
we're going to get into how we can

415
00:34:01.359 --> 00:34:04.960
use it in other industries, including
healthcare. Stay with us, We'll be

416
00:34:05.079 --> 00:34:23.920
right back. Doctor Elise Cortez is
a management consultant specializing in meaning and purpose.

417
00:34:24.039 --> 00:34:29.599
An inspirational speaker and author, she
helps companies visioneer for a greater purpose

418
00:34:29.679 --> 00:34:35.239
among stakeholders and develop purpose inspired leadership
and meaning infused cultures that elevate fulfillment,

419
00:34:35.400 --> 00:34:38.480
performance, and commitment within the workforce. To learn more or to invite a

420
00:34:38.559 --> 00:34:43.800
lease to speak to your organization,
please visit her at Elise Cortez dot com.

421
00:34:44.239 --> 00:34:52.320
Let's talk about how to get your
employees working on purpose. This is

422
00:34:52.360 --> 00:34:57.760
working on purpose with doctor Elise Cortez
to reach our program today, or to

423
00:34:57.880 --> 00:35:01.800
open a conversation with a lease,
send an email to a lease Ali se

424
00:35:02.159 --> 00:35:13.239
at Eleas Cortez dot com. Now
back to working on purpose. Thanks for

425
00:35:13.280 --> 00:35:15.920
saying with us, and welcome back
to working on purpose if you're just joining

426
00:35:15.960 --> 00:35:19.199
the program now. My guest is
Travis James. He is the fractional CTO

427
00:35:19.280 --> 00:35:22.079
of Tribe Health Solutions. He has
over twenty years of experience and software development,

428
00:35:22.320 --> 00:35:28.280
machine learning, architecture, natural language
processing, and decentralized distributed systems and

429
00:35:28.400 --> 00:35:31.840
has been teaching us about how to
leverage chat GBT to increase effectiveness and to

430
00:35:31.920 --> 00:35:37.480
distinguish our individual contribution. I'm your
host talk to Elias Quartez. So I

431
00:35:37.559 --> 00:35:39.559
knew this on your LinkedIn profile that
you have a that you put you re

432
00:35:39.719 --> 00:35:44.679
shared an article that I thought was
pretty interesting. It was it was about

433
00:35:44.760 --> 00:35:49.360
how chat GBT for health provider for
help for help providers and how how can

434
00:35:49.400 --> 00:35:52.840
the chatbot make the professional jobs easier? So can you comment a bit about

435
00:35:52.920 --> 00:35:58.119
that? How does it make their
jobs easier? Sure? So, um,

436
00:35:58.440 --> 00:36:05.360
you know the cool thing they're about
chat about GPT technologies as as it

437
00:36:06.239 --> 00:36:10.480
specifically applies to healthcare really have to
do with the fact there's a shortage of

438
00:36:10.599 --> 00:36:15.400
healthcare workers in certain areas. To
give the example, there's a shortage is

439
00:36:15.400 --> 00:36:21.760
a dramatic shortage of mental health practitioners. And it's very difficult to find a

440
00:36:21.920 --> 00:36:25.840
mental health practitioner, especially for children, which is a passion of mine because

441
00:36:25.840 --> 00:36:30.880
I have a son who has schizophrenia, and it was a real difficult journey

442
00:36:30.960 --> 00:36:36.000
to find, um the right kind
of practitioners that would take the time to

443
00:36:36.159 --> 00:36:39.800
diagnose them properly and get them on
the right medication right. And so to

444
00:36:39.880 --> 00:36:45.400
give an example where I've done.
One thing I've done personally with GPT technologies

445
00:36:45.679 --> 00:36:53.559
is I've used them, along with
certain outputs from assessments to determine, um,

446
00:36:54.239 --> 00:37:00.079
what type of generation second generation antipsychotics
would be better to prescribe for my

447
00:37:00.119 --> 00:37:06.960
son for schizophrenia. And there's some
thirty different variations that all act on the

448
00:37:07.079 --> 00:37:10.519
same receptors and everything in different ways
have different side effects. But what I

449
00:37:10.599 --> 00:37:16.719
found alarming was that these statistics are
horrible with regard to getting the right medication

450
00:37:17.119 --> 00:37:22.119
applied. There's only on average,
there's only a twenty percent chance you're going

