WEBVTT

1
00:00:03.720 --> 00:00:08.599
The information economy as a rod.
The world is teeming with innovation as new

2
00:00:08.640 --> 00:00:13.919
business models, we invent every industry
industry. Inside Analysis is your source of

3
00:00:14.000 --> 00:00:18.480
information and insight about how to make
the most of this exciting new era.

4
00:00:18.800 --> 00:00:24.079
Learn more and inside analysis dot Comside
Analysis dot com. And now here's your

5
00:00:24.120 --> 00:00:35.560
host, Eric Kavanaugh. All Right, ladies and gentleman, welcome back to

6
00:00:35.600 --> 00:00:39.280
the Gilman Coast to Coast radio show
all of doctor Information of Economy. It's

7
00:00:39.280 --> 00:00:43.799
time for Inside Analysis or should be
Eric Kavanaugh with my European Chappel Gang or

8
00:00:43.840 --> 00:00:48.359
Eve Mulferents out there. And we
were just chit chatting for our usual Monday

9
00:00:48.520 --> 00:00:51.560
catch up and we picked up on
a couple of things and I want to

10
00:00:51.840 --> 00:00:56.439
really dive into this because there's some
interesting ideas that are being hashed out right

11
00:00:56.479 --> 00:01:00.600
now in our industry and it team
varying format and tumultuous time. And I

12
00:01:00.600 --> 00:01:04.840
would have to say, and what
sputned this all was our show last week

13
00:01:04.879 --> 00:01:10.159
which was great with Jamak de Gani
from next Data and also Doug Kimball and

14
00:01:10.200 --> 00:01:15.200
Simmitpou Fromanto Text and Onto Text of
course does knowledge graphs. Next Data,

15
00:01:15.280 --> 00:01:19.400
which is a Jamak's new company.
She came from thought Works, where she

16
00:01:19.439 --> 00:01:23.239
came up with this whole concept of
data mesh. At this company, now

17
00:01:23.280 --> 00:01:26.799
she's building the data mesh. But
it's a small company and they're from what

18
00:01:26.840 --> 00:01:30.239
I can tell, I think they're
self funded, so they they're working I'm

19
00:01:30.280 --> 00:01:33.079
sure, around the clock to try
to build this stuff out. And the

20
00:01:33.079 --> 00:01:38.359
analogy that she gave was that they're
trying to do for data what Kubernets did

21
00:01:38.480 --> 00:01:45.000
for processing. So, of course
Kubernets containerization, container orchestration you sort of

22
00:01:45.519 --> 00:01:49.879
break down the monolithic approach to rating
system but sort of a de facto operating

23
00:01:49.920 --> 00:01:56.959
system, this container orchestration into lots
of little bitty pieces that then interact almost

24
00:01:56.959 --> 00:02:00.760
like legos, connecting as they need
to and phemeral way to get something done.

25
00:02:00.760 --> 00:02:05.200
Well, wouldn't it be nice to
have something like that for data?

26
00:02:05.280 --> 00:02:07.319
And you know, I really started
thinking to myself a number of different things,

27
00:02:07.360 --> 00:02:09.800
and Eve, I'll bring you into
comment on this in just one second.

28
00:02:09.800 --> 00:02:13.840
But you know, when I did
this show last week, I did

29
00:02:13.879 --> 00:02:16.919
a search for what is data mesh
on Google? And oh my goodness,

30
00:02:17.000 --> 00:02:23.520
there are like thirteen, fourteen,
fifteen different companies all sponsoring content on Google

31
00:02:23.599 --> 00:02:27.639
right now and the Google search engine
for what is data mesh? So you

32
00:02:27.719 --> 00:02:30.919
got like fifteen sixteen and there may
be twenty. I stopped counting at a

33
00:02:30.000 --> 00:02:32.680
certain point. They're all over the
map, and I'm thinking to myself,

34
00:02:32.800 --> 00:02:37.280
is this thing just off to the
races now? Is it is? All

35
00:02:37.280 --> 00:02:40.439
control of data mesh been lost because
you have all these folks doing their own

36
00:02:40.479 --> 00:02:44.960
things. So what is it?
And data mesh is different than big data,

37
00:02:45.039 --> 00:02:50.319
Like big data took off like Gangbusters
ten fifteen years ago because of all

38
00:02:50.400 --> 00:02:54.000
these giant social engines spinning out all
sorts of data, because of all the

39
00:02:54.000 --> 00:02:59.000
different ways they were able to capture
data. Now that isn't just transactional.

40
00:02:59.240 --> 00:03:02.280
That is systems data for example,
machine data basically, So that was a

41
00:03:02.280 --> 00:03:06.080
little bit different. Data mesh is
a much more or it should be a

42
00:03:06.159 --> 00:03:12.560
much more cogent and cohesive approach to
working with data. But then you have

43
00:03:12.599 --> 00:03:15.759
all these foundational models coming out,
and there's data modeling, and I remember

44
00:03:15.759 --> 00:03:20.599
as I was trying to understand data
modeling, really it's all about performance and

45
00:03:20.680 --> 00:03:24.080
understanding the intersection of your data and
whatever you want to do with it on

46
00:03:24.120 --> 00:03:29.159
the production side. So that kind
of governs Is it a snowflake model?

47
00:03:29.199 --> 00:03:30.960
What is going to be you know, the actual data model that you put

48
00:03:31.000 --> 00:03:36.800
into play. And that struck an
idea with you, Eve about your thoughts

49
00:03:36.840 --> 00:03:38.800
way back when with data models.
But what do you think first about this?

50
00:03:39.039 --> 00:03:44.080
Is it something specific or is it
just amorphous? What do you think

51
00:03:44.120 --> 00:03:47.919
about that? And how does it
relate to data models and structure data and

52
00:03:47.960 --> 00:03:53.000
all that stuff. Oh well,
that's a that's a whole bunch of eric

53
00:03:54.360 --> 00:04:00.159
data as a concept. What for
me is most strike is the concept of

54
00:04:00.240 --> 00:04:04.520
data products, which means you're trying
to create some value out of your data

55
00:04:04.639 --> 00:04:09.919
and focus on that small element.
That's one thing. An important other part

56
00:04:10.039 --> 00:04:14.159
is the governance, that security of
those data products which you provide so it

57
00:04:14.199 --> 00:04:18.120
doesn't travel to each and everyone and
it's freely available, you have it in

58
00:04:18.160 --> 00:04:24.279
the controlled environment. I think those
are very important principles of a data mesh.

59
00:04:24.319 --> 00:04:29.160
And you stated it pretty well in
how the history of the big data

60
00:04:29.199 --> 00:04:38.360
and everything came along in such a
way we thought big data was hot together

61
00:04:38.399 --> 00:04:41.680
with data science, where we thought, okay, let's throw all the data

62
00:04:41.720 --> 00:04:45.639
into a data leak. It's cheap, we can capture all the data and

63
00:04:45.920 --> 00:04:50.319
very lightly with data science and algorithms
and pattern mining, we can find some

64
00:04:50.439 --> 00:04:55.040
hidden gems into that data leak.
But then we thought, well, we

65
00:04:55.120 --> 00:04:58.279
saw that it was too much.
We didn't have any context about that.

66
00:04:58.639 --> 00:05:03.079
So data scientists had to have the
knowledge of the business to find those hidden

67
00:05:03.120 --> 00:05:08.279
gems. So a lot of communication
with the business to understand what is meaningful,

68
00:05:08.360 --> 00:05:12.040
what is worthful in your data,
and then pulling that out and giving

69
00:05:12.040 --> 00:05:15.160
the insights. But that still is
a lot of work preparing the data and

70
00:05:15.279 --> 00:05:18.759
making it available. And I think
we're now with the foundational models at the

71
00:05:18.839 --> 00:05:24.439
stage where you can indeed throw all
that data at the foundational model and have

72
00:05:24.560 --> 00:05:30.839
the foundational model give you the most
interesting connections between your data elements. If

73
00:05:30.839 --> 00:05:34.519
we call foundational model, it's a
large language model. These are words and

74
00:05:34.600 --> 00:05:41.240
how the words are connected. And
what changpt as a language model does is

75
00:05:41.519 --> 00:05:46.079
it tries to predict the next possible
word which is following in a sentence.

76
00:05:46.199 --> 00:05:49.279
So we analyzed all these data and
says, hey, very likely this is

77
00:05:49.279 --> 00:05:53.959
the next word, or this is
in the most context. These are the

78
00:05:53.959 --> 00:05:58.279
words that tie together, and if
you look into a business context, and

79
00:05:58.399 --> 00:06:01.040
the data what you have. This
could be meaningful that we can take these

80
00:06:01.279 --> 00:06:06.680
foundation models and try to navigate through
all the data what we have and say,

81
00:06:06.720 --> 00:06:12.000
in the most of the cases,
this customer is available in this country,

82
00:06:12.000 --> 00:06:16.160
for example, and then highlights all
the valuable information what you have,

83
00:06:16.519 --> 00:06:25.720
so understand the understandment and understanding your
data elements by applying foundational models on top

84
00:06:25.759 --> 00:06:30.639
of that, and then an extra
layer of governing and securing that part.

85
00:06:30.000 --> 00:06:33.920
I think that's interesting what you say, I know this little piece of information

86
00:06:34.000 --> 00:06:36.879
and this is valuable for me,
and you're, for example, in the

87
00:06:36.959 --> 00:06:43.759
marketing department, and another person is
in the sales department and has as well

88
00:06:43.800 --> 00:06:46.959
that valuable piece, but you have
a hard time tying to together. And

89
00:06:47.040 --> 00:06:50.879
if the foundation model comes along and
say hey, you're missing this valuable piece,

90
00:06:51.000 --> 00:06:56.720
connect it all together, then you're
creating that datamash and you're governing that

91
00:06:56.920 --> 00:07:00.160
all around. I think that's that's
definitely a possible ability where we could go.

92
00:07:00.439 --> 00:07:03.839
And with all the technology that now
is available, we are coming into

93
00:07:03.920 --> 00:07:10.040
that self driving data management. If
you look at how cars operate, they

94
00:07:10.079 --> 00:07:15.279
give signals to each other and they're
aware of their environment, and if you

95
00:07:15.319 --> 00:07:18.639
can put that in place on top
of all your data. That would be

96
00:07:18.800 --> 00:07:25.000
for me, well, crazy,
it's the technology and all the all the

97
00:07:25.160 --> 00:07:29.160
algorithms that are doing the work for
me. What are we doing? Typically,

98
00:07:29.160 --> 00:07:31.279
if you're doing data modeling, you're
looking at the data and you say,

99
00:07:31.319 --> 00:07:34.279
how would we split it up?
Like you said, for having the

100
00:07:34.360 --> 00:07:40.959
performance, you had vertical partitioning,
yet horizontal partitioning depending on how you will

101
00:07:41.199 --> 00:07:46.959
query your data. That's why dimensional
models and snowflake models came to be as

102
00:07:46.959 --> 00:07:49.839
well, because they had a different
behavior on top of the data compared to

103
00:07:50.879 --> 00:07:56.000
online transactional data, and you had
to store them in a different way.

104
00:07:57.160 --> 00:08:01.519
And that was just the expertise that
you made a good data model and a

105
00:08:01.560 --> 00:08:05.240
good data model I call it,
which is the best model fit for purpose?

