In Conversation with Christian Dippel: Developing Competitive Edge with AI

1
00:00:00,000 --> 00:00:01,625
CHRISTIAN DIPPEL:
Your job as a leader,

2
00:00:01,625 --> 00:00:03,230
whether it's at
the C-suite level

3
00:00:03,230 --> 00:00:06,120
or leading a strategic business
unit or whatever the case,

4
00:00:06,120 --> 00:00:09,080
I think it's going to become
more about asking good questions

5
00:00:09,080 --> 00:00:14,300
and less so about being the
person to provide the answers.

6
00:00:14,300 --> 00:00:16,410
ANNOUNCER: Welcome to
Learning In Action,

7
00:00:16,410 --> 00:00:19,740
where we explore fresh ideas
shaping leadership today.

8
00:00:19,740 --> 00:00:23,300
This episode is a recording of
our live session in conversation

9
00:00:23,300 --> 00:00:27,030
with Dr. Christian Dippel,
Associate Professor of Business,

10
00:00:27,030 --> 00:00:30,360
Economics and Public Policy
at Ivey Business School.

11
00:00:30,360 --> 00:00:33,590
If you're asking how to move
beyond the hype around AI

12
00:00:33,590 --> 00:00:36,180
and turn it into real
strategic impact,

13
00:00:36,180 --> 00:00:37,570
you're in the right place.

14
00:00:37,570 --> 00:00:40,280
Christian walks us through
spotting high potential AI

15
00:00:40,280 --> 00:00:43,620
opportunities, aligning them
with your broader strategy

16
00:00:43,620 --> 00:00:45,770
and bringing
organizational clarity

17
00:00:45,770 --> 00:00:48,180
to a fast-moving
technology wave.

18
00:00:48,180 --> 00:00:51,020
Whether you joined us live or
are tuning in for the first

19
00:00:51,020 --> 00:00:54,360
time, this episode delivers
"can't miss" insights.

20
00:00:54,360 --> 00:00:55,650
Let's get into it.

21
00:00:55,650 --> 00:00:57,000
HOST: Christian, welcome.

22
00:00:57,000 --> 00:00:58,950
Thank you for joining us today.

23
00:00:58,950 --> 00:01:00,908
CHRISTIAN DIPPEL: Thanks
so much for having me.

24
00:01:00,908 --> 00:01:03,660
HOST: Why don't we start
out on a more personal note,

25
00:01:03,660 --> 00:01:07,620
and if you could share a little
bit about your personal journey,

26
00:01:07,620 --> 00:01:12,500
which spans the rigor of
research and academics

27
00:01:12,500 --> 00:01:16,880
all the way through to practical
applications of organizations

28
00:01:16,880 --> 00:01:18,960
that you're directly
working with?

29
00:01:18,960 --> 00:01:21,470
CHRISTIAN DIPPEL: I've been a
professor for about 14 years

30
00:01:21,470 --> 00:01:24,500
now, graduated in 2011
and spent the first 10

31
00:01:24,500 --> 00:01:27,690
years of my academic life
at UCLA in California.

32
00:01:27,690 --> 00:01:29,760
Really, at the beginning
of my academic career,

33
00:01:29,760 --> 00:01:31,500
I wore three hats.

34
00:01:31,500 --> 00:01:33,420
I had an investment
banking background,

35
00:01:33,420 --> 00:01:35,070
so I had a bit of
that finance hat.

36
00:01:35,070 --> 00:01:38,910
My formal PhD training
was in global macro,

37
00:01:38,910 --> 00:01:40,550
really international
macroeconomics,

38
00:01:40,550 --> 00:01:42,150
and economic history.

39
00:01:42,150 --> 00:01:45,060
And so those were my
three hats that I wore.

40
00:01:45,060 --> 00:01:47,220
And then over time
through my career,

41
00:01:47,220 --> 00:01:50,553
two things influenced me
and pushed me, so to speak,

42
00:01:50,553 --> 00:01:53,220
a little bit towards the path of
what we're talking about today.

43
00:01:53,220 --> 00:01:57,260
The first is being at UCLA, it's
a very entrepreneurship-heavy

44
00:01:57,260 --> 00:01:58,050
business school.

45
00:01:58,050 --> 00:01:59,820
So among North American
business schools,

46
00:01:59,820 --> 00:02:01,850
it's got the highest
ratio of MBAs

47
00:02:01,850 --> 00:02:04,740
starting their own business
after MBA graduation.

48
00:02:04,740 --> 00:02:08,300
So you get a lot of this
Silicon Valley or Silicon Valley

49
00:02:08,300 --> 00:02:10,910
adjacent Southern
California tech

50
00:02:10,910 --> 00:02:13,430
entrepreneurship,
venture initiation stuff

51
00:02:13,430 --> 00:02:15,060
that you get exposed to.

52
00:02:15,060 --> 00:02:18,620
The other thing that happened
is I became more and more a data

53
00:02:18,620 --> 00:02:20,450
scientist, and I
got increasingly

54
00:02:20,450 --> 00:02:23,780
frustrated with the fact that
in academia, what often happens

55
00:02:23,780 --> 00:02:27,560
is people engage with these
incredibly interesting data

56
00:02:27,560 --> 00:02:31,100
science projects, but everything
is on the academic funding

57
00:02:31,100 --> 00:02:31,680
cycle.

58
00:02:31,680 --> 00:02:33,900
So you get a big grant
to build a cool thing.

59
00:02:33,900 --> 00:02:35,390
You build the
thing, then you try

60
00:02:35,390 --> 00:02:37,710
to write three or
four papers on it,

61
00:02:37,710 --> 00:02:39,630
and then you move on
to the next topic.

62
00:02:39,630 --> 00:02:42,560
So everything is optimized
for the paper writing, which

63
00:02:42,560 --> 00:02:44,600
is fine, but it's
really frustrating

64
00:02:44,600 --> 00:02:46,490
that you get these
cool data projects that

65
00:02:46,490 --> 00:02:48,785
then lead these zombie lives
on the internet, where you

66
00:02:48,785 --> 00:02:50,910
go on a website and you
know, this thing was built.

67
00:02:50,910 --> 00:02:51,660
It was so cool.

68
00:02:51,660 --> 00:02:54,170
And then it says,
last updated in 2019

69
00:02:54,170 --> 00:02:56,630
because the grant
ran out, the people

70
00:02:56,630 --> 00:02:58,800
got tenure or lost
interest or whatever.

71
00:02:58,800 --> 00:03:01,010
And so I was of
this mindset that I

72
00:03:01,010 --> 00:03:03,440
wanted to build data
science projects that

73
00:03:03,440 --> 00:03:05,660
have the ability to
self-sustain, which then

74
00:03:05,660 --> 00:03:08,390
naturally meant they had to
have a commercial element

75
00:03:08,390 --> 00:03:11,970
to themselves to break out
of this research cadence,

76
00:03:11,970 --> 00:03:12,510
if you will.

77
00:03:12,510 --> 00:03:15,660
And then I quickly find
out that if you build it,

78
00:03:15,660 --> 00:03:16,675
they won't just come.

79
00:03:16,675 --> 00:03:18,050
So in other words,
you can't just

80
00:03:18,050 --> 00:03:19,717
build the thing that
you're in love with

81
00:03:19,717 --> 00:03:22,050
and assume that it can
stand on its own two feet.

82
00:03:22,050 --> 00:03:24,990
And I had to quickly
learn a lot of strategy.

83
00:03:24,990 --> 00:03:28,040
And then, really, it sort of
became this natural confluence

84
00:03:28,040 --> 00:03:31,110
where I was building
projects in data science.

85
00:03:31,110 --> 00:03:32,550
I had to learn the strategy.

86
00:03:32,550 --> 00:03:34,740
And then right around the
time when that happened,

87
00:03:34,740 --> 00:03:35,850
AI came around.

88
00:03:35,850 --> 00:03:38,400
And so for me, strategy,
almost overnight,

89
00:03:38,400 --> 00:03:40,680
very quickly just
became AI strategy.

90
00:03:40,680 --> 00:03:43,260
So in other words, for
me, for what I was doing,

91
00:03:43,260 --> 00:03:46,140
there was no strategy
that wasn't AI strategy.

92
00:03:46,140 --> 00:03:48,950
HOST: Thank you for
the introduction.

93
00:03:48,950 --> 00:03:52,460
I'm smiling because it's two
comments here at the-- build it,

94
00:03:52,460 --> 00:03:55,530
and they will come doesn't
necessarily really happen.

95
00:03:55,530 --> 00:03:58,050
It's how do you go beyond
that, and how do you engage?

96
00:03:58,050 --> 00:04:03,060
And given the topic at hand,
even when you reference 2019,

97
00:04:03,060 --> 00:04:05,330
when it comes to AI, that
feels like it's centuries

98
00:04:05,330 --> 00:04:09,150
ago in terms of how fast
everything is moving.

99
00:04:09,150 --> 00:04:13,940
So let's go back to the
topic or even the headline.

100
00:04:13,940 --> 00:04:17,180
We talk about, creating
a competitive advantage,

101
00:04:17,180 --> 00:04:19,800
and how can I play a role?

102
00:04:19,800 --> 00:04:22,740
I actually talked about
moving away from buzzwords.

103
00:04:22,740 --> 00:04:26,030
Competitive advantage almost
feels like a buzzword,

104
00:04:26,030 --> 00:04:27,570
to some extent.

105
00:04:27,570 --> 00:04:29,810
It's obviously
something that we strive

106
00:04:29,810 --> 00:04:31,640
for when you're
leading organizations

107
00:04:31,640 --> 00:04:33,840
to create that edge.

108
00:04:33,840 --> 00:04:35,930
So how can we think
about it maybe

109
00:04:35,930 --> 00:04:39,210
in a more practical
or pragmatic term?

110
00:04:39,210 --> 00:04:41,730
CHRISTIAN DIPPEL: I got into
strategy in a roundabout way,

111
00:04:41,730 --> 00:04:42,313
you would say.

112
00:04:42,313 --> 00:04:43,860
I wasn't trained
as a strategist.

113
00:04:43,860 --> 00:04:45,960
I was trained as an
economist in global macro,

114
00:04:45,960 --> 00:04:47,820
and then the
strategy came later.