451
00:37:22.199 --> 00:37:27.199
through a this is going to a
practitioner that you have found after probably months

452
00:37:27.239 --> 00:37:31.920
of waiting, and they do an
assessment and they write a prescription and that

453
00:37:32.159 --> 00:37:37.840
prescription has a twenty percent chance of
working. Right. Wow, Now I

454
00:37:38.119 --> 00:37:43.800
have constructed one of the models that
we're that we're playing with where I can

455
00:37:43.920 --> 00:37:49.480
hit it around eighty percent at the
moment. With some specific training other things

456
00:37:49.519 --> 00:37:52.519
we're working on, we could get
that tone to where we because all the

457
00:37:52.679 --> 00:37:59.280
data that's needed to support you is
already out there and you actually have a

458
00:37:59.400 --> 00:38:02.519
lot of it, but it's not
actionable. And that's that's what Tribe Health

459
00:38:02.679 --> 00:38:07.199
is about, is getting your medical
records, can all your medical data and

460
00:38:07.280 --> 00:38:10.760
all the data about you put in
a form that is actionable to be able

461
00:38:10.800 --> 00:38:15.800
to help you, that tools can
be used to help you. And that

462
00:38:15.079 --> 00:38:19.559
is what's what's going to change,
Uh, you know, the face of

463
00:38:19.679 --> 00:38:22.320
medicine is being able to do that, especially since one of the uses of

464
00:38:23.239 --> 00:38:30.679
chat GPT is to train have fine
tune models that can take you through conversations

465
00:38:30.239 --> 00:38:35.920
with what would be normally a therapist
because it's such a shortage, right and

466
00:38:36.320 --> 00:38:42.119
that you know, you could provide
some measure of self care that's automated through

467
00:38:42.199 --> 00:38:47.280
AI that can bridge the gap right
between your current state and your two B

468
00:38:47.440 --> 00:38:52.519
condition at some level. And we
started that work a couple of years ago

469
00:38:52.840 --> 00:38:58.639
when we wrote an app that is
now used by UM it's called Corticode,

470
00:38:59.679 --> 00:39:05.119
so by a company called Lexipole that
is used by first responders all across the

471
00:39:05.199 --> 00:39:10.880
country. We talked about nine percent
of the jurisdictions inside of California use cortaco

472
00:39:12.280 --> 00:39:16.440
to support the mental health wellness of
their police officers. And in there we

473
00:39:16.599 --> 00:39:21.719
have a little mini assessment that you
can take. They say whether you or

474
00:39:21.760 --> 00:39:24.280
at risk for PTSD, for you
know, polar disorders and these other little

475
00:39:24.320 --> 00:39:30.559
things. And the data is kept
local in private, so that it cuts

476
00:39:30.800 --> 00:39:36.280
in the stigma that is normally associated
with pursuing mental health treatment and counseling.

477
00:39:36.480 --> 00:39:39.400
And that's where the box, as
they evolved in mental space, could really

478
00:39:39.440 --> 00:39:43.679
help out a lot, because it
cuts the stigma out. You're not talking

479
00:39:43.719 --> 00:39:49.000
to a person that where you're worried
that that person might tell your your superior

480
00:39:49.079 --> 00:39:52.599
officer or someone that you wouldn't want
to know. That you were there right

481
00:39:53.400 --> 00:39:59.360
all private and encrypted for you.
And that's what our platform provides you is

482
00:39:59.679 --> 00:40:02.719
a way to make your data actionable
in a number of ways, not just

483
00:40:02.960 --> 00:40:07.920
mental health, but also you know, when it comes to you chreology,

484
00:40:07.039 --> 00:40:10.239
when it comes to you know,
anything that can be predicted from your sleep,

485
00:40:10.880 --> 00:40:15.920
All those things can be done in
a way that you can be alerted

486
00:40:15.000 --> 00:40:20.159
first of certain things that you would
normally go to a doctor for a check

487
00:40:20.239 --> 00:40:23.039
up this heat and then tell you
that, hey, you might want to

488
00:40:23.119 --> 00:40:28.199
go to your chahalitius because you might
have um, you might be at risk

489
00:40:28.239 --> 00:40:30.719
for a stroke because we detected these
types of you know, these little patterns

