1
00:00:05,930 --> 00:00:07,579
Hello everyone, and welcome to

2
00:00:07,580 --> 00:00:09,049
another episode of The Operations

3
00:00:09,050 --> 00:00:10,879
Room, a podcast for CEOs.

4
00:00:10,880 --> 00:00:12,769
I am Brandon Bensinger, joined by my

5
00:00:12,770 --> 00:00:14,479
amazing costar Bethany Errors.

6
00:00:14,480 --> 00:00:16,128
How are things going, Bethany?

7
00:00:16,129 --> 00:00:17,419
I don't know.

8
00:00:17,420 --> 00:00:19,309
So one thing is I don't

9
00:00:19,310 --> 00:00:21,199
think Instagram is good for me.

10
00:00:21,200 --> 00:00:22,609
I suspect Instagram isn't good for

11
00:00:22,610 --> 00:00:24,619
anybody, but yet despite that,

12
00:00:24,620 --> 00:00:26,539
I still am on it

13
00:00:26,540 --> 00:00:28,369
and I'm still on it probably too

14
00:00:28,370 --> 00:00:29,149
much.

15
00:00:29,150 --> 00:00:30,829
And it really feeds my health

16
00:00:30,830 --> 00:00:32,179
anxiety.

17
00:00:32,180 --> 00:00:34,189
Health news and all the ways

18
00:00:34,190 --> 00:00:36,049
that we are killing ourselves and

19
00:00:36,050 --> 00:00:37,519
all of the rules that I'm not

20
00:00:37,520 --> 00:00:39,439
following and all

21
00:00:39,440 --> 00:00:41,239
of the reasons why when I die, it's

22
00:00:41,240 --> 00:00:42,559
going to be my fault.

23
00:00:42,560 --> 00:00:44,329
And even though I'm aware of it and

24
00:00:44,330 --> 00:00:46,309
I have blocked as many health

25
00:00:46,310 --> 00:00:48,139
news things as I

26
00:00:48,140 --> 00:00:49,999
can, any time it

27
00:00:50,000 --> 00:00:51,289
comes up, I can't help but watch.

28
00:00:51,290 --> 00:00:52,759
And then the algorithms like, what

29
00:00:52,760 --> 00:00:54,949
she cares about is health news.

30
00:00:54,950 --> 00:00:57,259
And then it just comes back in.

31
00:00:57,260 --> 00:00:59,239
So the most recent

32
00:00:59,240 --> 00:01:01,549
one was actually not specifically

33
00:01:01,550 --> 00:01:03,589
a health news

34
00:01:03,590 --> 00:01:05,869
influencer account.

35
00:01:05,870 --> 00:01:08,089
It was a data account

36
00:01:08,090 --> 00:01:10,189
and it was about how women

37
00:01:10,190 --> 00:01:11,599
are I don't know if it's dying of

38
00:01:11,600 --> 00:01:13,759
cancer or just getting cancer

39
00:01:13,760 --> 00:01:15,619
at a tremendously higher rate

40
00:01:15,620 --> 00:01:16,609
than men.

41
00:01:16,610 --> 00:01:17,959
And then you had comments from

42
00:01:17,960 --> 00:01:20,029
random people trying to decide

43
00:01:20,030 --> 00:01:21,409
why it is that women are getting

44
00:01:21,410 --> 00:01:22,669
cancer at much higher rates than

45
00:01:22,670 --> 00:01:24,919
men. And those comments have seeped

46
00:01:24,920 --> 00:01:26,839
into my brain and now I'm

47
00:01:26,840 --> 00:01:27,799
freaking out.

48
00:01:27,800 --> 00:01:29,719
So it was that

49
00:01:29,720 --> 00:01:31,789
women wear a lot

50
00:01:31,790 --> 00:01:33,649
more cosmetics than men

51
00:01:33,650 --> 00:01:35,449
and they're not properly tested and

52
00:01:35,450 --> 00:01:36,469
a lot of them are made with

53
00:01:36,470 --> 00:01:37,759
petroleum products.

54
00:01:37,760 --> 00:01:39,259
And so we're just like smothering

55
00:01:39,260 --> 00:01:41,329
oil all over our bodies.

56
00:01:41,330 --> 00:01:42,949
And that causes issues.

57
00:01:42,950 --> 00:01:44,929
Men tend to wear more

58
00:01:44,930 --> 00:01:46,909
cotton than women, and women wear

59
00:01:46,910 --> 00:01:48,739
more special fabrics, and all

60
00:01:48,740 --> 00:01:50,689
of the special fabrics are made

61
00:01:50,690 --> 00:01:51,709
out of plastic.

62
00:01:51,710 --> 00:01:52,969
And as they degrade, you get

63
00:01:52,970 --> 00:01:55,099
microplastics in your body,

64
00:01:55,100 --> 00:01:57,349
sports bras directly

65
00:01:57,350 --> 00:01:58,459
linked to breast cancer.