115
00:04:47,820 --> 00:04:52,820
And when I was at UCLA,
I had this very renowned

116
00:04:52,820 --> 00:04:56,750
senior strategy colleague who
retired midway through my stint

117
00:04:56,750 --> 00:05:00,028
at UCLA, Richard Rumelt. And
he used to tell the story of--

118
00:05:00,028 --> 00:05:02,570
he would have another senior
colleague come into his strategy

119
00:05:02,570 --> 00:05:06,210
class to maybe get into
teaching strategy himself.

120
00:05:06,210 --> 00:05:09,440
And after sitting there for 10
classes, he walked up to Richard

121
00:05:09,440 --> 00:05:12,380
and said, I think I can
summarize your entire strategy

122
00:05:12,380 --> 00:05:15,000
class in one question,
what's going on here?

123
00:05:15,000 --> 00:05:17,460
The point that he was making
is, at the end of the day,

124
00:05:17,460 --> 00:05:18,350
strategy is--

125
00:05:18,350 --> 00:05:21,890
it's a fairly unstructured
thing in the sense

126
00:05:21,890 --> 00:05:23,520
that it's so context-specific.

127
00:05:23,520 --> 00:05:25,580
And this was an interesting
experience for me

128
00:05:25,580 --> 00:05:28,440
as a macroeconomist who is
really a data scientist.

129
00:05:28,440 --> 00:05:30,507
You're always in the
business of generalities,

130
00:05:30,507 --> 00:05:32,090
so you're always
trying to figure out,

131
00:05:32,090 --> 00:05:33,780
what does the average do?

132
00:05:33,780 --> 00:05:35,713
What are like the
patterns in the data?

133
00:05:35,713 --> 00:05:37,130
But when you start
to think about,

134
00:05:37,130 --> 00:05:40,850
what is going to make my one
thing successful in this sea

135
00:05:40,850 --> 00:05:43,220
of other things around it,
it's like your whole way

136
00:05:43,220 --> 00:05:45,380
of thinking totally
pivots into-- everything

137
00:05:45,380 --> 00:05:46,800
becomes about context.

138
00:05:46,800 --> 00:05:49,200
And the data is a
background thing.

139
00:05:49,200 --> 00:05:50,820
You need to know
some business trends.

140
00:05:50,820 --> 00:05:53,445
You need to have a sense
for consumer behavior,

141
00:05:53,445 --> 00:05:55,820
competitive landscape behavior,
all these things that can

142
00:05:55,820 --> 00:05:57,068
be expressed quantitatively.

143
00:05:57,068 --> 00:05:58,610
But at the end of
the day, it becomes

144
00:05:58,610 --> 00:06:02,510
this highly context-specific,
very qualitative endeavor

145
00:06:02,510 --> 00:06:06,000
and really thinking deeply
about, what is going on here?

146
00:06:06,000 --> 00:06:08,330
The answer to the question,
what is going on here,

147
00:06:08,330 --> 00:06:10,860
is going to be different
for every kind of business.

148
00:06:10,860 --> 00:06:13,100
There will be businesses
where AI really hits you

149
00:06:13,100 --> 00:06:14,262
on the process innovation.

150
00:06:14,262 --> 00:06:16,220
There'll be businesses
where it really hits you

151
00:06:16,220 --> 00:06:17,460
on the product innovation.

152
00:06:17,460 --> 00:06:18,950
And I think it's really about--

153
00:06:18,950 --> 00:06:20,750
the way that I think
about it is really

154
00:06:20,750 --> 00:06:23,870
to not be too heavy on
the playbooks, like, let's

155
00:06:23,870 --> 00:06:25,550
apply this framework
or that framework.

156
00:06:25,550 --> 00:06:28,280
Frameworks have their use, for
sure, but nothing, I think,

157
00:06:28,280 --> 00:06:31,350
replaces deeply thinking
about your situation.

158
00:06:31,350 --> 00:06:33,570
HOST: Great way to frame it.

159
00:06:33,570 --> 00:06:36,870
I find sometimes when we
use these bigger terms--

160
00:06:36,870 --> 00:06:38,690
and even I had a
recent conversation

161
00:06:38,690 --> 00:06:41,690
with another faculty where
we dug in specifically

162
00:06:41,690 --> 00:06:44,070
on strategy and
scenario planning

163
00:06:44,070 --> 00:06:48,540
and, how do you stay ahead in
today's rapidly changing world?

164
00:06:48,540 --> 00:06:51,270
It's sort of broke
it down to say--

165
00:06:51,270 --> 00:06:54,710
but it doesn't have to be
this big grandiose piece where

166
00:06:54,710 --> 00:06:57,200
you almost run the risk of
making it bigger than it is

167
00:06:57,200 --> 00:06:58,680
and creating that inertia.

168
00:06:58,680 --> 00:07:01,740
It's like, OK, so
what's going on today?

169
00:07:01,740 --> 00:07:03,090
What's going on tomorrow?

170
00:07:03,090 --> 00:07:06,540
Continue to ask myself
these questions.

171
00:07:06,540 --> 00:07:08,540
So if we go to
your work, we look

172
00:07:08,540 --> 00:07:12,600
at drawing on political economy,
institutional economics.

173
00:07:12,600 --> 00:07:17,870
How do these lenses
illustrate the strategic role

174
00:07:17,870 --> 00:07:19,560
that AI can play?

175
00:07:19,560 --> 00:07:22,010
And maybe I'll put
in a second part

176
00:07:22,010 --> 00:07:25,610
to this question around
hidden assumptions

177
00:07:25,610 --> 00:07:28,763
that might be at play as well.

178
00:07:28,763 --> 00:07:29,930
What are your thoughts here?

179
00:07:29,930 --> 00:07:31,305
CHRISTIAN DIPPEL:
In my ventures,

180
00:07:31,305 --> 00:07:32,600
let's say-- let's start there.

181
00:07:32,600 --> 00:07:36,200
Those ventures, as well as
my research, really are in--

182
00:07:36,200 --> 00:07:40,410
political economy people refer
to that as government relations.

183
00:07:40,410 --> 00:07:41,820
It's really similar things.

184
00:07:41,820 --> 00:07:44,400
And I saw that transition of--

185
00:07:44,400 --> 00:07:46,860
2011 was when I got my PhD.

186
00:07:46,860 --> 00:07:50,840
2014, '15 is when it
was cutting-edge to be

187
00:07:50,840 --> 00:07:54,140
using keyword extraction
to characterize

188
00:07:54,140 --> 00:07:55,710
certain bodies of text.

189
00:07:55,710 --> 00:07:58,880
By 2016, you can show off
with your machine learning

190
00:07:58,880 --> 00:07:59,990
algorithms.

191
00:07:59,990 --> 00:08:03,090
By 2017, you do
sentiment analysis.

192
00:08:03,090 --> 00:08:05,550
And so the bar kept getting
raised but in a somewhat,

193
00:08:05,550 --> 00:08:07,020
I would say, incremental manner.

194
00:08:07,020 --> 00:08:09,350
But it's really 2023
when it hits you.

195
00:08:09,350 --> 00:08:11,280
Suddenly, there's
this level shift.

196
00:08:11,280 --> 00:08:12,860
There's a million
things that you

197
00:08:12,860 --> 00:08:14,850
can do that you
couldn't previously do.

198
00:08:14,850 --> 00:08:18,260
And from a business perspective,
anyone with a lot of exposure

199
00:08:18,260 --> 00:08:20,390
to data or a lot
of in-house data

200
00:08:20,390 --> 00:08:23,570
or needing to grapple with a lot
of data-- there's text as data

201
00:08:23,570 --> 00:08:24,960
and there's quantitative data.

202
00:08:24,960 --> 00:08:27,500
I would say the
manifestation of AI

203
00:08:27,500 --> 00:08:31,170
that we're talking about mostly
today, which is not agentic,

204
00:08:31,170 --> 00:08:32,990
but let's say just
like LLMs, that's

205
00:08:32,990 --> 00:08:34,470
mostly about text as data.

206
00:08:34,470 --> 00:08:36,679
And text as data is really
an application where,

207
00:08:36,679 --> 00:08:41,000
if your business interfaces with
text as data, which mine did--

208
00:08:41,000 --> 00:08:42,830
with text as data,
you're constantly

209
00:08:42,830 --> 00:08:46,250
in this world of not being able
to see the forest for the trees

210
00:08:46,250 --> 00:08:48,030
because there's
just so much text.

211
00:08:48,030 --> 00:08:49,830
So if you're interfacing
with government,

212
00:08:49,830 --> 00:08:52,010
there's hundreds
of bills every year

213
00:08:52,010 --> 00:08:55,970
that affect hundreds of X,
which in turn affect thousands

214
00:08:55,970 --> 00:08:57,180
of regulations.

215
00:08:57,180 --> 00:09:00,230
There's thousands of pieces
of formal commenting, lobby

216
00:09:00,230 --> 00:09:02,070
meeting, committee
meeting transcripts.

217
00:09:02,070 --> 00:09:04,520
There's just such
a big ocean of text

218
00:09:04,520 --> 00:09:07,830
that no traditional method can
really get a good handle on,

219
00:09:07,830 --> 00:09:12,560
and that's really where AI it
lowers the bar so much in terms

220
00:09:12,560 --> 00:09:14,630
of even a very
small business being

221
00:09:14,630 --> 00:09:17,090
able to do so much
with a lot of data,

222
00:09:17,090 --> 00:09:18,530
where you went
from being able to,

223
00:09:18,530 --> 00:09:20,750
do more or less, nothing
at all five years ago

224
00:09:20,750 --> 00:09:24,080
to being able to do
enterprise-level applications

225
00:09:24,080 --> 00:09:27,560
with a tiny team if you are
in this kind of very data

226
00:09:27,560 --> 00:09:29,780
heavy environment,
which, in my case,

227
00:09:29,780 --> 00:09:32,930
in the political economy,
institutional economics type

228
00:09:32,930 --> 00:09:34,710
environment, we work.

229
00:09:34,710 --> 00:09:37,430
HOST: I have
conversations frequently

230
00:09:37,430 --> 00:09:41,990
with leaders who are grappling
with almost too much.

231
00:09:41,990 --> 00:09:44,100
There's so much information.

232
00:09:44,100 --> 00:09:45,530
There's so much data.