106
00:08:07.639 --> 00:08:11.600
And yeah, now now with the
knowledge graphs and all of that together

107
00:08:11.879 --> 00:08:16.759
for me, that's that's a valuable
concept that we can get so much faster

108
00:08:18.040 --> 00:08:22.399
insights in our data. Then you
could do as as a human being.

109
00:08:22.600 --> 00:08:24.720
Otherwise you had to go through all
the data, you had to cleanse the

110
00:08:24.839 --> 00:08:30.639
data, you had to optimize it, and then you could start doing that.

111
00:08:30.879 --> 00:08:33.559
And this is in a certain way
be done by by a lot of

112
00:08:33.559 --> 00:08:39.200
the foundational models and the algorithms,
we still have the issue of data quality.

113
00:08:39.240 --> 00:08:41.679
I mean, if you put in
bad data into the algorithms, you

114
00:08:41.799 --> 00:08:48.519
get bad insights, and that's still
a big concern. By using foundational models.

115
00:08:50.000 --> 00:08:52.960
Yeah, well yeah, so the
hallucinations that these things come up with

116
00:08:54.120 --> 00:08:58.360
me and again they are predictive engines
and they have a corpus of text on

117
00:08:58.440 --> 00:09:03.039
which they've been trained. Now the
next phase, and I guess I want

118
00:09:03.039 --> 00:09:05.879
to see if we can go to
this certain place here, the intersection of

119
00:09:05.919 --> 00:09:11.320
the nexus of large language models or
foundational models and data mesh. That could

120
00:09:11.360 --> 00:09:16.480
get very interesting. And you know, you think about how container orchestration works.

121
00:09:16.559 --> 00:09:20.440
You've got these containers. They have
a little system process inside and they

122
00:09:20.440 --> 00:09:22.639
are self describing. So they land
somewhere, they open up and they go,

123
00:09:22.639 --> 00:09:26.960
oh, here's what I've got,
and it gets it gets dropped into

124
00:09:26.679 --> 00:09:31.840
the production line, basically into the
processing line. It does its thing,

125
00:09:31.879 --> 00:09:35.440
and then it goes away, and
then another pod opens up and it does

126
00:09:35.440 --> 00:09:37.039
its thing and then it goes away. Well, you know, if you

127
00:09:37.080 --> 00:09:39.919
could do that with data, that
would get interesting. So you'd have some

128
00:09:41.000 --> 00:09:43.039
sort of wrapper around it. That
says, you know who I am,

129
00:09:43.039 --> 00:09:46.080
this is the system I came from, what kind of data I have?

130
00:09:46.639 --> 00:09:50.039
And you land and you do something
and you kind of go on your way.

131
00:09:50.480 --> 00:09:54.399
I guess that's that's the vision.
What I'm trying to figure out is,

132
00:09:54.759 --> 00:09:58.960
you know, what are some of
the functions that are baked into these

133
00:09:58.399 --> 00:10:05.279
foundations models? Because you can just
think about all the gray area in business

134
00:10:05.279 --> 00:10:09.759
intelligence. So we spend all this
time pulling data from transactional systems, putting

135
00:10:09.759 --> 00:10:13.159
it into this model, and then
you query the model to understand what's going

136
00:10:13.200 --> 00:10:18.360
on. Well, that query side
is at least half the battle, right

137
00:10:18.440 --> 00:10:22.279
knowing which question to ask. It's
like the prompt when you're dealing with large

138
00:10:22.320 --> 00:10:24.440
language models. The prompt is really
important. You have to be very specific

139
00:10:24.440 --> 00:10:28.840
in telling you what you wanted to
give you. It's not like Google,

140
00:10:28.879 --> 00:10:31.080
which just gives you a whole bunch
of stuff to look up. Although you'll

141
00:10:31.120 --> 00:10:35.720
notice Google changed, I don't know, maybe six or twelve months ago,

142
00:10:35.360 --> 00:10:39.000
where a lot of times if you
ask it for a definition of something,

143
00:10:39.399 --> 00:10:43.000
what it'll do is it'll go to
a page that it's founded indexed, and

144
00:10:43.039 --> 00:10:48.240
it'll present that definition to you and
then a link to that page. So

145
00:10:48.759 --> 00:10:52.360
it's they're trying to get to the
place where they're really telling you what something

146
00:10:52.559 --> 00:10:56.720
is, at least in terms of
definition. But what's cool about a foundational

147
00:10:56.759 --> 00:11:00.279
model, of course, is you
can see all kinds of different things.

148
00:11:00.320 --> 00:11:03.080
You can give me, you know, tell me twenty of the most innovative

149
00:11:03.120 --> 00:11:07.679
vendors, and data science and is
going to give you its whole scoring,

150
00:11:07.000 --> 00:11:09.720
which it's doing on the fly.
Right. It's not like it creative as

151
00:11:09.799 --> 00:11:13.399
index and it's waiting for you to
ask for it. No, it's it's

152
00:11:13.440 --> 00:11:16.240
just waiting for you to prompt it
to do something and then coming up with

153
00:11:16.840 --> 00:11:20.840
whatever makes most sense. So like, if you think about that, these

154
00:11:20.879 --> 00:11:24.600
this entire industry of data management,
of trying to understand what's happening in the

155
00:11:24.600 --> 00:11:31.039
world, where what decisions should I
make? These foundational models themselves have greatly

156
00:11:31.080 --> 00:11:35.039
disrupted the whole process. And that's
why every vendor and their grandmother are all

157
00:11:35.039 --> 00:11:39.159
talking about lms. It's like they're
talking about two things, data mesh and

158
00:11:39.159 --> 00:11:41.440
the LMS all day long. And
like, you know, we're going to

159
00:11:41.519 --> 00:11:46.120
do a show on data science in
August, but it's like you knows data

160
00:11:46.200 --> 00:11:50.559
science fading. Now, are we
gonna stop talking less about data science and

161
00:11:50.559 --> 00:11:52.399
more about data mesh and foundational models. I don't know, what do you

162
00:11:52.440 --> 00:11:58.000
think? Yeah, talking about data
mesh because well, every every five to

163
00:11:58.120 --> 00:12:05.440
seven years we need to reinvent something
hot word where we can get about and

164
00:12:05.000 --> 00:12:11.240
che GPTs all around the place,
a large language models, foundational models,

165
00:12:11.279 --> 00:12:15.679
because we see there is a big
potential in how you can put that into

166
00:12:15.720 --> 00:12:18.600
place. I think, like you
said, you're listening, you're asking give

167
00:12:18.639 --> 00:12:22.279
me a top ten list of of
data science companies, and it's all about

168
00:12:22.399 --> 00:12:26.480
the contextualization of what you have in
such a way. That's why you said

169
00:12:26.480 --> 00:12:31.840
as well, the prompts need to
be very specific to get the output what

170
00:12:31.919 --> 00:12:35.279
you're expecting in such a way.
And that was back in the days as

171
00:12:35.320 --> 00:12:39.799
well. Big challenge on building data
models, on building data warehouses. For

172
00:12:39.840 --> 00:12:43.960
example, if you have Paris,
it's available, well, it's in three

173
00:12:43.000 --> 00:12:46.519
places that I know of. There
are cities called Paris all over the world.

174
00:12:48.039 --> 00:12:54.159
All this contextualization, I think,
with such a vast data set with

175
00:12:54.240 --> 00:12:58.559
a lot of value in there,
you can make up your mind that at

176
00:12:58.600 --> 00:13:05.240
this parents in the States or spouse
in France and having that capability into the

177
00:13:05.320 --> 00:13:09.279
models, into your data warehouse.
That's that's really the future where we're going

178
00:13:09.759 --> 00:13:16.039
for me. Yeah, yeah,
this is uh, it's it's fascinating.

179
00:13:16.039 --> 00:13:18.000
I mean from a business perspective,
I think the bottom line is that the

180
00:13:18.799 --> 00:13:24.639
song remains the same and that we
still need to be looking at these different

181
00:13:24.639 --> 00:13:28.120
models and understanding. But you know, the other takeaway I had from a

182
00:13:28.320 --> 00:13:35.840
Terra Data conference recently. They were
talking about data warehouse, data lakehouse,

183
00:13:35.279 --> 00:13:39.960
lakehouse architecture, things of this nature, and I thought to myself, you

184
00:13:39.960 --> 00:13:41.960
know, there's another whole set of
applications that are coming down the pike.

185
00:13:43.080 --> 00:13:48.200
They're AI driven, and that are
really going to change how the apps interact

186
00:13:48.240 --> 00:13:50.799
with the data. So you know, if you look at Snowflake and what

187
00:13:50.879 --> 00:13:54.279
they're doing with their new transactional apps, as they're saying, oh, why

188
00:13:54.320 --> 00:14:00.000
don't you just build your transactional apps
inside our environments and that way they can

189
00:14:00.080 --> 00:14:05.840
spin right out of your curated,
trusted data. So that's clearly a trajectory

190
00:14:05.960 --> 00:14:09.399
for them. But you know,
then, like I'm saying, there's this

191
00:14:09.440 --> 00:14:13.200
other side of the equation where I
think that there's a lot of heavy lifting

192
00:14:13.240 --> 00:14:16.759
we've been doing for a lot of
time. Now that's probably not going to

193
00:14:16.799 --> 00:14:20.480
be as necessary anymore, just the
amount of raw data that we move to

194
00:14:20.519 --> 00:14:22.960
get into place to be able to
analyze. I think some of the use

195
00:14:24.000 --> 00:14:28.080
cases that were fed by that model
are going to be fed other ways and

196
00:14:28.120 --> 00:14:31.960
we won't need to be doing all
this heavy lifting on the data management side.

197
00:14:31.960 --> 00:14:33.320
But maybe I'm wrong. What do
you think about that? Well,

198
00:14:33.360 --> 00:14:37.639
we still have to make very likely
a distinction between structured and unstructured data.

199
00:14:37.679 --> 00:14:41.759
And in the unstructured data, that's
the part where we're talking about foundational models.

200
00:14:41.799 --> 00:14:46.320
It can help us as well in
the structured part. But typically where

201
00:14:46.320 --> 00:14:48.480
you say, well, I think
there is a lot of value in these

202
00:14:48.559 --> 00:14:52.600
data sets and I don't know how
to connect it. Throw it in your

203
00:14:52.679 --> 00:14:56.759
lakehouse. And then on top of
that you can get the contextualization a few

204
00:14:56.840 --> 00:15:01.720
years back again, and still the
data catalog gets a lot of attention.

205
00:15:03.000 --> 00:15:07.039
It's indexing the data what you have, labeling it, understanding your data.

206
00:15:07.240 --> 00:15:15.720
It's typically what what what in the
video tex all the all the all the

207
00:15:15.799 --> 00:15:18.879
janitaries who are doing they were indexing
the books so you could find it based

208
00:15:18.919 --> 00:15:24.039
upon author, based upon title,
based upon the type of book. And

209
00:15:24.159 --> 00:15:28.120
this is what the foundational models are
doing for us right now with all the

210
00:15:28.200 --> 00:15:31.720
data what we have around, and
that's where I think it's going. It's

211
00:15:33.559 --> 00:15:39.279
merging more transactional data together with unstructured
data and analytical data. And if you

212
00:15:39.360 --> 00:15:45.559
see you mentioned just snowflake how they
are doing that. Another technology what I

213
00:15:45.559 --> 00:15:48.799
saw, is capable of doing that
a single store. They have the analytical

214
00:15:48.879 --> 00:15:52.679
and the operational engine into one database, so you can't move all your data

215
00:15:52.679 --> 00:15:58.879
all around all the right. Depending
on which type of question you're asking,

216
00:15:58.559 --> 00:16:04.720
the technology is formach enough to give
you the answer. And that's already a

217
00:16:04.799 --> 00:16:10.120
big step forward what we've seen in
the new technologies on how they're developing.