490
00:40:31.719 --> 00:40:37.599
and things that you would have to
otherwise wait for a negative event to occur

491
00:40:38.000 --> 00:40:42.719
that then gets you into the healthcare
system to then maybe find it or maybe

492
00:40:42.760 --> 00:40:49.079
not. Wow, Travis, You
know I very frequently on the show when

493
00:40:49.079 --> 00:40:51.199
I'm talking to my guest, I
get so present too. You know,

494
00:40:51.320 --> 00:40:53.159
we all have one precious life.
What are you doing with your one precious

495
00:40:53.199 --> 00:40:55.920
life to make a difference? And
I really applaud when you're doing with your

496
00:40:55.960 --> 00:41:00.880
one precious life to make a difference. This is just amazing. So now

497
00:41:00.920 --> 00:41:04.079
I'm going to ask a very selfish
question. So you know that I'm a

498
00:41:04.119 --> 00:41:07.639
management consultant, and of course the
work that I do is so much about

499
00:41:07.719 --> 00:41:12.079
speaking and teaching and such, and
it's obviously it's got a psychological component to

500
00:41:12.159 --> 00:41:15.840
it's got a spiritual component to it. How can I better use this kind

501
00:41:15.880 --> 00:41:20.960
of technology to better serve my clients? Well, you know that is another

502
00:41:21.079 --> 00:41:27.920
place where I believe this will have
in the next generations GPPT five and beyond

503
00:41:28.840 --> 00:41:35.000
will have great impact on your ability
clone yourself too. That sounds good.

504
00:41:35.039 --> 00:41:37.000
I can I can handle that.
That sounds good. So so you take

505
00:41:37.079 --> 00:41:42.719
all of your expertise in your field, all your experiences or your knowledge,

506
00:41:43.239 --> 00:41:47.199
and let's say you can embed all
of that into a model that is trained

507
00:41:49.400 --> 00:41:57.400
similarly to chat GPT to be chat
GPT dash doctor Cortest, and you can

508
00:41:57.440 --> 00:42:06.559
actually offer up a chat where where
a person can talk to you get all

509
00:42:06.639 --> 00:42:09.280
of your answers as if you the
one giving them, because it's been trained

510
00:42:09.320 --> 00:42:13.360
on all of your logic, your
behavior, all you think you've ever written

511
00:42:14.639 --> 00:42:19.119
um and the parameters around how you
come to certain types of conclusions, and

512
00:42:19.559 --> 00:42:22.760
it would do a very good job
of giving the same kinds of answers.

513
00:42:22.840 --> 00:42:27.000
And then you can set in and
into it parameters that say if if your

514
00:42:27.400 --> 00:42:30.199
if your tolerance level as far as
the you know, probability of the answer

515
00:42:30.280 --> 00:42:35.880
being correct gets to x amount that
you drop out and say, well,

516
00:42:35.960 --> 00:42:38.400
I really can't answer that for you, and um, and then we will

517
00:42:38.719 --> 00:42:43.639
let you pick it up with you
or someone you designate to be able to

518
00:42:43.840 --> 00:42:47.599
get into the human version of your
ecosystem. And that's where I think it

519
00:42:47.679 --> 00:42:53.239
allows you to Um, it would
allow you to make more money off your

520
00:42:53.280 --> 00:42:59.559
content because now people don't just read
your content, they can have a conversation

521
00:43:00.079 --> 00:43:05.880
with you about that content and with
in relation to everything else in the world

522
00:43:05.920 --> 00:43:09.679
has ever been written in that subject. And uh, and then allow you

523
00:43:10.320 --> 00:43:16.920
to be able to provide a mechanism
for sharing your knowledge with others without having

524
00:43:17.000 --> 00:43:21.719
be the one that they're doing it. Oh my gosh, job is that

525
00:43:21.840 --> 00:43:23.320
so exciting? Thank you so much
for that question. That was so cool.

526
00:43:23.719 --> 00:43:28.320
Okay, now we can't really not
address the big elephant to the room

527
00:43:28.360 --> 00:43:31.039
for some people that are really just
frightened silly by this technology here. They're

528
00:43:31.039 --> 00:43:36.320
worried about job losses various fields,
you know, they're worried lots of things.