233
00:09:45,530 --> 00:09:48,890
I don't even know
where to start.

234
00:09:48,890 --> 00:09:50,390
So I'm keen.

235
00:09:50,390 --> 00:09:52,830
I want to take advantage
of all the data.

236
00:09:52,830 --> 00:09:55,280
I want to be able to
analyze and look at it

237
00:09:55,280 --> 00:09:57,080
and get insights
that maybe I wasn't

238
00:09:57,080 --> 00:10:02,040
able to get before or get them
in a much more expedited manner.

239
00:10:02,040 --> 00:10:07,160
Do you have an example of an
organization or maybe someone

240
00:10:07,160 --> 00:10:09,950
that you work with
that's doing this well?

241
00:10:09,950 --> 00:10:12,990
And how do they start,
or what are they doing?

242
00:10:12,990 --> 00:10:16,280
Are there any
secrets but are not

243
00:10:16,280 --> 00:10:18,890
really secret that we
could share to just help

244
00:10:18,890 --> 00:10:22,370
those that are listening in who
want to move forward, know where

245
00:10:22,370 --> 00:10:23,220
to start?

246
00:10:23,220 --> 00:10:25,070
CHRISTIAN DIPPEL:
There's general, almost

247
00:10:25,070 --> 00:10:27,110
irony you could say, where--

248
00:10:27,110 --> 00:10:30,310
I always think, in strategy,
there's two buckets of strategy,

249
00:10:30,310 --> 00:10:32,310
or at least that's a frame
you can impose on it.

250
00:10:32,310 --> 00:10:34,070
You can say there's
quantitative decision

251
00:10:34,070 --> 00:10:37,980
making, which is operating
in data-rich environments.

252
00:10:37,980 --> 00:10:41,937
So a typical example of that
would be pricing decisions.

253
00:10:41,937 --> 00:10:43,520
For the most part,
if you're operating

254
00:10:43,520 --> 00:10:45,480
in a market that's
well established,

255
00:10:45,480 --> 00:10:48,630
you can make very
data-informed decisions

256
00:10:48,630 --> 00:10:49,930
about pricing decisions.

257
00:10:49,930 --> 00:10:52,870
You can call that strategy,
but it's very quantitative,

258
00:10:52,870 --> 00:10:54,190
very data-driven strategy.

259
00:10:54,190 --> 00:10:58,360
And then there's very
imagination-driven strategy,

260
00:10:58,360 --> 00:11:01,410
which is the strategy you
apply to settings for markets

261
00:11:01,410 --> 00:11:03,390
that don't exist yet.

262
00:11:03,390 --> 00:11:05,790
Who knows what the
market for space travel

263
00:11:05,790 --> 00:11:07,510
is going to look
like in 10 years?

264
00:11:07,510 --> 00:11:08,710
We can't know this.

265
00:11:08,710 --> 00:11:09,850
We can only imagine it.

266
00:11:09,850 --> 00:11:13,080
And I think, one of the
tensions that people

267
00:11:13,080 --> 00:11:14,970
seem to have when
it comes to AI is--

268
00:11:14,970 --> 00:11:16,630
AI is all about the data.

269
00:11:16,630 --> 00:11:19,230
It's all about ingesting
and transforming the data.

270
00:11:19,230 --> 00:11:21,090
Yet when we think
about, what are

271
00:11:21,090 --> 00:11:23,940
going to be the repercussions
of AI for my business

272
00:11:23,940 --> 00:11:27,900
or in my industry-- it's really
a very qualitative exercise

273
00:11:27,900 --> 00:11:30,870
of imagining the future
that is not-- inherently,

274
00:11:30,870 --> 00:11:33,550
because it is the future,
can't be very data-informed.

275
00:11:33,550 --> 00:11:35,225
And because you can't
really interpolate

276
00:11:35,225 --> 00:11:37,600
from what's been going on in
the last two or three years,

277
00:11:37,600 --> 00:11:39,070
there's just not enough data.

278
00:11:39,070 --> 00:11:40,360
It's not established enough.

279
00:11:40,360 --> 00:11:41,410
It's too exponential.

280
00:11:41,410 --> 00:11:44,550
The process is too to
accelerating to really use

281
00:11:44,550 --> 00:11:47,415
data to extrapolate into what's
that going to look like in two

282
00:11:47,415 --> 00:11:48,040
or three years.

283
00:11:48,040 --> 00:11:49,870
So really, you just
need to think about it.

284
00:11:49,870 --> 00:11:51,100
HOST: Yeah, for sure.

285
00:11:51,100 --> 00:11:54,390
You mentioned earlier around
frameworks, which sometimes

286
00:11:54,390 --> 00:11:57,070
can be a challenge
if we over rely,

287
00:11:57,070 --> 00:12:02,080
also can be an enabler as a
way to get things structured.

288
00:12:02,080 --> 00:12:06,180
So let's play out
strategy and maybe

289
00:12:06,180 --> 00:12:08,670
sort of bold or future thinking.

290
00:12:08,670 --> 00:12:12,180
Frameworks that either you've
used or approaches that you've

291
00:12:12,180 --> 00:12:16,320
seen that help leaders
process their thinking

292
00:12:16,320 --> 00:12:19,230
or start to get things
organized so that they

293
00:12:19,230 --> 00:12:23,400
can make better decisions,
they can lean into the data

294
00:12:23,400 --> 00:12:25,330
to create insights--

295
00:12:25,330 --> 00:12:28,240
CHRISTIAN DIPPEL: So strategy
is heavy on frameworks

296
00:12:28,240 --> 00:12:30,010
and I think for good reason.

297
00:12:30,010 --> 00:12:32,370
I think frameworks are
always a little bit

298
00:12:32,370 --> 00:12:34,410
of a double edged
sword in the sense

299
00:12:34,410 --> 00:12:36,880
that a framework
offers you a playbook.

300
00:12:36,880 --> 00:12:39,340
And I think at the
end of the day,

301
00:12:39,340 --> 00:12:42,010
you need to think deeply
about your situation.

302
00:12:42,010 --> 00:12:44,040
And I think playbooks
or frameworks

303
00:12:44,040 --> 00:12:47,800
can help you get there
quicker, or they can help you--

304
00:12:47,800 --> 00:12:49,440
It's almost like,
you know, you're

305
00:12:49,440 --> 00:12:52,260
trying on a Phillips screwdriver
and a flathead screwdriver,

306
00:12:52,260 --> 00:12:54,130
and you see which
one fits the screw.

307
00:12:54,130 --> 00:12:56,280
So you have tools
at your disposal

308
00:12:56,280 --> 00:12:58,030
that helped you
solve the problem.

309
00:12:58,030 --> 00:13:01,660
Where I caution always
when I work with clients,

310
00:13:01,660 --> 00:13:03,360
when I work with
students in exec ed,

311
00:13:03,360 --> 00:13:06,510
is it also satisfies
a little bit our bias

312
00:13:06,510 --> 00:13:09,210
towards not having to think
deeply and just having

313
00:13:09,210 --> 00:13:10,510
a playbook to play.

314
00:13:10,510 --> 00:13:13,980
If someone just hands me my five
forces canvas or my business

315
00:13:13,980 --> 00:13:16,270
model canvas and I
can fill in the thing,

316
00:13:16,270 --> 00:13:18,340
I can feel like I've
achieved something.

317
00:13:18,340 --> 00:13:21,750
But I think ultimately, a person
that doesn't have those tools

318
00:13:21,750 --> 00:13:25,020
but is really willing to sit
there still and think deeply

319
00:13:25,020 --> 00:13:26,820
about their situation,
they will come up

320
00:13:26,820 --> 00:13:28,770
with better solutions
than the person that

321
00:13:28,770 --> 00:13:31,810
is just superficially filling
in their business model canvas.

322
00:13:31,810 --> 00:13:34,500
And that's not a criticism
on the business model canvas.

323
00:13:34,500 --> 00:13:37,920
It's human nature that I think
it can be a little bit offering

324
00:13:37,920 --> 00:13:40,800
us sort of this like
fake ticking boxes sense

325
00:13:40,800 --> 00:13:42,580
of getting a handle
on a situation.

326
00:13:42,580 --> 00:13:44,050
So I love frameworks.

327
00:13:44,050 --> 00:13:45,130
They're very useful.

328
00:13:45,130 --> 00:13:49,840
But I always caution students or
in consulting work with clients.

329
00:13:49,840 --> 00:13:51,040
They're a means to an end.

330
00:13:51,040 --> 00:13:52,450
They're not the end in itself.

331
00:13:52,450 --> 00:13:55,750
HOST: I like your analogy
around the screwdrivers.

332
00:13:55,750 --> 00:14:00,450
You know, we sometimes
confuse motion with progress.

333
00:14:00,450 --> 00:14:05,010
So someone can actually be
moving but not necessarily

334
00:14:05,010 --> 00:14:05,980
advancing.

335
00:14:05,980 --> 00:14:07,300
But it satisfies this.

336
00:14:07,300 --> 00:14:07,900
Oh, I'm busy.

337
00:14:07,900 --> 00:14:09,490
I'm doing things, right?

338
00:14:09,490 --> 00:14:10,990
I'm making stuff happen.

339
00:14:10,990 --> 00:14:14,920
How would you help someone
sort of shift between the two?

340
00:14:14,920 --> 00:14:17,800
So are there any
cues that I can say,

341
00:14:17,800 --> 00:14:21,090
like maybe I'm not doing
enough of the hard work,

342
00:14:21,090 --> 00:14:23,500
and I'm sort of
relying too much on.

343
00:14:23,500 --> 00:14:25,240
OK, I filled out the sheet.

344
00:14:25,240 --> 00:14:28,513
I've done work, pat
myself on the back.

345
00:14:28,513 --> 00:14:30,180
What are some things
I should be holding

346
00:14:30,180 --> 00:14:33,100
myself accountable to maybe
or paying attention to?

347
00:14:33,100 --> 00:14:34,440
CHRISTIAN DIPPEL: If you look
back in history, and that's

348
00:14:34,440 --> 00:14:36,232
always like when you
look back at the past,

349
00:14:36,232 --> 00:14:39,870
things appear very obvious, but
they're not obvious at the time

350
00:14:39,870 --> 00:14:43,810
or for us looking forward for
a new technological disruption.