218
00:16:11.360 --> 00:16:15.759
So yeah, it's exciting times to
see where all of this is going into

219
00:16:15.759 --> 00:16:21.519
the data management space, but finally
having our digital footprint transferred into real business

220
00:16:21.519 --> 00:16:25.840
insights. Yeah. Well, and
so you bring up data catalogs, and

221
00:16:25.879 --> 00:16:30.679
I'm almost wondering, you know,
does the foundational model become your data catalog?

222
00:16:30.399 --> 00:16:36.879
It seems to me that's what it's
probably best at doing is understanding entities

223
00:16:36.919 --> 00:16:40.879
and the relationships between them, and
that's really what a data catalog is supposed

224
00:16:40.919 --> 00:16:45.559
to be doing for you is explaining
the entities and helping you understand where they

225
00:16:45.600 --> 00:16:48.960
go and what they do. I
mean, you know, I think what

226
00:16:48.600 --> 00:16:53.600
we're going to see here are these
sort of single instance large language models who

227
00:16:53.600 --> 00:16:59.120
as I talking to someone just the
other day who was talking about how that

228
00:16:59.279 --> 00:17:03.200
architecture works and that you know,
you can just set up your own instance

229
00:17:03.279 --> 00:17:07.519
of an LM in Microsoft as your
and you can do all kind of fun

230
00:17:07.559 --> 00:17:11.200
stuff with it, and then it's
just like a database that you've set up,

231
00:17:11.559 --> 00:17:15.000
so it's already you know, to
a point where it is instantiated just

232
00:17:15.119 --> 00:17:18.880
for your company, right, and
so you can control what goes into the

233
00:17:18.880 --> 00:17:23.519
rest of the model what doesn't etc
to me. That's you know, that's

234
00:17:23.599 --> 00:17:27.519
probably the holy grail. But I'm
sure there's a lot of loose ends to

235
00:17:27.640 --> 00:17:30.160
tie. You know. That's the
problem, right, It's any any new

236
00:17:30.200 --> 00:17:33.279
solution can solve eighty percent of your
problems, but the last twenty percent it's

237
00:17:33.279 --> 00:17:37.279
always going to be a pain.
And that brings us to the end of

238
00:17:37.279 --> 00:17:38.799
our first segment. Don't touch dot
doot, We'll be right back. You're

239
00:17:38.799 --> 00:17:42.240
listening to the only Coast to coast
Radio show all about the information economy.

240
00:17:42.240 --> 00:18:03.400
It's called inside enough. Can your
IRA stand up to the next financial crisis

241
00:18:03.480 --> 00:18:07.240
that our top economists are saying is
at our doorsteps. By allocating a percentage

242
00:18:07.240 --> 00:18:12.000
of your IRA into physical gold and
silver with a tax free rollover, you

243
00:18:12.000 --> 00:18:18.039
can diversify in safeguard your holdings from
turbulent markets and economic downturns by putting your

244
00:18:18.079 --> 00:18:23.119
IRA back. Find out how to
safeguard your ass with a tax free rollover

245
00:18:23.240 --> 00:18:29.559
with a Genesis Gold Ira, the
only IRA that can hold physical precious metals.

246
00:18:29.759 --> 00:18:33.240
Call now for your free gold and
silver report. Protect your IRA today

247
00:18:33.400 --> 00:18:38.079
with one simple phone call and learn
how to qualify for up to ten thousand

248
00:18:38.119 --> 00:18:45.720
Genesis Gold Group empowering faith driven Stewardship
eight hundred six four four eight six one

249
00:18:45.839 --> 00:18:51.519
one eight hundred six four four eight
six one one eight hundred six four one

250
00:18:51.599 --> 00:18:56.599
one. That's eight hundred six four
four eighty six eleven. Can your IRA

251
00:18:56.880 --> 00:19:00.319
stand up to the next financial crisis
that our top economister saying is that our

252
00:19:00.400 --> 00:19:06.759
doorsteps. By allocating a percentage of
your IRA into physical gold and silver with

253
00:19:06.839 --> 00:19:11.000
a tax free rollover, you can
diversify in safeguard your holdings from turbulent markets

254
00:19:11.000 --> 00:19:15.279
and ECODO. Putting your IRA back
on the gold standard. Find out how

255
00:19:15.279 --> 00:19:21.279
to safeguard your assets with a tax
free rollover with a Genesis Gold IRA,

256
00:19:21.680 --> 00:19:26.319
the only IRA that can hold physical
precious metals. Call now for your free

257
00:19:26.359 --> 00:19:30.680
gold and silver report. Protect your
IRA today with one simple phone call and

258
00:19:30.839 --> 00:19:36.839
learn how to qualify for up to
ten thousand dollars in free silver called Genesis

259
00:19:36.839 --> 00:19:41.920
Gold Group Empowering faith driven Stewardship.
Eight hundred six four four eight six one

260
00:19:42.039 --> 00:19:48.799
one eight hundred six four four eight
six one one eight hundred six four four

261
00:19:48.119 --> 00:19:52.880
eight six one one. That's eight
hundred six four four eighty six eleven.

262
00:19:56.240 --> 00:20:00.279
When a player's sudden cardiac event brought
a national football game to a halt,

263
00:20:00.440 --> 00:20:06.000
it's shown a spotlight on the importance
of CPR readiness. Now, with youth

264
00:20:06.039 --> 00:20:10.920
sports in full swing, the American
Heart Association is rallying parents and coaches to

265
00:20:10.960 --> 00:20:15.200
be ready in an emergency. To
be ready, learn hands only CPR.

266
00:20:15.440 --> 00:20:18.680
It's a skill anyone can learn in
minutes. Just visit Heart dot org slash

267
00:20:18.720 --> 00:20:25.200
hands only CPR. Hands only CPR
is nationally supported by Elephant's Health Foundation.

268
00:20:26.079 --> 00:20:30.279
Each year three hundred and fifty thousand
Americans die from a cardiac arrest. When

269
00:20:30.319 --> 00:20:34.519
seconds matter most, CPR can be
the difference in whether a friend or family

270
00:20:34.519 --> 00:20:41.440
member survives. That's why the American
Heart Association is challenging every household to elect

271
00:20:41.480 --> 00:20:45.599
at least one person to learn a
CPR. If you have ninety seconds,

272
00:20:45.960 --> 00:20:49.799
you can be your family CPR hero. Just watch the American Heart Association's hands

273
00:20:49.799 --> 00:20:57.359
only CPR video at heart dot org
and become a hero. When a player's

274
00:20:57.400 --> 00:21:02.519
sudden cardiac event brought a national football
game to a halt, it's shown a

275
00:21:02.599 --> 00:21:07.279
spotlight on the importance of CPR readiness. Now, with youth sports in full

276
00:21:07.319 --> 00:21:11.960
swing, the American Heart Association is
rallying parents and coaches to be ready in

277
00:21:11.000 --> 00:21:15.880
an emergency. To be ready learn
hands only CPR. It's a skill anyone

278
00:21:15.920 --> 00:21:21.319
can learn in minutes. Just visit
Heart dot org slash hands only CPR.

279
00:21:21.640 --> 00:21:26.480
Hands only CPR is nationally supported by
Elephant's Health Foundation. Now you can fly

280
00:21:26.640 --> 00:21:32.359
anywhere in the world and pay discount
prizes on your airline tickets. Book a

281
00:21:32.400 --> 00:21:36.400
flight today to London, Paris,
Madrid, or anywhere else you want to

282
00:21:36.400 --> 00:21:41.200
go and pay a lot less guarantee. Call the International Travel Department right now

283
00:21:41.240 --> 00:21:47.319
at low cost airlines eight hundred two
nine eight five seven eight three. Eight

284
00:21:47.400 --> 00:21:51.759
hundred two nine eight five seven eight
three. That's eight hundred two nine eight

285
00:21:51.880 --> 00:21:59.960
fifty seven eighty three. Welcome back
to Inside Analysis. Here's your host,

286
00:22:00.519 --> 00:22:06.960
Eric Tavanaugh, right ful, pfect, your anci anouts talking to the legendary

287
00:22:07.000 --> 00:22:11.799
eat mookers at SEVENS data over from
Belgium. And we're noticing that people are

288
00:22:11.799 --> 00:22:15.880
waking up now. You know,
I mentioned at the ter Data conference my

289
00:22:15.000 --> 00:22:22.359
input about different architectures, the data
warehouse versus the data lake versus the lakehouse,

290
00:22:22.359 --> 00:22:26.960
and these different architectures, and that
is important stuff less important in a

291
00:22:26.079 --> 00:22:32.720
number of different ways going forward,
just because this new generation of apps is

292
00:22:32.759 --> 00:22:36.519
going to be able to get signal
from almost anywhere. And I don't think

293
00:22:36.519 --> 00:22:38.680
you're going to need to do all
the massaging of the data and all the

294
00:22:38.720 --> 00:22:41.599
management and movement of the data.
I think you're going to be able to

295
00:22:41.640 --> 00:22:48.680
get signal from just a couple threads. Think click stream analysis for example.

296
00:22:48.759 --> 00:22:52.200
Just to throw something out there,
there's so much that we can now gather

297
00:22:52.359 --> 00:22:55.640
from web browsers as people are moving
around where their mouse goes, what they

298
00:22:55.759 --> 00:23:00.200
click on, where they go,
what they do to ascertain intents and then

299
00:23:00.359 --> 00:23:03.559
serve up some interesting piece of contents. And you know, I have to

300
00:23:03.599 --> 00:23:07.400
think that Amazon and some of the
other real big marketplaces are doing a lot

301
00:23:07.400 --> 00:23:11.359
of stuff in this space. You
kind of see little bits and pieces of

302
00:23:11.400 --> 00:23:15.519
it of the changing of the guard, if you will, about how things

303
00:23:15.599 --> 00:23:19.000
are done. They're getting more clever
about abandoned carts, for example. Many

304
00:23:19.000 --> 00:23:22.400
more companies are getting clever about that
noticing, hey, you know, are

305
00:23:22.400 --> 00:23:26.680
you still looking for a new mattress
kind of thing? And there are engines

306
00:23:26.720 --> 00:23:30.920
that drive all these applications, right, and there's a lot of legacy technology

307
00:23:30.960 --> 00:23:34.400
out there doing it. But I'm
positive that you're going to see at least

308
00:23:34.400 --> 00:23:40.720
the major players like Amazon and the
other e commerce engines kind of jumping on

309
00:23:40.880 --> 00:23:47.359
this stuff and leveraging a new generation
of analytical capability driven by AI, driven

310
00:23:47.359 --> 00:23:52.279
by algorithms, and again the nexus
to maybe it's data mesh, maybe it's

311
00:23:52.319 --> 00:23:56.359
just leveraging these foundational models. But
the point is there are new ways of

312
00:23:56.400 --> 00:23:59.559
doing things. They're going to be
leaner, they're going to be more cost

313
00:23:59.559 --> 00:24:02.440
efficient, and I think they're going
to be more accurate. What was the

314
00:24:02.519 --> 00:24:04.240
one I saw a stat that kind
of blew my mind to end. I'm

315
00:24:04.240 --> 00:24:07.480
trying to remember who it was,
but they basically said that they're like ninety

316
00:24:07.480 --> 00:24:11.440
two percent accurate in terms of their
predictions of what someone's going to do.