66
00:01:58,460 --> 00:02:00,379
I mean, all of this is not real.

67
00:02:00,380 --> 00:02:02,179
And these are people in the comment

68
00:02:02,180 --> 00:02:04,039
section postulating

69
00:02:04,040 --> 00:02:05,599
as to why women have higher rates of

70
00:02:05,600 --> 00:02:06,499
cancer.

71
00:02:06,500 --> 00:02:08,418
But my ability

72
00:02:08,419 --> 00:02:10,399
to cheat death and do

73
00:02:10,400 --> 00:02:12,499
all of the right things,

74
00:02:12,500 --> 00:02:14,359
follow all the rules

75
00:02:14,360 --> 00:02:16,039
and never die.

76
00:02:16,040 --> 00:02:16,999
That's a lot of pressure on

77
00:02:17,000 --> 00:02:18,379
somebody's shoulders.

78
00:02:18,380 --> 00:02:19,759
At some point you will die,

79
00:02:19,760 --> 00:02:21,079
Bethenny, I suspect.

80
00:02:21,080 --> 00:02:23,089
No, no. If I follow all the rules,

81
00:02:23,090 --> 00:02:24,409
I'll be fine forever.

82
00:02:24,410 --> 00:02:26,359
So sometimes Instagram's not bad.

83
00:02:26,360 --> 00:02:28,459
I was served a video

84
00:02:28,460 --> 00:02:29,959
earlier this week That was Jane

85
00:02:29,960 --> 00:02:32,299
Fonda talking about old age.

86
00:02:32,300 --> 00:02:33,559
And the opening of the video was the

87
00:02:33,560 --> 00:02:35,449
best because she said,

88
00:02:35,450 --> 00:02:36,589
I am Jane Fonda.

89
00:02:36,590 --> 00:02:38,089
You may know me from being Jane

90
00:02:38,090 --> 00:02:39,740
Fonda for the last 86 years.

91
00:02:41,840 --> 00:02:43,759
Exactly. And growing up, it's

92
00:02:43,760 --> 00:02:45,349
literally impossible somehow to not

93
00:02:45,350 --> 00:02:46,579
know who she is.

94
00:02:46,580 --> 00:02:48,619
In the video, she it's like really.

95
00:02:48,620 --> 00:02:49,849
Well, production is video.

96
00:02:49,850 --> 00:02:51,199
And she's working with a scientist

97
00:02:51,200 --> 00:02:52,519
in the scientist is talking about

98
00:02:52,520 --> 00:02:53,749
the facts. And then Jane Fonda is

99
00:02:53,750 --> 00:02:55,519
adding in her bits.

100
00:02:55,520 --> 00:02:57,319
And one of the things she said like

101
00:02:57,320 --> 00:02:59,239
trying to banish senior moments

102
00:02:59,240 --> 00:03:00,049
as a thing.

103
00:03:00,050 --> 00:03:00,979
And it's just like what we're

104
00:03:00,980 --> 00:03:02,239
telling ourselves is that we're

105
00:03:02,240 --> 00:03:03,259
losing our memory. Like when we

106
00:03:03,260 --> 00:03:04,999
can't find our car keys when we're

107
00:03:05,000 --> 00:03:05,989
17.

108
00:03:05,990 --> 00:03:07,579
We're not saying, it's a senior

109
00:03:07,580 --> 00:03:08,689
moment, we just can't find our car

110
00:03:08,690 --> 00:03:10,999
keys. And to embrace

111
00:03:11,000 --> 00:03:13,219
aging and not be so afraid of it

112
00:03:13,220 --> 00:03:15,139
and also say that we all focus

113
00:03:15,140 --> 00:03:17,119
on ending up in nursing homes

114
00:03:17,120 --> 00:03:18,349
and freaking out about nursing

115
00:03:18,350 --> 00:03:19,189
homes.

116
00:03:19,190 --> 00:03:21,559
But actually only 3%

117
00:03:21,560 --> 00:03:22,909
of the population will end up in

118
00:03:22,910 --> 00:03:23,779
them for that.

119
00:03:23,780 --> 00:03:25,129
True? 3%.

120
00:03:25,130 --> 00:03:26,479
I mean, that's true according to

121
00:03:26,480 --> 00:03:27,799
Instagram and Jane Fonda.

122
00:03:27,800 --> 00:03:29,389
I have no idea if that's true.

123
00:03:29,390 --> 00:03:31,609
We could go ahead and verify that.

124
00:03:31,610 --> 00:03:32,929
Jane Fonda said it has got to be

125
00:03:32,930 --> 00:03:33,799
true.

126
00:03:33,800 --> 00:03:35,809
Jane Fonda and the scientist

127
00:03:35,810 --> 00:03:37,639
said it. So, you know,

128
00:03:37,640 --> 00:03:39,229
I choose to believe that that is

129
00:03:39,230 --> 00:03:41,059
true, that 3% of us are going to end

130
00:03:41,060 --> 00:03:42,289
up in nursing homes and the rest of

131
00:03:42,290 --> 00:03:43,579
us are going to have some

132
00:03:43,580 --> 00:03:45,289
combination of home care or just

133
00:03:45,290 --> 00:03:47,629
dying doing okay.