351
00:14:43,810 --> 00:14:45,820
But with past
technological disruptions,

352
00:14:45,820 --> 00:14:49,090
you often have these first order
effects that are quite obvious.

353
00:14:49,090 --> 00:14:52,410
The radio comes around, and it
totally disrupts the newspaper

354
00:14:52,410 --> 00:14:53,230
industry.

355
00:14:53,230 --> 00:14:56,880
TV comes around, and it
completely disrupts the radio

356
00:14:56,880 --> 00:14:58,000
station industry.

357
00:14:58,000 --> 00:15:00,810
The railway comes
around, and it kills off

358
00:15:00,810 --> 00:15:02,950
the Pony Express in
the American West,

359
00:15:02,950 --> 00:15:04,860
and it completely
destroys anyone

360
00:15:04,860 --> 00:15:07,860
who is invested in canal
building, which 10 years prior,

361
00:15:07,860 --> 00:15:09,180
was incredibly profitable.

362
00:15:09,180 --> 00:15:13,150
Because these are sort of first
order obvious replacements.

363
00:15:13,150 --> 00:15:16,650
And many in what some
management people

364
00:15:16,650 --> 00:15:18,220
like to call fast history.

365
00:15:18,220 --> 00:15:21,030
So recent technological
history, whether that's

366
00:15:21,030 --> 00:15:23,260
in microchips or
PCs or whatever.

367
00:15:23,260 --> 00:15:26,320
You get these examples like
Netflix killing Blockbuster.

368
00:15:26,320 --> 00:15:27,990
And they clearly,
Blockbuster just

369
00:15:27,990 --> 00:15:30,360
didn't think about
the first order effect

370
00:15:30,360 --> 00:15:32,470
of home delivery of DVDs.

371
00:15:32,470 --> 00:15:34,530
And then subsequent
to that, the long game

372
00:15:34,530 --> 00:15:36,085
was always going
to be streaming.

373
00:15:36,085 --> 00:15:37,210
They didn't think about it.

374
00:15:37,210 --> 00:15:38,170
They got killed by it.

375
00:15:38,170 --> 00:15:40,420
Kodak getting
killed by digicams.

376
00:15:40,420 --> 00:15:42,120
But then oftentimes,
when you think

377
00:15:42,120 --> 00:15:45,750
about the really fundamental
technological changes that

378
00:15:45,750 --> 00:15:49,290
really changed the way we
live, the second order effects

379
00:15:49,290 --> 00:15:51,210
are actually much
more interesting

380
00:15:51,210 --> 00:15:53,102
or, well, depending--
if you're Kodak,

381
00:15:53,102 --> 00:15:54,310
they're not more interesting.

382
00:15:54,310 --> 00:15:55,510
Depending on what
business you run,

383
00:15:55,510 --> 00:15:57,460
you might be more affected
by the first order effect.

384
00:15:57,460 --> 00:15:59,400
But what I'm trying to get
at is you might be operating

385
00:15:59,400 --> 00:16:02,350
in a business where you say, AI
doesn't affect me all that much.

386
00:16:02,350 --> 00:16:04,900
But when you think through
the second order effects,

387
00:16:04,900 --> 00:16:07,620
think about something like
the tractor coming around

388
00:16:07,620 --> 00:16:11,080
in the 1920s completely
changed agriculture,

389
00:16:11,080 --> 00:16:14,257
completely changed employment
opportunities in agriculture.

390
00:16:14,257 --> 00:16:15,840
Those are like the
first order effects

391
00:16:15,840 --> 00:16:17,410
that people could
see right away.

392
00:16:17,410 --> 00:16:19,260
But then within 10
years, what happened

393
00:16:19,260 --> 00:16:22,110
is you got massive
rural urban migration,

394
00:16:22,110 --> 00:16:25,420
and you got complete changing
of the urban landscape,

395
00:16:25,420 --> 00:16:28,150
the rise of superstar
cities, urban manufacturing,

396
00:16:28,150 --> 00:16:30,960
the rise of the service sector,
and just this like infinite

397
00:16:30,960 --> 00:16:33,498
plethora of new
opportunities coming around.

398
00:16:33,498 --> 00:16:35,040
And that's the part
where I sometimes

399
00:16:35,040 --> 00:16:37,110
think, because I
was fortunate enough

400
00:16:37,110 --> 00:16:39,820
to have some training as
an economic historian,

401
00:16:39,820 --> 00:16:41,770
you have a bit of this
longer run lens of,

402
00:16:41,770 --> 00:16:44,970
OK, let's try to really think
through what this thing is going

403
00:16:44,970 --> 00:16:47,020
to do in five
years, in 10 years,

404
00:16:47,020 --> 00:16:49,480
and not just get hung
up on the difference of

405
00:16:49,480 --> 00:16:52,050
does my competitor
have a chatbot today,

406
00:16:52,050 --> 00:16:54,240
and I don't have a
chatbot, and what

407
00:16:54,240 --> 00:16:56,880
that's going to do to my
perception of my brand

408
00:16:56,880 --> 00:16:58,040
in the marketplace today?

409
00:16:58,040 --> 00:17:08,130


410
00:17:08,130 --> 00:17:11,260
HOST: So let's build on
that last point, right?

411
00:17:11,260 --> 00:17:14,210
And I think most leaders
endeavor to learn from the past.

412
00:17:14,210 --> 00:17:14,710
Right.

413
00:17:14,710 --> 00:17:17,579
We can't change it,
but we can certainly

414
00:17:17,579 --> 00:17:18,940
glean insights from it.

415
00:17:18,940 --> 00:17:22,380
So how do you challenge
and get someone

416
00:17:22,380 --> 00:17:27,132
to think five and 10 years
when we're talking about AI?

417
00:17:27,132 --> 00:17:28,590
And I'm thinking,
jeez, something's

418
00:17:28,590 --> 00:17:30,640
going to change in
five or 10 days.

419
00:17:30,640 --> 00:17:32,820
And that idea of,
oh, I don't have

420
00:17:32,820 --> 00:17:35,920
a chatbot that is on my site.

421
00:17:35,920 --> 00:17:37,500
And if someone goes
to my site, I'm

422
00:17:37,500 --> 00:17:39,690
going to look like
I'm behind the times.

423
00:17:39,690 --> 00:17:41,940
So how do you
toggle between what

424
00:17:41,940 --> 00:17:45,460
do I need to do right
now while anticipating?

425
00:17:45,460 --> 00:17:48,370
Because if I don't think
five years plus out,

426
00:17:48,370 --> 00:17:51,090
I might be in for a much
bigger surprise than not having

427
00:17:51,090 --> 00:17:52,202
a chatbot today.

428
00:17:52,202 --> 00:17:53,910
CHRISTIAN DIPPEL:
There's a lot of myopia

429
00:17:53,910 --> 00:17:57,310
in how humans think in general.

430
00:17:57,310 --> 00:17:59,460
I think the value
of trying to bring

431
00:17:59,460 --> 00:18:02,460
a little bit of economic
history into whether that's

432
00:18:02,460 --> 00:18:05,240
a one on one consultation or
whether that's a classroom

433
00:18:05,240 --> 00:18:05,740
setting.

434
00:18:05,740 --> 00:18:07,907
Bringing a little bit of
that historical perspective

435
00:18:07,907 --> 00:18:09,970
can allow you to do a
little bit of back casting.

436
00:18:09,970 --> 00:18:13,170
You can say, OK, what is
it that people didn't see

437
00:18:13,170 --> 00:18:18,030
in the '60s or in the '70s about
what the arrival of all sorts

438
00:18:18,030 --> 00:18:21,660
of home appliances would
do to female labor force

439
00:18:21,660 --> 00:18:24,730
participation, and then female
educational attainments,

440
00:18:24,730 --> 00:18:27,040
and then changing
gender roles 10, 20,

441
00:18:27,040 --> 00:18:30,070
30 years out from those
technological innovations?

442
00:18:30,070 --> 00:18:32,340
What are the kinds of
biases that are making

443
00:18:32,340 --> 00:18:34,450
it hard to see out like that?

444
00:18:34,450 --> 00:18:37,380
Really trying to, I think,
lean into just sitting

445
00:18:37,380 --> 00:18:41,020
back and imagining what is my
industry going to look like,

446
00:18:41,020 --> 00:18:43,200
either in free form or
doing that through something

447
00:18:43,200 --> 00:18:44,877
like Porter's Five
Forces framework.

448
00:18:44,877 --> 00:18:46,210
I think that's really important.

449
00:18:46,210 --> 00:18:48,180
HOST: So how do you
feel about using

450
00:18:48,180 --> 00:18:53,260
AI to cross-check your own
human-based deep thinking,

451
00:18:53,260 --> 00:18:53,760
right?

452
00:18:53,760 --> 00:18:57,610
So you've sort of challenged
to say, carve space.

453
00:18:57,610 --> 00:18:59,050
Push yourself.

454
00:18:59,050 --> 00:19:01,110
Exercise that what's
going to happen

455
00:19:01,110 --> 00:19:04,660
to my industry in my space
five years and 10 years?

456
00:19:04,660 --> 00:19:06,700
How can AI help me do that?

457
00:19:06,700 --> 00:19:08,500
So I'm still going
to do the hard work,

458
00:19:08,500 --> 00:19:12,720
but maybe it can fact check me
or cross-reference me or keep me

459
00:19:12,720 --> 00:19:13,780
on track.

460
00:19:13,780 --> 00:19:15,780
CHRISTIAN DIPPEL: AI means
many different things

461
00:19:15,780 --> 00:19:19,390
to different people, different
strategic business units.

462
00:19:19,390 --> 00:19:21,300
My own perspective
and a little bit

463
00:19:21,300 --> 00:19:23,730
that's shaped our
conversations so far

464
00:19:23,730 --> 00:19:26,280
is when you're operating
in a kind of business that

465
00:19:26,280 --> 00:19:30,090
is more data-rich, there are
a lot of operational things

466
00:19:30,090 --> 00:19:31,690
that you can do with AI.

467
00:19:31,690 --> 00:19:34,380
One thing that I'm finding in
just some work with clients

468
00:19:34,380 --> 00:19:37,470
is there's a lot of small
and medium-sized businesses

469
00:19:37,470 --> 00:19:39,360
that have a ton of data.