317
00:24:11.480 --> 00:24:14.960
And this is a oh, I
know it was. It was one I

318
00:24:15.039 --> 00:24:18.359
was talking to Mendix for the folks
at Mendix, and what they're doing,

319
00:24:18.519 --> 00:24:22.000
of course, is digital transformation.
They've come up. It's a really I

320
00:24:22.119 --> 00:24:25.079
love Mendix and they're they're a version
ten, so they've been around for a

321
00:24:25.079 --> 00:24:29.720
long time. What she was saying
is that their AI is now making suggestions

322
00:24:29.759 --> 00:24:32.720
to people as they're going through and
designing these workflows, saying, hey,

323
00:24:32.759 --> 00:24:33.400
do you want to add this button? Do you want to do this?

324
00:24:33.440 --> 00:24:37.559
Do you want to do that?
And she said it's like ninety two percent

325
00:24:37.640 --> 00:24:40.519
of the time. I think,
does the number the user says yes,

326
00:24:40.559 --> 00:24:44.079
I want to do that. So
you had mentioned how the foundational models,

327
00:24:44.079 --> 00:24:45.480
the language models, they're trying to
figure out what word are you going to

328
00:24:45.599 --> 00:24:49.559
use next right, what makes most
sense or what paragraph, what subject area,

329
00:24:49.599 --> 00:24:52.160
et cetera. And it's a cascading
you can tell sort of a cascading

330
00:24:52.200 --> 00:24:56.920
engine where it's like ri, you
know what is the general theme? Okay,

331
00:24:56.960 --> 00:24:59.200
I got it. You can actually
tell when you you know, you

332
00:24:59.279 --> 00:25:03.119
prompt and you see weight and like
m it's like thinking and then now how

333
00:25:03.200 --> 00:25:07.079
it comes right so and then of
course it stops, you can go continue

334
00:25:07.119 --> 00:25:11.680
to generate, and that obviously is
uh, you know, a sort of

335
00:25:11.720 --> 00:25:14.480
guard rail in the system to keep
it from just going on and on and

336
00:25:14.519 --> 00:25:15.640
on. They don't want to go
on and on and on. They want

337
00:25:15.640 --> 00:25:19.000
it to be as cost efficient as
possible. Right, But my point is

338
00:25:19.000 --> 00:25:22.000
if you, if you kind of
metaphorically do the math here and realize,

339
00:25:22.359 --> 00:25:26.160
all right, these foundational models know
a lot already. They have a whole

340
00:25:26.160 --> 00:25:30.039
lot of understanding about the relationship of
concepts and objects and places and things and

341
00:25:30.119 --> 00:25:37.000
industries. And you you built that
onto a traditional or you underpin a traditional

342
00:25:37.079 --> 00:25:41.119
business intelligence system or an analytics system, and things get really interesting, really

343
00:25:41.160 --> 00:25:45.079
fast. So I think that this
next generation of AI technology, it's going

344
00:25:45.160 --> 00:25:47.720
to be there's gonna be a lot
of lightweight stuff that kind of sits on

345
00:25:47.759 --> 00:25:51.359
the top and just looks for a
signal and then does something. And I

346
00:25:51.680 --> 00:25:53.960
think that's gonna be a huge change
in how all this stuff gets done.

347
00:25:53.960 --> 00:25:57.400
But what do you think there's kind
of personal assistance? What you get that

348
00:25:57.599 --> 00:26:03.680
if you say they listen to the
signals. But yeah, it's it's it's

349
00:26:03.720 --> 00:26:11.000
a rich context aware system what we
now have. And with our human brains

350
00:26:11.160 --> 00:26:15.759
we have a kind of limitation of
how much context or attention points we can

351
00:26:15.799 --> 00:26:18.640
get at the same time. And
these models can have a few millions at

352
00:26:18.640 --> 00:26:22.200
the same time. So I think
that's that's where you say where the power

353
00:26:22.359 --> 00:26:27.200
is in having the models and in
the AI that can help us in really

354
00:26:27.720 --> 00:26:33.519
assisting in what we are doing.
And it's amazing, like you say,

355
00:26:33.240 --> 00:26:37.319
nine predictability. So I was thinking
when you were saying that, are we

356
00:26:37.920 --> 00:26:41.880
human beings all in the same behavior? Are we acting in the same way.

357
00:26:41.960 --> 00:26:47.279
If you take about one hundred parameters
and can classify any human being by

358
00:26:47.279 --> 00:26:51.720
the one hundred parameters, very likely
any pursue percent of the people fit into

359
00:26:51.799 --> 00:26:56.039
that that segment. That's really really
amazing to see. I was thinking,

360
00:26:56.079 --> 00:26:57.519
well, at the same time,
at the small ends, if you see

361
00:26:57.559 --> 00:27:00.680
them moving across the path, yeah, yeah, yeah, right, So

362
00:27:00.720 --> 00:27:06.319
it's an ecosystem that works together,
and I'm always thinking of the AI systems

363
00:27:06.359 --> 00:27:10.640
with the neural networks below there on
how they operates. And it's in a

364
00:27:10.680 --> 00:27:15.759
certain way that we are mimicking just
our environment around us, the biology and

365
00:27:15.839 --> 00:27:21.039
the systems that are already in place, and we're finding a new way of

366
00:27:21.119 --> 00:27:26.319
developing systems that will help us in
augment what we are doing. And yeah,

367
00:27:26.400 --> 00:27:33.240
it's it's it's it's an exciting time
whereas see a lot of U applications

368
00:27:33.279 --> 00:27:36.680
of AI will be and that's thanks
to the rich environment what you have and

369
00:27:36.759 --> 00:27:41.160
the scale what you can do.
Another example is in the healthcare diagnosis,

370
00:27:41.359 --> 00:27:45.880
where you have the system that analyzes
all analysis of the doctor and gathers that

371
00:27:47.039 --> 00:27:52.200
together and then connected to the various
diseases, what you have seeing, connected

372
00:27:52.319 --> 00:27:57.599
to the various patients, and then
can really assist you in taking a certain

373
00:27:57.640 --> 00:28:03.440
decision based upon all the science.
And that's very likely where we are going.

374
00:28:03.000 --> 00:28:08.759
Still aside the privacy and everything what
we have to take into account be

375
00:28:08.839 --> 00:28:15.680
careful with. But yeah, we
see that developing in very specific AI applications.

376
00:28:15.880 --> 00:28:19.960
We will what will be built in
the future to help you and having

377
00:28:21.000 --> 00:28:26.599
those little satellites, those little specific
assistance that help you decide on soca.

378
00:28:26.759 --> 00:28:32.559
Well, like so the gentleman lou
Simon was his name from Optima he made.

379
00:28:32.559 --> 00:28:33.799
He's a super smart guy. Very
impressed with these folks. I want

380
00:28:33.839 --> 00:28:37.440
to learn more about what they're doing. But he was saying, for example,

381
00:28:37.559 --> 00:28:41.279
like you just said, in the
healthcare space, when you consider the

382
00:28:41.319 --> 00:28:47.599
corpus of data that we already have
scientific data on liver disease, on lung

383
00:28:47.640 --> 00:28:51.559
disease, on all these different on
heart disease for example. You know,

384
00:28:51.880 --> 00:28:56.400
his point was, these foundational models
that are trained on real world medical data,

385
00:28:56.640 --> 00:29:02.200
they are going to be so much
more powerful and knowledgeable than any individual

386
00:29:02.279 --> 00:29:06.640
doctor. And as he pointed out, today is the dumbest it'll ever be.

387
00:29:06.839 --> 00:29:08.799
So it's going to be learning day
after day after day after day.

388
00:29:10.400 --> 00:29:12.920
Now, you know, things do
change, Environmental conditions change, I mean,

389
00:29:14.039 --> 00:29:17.680
lots of things can change over time. But nonetheless, you know,

390
00:29:17.759 --> 00:29:23.759
you think about how that's going to
just transform the healthcare industry and the time

391
00:29:23.799 --> 00:29:27.880
we spend working on stuff. I
mean healthcare from my perspective, I mean,

392
00:29:27.880 --> 00:29:33.200
there's tremendously advanced aspects of healthcare and
science in big pharma and in other

393
00:29:33.279 --> 00:29:41.519
places but then there's this just monster
of a bureaucracy and legacy mindsets and processes

394
00:29:41.519 --> 00:29:45.480
and workflows, and you know,
bureaucracy I think is the is the appropriate

395
00:29:45.599 --> 00:29:48.480
term, and you know you're just
kind of crashing up against that wall.

396
00:29:48.960 --> 00:29:52.000
You know. Just to give an
example here, they just built a new

397
00:29:52.079 --> 00:29:55.759
VA hospital in New Orleans a number
of years ago. It just got to

398
00:29:55.839 --> 00:30:00.599
finished, and I wonder how much
that cost because it's a huge building.

399
00:30:00.640 --> 00:30:06.200
I mean, it's just absolutely gargantuan
and it's like, okay, so one

400
00:30:06.240 --> 00:30:10.920
billion dollar project broke ground in June
twenty ten and features one point six million

401
00:30:10.960 --> 00:30:14.599
square feet in eight buildings spread it
over thirty one acre campus, to which

402
00:30:14.640 --> 00:30:18.839
I ask, why, like why
do we need these huge buildings? I

403
00:30:18.920 --> 00:30:22.880
almost think the size of the buildings
are meant to be a metaphor for how

404
00:30:22.920 --> 00:30:26.599
important it is to us. We
don't need these huge structures anymore. They're

405
00:30:26.640 --> 00:30:30.880
incredibly expensive to air condition. I
mean, can you imagine the AC built

406
00:30:30.920 --> 00:30:33.680
for these places? And that's why
it costs two thousand dollars for an MRI

407
00:30:33.200 --> 00:30:37.160
And you think about how many MRI
machines you could have built for a billion

408
00:30:37.160 --> 00:30:41.039
dollars and it's a whole heck of
a lot that you would just then have

409
00:30:41.119 --> 00:30:45.079
in small little office buildings around town, because now you're forcing people to drive

410
00:30:45.160 --> 00:30:47.559
in. It's almost like the Lakehouse, right, You're just forcing them to

411
00:30:47.599 --> 00:30:52.759
commando the Lakehouse. Et other data
back out hydrate or ETL data EEL reverse

412
00:30:52.799 --> 00:30:56.799
ETL. Right. Anyway, what
I'm kind of speaking to is the this

413
00:30:56.200 --> 00:31:03.039
mountain of inefficiency that's baked into system
today, including the actual structures like look

414
00:31:03.039 --> 00:31:07.519
at downtown San Francisco, look at
Salesforce Tower. I mean, wow,

415
00:31:07.599 --> 00:31:11.799
what a you know, a symbol
of the old ways of doing things and