529
00:43:36.519 --> 00:43:38.920
It's going to take over, it's
going to people will become obsolete,

530
00:43:39.519 --> 00:43:45.239
how do you how do you respond
to those kinds of big concerns? Well,

531
00:43:45.679 --> 00:43:51.719
um, and it depends on what
your purpose as a person is as

532
00:43:51.760 --> 00:43:58.960
to whether you need to worry um
and and depending on your purpose, UM,

533
00:43:59.119 --> 00:44:01.239
it may not you probably need to
worry, but you probably need to

534
00:44:01.280 --> 00:44:07.039
adjust you think in a way.
You know, if because see chat,

535
00:44:07.119 --> 00:44:13.079
GPT can't replace a human it can, it can be. It can perform

536
00:44:13.159 --> 00:44:15.719
a lot of human tasks, artomate
a lot of them, but it never

537
00:44:15.880 --> 00:44:19.840
takes away a job. It would
change the way you would do a job.

538
00:44:20.800 --> 00:44:25.039
And if you're chounting on um,
you know, gas engines. You

539
00:44:25.079 --> 00:44:30.719
know, if you're you're banking on
gas engines for for as an example being

540
00:44:30.800 --> 00:44:36.559
the way we're going to build cars
forever and um and you feel entitled in

541
00:44:36.719 --> 00:44:42.119
being able to UM to build and
service gas engines, Well, that's not

542
00:44:42.199 --> 00:44:50.239
a reasonable expectation. So because we
all evolve, everything evolves, you always

543
00:44:50.280 --> 00:44:53.760
need a human for something. And
what your job would probably change to be

544
00:44:54.400 --> 00:45:01.360
is being an input mechanism to this
engine to to continue to train it rather

545
00:45:01.440 --> 00:45:06.480
than actually doing that engine's job,
which your job is also will work more

546
00:45:06.519 --> 00:45:10.239
money, And that's what I've always
said, and that's I've talked about this

547
00:45:10.280 --> 00:45:13.440
in my first book as well,
so I aligned very much with that.

548
00:45:13.599 --> 00:45:15.000
So now if we can take it
one step further, and this is where

549
00:45:15.079 --> 00:45:20.159
I get really excited about this,
as well as somewhere in a conversation or

550
00:45:20.199 --> 00:45:22.960
something that I read here something about
the notion of how we can actually use

551
00:45:23.039 --> 00:45:28.440
technologies like chat GPT to enhance our
own individual distinction in the way that we

552
00:45:28.559 --> 00:45:31.480
work and express ourselves. How can
we do that well? I can give

553
00:45:31.480 --> 00:45:37.400
you a personal example on that.
I think you know, as a lot

554
00:45:37.440 --> 00:45:42.920
of my friends and family do for
about fifteen years or more actually Netclace or

555
00:45:43.000 --> 00:45:46.119
seventeen. Every morning when I do
my personal reading and personal development, I

556
00:45:46.239 --> 00:45:52.199
pick a quote out of that reading
or a quote that someone shared in the

557
00:45:52.320 --> 00:45:55.840
context of that reading, and that
strikes a chord with me. It's something

558
00:45:55.920 --> 00:46:00.440
I'm trying to work on, or
is something that means something to me,

559
00:46:00.719 --> 00:46:07.199
and I send it out as a
text quote to to everyone UH on my

560
00:46:07.679 --> 00:46:12.239
current list and I probably have like
forty people on that list termedly now um.

561
00:46:12.519 --> 00:46:15.840
Over the years. At first it
was just text I would send UH.

562
00:46:15.960 --> 00:46:20.480
Then I started mixing in images and
I would go out to something like

563
00:46:21.039 --> 00:46:24.480
Getty's Images or something like that,
pick an image that I think would would

564
00:46:24.559 --> 00:46:28.519
fit the quote and everything, because
I found it was it was a powerful

565
00:46:28.599 --> 00:46:32.400
combination. You're hitting two s two
senses there, you know, UM or

566
00:46:32.480 --> 00:46:38.119
two types of thought patterns when it
comes to art and UH and then text

567
00:46:38.800 --> 00:46:43.199
uh, you know, to some
really cement the idea of the quote.