470
00:19:39,360 --> 00:19:41,230
That could be
manufacturing data.

471
00:19:41,230 --> 00:19:43,680
It could be order
processing data, client

472
00:19:43,680 --> 00:19:45,940
facing data, all types of data.

473
00:19:45,940 --> 00:19:50,200
But up until now, it's
never been worth hiring,

474
00:19:50,200 --> 00:19:52,210
let's say, two data scientists.

475
00:19:52,210 --> 00:19:54,330
Like if you're not like
a data-driven enterprise

476
00:19:54,330 --> 00:19:56,860
per se, you just sort of
accidentally, you might say,

477
00:19:56,860 --> 00:19:59,410
or just as a side product
of your business operations,

478
00:19:59,410 --> 00:20:01,170
I think a lot of
enterprises are sort

479
00:20:01,170 --> 00:20:03,750
of moving into a world
where you've accumulated

480
00:20:03,750 --> 00:20:05,530
an incredible treasure
trove of data,

481
00:20:05,530 --> 00:20:07,330
have never really
done anything with it,

482
00:20:07,330 --> 00:20:10,750
and now, it's a matter
of can I build a back end

483
00:20:10,750 --> 00:20:13,620
maybe in a very
sandboxed way that

484
00:20:13,620 --> 00:20:16,710
doesn't take a lot of resources,
where I can really come up

485
00:20:16,710 --> 00:20:20,310
with new business offerings
or just really process

486
00:20:20,310 --> 00:20:22,410
optimization,
maybe knowing which

487
00:20:22,410 --> 00:20:25,150
clients are due a maintenance,
which clients are due,

488
00:20:25,150 --> 00:20:26,110
a new product,

489
00:20:26,110 --> 00:20:28,140
having an automated
email pipeline

490
00:20:28,140 --> 00:20:30,480
that automates part
of the outreach

491
00:20:30,480 --> 00:20:32,340
or standardizes some
of the procedures

492
00:20:32,340 --> 00:20:33,940
we haven't standardized
in the past.

493
00:20:33,940 --> 00:20:37,180
So that's the part where
even as a small business,

494
00:20:37,180 --> 00:20:40,023
you can use AI to
automate a lot of things.

495
00:20:40,023 --> 00:20:41,940
The thing that I think
you're alluding to more

496
00:20:41,940 --> 00:20:46,390
is, is almost like a slightly
more individualistic enterprise,

497
00:20:46,390 --> 00:20:49,440
which is you might just
sit there at your desk,

498
00:20:49,440 --> 00:20:50,910
and you have to make a decision.

499
00:20:50,910 --> 00:20:54,610
And you can go speak to your
colleagues at the water cooler,

500
00:20:54,610 --> 00:20:55,600
and that's useful.

501
00:20:55,600 --> 00:20:58,080
But maybe you're
worried that everyone

502
00:20:58,080 --> 00:21:00,660
is drinking the same
Kool-Aid to an extent when

503
00:21:00,660 --> 00:21:03,450
it comes to certain
strategic business decisions.

504
00:21:03,450 --> 00:21:04,950
Now you have the
ability, and that's

505
00:21:04,950 --> 00:21:07,350
a very individualistic
exercise that

506
00:21:07,350 --> 00:21:10,170
is almost something that I
think in a business culture can

507
00:21:10,170 --> 00:21:11,380
be encouraged.

508
00:21:11,380 --> 00:21:13,800
But early on, it's not
really actually that clear

509
00:21:13,800 --> 00:21:15,670
how to even standardize it.

510
00:21:15,670 --> 00:21:17,700
Encouraging could be
something as simple

511
00:21:17,700 --> 00:21:22,380
as the company will pay for
your $20 a month subscription

512
00:21:22,380 --> 00:21:25,230
to one of the models
to have a more premium

513
00:21:25,230 --> 00:21:28,530
access to the capabilities, like
deeper thinking, deeper research

514
00:21:28,530 --> 00:21:29,138
capabilities.

515
00:21:29,138 --> 00:21:30,930
That's not a lot of
money you need to spend

516
00:21:30,930 --> 00:21:32,860
to use those kind of tools.

517
00:21:32,860 --> 00:21:36,850
But what they allow you to do is
really check your own thinking,

518
00:21:36,850 --> 00:21:38,238
scrutinize your own thinking.

519
00:21:38,238 --> 00:21:40,030
And of course, any
individual can use that.

520
00:21:40,030 --> 00:21:42,520
You can use that to
prepare for job interviews.

521
00:21:42,520 --> 00:21:45,220
You can use that to prepare
for this particular interview.

522
00:21:45,220 --> 00:21:46,950
I didn't do it, but
I could have tried

523
00:21:46,950 --> 00:21:49,860
to characterize Brian
Benjamin to a chatbot

524
00:21:49,860 --> 00:21:52,480
and ask what kind of questions,
is he going to ask me?

525
00:21:52,480 --> 00:21:54,460
And so there's a lot of
capability to do that.

526
00:21:54,460 --> 00:21:57,940
And one area where I see that
a lot is in the classroom now.

527
00:21:57,940 --> 00:22:01,530
Just in the last, let's say,
12 months, the ability to just

528
00:22:01,530 --> 00:22:05,640
use really any LLM, I was
hesitant to just say ChatGPT.

529
00:22:05,640 --> 00:22:08,520
ChatGPT is becoming a little
bit like Xerox for copying.

530
00:22:08,520 --> 00:22:10,780
And getting a really
good strategic partner.

531
00:22:10,780 --> 00:22:14,250
So when I teach geopolitics, I
might throw a question out there

532
00:22:14,250 --> 00:22:16,650
like, OK, imagine
there's a scenario where

533
00:22:16,650 --> 00:22:19,690
we are an electric vehicle
producer in Ontario,

534
00:22:19,690 --> 00:22:23,010
and there's a civil war in the
Democratic Republic of Congo

535
00:22:23,010 --> 00:22:24,400
where all the cobalt comes from.

536
00:22:24,400 --> 00:22:26,850
Chances are zero people
in the audience will

537
00:22:26,850 --> 00:22:28,090
have deep knowledge on this.

538
00:22:28,090 --> 00:22:30,070
But you can do a
scenario analysis.

539
00:22:30,070 --> 00:22:32,040
You can describe
the thing to ChatGPT

540
00:22:32,040 --> 00:22:34,587
and do a scenario analysis
within five minutes.

541
00:22:34,587 --> 00:22:36,420
And within five minutes,
we can have a group

542
00:22:36,420 --> 00:22:38,280
strategic discussion.

543
00:22:38,280 --> 00:22:41,910
So we get elevated in
our industry knowledge

544
00:22:41,910 --> 00:22:45,180
and in our sort of high level
strategic thinking and an issue.

545
00:22:45,180 --> 00:22:47,770
Within minutes, you can
get elevated so much.

546
00:22:47,770 --> 00:22:50,860
And then of course, that
can't be all and say all.

547
00:22:50,860 --> 00:22:54,910
There needs to be additional
guardrails around the process,

548
00:22:54,910 --> 00:22:57,660
and there's all sorts of things
you need to guard against and be

549
00:22:57,660 --> 00:22:58,300
worried about.

550
00:22:58,300 --> 00:23:01,990
But as a quick way of just
getting deep into a topic,

551
00:23:01,990 --> 00:23:04,360
scrutinizing your own
thinking, it's just incredible.

552
00:23:04,360 --> 00:23:06,870
Predictive nature can
be an interesting tool

553
00:23:06,870 --> 00:23:10,450
to grab from the
past and context.

554
00:23:10,450 --> 00:23:11,800
And so forth.

555
00:23:11,800 --> 00:23:17,370
How can AI be used to help
me imagine 10 years out

556
00:23:17,370 --> 00:23:20,790
if I am trying to see
where my industry is going

557
00:23:20,790 --> 00:23:21,970
or what's going to happen?

558
00:23:21,970 --> 00:23:24,183
Is it a tool I can use
in that kind of capacity?

559
00:23:24,183 --> 00:23:26,350
CHRISTIAN DIPPEL: You often
hear this one liner when

560
00:23:26,350 --> 00:23:28,725
it comes to AI, and I think
there's a lot of truth in it,

561
00:23:28,725 --> 00:23:32,000
that it really raises the
value of good questions.

562
00:23:32,000 --> 00:23:34,190
In other words, your
job as a leader,

563
00:23:34,190 --> 00:23:36,340
whether it's at
the c-suite level

564
00:23:36,340 --> 00:23:40,550
or leading a strategic business
unit or whatever the case,

565
00:23:40,550 --> 00:23:43,690
I think it's going to become
more about asking good questions

566
00:23:43,690 --> 00:23:47,810
and less so about being the
person to provide the answers.

567
00:23:47,810 --> 00:23:50,540
And so there's a lot of
debates in the AI space.

568
00:23:50,540 --> 00:23:52,690
Like with my companies,
I interface more

569
00:23:52,690 --> 00:23:54,220
with the sort of
technical elements

570
00:23:54,220 --> 00:23:56,860
of prompt engineering
and automating

571
00:23:56,860 --> 00:24:01,030
prompt engineering through the
various LLMs, APIs, et cetera.

572
00:24:01,030 --> 00:24:03,790
So asking good
questions is becoming

573
00:24:03,790 --> 00:24:04,850
more and more important.

574
00:24:04,850 --> 00:24:07,730
Some of those questions
will be automated questions.

575
00:24:07,730 --> 00:24:10,840
So if you just have a data
feed that operationally matters

576
00:24:10,840 --> 00:24:13,450
for your company, you want
to have an automated pipeline

577
00:24:13,450 --> 00:24:15,850
in the back end that
queries the data feed

578
00:24:15,850 --> 00:24:17,990
and generates the
answer from the data,

579
00:24:17,990 --> 00:24:19,760
and it's all about
asking good questions.

580
00:24:19,760 --> 00:24:21,427
That's sort of like
the engineering side

581
00:24:21,427 --> 00:24:22,490
of asking good questions.