416
00:31:11.839 --> 00:31:14.960
now they're all trying to do RTO. I saw, I'm getting a tangent

417
00:31:15.000 --> 00:31:18.599
here, but I'm getting excited.
I saw an article came out and it

418
00:31:18.680 --> 00:31:22.920
was you may have seen this.
It was like an artist transition of what

419
00:31:22.039 --> 00:31:26.640
all these homeworkers are gonna look like
in thirty years. You're gonna be hunched

420
00:31:26.680 --> 00:31:27.839
over, and you're gonna have a
big belly, and you're gonna have bad

421
00:31:27.880 --> 00:31:33.519
blood pressure and all this stuff there. Which big tech company that's pushing RTO

422
00:31:33.680 --> 00:31:37.440
got that story placed because it's like
I can do exercise at home. I

423
00:31:37.440 --> 00:31:40.319
could do whatever I want at home. I don't have to just sit in

424
00:31:40.359 --> 00:31:42.000
front of my desk. So my
point is like, they are these there

425
00:31:42.000 --> 00:31:45.839
are these old constructs, and we
see them in our business too, like

426
00:31:45.839 --> 00:31:49.440
the data warehouses, the monolithic applications. SAP is still a modelith right,

427
00:31:49.480 --> 00:31:52.720
even they're like, oh, I
guess we have to do something with this

428
00:31:52.839 --> 00:31:56.279
Kubernetes thing. I don't know.
Man, the bigger they come, the

429
00:31:56.319 --> 00:31:57.799
harder they fall, is what I'm
saying. What do you think. Yeah,

430
00:31:59.359 --> 00:32:02.880
SAP is moving in a different direction
with still a monolith, but they're

431
00:32:02.880 --> 00:32:07.839
they're they're following up and like they
once said that they're an oil tanker,

432
00:32:07.839 --> 00:32:12.319
and it takes a while before they
get of course, but once they they're

433
00:32:12.440 --> 00:32:15.559
up, they hid it and they
moved forward. I was thinking about,

434
00:32:15.839 --> 00:32:19.359
like you said, with all the
massive buildings of the hospital, it's it's

435
00:32:19.400 --> 00:32:22.319
it's moving in a different way of
treatment. We do more preventive healthcare.

436
00:32:22.559 --> 00:32:29.200
If you can trace back in the
yeah, by by the source of why

437
00:32:29.240 --> 00:32:32.839
you have a disease and can start
treating that as of your ten years or

438
00:32:32.880 --> 00:32:37.039
five years or whatever, then you
prevent of having the disease off the wall

439
00:32:37.079 --> 00:32:42.039
in the end and then it's less
less expensive. And with all the data

440
00:32:42.079 --> 00:32:46.119
that we are now tracing, we're
very much capable of doing that preventive healthcare

441
00:32:46.240 --> 00:32:51.279
instead of doing the aftercare. And
see what you've got the symptoms, and

442
00:32:51.319 --> 00:32:53.839
we treat the symptoms. We got
something generic that's good for am, B

443
00:32:54.000 --> 00:32:58.079
and C, and we kind of
treat them all in the same way.

444
00:32:58.279 --> 00:33:02.400
It's the problem with and tell my
people are getting used to the antibiotics and

445
00:33:02.440 --> 00:33:09.200
are right anymore. But if you
see that that just the these institutions,

446
00:33:09.240 --> 00:33:14.440
they earned so much money on just
this way of working, so it will

447
00:33:14.480 --> 00:33:21.960
be hard to cut down these processes
because they're so established in just treating um

448
00:33:22.680 --> 00:33:25.799
the results and not doing to preventive
healthcare. Right now, that's a good

449
00:33:25.799 --> 00:33:30.960
point. And the organizational structures,
you know, that's where I'm I'm seeing

450
00:33:30.960 --> 00:33:34.960
there's going to be tremendous change here. We've just seen it in our industry.

451
00:33:35.000 --> 00:33:36.880
I mean, you know, let's
be honest, it's been a blood

452
00:33:36.880 --> 00:33:40.119
bath the last six or seven months
or so. You know. My my

453
00:33:40.839 --> 00:33:46.079
cynical side thinks that certain companies waited
until after the November elections here in the

454
00:33:46.119 --> 00:33:51.200
States to announce their layoffs because they're
very politically motivated. They didn't want the

455
00:33:51.240 --> 00:33:53.160
world to think that there is a
recession coming. And it's like, Okay,

456
00:33:53.160 --> 00:33:57.880
everything was fine until remember ninth and
then all we've gotta lay off ten

457
00:33:57.880 --> 00:34:00.519
percent of our workforce. We're gonna
lay off thirty percent of or workforce.

458
00:34:00.519 --> 00:34:04.119
I mean, I know companies they
let out thirty three percent of their workforce.

459
00:34:04.200 --> 00:34:06.960
You know, that's a lot of
people to lay off. And then

460
00:34:07.319 --> 00:34:08.760
what do you do then, Like, how do you pick up the pieces

461
00:34:08.760 --> 00:34:12.880
and get back to work? You
know, And the answer is going to

462
00:34:13.000 --> 00:34:16.320
be stuff like foundational models, using
that to bring out some of your marketing

463
00:34:16.360 --> 00:34:20.239
copy. You know, they're they're
just going to be different ways of doing

464
00:34:20.280 --> 00:34:22.280
things. But I think it's gonna
be disruptive. I think it's going to

465
00:34:22.320 --> 00:34:27.039
be disruptive for a while. You
know, one of my longtime clients,

466
00:34:27.159 --> 00:34:30.320
she's been working on an acquisition.
They just got acquired, and he referred

467
00:34:30.360 --> 00:34:35.639
to certain this environment today is a
nuclear winter. It's like, oh no,

468
00:34:36.519 --> 00:34:39.000
that doesn't sound good at all.
But necessity breeds innovation, right,

469
00:34:39.079 --> 00:34:43.519
because that we're gonna have to change. We're gonna have to change, and

470
00:34:43.800 --> 00:34:46.079
I think it is going to be
the agile companies to figure out how to

471
00:34:46.119 --> 00:34:50.079
carve out the niches and and make
some money. What do you think,

472
00:34:50.440 --> 00:34:53.079
Yeah, they're still there. They
will carve up the niches. But still

473
00:34:53.119 --> 00:34:57.039
the big companies have still big pockets, so they can do a lot of

474
00:34:57.800 --> 00:35:00.480
They have more power to do these
things. Like you were just referring before,

475
00:35:00.599 --> 00:35:07.199
Tata Data investing a quarter rebellion to
transform their legacy system into a cloud

476
00:35:07.239 --> 00:35:13.119
based system. And that's that's very
hard as a smaller company to compete against

477
00:35:13.400 --> 00:35:16.719
decoding. But what I see and
talking by or to a lot of peers

478
00:35:16.719 --> 00:35:23.400
and experts as well as skills that
we now very much need or data scientists,

479
00:35:23.519 --> 00:35:29.519
that's exactly critical thinking. It's questioning
what we are so used to see,

480
00:35:29.599 --> 00:35:32.440
what we are used to do,
and that critical thinking will be a

481
00:35:32.559 --> 00:35:37.119
very very important aspect, a soft
skill that a lot of people will need

482
00:35:37.159 --> 00:35:40.960
in the future. It's as well
with with the large language models, we

483
00:35:42.079 --> 00:35:46.079
will learn in a different way than
traditionally. Traditionally, or you are educated,

484
00:35:46.280 --> 00:35:50.039
you read a lot of books,
you stored all the information in your

485
00:35:50.039 --> 00:35:52.480
head, and that's where you are
smart. If you had all that knowledge

486
00:35:52.480 --> 00:35:55.599
stored. Now all the knowledge is
available, but you have you need to

487
00:35:55.639 --> 00:36:00.000
have the capabilities of questioning what you
get as the answer, a questioning of

488
00:36:00.440 --> 00:36:05.159
how do things work, how do
they fit within what we already know?

489
00:36:05.559 --> 00:36:07.960
And that's that's going to be transformational
in like you say, the type of

490
00:36:08.039 --> 00:36:13.679
companies that we will see in the
future, and when you will be capable

491
00:36:13.800 --> 00:36:19.239
of having that kind of culture,
that environment of doing things in this type

492
00:36:19.280 --> 00:36:23.679
of way of being really critical about
things and then moving forward, that's that's

493
00:36:23.719 --> 00:36:29.079
going to be transformational. Yeah,
it is. And I think that basically

494
00:36:29.400 --> 00:36:32.719
it's every developer, for example,
is going to become a ten X developer

495
00:36:34.039 --> 00:36:36.760
or they talk about that the ten
X developer. I mean, who was

496
00:36:36.800 --> 00:36:40.360
I I came at which vendor it
was? It might have been get hub.

497
00:36:42.039 --> 00:36:44.400
I thought it might have been someone
else, But on their website they

498
00:36:44.440 --> 00:36:47.400
said, you know now, and
this is huge. Now, these engines

499
00:36:47.440 --> 00:36:51.119
can tell you what your code means. Right. It's not just that they

500
00:36:51.119 --> 00:36:53.039
can write code for you, it's
that you can feed them code and say,

501
00:36:53.079 --> 00:36:55.519
hey, what is this doing,
and it'll come back and tell you,

502
00:36:55.519 --> 00:36:59.039
Oh, what it's doing is it's
reaching into this system and pulling that

503
00:36:59.119 --> 00:37:00.800
score and then it's doing this,
and it's doing that. You're just like,

504
00:37:00.920 --> 00:37:05.800
oh my goodness that see I was
talking to the I think it was

505
00:37:05.840 --> 00:37:09.239
the guy from answer Rocket, and
this was a sort of an epiphany for

506
00:37:09.320 --> 00:37:15.760
me. Think about the coball developers
and how much application has been built in

507
00:37:15.840 --> 00:37:19.760
coball from thirty and forty years ago, and there are no coball developers anymore.

508
00:37:19.760 --> 00:37:22.039
Well, yeah there are. It's
called chat chypt or it's called co

509
00:37:22.280 --> 00:37:25.039
pilot or whatever. So it's now
you can feed this stuff in there.

510
00:37:25.199 --> 00:37:28.159
Throw what the hell is it doing? Oh that's what it was doing.

511
00:37:28.239 --> 00:37:30.440
Oh good, Just changes change that. That's what it's going to be.

512
00:37:30.519 --> 00:37:34.280
It's going to be kind of changing
things. You're gonna use these engines to

513
00:37:34.360 --> 00:37:37.880
create the for the bulk, but
the clay model, and then the human

514
00:37:37.000 --> 00:37:39.559
is going to go in and do
the fine tuckey and chain in music shop

515
00:37:39.599 --> 00:37:44.079
things like that to get it just
right. But both don't set that would

516
00:37:44.079 --> 00:37:51.480
be right back. And do you
own an annuity either fixed rate, indexed

517
00:37:51.559 --> 00:37:54.519
or variable? Are you paying high
fees and getting low returns? If so,

518
00:37:54.920 --> 00:37:59.280
Annuity General would like you to have
this free book to learn the pitfalls

519
00:37:59.280 --> 00:38:04.000
and mistakes buying an annuity. The
Annuity Dues and Don'ts for Baby Boomers contains

520
00:38:04.000 --> 00:38:07.719
the little known truths about annuities,
like how to help reduce your fees and

521
00:38:07.800 --> 00:38:12.840
increase retirement income. And it's free, that's right free. As a bonus,

522
00:38:12.920 --> 00:38:15.360
we'll also throw in a free annuity
rate report just for calling. We

523
00:38:15.519 --> 00:38:22.199
researched over one thousand annuities and summarized
rates and benefits from financially strong insurers.