568
00:46:44.000 --> 00:46:49.920
Well, since then, I've gone
another level where I use chat GPT to

569
00:46:50.079 --> 00:46:55.000
actually generate a style of prompt that
I then passed to a stable diffusion model

570
00:46:55.440 --> 00:47:00.800
to generate an image that provides an
idea of the quote that has never existed

571
00:47:00.840 --> 00:47:04.760
before, and then I put the
text of the quote on top of that.

572
00:47:06.159 --> 00:47:12.599
And that I have found to be
the most powerful expression so far of

573
00:47:12.760 --> 00:47:21.119
the ideas with these quotes, because
it it provides a visual that is tied

574
00:47:21.280 --> 00:47:25.400
to the essence of the quote.
And I have just in the past year

575
00:47:25.920 --> 00:47:30.599
evolved how I generate the prompts with
those things and the chain that I use

576
00:47:30.679 --> 00:47:37.000
for those things in order to generate
specifically the types of images the imagery that

577
00:47:37.280 --> 00:47:40.400
you want. And you know the
future of this is to be able to

578
00:47:40.760 --> 00:47:46.199
use that in mental health treatment and
also just encouraging people even in the workplace,

579
00:47:46.920 --> 00:47:50.920
and you can take to take into
account because the generator of model,

580
00:47:51.440 --> 00:47:53.480
you don't have to generate the same
image for everyone. You can generate a

581
00:47:53.559 --> 00:47:59.840
different image person based on their particular
situation, which you know about them,

582
00:48:00.079 --> 00:48:04.039
or what or what the engine knows
about them because of what they what information

583
00:48:04.119 --> 00:48:07.719
they provided, and you can end
up with really powerful personalized representations of an

584
00:48:07.840 --> 00:48:14.719
idea that you want to express to
those people that would then really help them.

585
00:48:15.280 --> 00:48:19.480
That's so cool. Well, we
ran out of time already here,

586
00:48:19.519 --> 00:48:21.480
but I want to give you the
chance to close here. So Travis,

587
00:48:21.519 --> 00:48:23.119
you knew the show was listened to
or watched by people around the world.

588
00:48:23.239 --> 00:48:27.440
We care about eating the elevating experience
of work in our place in it.

589
00:48:27.639 --> 00:48:30.039
What would like to leave people with
today? I would like to leave people

590
00:48:30.079 --> 00:48:38.039
with the idea that technology sometimes gets
a bad rap of being impersonal and really

591
00:48:38.159 --> 00:48:45.159
less than artistic. And I would
say it's it's actually in reverse that that

592
00:48:45.519 --> 00:48:52.239
technology, when used properly, can
take into account all of the factors that

593
00:48:52.440 --> 00:48:58.119
make each and every person unique and
give them precisely the answers that they need

594
00:48:58.760 --> 00:49:01.679
when they need them. In the
context that they need them. What a

595
00:49:01.719 --> 00:49:05.920
beautiful way to finish. Travis,
You've sold me for sure. Thank you

596
00:49:06.039 --> 00:49:07.400
very much for being Working on Purpose. It's been great to have you and

597
00:49:07.480 --> 00:49:10.400
great to learn from you and be
inspired by you. Thank you, Thank

598
00:49:10.440 --> 00:49:15.000
you so much. Listeners be worth
the world. Learn more about Travis James

599
00:49:15.079 --> 00:49:17.079
or the working his team are doing
at Tribe Health Solutions. Just go to

600
00:49:17.320 --> 00:49:22.880
Tribe Health Solutions dot com. See
you next week for another elevating conversation,

601
00:49:22.920 --> 00:49:24.559
and remember that works in an integral
and important part of our lives. It

602
00:49:24.599 --> 00:49:28.599
can be one of the best adventures
and means of realizing our potential and making

603
00:49:28.639 --> 00:49:35.000
the impact we crave. So let's
work on Purpose. We hope you've enjoyed

604
00:49:35.079 --> 00:49:38.079
this week's program. Be sure to
tune into Working on Purpose featuring your host,

605
00:49:38.239 --> 00:49:43.519
doctor Elise Cortez, each week on
W four C. Why Together We'll

606
00:49:43.559 --> 00:49:49.320
create a world where business operates conscientiously. Leadership inspires and passion performance and employees

607
00:49:49.360 --> 00:49:52.239
are fulfilled in work that provides the
meaning and purpose they crave. See you

608
00:49:52.360 --> 00:49:54.199
there, Let's work on Purpose.