582
00:24:22,490 --> 00:24:24,420
But I think at the
more qualitative level,

583
00:24:24,420 --> 00:24:26,170
if you're just a person
making a decision,

584
00:24:26,170 --> 00:24:27,670
you're sitting at
your desk, and you

585
00:24:27,670 --> 00:24:31,590
can use AI as a copilot in a
non-technical sense, so not

586
00:24:31,590 --> 00:24:34,810
as an explicit copilot but just
as someone to run ideas by.

587
00:24:34,810 --> 00:24:37,020
So much of the way
we ask questions

588
00:24:37,020 --> 00:24:38,740
is informed by our context.

589
00:24:38,740 --> 00:24:40,330
And I'll give you one example.

590
00:24:40,330 --> 00:24:43,410
When I teach an executive
education class in Canada,

591
00:24:43,410 --> 00:24:46,350
and we talk about how is
the labor market going

592
00:24:46,350 --> 00:24:48,400
to be impacted by AI?

593
00:24:48,400 --> 00:24:50,850
One thing that you often
get is a lot of people

594
00:24:50,850 --> 00:24:53,370
assume that one
area that is going

595
00:24:53,370 --> 00:24:55,650
to be relatively
unaffected by AI

596
00:24:55,650 --> 00:24:58,740
is old age care because
that is something

597
00:24:58,740 --> 00:25:00,730
that requires the human touch.

598
00:25:00,730 --> 00:25:03,005
And that might be entirely true.

599
00:25:03,005 --> 00:25:05,130
But what I found really
interesting is the contrast

600
00:25:05,130 --> 00:25:08,170
to if you talk to people in East
Asia, let's say in Hong Kong,

601
00:25:08,170 --> 00:25:11,970
about AI, that is their
number one use case for AI,

602
00:25:11,970 --> 00:25:13,210
is old age care.

603
00:25:13,210 --> 00:25:16,050
Because the demographics are
so fundamentally different

604
00:25:16,050 --> 00:25:17,620
in those two geographies.

605
00:25:17,620 --> 00:25:19,930
We have relatively
young demographics.

606
00:25:19,930 --> 00:25:22,390
Of course, we talk about
aging and things like that.

607
00:25:22,390 --> 00:25:25,840
But it's not a very pressing
concern in a macroeconomic sense

608
00:25:25,840 --> 00:25:28,510
when you look at our population
trees in North America.

609
00:25:28,510 --> 00:25:30,210
When you look at
population trees

610
00:25:30,210 --> 00:25:34,240
in Hong Kong, Japan,
Singapore, oh, my gosh,

611
00:25:34,240 --> 00:25:35,890
they look very
different from ours.

612
00:25:35,890 --> 00:25:40,050
And old age care is viewed as
the number one-- well, maybe not

613
00:25:40,050 --> 00:25:42,030
the number one, but
one of the number one

614
00:25:42,030 --> 00:25:45,600
use cases where I can actually
help in the form of literally

615
00:25:45,600 --> 00:25:46,990
AI-informed robotics.

616
00:25:46,990 --> 00:25:50,790
And so it's just striking
how people's predictions

617
00:25:50,790 --> 00:25:53,490
about future use
cases are so socially

618
00:25:53,490 --> 00:25:55,060
informed by the context.

619
00:25:55,060 --> 00:25:57,940
And I think AI can help you
question your own assumptions.

620
00:25:57,940 --> 00:25:59,620
In when you're asking
those questions.

621
00:25:59,620 --> 00:26:00,480
HOST: You're right.

622
00:26:00,480 --> 00:26:03,280
We kind of naturally
go to what we know.

623
00:26:03,280 --> 00:26:04,920
And one would hope
that we can learn

624
00:26:04,920 --> 00:26:08,400
from each other because as
things shift in North America

625
00:26:08,400 --> 00:26:12,378
and aging does become an even
bigger sort of prominent piece,

626
00:26:12,378 --> 00:26:14,670
what have we learned from
other parts of the world that

627
00:26:14,670 --> 00:26:16,270
have already been there first?

628
00:26:16,270 --> 00:26:21,180
I wrote down, you're
asking good questions,

629
00:26:21,180 --> 00:26:23,500
and we talk about
prompt engineers.

630
00:26:23,500 --> 00:26:26,460
I'm fortunate to work with
a lot of exceptional coaches

631
00:26:26,460 --> 00:26:27,730
here at Ivey.

632
00:26:27,730 --> 00:26:30,210
So I've had the
pleasure of being

633
00:26:30,210 --> 00:26:32,040
asked really good questions.

634
00:26:32,040 --> 00:26:37,780
Very specific, often open-ended,
leaving space for conversation.

635
00:26:37,780 --> 00:26:40,720
And so I think that your
comment and maybe the art

636
00:26:40,720 --> 00:26:44,820
of good questions is going to
be an even more prominent skill

637
00:26:44,820 --> 00:26:48,540
that leaders are going to
need to hone and to build

638
00:26:48,540 --> 00:26:50,110
and to refine.

639
00:26:50,110 --> 00:26:53,250
So I'm going to go back a
little bit to the conversation

640
00:26:53,250 --> 00:26:56,860
around sort of executives and
maybe even park executives.

641
00:26:56,860 --> 00:27:00,060
I think leaders at multiple
levels in the organization

642
00:27:00,060 --> 00:27:02,530
are involved in
strategic planning.

643
00:27:02,530 --> 00:27:03,160
So yes.

644
00:27:03,160 --> 00:27:05,050
It can be organizational
strategic planning,

645
00:27:05,050 --> 00:27:08,650
but it also be departmental or
even team strategic planning.

646
00:27:08,650 --> 00:27:11,850
So when we hear about AI
initiatives, most strategic

647
00:27:11,850 --> 00:27:14,670
planning exercises,
we got to leverage AI

648
00:27:14,670 --> 00:27:16,540
to some capacity or another.

649
00:27:16,540 --> 00:27:21,250
So let's toggle between is this
sort of a process innovation?

650
00:27:21,250 --> 00:27:24,790
So can AI save me time?

651
00:27:24,790 --> 00:27:26,170
Can it save me money?

652
00:27:26,170 --> 00:27:28,800
Can it create speed
and efficiency

653
00:27:28,800 --> 00:27:33,030
with you sort of reshaping
maybe our business model

654
00:27:33,030 --> 00:27:35,650
or our team model altogether?

655
00:27:35,650 --> 00:27:38,080
CHRISTIAN DIPPEL: I used to
have a little bit this idea,

656
00:27:38,080 --> 00:27:40,210
I think, early on in
thinking about AI,

657
00:27:40,210 --> 00:27:42,720
that product
innovation is somehow

658
00:27:42,720 --> 00:27:44,770
better than process innovation.

659
00:27:44,770 --> 00:27:47,500
Process innovation, it
feels a little boring.

660
00:27:47,500 --> 00:27:50,920
It's sort of like you're
wringing the towel dry,

661
00:27:50,920 --> 00:27:52,840
like getting the
efficiencies you can get,

662
00:27:52,840 --> 00:27:55,390
but product innovation is
really the exciting part.

663
00:27:55,390 --> 00:27:57,550
But I've really changed
my thinking on that.

664
00:27:57,550 --> 00:28:00,510
And actually, some of our
colleagues in Hong Kong

665
00:28:00,510 --> 00:28:03,070
had a helping hand
in a discussion.

666
00:28:03,070 --> 00:28:05,070
I remember well where
the framework that

667
00:28:05,070 --> 00:28:06,840
was applied in that
conversation was more

668
00:28:06,840 --> 00:28:10,830
about where is AI a
homogenizer, and where

669
00:28:10,830 --> 00:28:12,220
is it a differentiator?

670
00:28:12,220 --> 00:28:17,040
And the chatbot is a great
example of a product innovation

671
00:28:17,040 --> 00:28:19,600
that is not going to
be a differentiator.

672
00:28:19,600 --> 00:28:25,420
Chatbot technology is a very
commoditized type of technology.

673
00:28:25,420 --> 00:28:27,340
I've been involved
in building chatbots.

674
00:28:27,340 --> 00:28:29,200
It's a very
standardized product.

675
00:28:29,200 --> 00:28:31,630
And yes, there's better
ones and less good ones.

676
00:28:31,630 --> 00:28:34,030
And yes, we've seen
with Air Canada,

677
00:28:34,030 --> 00:28:36,240
we've seen examples
of it horribly

678
00:28:36,240 --> 00:28:38,460
backfiring because
they didn't execute

679
00:28:38,460 --> 00:28:39,952
very well on their chatbot.

680
00:28:39,952 --> 00:28:42,160
But at the end of the day,
it's not a differentiator.

681
00:28:42,160 --> 00:28:45,730
No business will stand
out for having a chatbot.

682
00:28:45,730 --> 00:28:48,640
Some businesses might be
worried that in the short run,

683
00:28:48,640 --> 00:28:51,190
they might stand out
negatively for not having one.

684
00:28:51,190 --> 00:28:53,890
I think those are
relatively minor concerns.

685
00:28:53,890 --> 00:28:56,550
At the same time, sometimes,
the process innovation

686
00:28:56,550 --> 00:29:00,640
is your real differentiator
and your real business model.

687
00:29:00,640 --> 00:29:04,360
So when you think about famous
examples like IKEA, Walmart,

688
00:29:04,360 --> 00:29:09,160
to a lesser extent, the network
dynamics of Netflix or Amazon,

689
00:29:09,160 --> 00:29:11,760
those are really all businesses
where the business model

690
00:29:11,760 --> 00:29:14,550
is on the process innovation
more than the product

691
00:29:14,550 --> 00:29:15,400
innovation.

692
00:29:15,400 --> 00:29:19,540
And so that is really an area
where my thinking has changed

693
00:29:19,540 --> 00:29:22,390
from thinking product
innovation somehow

694
00:29:22,390 --> 00:29:25,930
better than process innovation,
to really being agnostic

695
00:29:25,930 --> 00:29:28,540
about the two and
thinking about where is

696
00:29:28,540 --> 00:29:30,350
the true competitive advantage.

697
00:29:30,350 --> 00:29:33,310
Is it plausible for us to build
a product that really sets us

698
00:29:33,310 --> 00:29:35,170
apart, or is it maybe
the case that there's

699
00:29:35,170 --> 00:29:36,880
something in our
processes that is

700
00:29:36,880 --> 00:29:39,640
going to really be able to
leverage AI in a way that

701
00:29:39,640 --> 00:29:40,730
sets us apart?