524
00:38:22.639 --> 00:38:25.800
You get Annuity Dudes and Don'ts for
Baby Boomers and the Annuity Rate Report,

525
00:38:25.840 --> 00:38:31.360
both absolutely free for calling Annuity General
Today. Hurry supplies are limited. Call

526
00:38:31.480 --> 00:38:37.639
now eight hundred two four or five
one six nine seven, eight hundred two

527
00:38:37.760 --> 00:38:42.639
four, five one six nine seven, eight hundred two four or five one

528
00:38:42.760 --> 00:38:47.239
six nine seven. That's eight hundred
two four or five sixteen ninety seven.

529
00:38:49.039 --> 00:38:52.000
Do you own a timeshare? We'll
face the facts. You made a mistake,

530
00:38:52.159 --> 00:38:57.239
you made a bad purchase. A
timeshare is not an investment. It's

531
00:38:57.239 --> 00:39:01.199
a money pit that continues forever you
use your time share. That's great,

532
00:39:01.440 --> 00:39:07.119
But if you don't and you want
illegally get out of your contract, call

533
00:39:07.159 --> 00:39:10.760
my friends right now at the time
Share Exit hotline. They're an experienced team

534
00:39:10.800 --> 00:39:15.199
of lawyers who help good people like
you get out of a timeshare contract that

535
00:39:15.320 --> 00:39:20.760
they just don't want. Don't throw
away your money on maintenance fees. Use

536
00:39:20.800 --> 00:39:23.239
it for things you really want.
We can help you end your time share

537
00:39:23.280 --> 00:39:28.960
contract and stop the money drain immediately. If you are ready to move on

538
00:39:29.000 --> 00:39:32.559
with your time share, call our
team right now. Cancel your time share

539
00:39:32.599 --> 00:39:38.119
now with a free call eight hundred
two ninezero six seven O five eight hundred

540
00:39:38.159 --> 00:39:44.280
two nine zero six seven O five
eight hundred two nine zero six seven O

541
00:39:44.559 --> 00:39:50.280
five. That's eight hundred two nine
oh sixty seven. Do you need to

542
00:39:50.320 --> 00:39:53.400
get your hands on some extra money
right now? Maybe twenty five thousand or

543
00:39:53.480 --> 00:39:58.039
more if you're a homeowner. Now
it's a perfect time to get cash out

544
00:39:58.119 --> 00:40:01.199
while homes in many neighborhoods like yours
have gone up in value. You can

545
00:40:01.320 --> 00:40:05.880
use the money for anything it's yours. You can buy an investment property,

546
00:40:05.960 --> 00:40:09.000
payoff higher interest debt, or make
home improvements. If you need twenty five

547
00:40:09.079 --> 00:40:14.639
thousand, fifty thousand or more.
Now is the time home values are up,

548
00:40:14.719 --> 00:40:16.599
and so is your equity. We
offer you a way to use it.

549
00:40:16.679 --> 00:40:21.360
No need to use your savings called
New American Funding now and see how

550
00:40:21.440 --> 00:40:24.480
much cash out you can get.
Call eight hundred seven one h three seven

551
00:40:24.679 --> 00:40:30.519
three nine, eight hundred seven one
h three seven three nine, eight hundred

552
00:40:30.679 --> 00:40:36.000
seven one h three seven three nine. That's eight thirty nine NMLS sixty six

553
00:40:36.119 --> 00:40:39.199
h six Www dot MLS, consumer
xs dot org. This is not an

554
00:40:39.199 --> 00:40:43.639
offer or commitment to lend. Subject
to borrow or improperty qualifications. Not all

555
00:40:43.679 --> 00:40:49.679
borrowers will qualify. Terms and conditions
apply equal housing opportunity. When a player's

556
00:40:49.719 --> 00:40:52.840
sudden cardiac event brought a national football
game to a halt, it's shown a

557
00:40:52.880 --> 00:40:58.599
spotlight on the importance of CPR readiness. Now, with youth sports in full

558
00:40:58.639 --> 00:41:02.199
swing, the American Heart Association is
rallying parents and coaches to be ready in

559
00:41:02.239 --> 00:41:07.719
an emergency. To be ready,
learn hands only CPR. It's a skill

560
00:41:07.760 --> 00:41:12.559
anyone can learn in minutes. Just
visit haart dot org slash hands only CPR.

561
00:41:12.880 --> 00:41:17.480
Hands only CPR is nationally supported by
Elephant's Health Foundation. Now you can

562
00:41:17.519 --> 00:41:22.320
fly anywhere in the world and pay
discount prizes on your airline tickets. Book

563
00:41:22.320 --> 00:41:27.320
a flight today to London, Paris, Madrid, or anywhere else you want

564
00:41:27.320 --> 00:41:31.880
to go and pay a lot less
guarantee, call the International Travel Department right

565
00:41:31.920 --> 00:41:37.880
now at low cost airlines eight hundred
two nine eight five seven eight three.

566
00:41:37.119 --> 00:41:42.559
Eight hundred two nine eight five seven
eight three. That's eight hundred two nine

567
00:41:42.599 --> 00:41:51.159
eight fifty seven eighty three. Welcome
back to Inside Analysis. Here's your host,

568
00:41:51.639 --> 00:41:57.599
Eric Kavanaugh. Fright, folks back
here in front of our three four

569
00:41:57.719 --> 00:42:00.480
and all kind of taking here today. Folk, Like I said, I

570
00:42:00.480 --> 00:42:04.320
had it to epiphany with coball,
coball, all these people are worried,

571
00:42:04.360 --> 00:42:07.840
and no coball there are barred.
It's called copilot in particular, which is

572
00:42:07.920 --> 00:42:13.039
a forgithub. Right, And you
just mentioned in the break there that now

573
00:42:13.039 --> 00:42:15.000
you can create these snake applications.
What's that all about? Yeah, you

574
00:42:15.079 --> 00:42:21.079
get the large language models that are
trained to build applications for you called it's

575
00:42:21.079 --> 00:42:23.679
a plugin on a check GPT called
to a GPT engineer. You install it

576
00:42:23.719 --> 00:42:28.320
on your local machine and you say, for example, I want to build

577
00:42:28.679 --> 00:42:31.440
that snake application. So it builds
all the male modules, all the libraries,

578
00:42:31.480 --> 00:42:35.480
all the environments, whatever you need, and you get your base application.

579
00:42:35.679 --> 00:42:37.400
I think in a few minutes that
you have that application, then you

580
00:42:37.440 --> 00:42:42.960
can start weak again and optimizing.
I was thinking of building a few applications

581
00:42:42.960 --> 00:42:45.840
in that way. Apparently the second
version is well, what I saw on

582
00:42:45.880 --> 00:42:50.760
the video is already in say the
topic of application, and that you can

583
00:42:50.880 --> 00:42:53.159
can build an even of such a
high quality that you think, hey,

584
00:42:53.159 --> 00:43:00.360
guys, everybody can become a programmer. Simply said so yeah, next to

585
00:43:00.880 --> 00:43:05.719
which is true. I mean still, I think for a while now it's

586
00:43:05.719 --> 00:43:09.039
going to take human involvements to put
the finishing touches on these things. That

587
00:43:09.119 --> 00:43:12.440
last twenty percent. We just already
talked about it, the preto principle,

588
00:43:12.519 --> 00:43:15.920
right, eighty percent is going to
be done by these engines. The last

589
00:43:15.920 --> 00:43:19.400
twenty percent it's going to be done
checking things and fine tuning. But I

590
00:43:19.400 --> 00:43:22.519
mean even simple applications like I use
constant contact for email, and I noticed,

591
00:43:22.599 --> 00:43:27.599
as of three or four months ago, when you would when you would

592
00:43:27.719 --> 00:43:30.159
use your own htmail may be older
than this, but when you use your

593
00:43:30.159 --> 00:43:32.239
own htmail that a vendor would give
us to promote their stuff, you just

594
00:43:32.280 --> 00:43:36.599
plug that into the window and it'll
there's a button at the top it says

595
00:43:36.760 --> 00:43:39.880
click here to have the errors fixed
for you. I'm like okay, click

596
00:43:39.960 --> 00:43:44.760
boom, errors fixed. You're like
that's fun. Why not like you know,

597
00:43:44.800 --> 00:43:52.639
click here to fix my financial future? Like okay, alreadio right it

598
00:43:52.719 --> 00:43:55.559
sometimes it works so well. But
the point being like, oh my god,

599
00:43:55.639 --> 00:43:59.920
that is a huge or league game
changer. It just being able to

600
00:44:00.159 --> 00:44:02.480
analyze the legacy code. So the
stuff that's you know, running some of

601
00:44:02.480 --> 00:44:07.519
these really old systems that still runs. One of my friends, buddy of

602
00:44:07.519 --> 00:44:12.320
mind is a smart techies like Unix
systems, because I know Unix systems have

603
00:44:12.400 --> 00:44:15.400
been running for thirty years, never
restarted, never had a problem. Like

604
00:44:16.079 --> 00:44:20.280
analytics was based on Unix, wasn't
it. Yeah? Yeah, And if

605
00:44:20.280 --> 00:44:22.760
you see how that works with with
the large language models, you have now

606
00:44:22.800 --> 00:44:25.519
so many plugins where you can say, okay, give me this overview,

607
00:44:25.559 --> 00:44:30.039
put it in a Mermaid graph,
and you have you have your diagrams available.

608
00:44:30.119 --> 00:44:32.400
So imagine that you put that to
your cobal codes and say, just

609
00:44:32.480 --> 00:44:36.880
give me all the interactions with the
systems. You don't have to go into

610
00:44:36.920 --> 00:44:39.159
the code, understand the code,
then take it out, put it in

611
00:44:39.199 --> 00:44:44.440
an over diagram type of system.
Uh. That already is a lot of

612
00:44:44.440 --> 00:44:49.639
work, but imagine just maintaining that
the diagrams as such, if you change

613
00:44:49.679 --> 00:44:53.280
something in the goat all your diagrams
and the visualization, they change it along

614
00:44:53.280 --> 00:44:57.920
with what you are doing. So
that's that's really powerful where you say you

615
00:44:58.000 --> 00:45:01.599
have to maintain it in one place
and the rest will follow based upon understanding

616
00:45:01.800 --> 00:45:06.599
of what you all are doing.
And I was thinking of another example of

617
00:45:06.639 --> 00:45:10.840
all these insane SQL queries that I
saw from three four pages long, which

618
00:45:10.840 --> 00:45:15.960
takes you about five days to understand
what they're really trying to do with EsCl.

619
00:45:15.599 --> 00:45:20.280
Throw it in the large language models
and have an understanding of how to

620
00:45:20.360 --> 00:45:24.119
simplify it, how to break it
down and debug it and really finding the

621
00:45:24.199 --> 00:45:30.599
issues with all these queries. Yeah, I mean so many possibilities what we

622
00:45:30.679 --> 00:45:37.880
have with the models and how to
have that content, which understanding of systems

623
00:45:37.199 --> 00:45:42.119
in place and not our only systems. When we're talking about if you think

624
00:45:42.159 --> 00:45:45.719
about how health desks and service desks
can be optimized, now you have to

625
00:45:45.760 --> 00:45:51.079
think about all the questions that your
customers can ask. Now you guide it

626
00:45:51.119 --> 00:45:55.320
in a certain way and the model
learns by having new questions or related questions

627
00:45:55.360 --> 00:46:00.280
and giving those insights to customers,
or you just feed it the model base

628
00:46:00.519 --> 00:46:02.920
what you already have, and they
will pop out the questions that very likely

629
00:46:04.280 --> 00:46:08.760
can be asked by your customers if
they're communicating with the chambob. So you

630
00:46:08.800 --> 00:46:15.760
can get a very personalized assistance if
you're calling your energy company or your bank

631
00:46:15.920 --> 00:46:21.519
or whatever, and they will be
very knowledgeable compared to what the limitations we

632
00:46:21.599 --> 00:46:25.079
have as a human being. Right, yeah, and wow, I mean

633
00:46:25.320 --> 00:46:28.559
it is going to change. I
think you are going to see it's probably

634
00:46:28.599 --> 00:46:34.800
happening already. Prompt engineers experts in
large language models to do to examine your

635
00:46:34.800 --> 00:46:37.679
code base. I mean that that
would be if I'm a CIO, That's

636
00:46:37.679 --> 00:46:38.960
probably one of the first things I'm
going to do is figure out, all

637
00:46:39.000 --> 00:46:43.599
right, how can we instance you
at our own large language model. You

638
00:46:43.599 --> 00:46:45.960
know. The guy I was talking
to is like, just do it in

639
00:46:45.960 --> 00:46:47.320
the cloud and as your you set
it up and it's off to the races.

640
00:46:47.679 --> 00:46:51.960
You can start playing around with it
and just start examining your code base.

641
00:46:52.079 --> 00:46:54.280
Look at wherever you're having the most
problems. Is a customer service?

642
00:46:54.880 --> 00:47:00.000
Is it downtime for example? Is
it just performance? Our system is running

643
00:47:00.039 --> 00:47:02.039
slow today. That's where I think
it's going to get very interesting is when,

644
00:47:02.119 --> 00:47:05.519
as you just suggested, and maybe
you have to get this plug in,

645
00:47:05.920 --> 00:47:09.400
this engineer or some other plug in, But basically I want this system.

646
00:47:09.760 --> 00:47:14.679
I want this large language model to
examine all the log files for all

647
00:47:14.719 --> 00:47:19.000
my applications for you know, the
last day, for example, and just

648
00:47:19.079 --> 00:47:22.400
tell me what the hell is going
on. That gets very interesting, very

649
00:47:22.480 --> 00:47:27.719
very quickly, because this machine to
machine conversation, even though it is complex

650
00:47:27.760 --> 00:47:30.239
to a certain extent, it's not
as complex as the English language or the

651
00:47:30.239 --> 00:47:36.519
French language. Rights it's pretty simple. What's happening. They're just sending calls

652
00:47:36.559 --> 00:47:39.199
back and forth and sending packets back
and forth. So you know, I

653
00:47:39.320 --> 00:47:44.679
wonder, are we going to be
able to re architect information landscapes? And

654
00:47:44.760 --> 00:47:47.320
I think the answer is yes.
What do you think, Yeah, well,

655
00:47:47.440 --> 00:47:51.880
re architecture. I think we can
build sub optimal systems. That's the

656
00:47:51.920 --> 00:47:57.360
advantage. In a certain way.
We can put up sub optimal systems and

657
00:47:57.440 --> 00:48:00.559
the technology or the algorithms will help
us optimize that. So we can learn

658
00:48:00.599 --> 00:48:06.079
from the from from the algorithms,
how it is optimizing your systems for future

659
00:48:06.199 --> 00:48:09.559
architecture. So learning from your mistakes
thanks to the algorithms and the inside and

660
00:48:09.760 --> 00:48:14.559
was thinking of an example. Keybo
is one of those those players in the

661
00:48:14.639 --> 00:48:19.199
market. What they do they look
They now specifically work with Snowflake, and

662
00:48:19.239 --> 00:48:22.000
the big problem was everybody was putting
everything in the Snowflake databases in the cloud,

663
00:48:22.360 --> 00:48:28.400
but their bill went pretty much well
ten poled up, so big surprise.

664
00:48:28.480 --> 00:48:30.840
So what they do They look at
at the log files and your query

665
00:48:30.920 --> 00:48:36.320
behavior and the usage of your databases, and based upon that, they know

666
00:48:36.440 --> 00:48:40.280
how to optimize the Snowflake system.
They know how to scale from from the

667
00:48:40.280 --> 00:48:45.440
medium large and extra large type of
systems. They know how to optimize the

668
00:48:45.559 --> 00:48:47.840
memory and so on and so forth. So based upon when you need which

669
00:48:47.840 --> 00:48:52.760
type of resources, they scale up
and scale down automatically. You even can,

670
00:48:52.880 --> 00:48:54.199
for example, say hey, by
the end of the month, we

671
00:48:54.639 --> 00:49:00.320
definitely need so much more compute power
because it's an end of must closure and

672
00:49:00.360 --> 00:49:04.159
I couldn't care about it, couldn't
care about the costs in a certain way,

673
00:49:04.239 --> 00:49:07.360
but we do want to have this
animal closure done into hours, for

674
00:49:07.400 --> 00:49:13.840
example, and coordinate the system and
tell the system this is how I want

675
00:49:13.880 --> 00:49:15.360
to work, and they do.
They have a lifting for you, just

676
00:49:15.400 --> 00:49:20.280
bringing your build down by by sixty
percent. So that's really insane. Why

677
00:49:20.320 --> 00:49:23.440
you say, okay, we can
focus upon building the suboptimal systems, because

678
00:49:23.519 --> 00:49:28.320
before we need to understand all the
ins and outs of the system to have

679
00:49:28.360 --> 00:49:31.239
it optimized to have the optimal architecture. It's going to take a long while

680
00:49:31.280 --> 00:49:35.280
because we don't have all the experts. And I think there there's going to

681
00:49:35.360 --> 00:49:38.039
be a lot of these systems which
are coming in. And I was thinking

682
00:49:38.039 --> 00:49:44.199
back in the days where yet the
data virtualization systems that have the knowledge of

683
00:49:44.239 --> 00:49:45.840
how to upt to you that performance
what you are looking for. So we

684
00:49:45.880 --> 00:49:52.159
were the data virtualization systems where a
rule based system on how to optimize systems.

685
00:49:52.199 --> 00:49:55.519
But now with the algorithms, they
really look at the behaviors and connect

686
00:49:55.679 --> 00:50:00.039
everything what is under the hood,
and then based upon these sites, optimize

687
00:50:00.079 --> 00:50:06.960
your system. So we're still at
the beginning off of having insane systems optimizing

688
00:50:07.000 --> 00:50:09.719
each other. And where it's going. I do wonder, like one would

689
00:50:09.760 --> 00:50:15.679
last topic to dive into, there's
obviously a cost for open AI. There's

690
00:50:15.679 --> 00:50:19.480
a cost for Microsoft when you're deploying
that environment. Of course they're charging you

691
00:50:19.639 --> 00:50:22.320
now, but I wonder when we're
gonna when when's that other shoe gonna drop?

692
00:50:22.719 --> 00:50:24.920
Because I have to think it's probably
gonna come, you know, probably

693
00:50:24.920 --> 00:50:29.719
the next six to nine months or
something specifically from open Ai because they gotta

694
00:50:29.719 --> 00:50:30.880
make money. I mean, they're
getting investments, right, They're gonna a

695
00:50:30.920 --> 00:50:34.239
lot of money to be able to
spread this stuff out. But at a

696
00:50:34.280 --> 00:50:37.199
certain point they're going to you know, change and it's the freemium model,

697
00:50:37.280 --> 00:50:37.960
right, you know, all of
a sudden, Okay, well I was

698
00:50:38.000 --> 00:50:40.079
free. Now it's gonna be X
amount of money. I mean, I

699
00:50:40.159 --> 00:50:43.920
wonder what's going to happen there.
It's going to be a cost per query,

700
00:50:44.079 --> 00:50:45.920
you know, what's like, what's
the story? That's my guess is

701
00:50:45.960 --> 00:50:49.760
that you know you're gonna be able
to do and they're still probably figuring out

702
00:50:49.760 --> 00:50:52.119
how to charge for all that.
But what do you think is that going

703
00:50:52.159 --> 00:50:53.400
to be three months? Six months? When is when is the other shoe

704
00:50:53.400 --> 00:50:58.559
gonna drop? Financially? Well,
I think it's it's coming pretty pretty quickly.

705
00:50:58.559 --> 00:51:00.320
If you see, you're already have
a check GPT plus on top of

706
00:51:00.360 --> 00:51:05.320
open Ai of all the extensions which
are comic, so we're going their direction.

707
00:51:05.960 --> 00:51:08.760
What we saw is that that based
upon how you are charges based upon

708
00:51:08.760 --> 00:51:14.599
a token tokens one token is one
hundred words something like that. And you

709
00:51:14.639 --> 00:51:19.000
see a lot of other language models, they are only a tenth of diffraction

710
00:51:19.159 --> 00:51:22.800
of an opening isis. So there
we see already that they're optimizing on how

711
00:51:22.840 --> 00:51:30.679
they can train their models to consume
less resources, might be more comfortable.

712
00:51:30.039 --> 00:51:34.239
And that's going to be a big
challenge and and and and and a focus

713
00:51:34.320 --> 00:51:38.960
for all the big models which are
out there and become profitable. But yeah,

714
00:51:39.000 --> 00:51:43.280
it's a trade off. Maybe Microsoft
says, yeah, we just keep

715
00:51:43.280 --> 00:51:46.960
on offering the large language models,
and we take our profit based upon all

716
00:51:46.960 --> 00:51:51.480
the services, what we offer on
the infrastructure, on the compute. That

717
00:51:51.599 --> 00:51:55.320
can be a model that is working
for them, and we can't anticipate what

718
00:51:55.440 --> 00:52:00.599
is their business model. But that's
what I've heard that some companies just say

719
00:52:00.760 --> 00:52:04.599
it's how Amazon grew. In fact, they offered a lot of thanks for

720
00:52:04.760 --> 00:52:07.239
free, but they took the profit
on top of the services and the infrastructure

721
00:52:07.280 --> 00:52:12.679
what they are offering, so they
moved a bit where the profit came from.