702
00:29:40,730 --> 00:29:44,020
HOST: We're often looking for
the bigger splash or the bigger

703
00:29:44,020 --> 00:29:46,060
bold innovation, and
maybe it's actually

704
00:29:46,060 --> 00:29:48,070
something a little
closer to home

705
00:29:48,070 --> 00:29:50,090
that we need to
pay attention to.

706
00:29:50,090 --> 00:29:52,690
I want to dig in
because you're involved

707
00:29:52,690 --> 00:29:56,870
in a lot of really cool ventures
and sort of consulting work.

708
00:29:56,870 --> 00:30:03,520
A couple of live examples of
where AI is being strategically

709
00:30:03,520 --> 00:30:04,160
applied.

710
00:30:04,160 --> 00:30:06,110
So it's not hype.

711
00:30:06,110 --> 00:30:08,410
So don't tell me
about the ChatGPT bot

712
00:30:08,410 --> 00:30:11,240
that Company A implemented.

713
00:30:11,240 --> 00:30:15,800
Something that be
interesting for our listeners

714
00:30:15,800 --> 00:30:18,690
to hear that you've been
involved with firsthand.

715
00:30:18,690 --> 00:30:20,690
CHRISTIAN DIPPEL: One
example that comes to mind

716
00:30:20,690 --> 00:30:23,270
is in one of my own
businesses, which is really

717
00:30:23,270 --> 00:30:25,940
a newsletter business, that
takes an enormous amount

718
00:30:25,940 --> 00:30:27,480
of government text.

719
00:30:27,480 --> 00:30:31,400
And we knew with that
idea that no one really

720
00:30:31,400 --> 00:30:34,040
wants to have a newsletter
that is everything

721
00:30:34,040 --> 00:30:35,637
that is going on in Ottawa.

722
00:30:35,637 --> 00:30:37,220
No one really needs
to know everything

723
00:30:37,220 --> 00:30:39,000
that is going on in Ottawa.

724
00:30:39,000 --> 00:30:41,070
The value really
comes from saying,

725
00:30:41,070 --> 00:30:42,770
if you're in oil
and gas, you want

726
00:30:42,770 --> 00:30:44,480
to have a newsletter
that says everything

727
00:30:44,480 --> 00:30:47,720
that's going on in Ottawa
that affects the oil and gas

728
00:30:47,720 --> 00:30:48,360
industry.

729
00:30:48,360 --> 00:30:50,712
If you're in utilities, you
want a newsletter for that.

730
00:30:50,712 --> 00:30:52,920
If you're in telecoms, you
want a newsletter in that.

731
00:30:52,920 --> 00:30:55,128
If you're in finance, you
want a newsletter for that.

732
00:30:55,128 --> 00:30:58,580
And so we had this idea, but we
had this forest for the trees

733
00:30:58,580 --> 00:31:00,770
problem, that it was,
well, government just

734
00:31:00,770 --> 00:31:04,760
generates too much text, like
the editorial team that it would

735
00:31:04,760 --> 00:31:09,410
require to like, parse all those
things into 20 different topics

736
00:31:09,410 --> 00:31:10,530
is just enormous.

737
00:31:10,530 --> 00:31:15,310
And it's a good actually example
of this process versus product

738
00:31:15,310 --> 00:31:18,610
innovation, and how fuzzy that
line sometimes can be with AI

739
00:31:18,610 --> 00:31:21,320
is we have a very,
very simple AI stack.

740
00:31:21,320 --> 00:31:24,820
All it does is take all the
data that we're ingesting

741
00:31:24,820 --> 00:31:28,190
and funnel it into
20 plus funnels.

742
00:31:28,190 --> 00:31:29,900
The technology is not that hard.

743
00:31:29,900 --> 00:31:32,150
It's not a particularly
sophisticated model.

744
00:31:32,150 --> 00:31:34,160
It's using very off
the shelf tools.

745
00:31:34,160 --> 00:31:36,310
But it allows you
to go from 0 to 1

746
00:31:36,310 --> 00:31:39,560
in being able to not being
able to do it at all,

747
00:31:39,560 --> 00:31:42,680
to being able to do it in
an extremely automated way.

748
00:31:42,680 --> 00:31:45,280
And so now, we're in a place
where I sit there on a Sunday

749
00:31:45,280 --> 00:31:49,040
night, editing 24 newsletters in
the space of an hour and a half.

750
00:31:49,040 --> 00:31:51,190
Everything else is AI
generated, and I just

751
00:31:51,190 --> 00:31:53,690
look through them, do a little
bit of polish, and that's it.

752
00:31:53,690 --> 00:31:56,710
So your capability
as a small business

753
00:31:56,710 --> 00:31:58,832
to do an enormous
amount, obviously,

754
00:31:58,832 --> 00:32:00,290
depending on your
line of business,

755
00:32:00,290 --> 00:32:02,900
it's harder in hardware
than it is in software.

756
00:32:02,900 --> 00:32:05,710
In certain applications, you get
completely supercharged in what

757
00:32:05,710 --> 00:32:07,510
you can do, and a
lot of it can be

758
00:32:07,510 --> 00:32:09,100
much more about the
process innovation

759
00:32:09,100 --> 00:32:10,400
than the product innovation.

760
00:32:10,400 --> 00:32:11,950
HOST: Yeah, interesting.

761
00:32:11,950 --> 00:32:13,760
You think about an
hour and a half.

762
00:32:13,760 --> 00:32:16,490
What would it have taken
without AI support?

763
00:32:16,490 --> 00:32:17,870
How long would it have been?

764
00:32:17,870 --> 00:32:18,370
CHRISTIAN DIPPEL:
Well, it's almost

765
00:32:18,370 --> 00:32:20,350
like it's your process
is changed in a way

766
00:32:20,350 --> 00:32:22,372
where it would have been
straight up impossible

767
00:32:22,372 --> 00:32:24,830
because you would have had to
see the forest for the trees.

768
00:32:24,830 --> 00:32:26,620
But there were too many trees,
and you could have never

769
00:32:26,620 --> 00:32:27,710
done it to begin with.

770
00:32:27,710 --> 00:32:29,920
Then there is a maybe
intermediate solution

771
00:32:29,920 --> 00:32:32,440
that's like six years
ago technology where

772
00:32:32,440 --> 00:32:34,220
you define a set of keywords.

773
00:32:34,220 --> 00:32:37,730
You parse through all the text
based on buckets of keywords

774
00:32:37,730 --> 00:32:39,802
and to try to do an
initial funneling.

775
00:32:39,802 --> 00:32:41,260
But oh, my gosh,
you would have had

776
00:32:41,260 --> 00:32:44,300
a lot of work left to do
to then make that good.

777
00:32:44,300 --> 00:32:45,910
And then you need
to still summarize

778
00:32:45,910 --> 00:32:48,190
all the content, which four
or five years ago, would

779
00:32:48,190 --> 00:32:49,880
have been very hard to do.

780
00:32:49,880 --> 00:32:51,805
There would have been a
lot of hallucinations,

781
00:32:51,805 --> 00:32:53,930
the quality of the text
would have been horrendous.

782
00:32:53,930 --> 00:32:55,960
And so all of those
problems just go away.

783
00:32:55,960 --> 00:32:58,700
And you're left with just
polishing, editing work.

784
00:32:58,700 --> 00:32:59,200
HOST: Yeah.

785
00:32:59,200 --> 00:33:01,540
So saving time on something
that you might not

786
00:33:01,540 --> 00:33:05,110
have even been able to do
because of the complexity

787
00:33:05,110 --> 00:33:05,830
involved.

788
00:33:05,830 --> 00:33:09,850
So what's one single takeaway
that you would leave, something

789
00:33:09,850 --> 00:33:13,030
that I can do as
a leader, when I'm

790
00:33:13,030 --> 00:33:16,720
thinking about AI in
the context of creating

791
00:33:16,720 --> 00:33:18,290
competitive advantage?

792
00:33:18,290 --> 00:33:20,590
CHRISTIAN DIPPEL: I
think one big takeaway

793
00:33:20,590 --> 00:33:25,750
that I think a lot of
businesses, especially

794
00:33:25,750 --> 00:33:28,760
businesses that interface
with Ivey a lot,

795
00:33:28,760 --> 00:33:30,910
I think have
traditionally operated

796
00:33:30,910 --> 00:33:33,580
on a model where a
wait and see approach

797
00:33:33,580 --> 00:33:37,070
is often the right approach when
it comes to new technologies.

798
00:33:37,070 --> 00:33:39,040
And you know, in academia,
the fancy language

799
00:33:39,040 --> 00:33:41,900
for that is like the real
option value of waiting.

800
00:33:41,900 --> 00:33:43,910
And oftentimes, with
new technologies,

801
00:33:43,910 --> 00:33:46,870
the idea is, well, some
standardized gold standard

802
00:33:46,870 --> 00:33:48,890
is going to emerge.

803
00:33:48,890 --> 00:33:51,700
And maybe I need to wait for
two or three years for the dust

804
00:33:51,700 --> 00:33:52,490
to settle.

805
00:33:52,490 --> 00:33:54,160
But then there'll
be an off the shelf

806
00:33:54,160 --> 00:33:56,840
product that will do
what I need to do.

807
00:33:56,840 --> 00:33:58,250
And so it's better to wait.

808
00:33:58,250 --> 00:34:00,620
And especially in
the world of atoms,

809
00:34:00,620 --> 00:34:02,630
like in the physical
hardware world,

810
00:34:02,630 --> 00:34:04,390
that is often the case, right?

811
00:34:04,390 --> 00:34:06,250
We all know Moore's law.

812
00:34:06,250 --> 00:34:09,110
The cost of computing
decreases exponentially.

813
00:34:09,110 --> 00:34:11,469
Let's say things like
the cost of solar panels

814
00:34:11,469 --> 00:34:13,510
has also been
decreasing exponentially

815
00:34:13,510 --> 00:34:14,510
over the last 10 years.

816
00:34:14,510 --> 00:34:16,219
So lots of technologies.

817
00:34:16,219 --> 00:34:18,800
Absolutely, it makes
sense to wait and see.

818
00:34:18,800 --> 00:34:21,159
I think AI is really
different, and it's

819
00:34:21,159 --> 00:34:22,630
different for two reasons.