722
00:52:12.800 --> 00:52:15.519
Wow, that's what you do,
folks in the real world. Well,

723
00:52:15.559 --> 00:52:19.119
you're talking every buddy Bobbers, you
know, inside Analysis semi email income

724
00:52:19.119 --> 00:52:22.360
on Inside Analysis dot com. You've
been listening to Inside Analysis, miss something

725
00:52:22.360 --> 00:52:29.679
today, yesterday, last week.
Check out our podcasts at WWWKCAA radio dot

726
00:52:29.679 --> 00:52:35.199
com. We leave no listener behind. Happy Fourth of July from this radio

727
00:52:35.239 --> 00:52:37.840
station. It's time to celebrate again
this year with the Native Sons of the

728
00:52:37.880 --> 00:52:43.559
Golden West Arrowhead Parlor and their family
safe fireworks. This year, the Native

729
00:52:43.599 --> 00:52:46.920
Sons have moved two blocks south of
the two ten Freeway to twenty five hundred

730
00:52:46.920 --> 00:52:51.360
North deal Rosa Avenue. That's twenty
five hundred North Del Rosa Avenue at the

731
00:52:51.400 --> 00:52:54.159
corner of Pamelo, next to the
Dental Center and across from Wabba Grill,

732
00:52:54.280 --> 00:52:59.639
Piccolo Peet's, Buzzing Bees, Flaming
Flowers, and the all time favorite sparkler

733
00:53:00.039 --> 00:53:02.400
spark on the Fourth of July,
just like the ones available at the Native

734
00:53:02.400 --> 00:53:07.119
Sons of the Golden West Fireworks Booth
in San Bernardino. The Golden Sons not

735
00:53:07.159 --> 00:53:12.159
only celebrate our nation, but at
great State of California through friendship, loyalty

736
00:53:12.239 --> 00:53:16.320
and charity. They support our communities
through volunteerism. Support them with your fireworks

737
00:53:16.320 --> 00:53:21.480
sale. The Native Sons Arrowhead Parlor
one ten would like to remind you that

738
00:53:21.519 --> 00:53:24.960
only Safe approves fireworks like the One
Soul, that their fireworks stand now through

739
00:53:24.960 --> 00:53:30.039
the fourth are allowed and no fireworks
are to be ignited above the two ten

740
00:53:30.119 --> 00:53:32.559
Freeway, and remember to keep fireworks
out of the hands of children. Keep

741
00:53:32.559 --> 00:53:37.519
it safe with the Native Sons of
the Golden West and San Bernardino and fireworks

742
00:53:37.519 --> 00:53:40.280
from their boat just south of their
two ten Freeway in San Bernardino. Be

743
00:53:40.360 --> 00:53:44.559
sure you don't miss out. Fireworks
stands are limited and the Native Suns is

744
00:53:44.599 --> 00:53:47.440
sold out every year on or before
the fourth for the past four years at

745
00:53:47.440 --> 00:53:57.760
twenty five hundred North Del Rosa Avenue
in San Bernardino. This segment sponsored by

746
00:53:57.760 --> 00:54:00.760
the generous support of the Dream Team. Looking for the keys to something bigger

747
00:54:00.800 --> 00:54:07.360
and better downsizing or relocating to the
perfect spot. Oscar Ramirez from Century twenty

748
00:54:07.360 --> 00:54:12.599
one Lowslower real Estate and Matt Flores
from Secure Choice Lending are here to help

749
00:54:12.639 --> 00:54:16.480
you sell or buy with their trusted
and experienced knowledge and advice. People are

750
00:54:16.480 --> 00:54:21.480
calling Oscar and Matt at nine five, one seventy five, one three two

751
00:54:21.559 --> 00:54:24.079
four nine. That's nine five,
one seven, five, one thirty two

752
00:54:24.320 --> 00:54:30.000
forty nine real estate and loan advisors
Oscar and Matt can give you a no

753
00:54:30.159 --> 00:54:34.360
cost consultation. You don't have to
buy anything. Matt and Oscar can help

754
00:54:34.360 --> 00:54:38.079
you figure your way through the complicated
real estate market. Email Oscar at loislower

755
00:54:38.199 --> 00:54:45.800
dot com or on Instagram at Oscar
Ramirez Garcia and Matt Flores at Secure Choice

756
00:54:45.880 --> 00:54:50.760
Lending dot com. Don't let today's
real estate pitfalls stop you from dreaming.

757
00:54:51.000 --> 00:54:55.599
Make your new home dreams come true. Dri number zero two zero seven zero

758
00:54:55.679 --> 00:55:05.480
three four four Zehabla espanol Tahebow Tea
Club's original pure power Darco Supertea helps build

759
00:55:05.519 --> 00:55:08.119
red corpusals in the blood, which
carry oxygen to organs and cells. Our

760
00:55:08.239 --> 00:55:13.719
organs and cells need oxygen to regenerate
themselves. The immune system needs oxygen to

761
00:55:13.800 --> 00:55:16.320
develop, and cancer dyes in oxygen. So the t is great for healthy

762
00:55:16.360 --> 00:55:21.119
people because it helps build the immune
system, and it can truly be miraculous

763
00:55:21.199 --> 00:55:24.320
for someone fighting a potentially life threatening
disease due to an infection, diabetes,

764
00:55:24.440 --> 00:55:29.360
or cancer. The tea is also
organic and naturally caffeine free. A one

765
00:55:29.440 --> 00:55:32.639
pound package of tea is forty nine
ninety five, which includes shipping to order,

766
00:55:32.719 --> 00:55:37.679
please visit Tehebow Tea Club dot com. Tihebow is spelled tea like tom,

767
00:55:37.000 --> 00:55:42.119
a, h ee b like boy
o. Then continue with the word

768
00:55:42.159 --> 00:55:45.280
t and then the word club.
The complete website is to Heebow Tea Club

769
00:55:45.360 --> 00:55:50.639
dot com or call us at eight
one eight six one zero eight zero eight

770
00:55:50.679 --> 00:55:55.039
eight Monday through Saturday nine am to
five pm California time. That's eight one

771
00:55:55.159 --> 00:56:02.639
eight six one zero eight zero eight
eight to Heebotclub dot com. Take time

772
00:56:02.679 --> 00:56:07.400
today and every day to thank our
veterans for their service. Our sponsor Inland

773
00:56:07.400 --> 00:56:10.199
Empire Roofing of Riverside, a second
generation owned and operated serving the area for

774
00:56:10.280 --> 00:56:15.519
over thirty five years. Inland Empire
Roofing specializes in roofing and blown in attic

775
00:56:15.559 --> 00:56:19.199
installation. You call your local roofing
experts at nine five one six eight four

776
00:56:19.360 --> 00:56:22.920
sixty eight zero zero again nine five
one six eight four sixty eight zero zero

777
00:56:23.039 --> 00:56:28.199
or visit Inland Empire Roofing code dot
com. Inland Empire Roofing on the air

778
00:56:28.320 --> 00:56:36.960
because they care. With sixty years
of fascinating facts. This is the man

779
00:56:37.239 --> 00:56:42.599
from yesterday and back in time.
We go to this time in nineteen eighty

780
00:56:42.639 --> 00:56:46.320
five, Coca Cola is still getting
tons of complaints. Will those who describe

781
00:56:46.400 --> 00:56:52.920
their new Coca Cola as lacking the
old formulas bite and zing. New coke

782
00:56:52.119 --> 00:56:58.039
was introduced to bring back the old
taste after the formula was changed. So

783
00:56:58.400 --> 00:57:00.199
new coke he was catching on the
tap is better. And you were there,

784
00:57:04.559 --> 00:57:07.639
you said the word. And from
this time. In nineteen sixty six,

785
00:57:07.880 --> 00:57:12.199
the Real Mayor of New Orleans,
Victor Shiro, helps Walt Disney open

786
00:57:12.280 --> 00:57:16.719
the new New Orleans Square and Disneyland. New Orleans Square, the new home

787
00:57:16.800 --> 00:57:21.639
of the Pirates is located at a
bend of the river, just a few

788
00:57:21.719 --> 00:57:24.039
steps from the landing where the Mark
Twain and Columbia dock. With its ornate

789
00:57:24.119 --> 00:57:29.760
woods and lacey iron grill work,
New Orleans Square stands above all the Lands

790
00:57:29.800 --> 00:57:34.159
and Disneylands or shire elegance of design. And from this time in nineteen seventy

791
00:57:34.199 --> 00:57:37.760
one, responding to a friend's appeal
for help, George Harrison would give his

792
00:57:37.880 --> 00:57:42.800
first concert in four years. The
friend was Rabbi Shankar, the Indian maestro

793
00:57:43.000 --> 00:57:45.840
of the sitar. The appeal was
to help raise money for the millions of

794
00:57:45.960 --> 00:58:02.480
refugees who fled to India from Bangladesh
with Moore at Man from yesterday dot Com.

795
00:58:07.639 --> 00:58:09.880
You've eaten lots of great food and
lots of great food at restaurants.

796
00:58:10.079 --> 00:58:15.960
Cowboy Burgers in Fontana and now on
Arlington and Riverside will fast become one of

797
00:58:15.039 --> 00:58:22.159
your favorites with their delicious, mouthwatering
burgers and breakfast burritos. Cowboy Burgers and

798
00:58:22.199 --> 00:58:25.840
Barbecue also serves fantastic smoke barbecue,
baby back ribs, dry tit chicken,

799
00:58:27.000 --> 00:58:30.480
pulled pork sandwiches, as well as
lunch and dinner plates. Everything is made

800
00:58:30.559 --> 00:58:34.679
from scratch, including their delicious side
dishes like cole slap, potato salad,

801
00:58:34.679 --> 00:58:37.760
barbecue beans, and much much more. Check out their rich, decadent chocolate

802
00:58:37.800 --> 00:58:42.519
brownies. Hi. I'm food critic
Gallenborgan and you can dine in, take

803
00:58:42.599 --> 00:58:45.599
food out, or have them cater
your next special event. I highly recommend

804
00:58:45.679 --> 00:58:52.480
Cowboy Burgers and Barbecue at their new
location at five five seven three Arlington Avenue

805
00:58:52.519 --> 00:58:55.400
in Riverside. Just look them up
on the Internet. That's Cowboy Burgers and

806
00:58:55.480 --> 00:59:01.000
Barbecue, happy eating and perfect for
the holidays. Cowboy Burgers in Barbecue is

807
00:59:01.079 --> 00:59:07.679
also available for catering. That's Cowboy
Burgers in Barbecue in Fontana and now in

808
00:59:07.800 --> 00:59:14.280
Riverside on Arlington, HI. This
is Steve Crawford, the retirement wealth coach

809
00:59:14.480 --> 00:59:19.360
and host of sports Bucks with Steve
Spots. Spots with Steve is a financial

810
00:59:19.440 --> 00:59:23.679
show that also looks at sports and
how it impacts our culture. We'll talk

811
00:59:23.760 --> 00:59:29.280
all things money in sports and give
you the right game plan to help coachu

812
00:59:29.440 --> 00:59:32.480
up to be successful in your retirement. So tune in every Saturday morning at

813
00:59:32.559 --> 00:59:37.960
seven am for Sports Bucks right here
at casey AA the station at least no

814
00:59:38.159 --> 00:59:45.960
listener behind Itistina CACAA, Lowlinda and
one oh six point five FM, K

815
00:59:46.119 --> 00:59:52.239
two ninety three c F Burrino Valley, NBC News Radio. I'm Chris Garraggio.

816
00:59:52.400 --> 00:59:55.880
Severe storms are taking aim at parts
of the South and Northeast during the

817
00:59:57.000 --> 01:00:00.679
extended Fourth of July weekend. More
than fifty million America ins are at risk

818
01:00:00.719 --> 01:00:05.639
of experiencing thunderstorms through Sunday night.
The weather threat spans areas like Philadelphia,

819
01:00:05.679 --> 01:00:08.360
Baltimore, Charlotte, and Washington,
d C. With the strongest storms targeting could