820
00:34:22,630 --> 00:34:25,047
I'm sure it's different for
more than two reasons, but two

821
00:34:25,047 --> 00:34:26,630
that are sort of
prominent in my head.

822
00:34:26,630 --> 00:34:28,820
One is it just
keeps accelerating.

823
00:34:28,820 --> 00:34:30,440
And so if you wait
and do nothing,

824
00:34:30,440 --> 00:34:33,850
the thing is going to accelerate
away from you in a way

825
00:34:33,850 --> 00:34:35,449
that you might
come to regret two

826
00:34:35,449 --> 00:34:36,730
or three years down the road.

827
00:34:36,730 --> 00:34:38,230
And then the other
thing is the cost

828
00:34:38,230 --> 00:34:40,659
of an option is just
so darn low compared

829
00:34:40,659 --> 00:34:42,560
to most other technologies.

830
00:34:42,560 --> 00:34:45,550
All it really takes is
start using the thing

831
00:34:45,550 --> 00:34:47,840
in a potentially
very low tech way,

832
00:34:47,840 --> 00:34:51,550
maybe designating two
or three critical people

833
00:34:51,550 --> 00:34:54,920
in your organizations to
say, explore some options.

834
00:34:54,920 --> 00:34:56,800
Free up one day a
week where you say

835
00:34:56,800 --> 00:35:00,400
today is like your moonshot
research day, where you just

836
00:35:00,400 --> 00:35:02,390
think about things
we can do with AI.

837
00:35:02,390 --> 00:35:06,070
So the barrier to
adoption is really so low

838
00:35:06,070 --> 00:35:07,540
that I think it
really makes sense

839
00:35:07,540 --> 00:35:11,380
to engage today and not wait,
even for organizations that

840
00:35:11,380 --> 00:35:14,470
have been well-served in
the past, with taking a wait

841
00:35:14,470 --> 00:35:17,883
and see approach when it comes
to adopting new technologies.

842
00:35:17,883 --> 00:35:19,300
And then the second
thing is going

843
00:35:19,300 --> 00:35:20,990
to be quicker because
I already said it.

844
00:35:20,990 --> 00:35:23,050
I think a lot of people
I find in conversations

845
00:35:23,050 --> 00:35:25,930
grapple with this tension,
that this idea that AI

846
00:35:25,930 --> 00:35:27,140
is all about data.

847
00:35:27,140 --> 00:35:29,500
But thinking about
what AI will do

848
00:35:29,500 --> 00:35:32,300
is all about
qualitative imagination.

849
00:35:32,300 --> 00:35:35,020
And I feel like that creates
a real barrier to engaging

850
00:35:35,020 --> 00:35:37,190
with it, and I think getting
comfortable with that,

851
00:35:37,190 --> 00:35:40,790
with the idea that, yes, AI is
all about transforming data,

852
00:35:40,790 --> 00:35:43,390
but that doesn't mean that
you can't just sit there

853
00:35:43,390 --> 00:35:45,400
and think carefully
about what AI

854
00:35:45,400 --> 00:35:48,170
will do to your business
or your industry.

855
00:35:48,170 --> 00:35:50,680
HOST: And Thank you for sort
of making the comparison

856
00:35:50,680 --> 00:35:54,350
to sometimes, it did make
sense to wait and see,

857
00:35:54,350 --> 00:35:57,020
let the dust settle, let
the kinks get worked out.

858
00:35:57,020 --> 00:35:59,200
And whereas other
cases, I sort of

859
00:35:59,200 --> 00:36:01,090
pictured this train
going down the tracks,

860
00:36:01,090 --> 00:36:03,730
and it's picking up speed,
and it's like, I'm not ever

861
00:36:03,730 --> 00:36:05,050
going to catch up.

862
00:36:05,050 --> 00:36:07,280
I gotta jump now and get on.

863
00:36:07,280 --> 00:36:11,260
For individuals, do
you feel it's worth

864
00:36:11,260 --> 00:36:13,910
paying for premium AI models.

865
00:36:13,910 --> 00:36:16,218
And if so, or if not,
do you have a favorite?

866
00:36:16,218 --> 00:36:18,760
CHRISTIAN DIPPEL: One of the
things that I found most amazing

867
00:36:18,760 --> 00:36:20,570
is in working with engineers.

868
00:36:20,570 --> 00:36:22,970
When you use something
like GitHub Copilot,

869
00:36:22,970 --> 00:36:25,690
GitHub Copilot is like
integrated with five or six

870
00:36:25,690 --> 00:36:29,290
different AI models, and just
seeing an engineer switching

871
00:36:29,290 --> 00:36:31,300
between models
and one model will

872
00:36:31,300 --> 00:36:33,800
be incredible at
changing your code.

873
00:36:33,800 --> 00:36:37,090
Another model will be incredible
at giving you code from scratch

874
00:36:37,090 --> 00:36:39,200
if you're just describing
what you want to do.

875
00:36:39,200 --> 00:36:41,710
And so there's a
lot of nuance, and I

876
00:36:41,710 --> 00:36:44,630
think the deeper down
that rabbit hole you go,

877
00:36:44,630 --> 00:36:46,123
the more those nuances matter.

878
00:36:46,123 --> 00:36:48,040
But then at the end of
the day, it's up to you

879
00:36:48,040 --> 00:36:50,420
to decide if those
nuances matter to you.

880
00:36:50,420 --> 00:36:53,470
So I routinely use
four different models

881
00:36:53,470 --> 00:36:54,980
just in conversation.

882
00:36:54,980 --> 00:36:56,780
I have premium
subscriptions to two.

883
00:36:56,780 --> 00:36:59,600
I don't have premium
subscriptions to the other two,

884
00:36:59,600 --> 00:37:00,980
and I just experiment.

885
00:37:00,980 --> 00:37:04,000
And certain models I find very
little difference between them

886
00:37:04,000 --> 00:37:05,488
depending on the context too.

887
00:37:05,488 --> 00:37:07,030
So if you just ask,
here's the things

888
00:37:07,030 --> 00:37:08,450
I have left over in my fridge.

889
00:37:08,450 --> 00:37:10,460
Give me some suggestions
for dinner tonight,

890
00:37:10,460 --> 00:37:13,220
it's not going to matter all
that much what you use it for.

891
00:37:13,220 --> 00:37:15,850
If you want to do like
really deep thinking

892
00:37:15,850 --> 00:37:18,400
analysis of an industry
profile, maybe there

893
00:37:18,400 --> 00:37:20,410
will be some real
differences between a premium

894
00:37:20,410 --> 00:37:22,070
model and a non-premium model.

895
00:37:22,070 --> 00:37:24,233
HOST: So what I take
away is it depends.

896
00:37:24,233 --> 00:37:25,150
CHRISTIAN DIPPEL: Yes.

897
00:37:25,150 --> 00:37:25,940
It depends.

898
00:37:25,940 --> 00:37:27,790
And maybe further
to that point, I

899
00:37:27,790 --> 00:37:30,220
don't think you need
professional advice

900
00:37:30,220 --> 00:37:34,260
on the decision of whether to
spend $20 a month on a premium

901
00:37:34,260 --> 00:37:34,760
model.

902
00:37:34,760 --> 00:37:37,910
Just spend $20 on a premium
model, and if you don't like it,

903
00:37:37,910 --> 00:37:39,470
unsubscribe a month later.

904
00:37:39,470 --> 00:37:41,608
It's not an
irreversible decision.

905
00:37:41,608 --> 00:37:43,900
HOST: You know, someone might
like this particular tool

906
00:37:43,900 --> 00:37:46,940
because they're using it for
this very specific reason,

907
00:37:46,940 --> 00:37:51,550
whereas someone else might have
success with a different tool

908
00:37:51,550 --> 00:37:54,320
because again, it's
a different purpose.

909
00:37:54,320 --> 00:37:55,910
And given how fast
things are going,

910
00:37:55,910 --> 00:37:59,170
I'm sure new tools will
be emerging all the time.

911
00:37:59,170 --> 00:38:02,020
CHRISTIAN DIPPEL: To your point,
in a classroom setting, if we

912
00:38:02,020 --> 00:38:03,895
have a session where we
actually say, OK, now

913
00:38:03,895 --> 00:38:05,540
everyone go and prompt an AI.

914
00:38:05,540 --> 00:38:07,220
People tend to use
very different ones.

915
00:38:07,220 --> 00:38:09,070
Some people are on
their work laptop.

916
00:38:09,070 --> 00:38:12,380
And so they often use
just Microsoft Copilot.

917
00:38:12,380 --> 00:38:14,170
Other people use
GPT, some people

918
00:38:14,170 --> 00:38:15,860
use Perplexity,
Claude, whatever.

919
00:38:15,860 --> 00:38:18,975
You get a pretty wide range
of results in the answers.

920
00:38:18,975 --> 00:38:20,350
But then at the
same time, if you

921
00:38:20,350 --> 00:38:22,510
ask the same question
of the same LLM

922
00:38:22,510 --> 00:38:25,640
5 minutes later, you can also
get a wide range of results.

923
00:38:25,640 --> 00:38:26,750
HOST: It's like us, right?

924
00:38:26,750 --> 00:38:28,610
If you ask me the same
question tomorrow,

925
00:38:28,610 --> 00:38:31,480
I might get a different
answer depending

926
00:38:31,480 --> 00:38:34,180
on what we were talking about.

927
00:38:34,180 --> 00:38:36,770
Thanks for listening
to Learning in Action.

928
00:38:36,770 --> 00:38:40,400
If Christian's insights sparked
new thinking on AI strategies,

929
00:38:40,400 --> 00:38:41,570
don't stop here.

930
00:38:41,570 --> 00:38:46,120
Head over to ivexeced.com
to explore upcoming programs

931
00:38:46,120 --> 00:38:48,440
and deepen your fluency with AI.

932
00:38:48,440 --> 00:38:51,520
And follow us on your
favorite social media platform

933
00:38:51,520 --> 00:38:55,210
to catch more real time thought
leadership and exclusive podcast

934
00:38:55,210 --> 00:38:55,940
drops.

935
00:38:55,940 --> 00:38:58,380
We'll see you next time.

936
00:38:58,380 --> 00:39:02,000