134
00:03:47,630 --> 00:03:49,339
So anyhow, that's the inner workings

135
00:03:49,340 --> 00:03:50,719
of my brain. So how am I doing,

136
00:03:50,720 --> 00:03:51,860
Brandon? I don't know.

137
00:03:55,880 --> 00:03:58,009
All right. So we have got an amazing

138
00:03:58,010 --> 00:03:59,779
topic for today, which is making AI

139
00:03:59,780 --> 00:04:01,429
happen in 2025 for your

140
00:04:01,430 --> 00:04:02,419
organization.

141
00:04:02,420 --> 00:04:03,919
We have an amazing guest for this,

142
00:04:03,920 --> 00:04:05,209
which is Charlie Cullen.

143
00:04:05,210 --> 00:04:07,579
He is an AI strategist and

144
00:04:07,580 --> 00:04:09,349
has singlehandedly developed a

145
00:04:09,350 --> 00:04:11,749
product called Koala using

146
00:04:11,750 --> 00:04:13,639
AI tools behind him in terms

147
00:04:13,640 --> 00:04:15,109
of development tools to make the app

148
00:04:15,110 --> 00:04:16,578
happen. Being a non techie, which is

149
00:04:16,579 --> 00:04:18,289
phenomenal and quite an amazing

150
00:04:18,290 --> 00:04:19,069
story.

151
00:04:19,070 --> 00:04:20,869
So before we get to Charlie, the

152
00:04:20,870 --> 00:04:22,559
first question I wanted to ask you

153
00:04:22,560 --> 00:04:24,439
bit of a broad question, which is

154
00:04:24,440 --> 00:04:26,359
what is the opportunity right now

155
00:04:26,360 --> 00:04:28,399
for scale ups to go horror

156
00:04:28,400 --> 00:04:30,679
on being AI first, as it were.

157
00:04:30,680 --> 00:04:31,669
I made a face when you asked that

158
00:04:31,670 --> 00:04:32,839
question because I was wondering if

159
00:04:32,840 --> 00:04:34,729
you meant I first in the product

160
00:04:34,730 --> 00:04:37,159
or I first within the company.

161
00:04:37,160 --> 00:04:38,449
I mean, I think there's the bigger

162
00:04:38,450 --> 00:04:40,429
existential question is, is

163
00:04:40,430 --> 00:04:41,689
your product shit and do you need to

164
00:04:41,690 --> 00:04:43,369
tear it up and just started again as

165
00:04:43,370 --> 00:04:44,269
an AI product.

166
00:04:44,270 --> 00:04:45,829
And if you're more on the startup

167
00:04:45,830 --> 00:04:47,329
and scale up side, I would highly

168
00:04:47,330 --> 00:04:48,619
encourage you to do that.

169
00:04:48,620 --> 00:04:49,759
I'm just pushing that one out.

170
00:04:49,760 --> 00:04:50,760
There is like

171
00:04:51,830 --> 00:04:54,049
a little moment of of

172
00:04:54,050 --> 00:04:55,969
compelling thinking or controversial

173
00:04:55,970 --> 00:04:59,059
thoughts, although it's interesting

174
00:04:59,060 --> 00:05:01,909
because everything is so new.

175
00:05:01,910 --> 00:05:03,559
There's a lot of scope for improved

176
00:05:03,560 --> 00:05:05,599
products. Like for me, the really

177
00:05:05,600 --> 00:05:07,489
big gap is

178
00:05:07,490 --> 00:05:09,349
a next generation Zapier

179
00:05:09,350 --> 00:05:11,419
like something that is really easy

180
00:05:11,420 --> 00:05:13,729
for non-technical users to you

181
00:05:13,730 --> 00:05:15,979
to use to automate

182
00:05:15,980 --> 00:05:17,989
the soul destroying work in

183
00:05:17,990 --> 00:05:18,889
their lives.

184
00:05:18,890 --> 00:05:20,659
And there are loads and loads of

185
00:05:20,660 --> 00:05:22,699
companies trying to figure out

186
00:05:22,700 --> 00:05:24,649
a tick, but they're still like

187
00:05:24,650 --> 00:05:26,809
technical for technical people

188
00:05:26,810 --> 00:05:28,699
rather than

189
00:05:28,700 --> 00:05:29,839
no code.

190
00:05:29,840 --> 00:05:31,789
Easy automation for

191
00:05:31,790 --> 00:05:33,079
non techies.

192
00:05:33,080 --> 00:05:34,759
So I'm very excited about that

193
00:05:34,760 --> 00:05:36,739
space. And once that.

194
00:05:36,740 --> 00:05:37,969
Unlocks.

195
00:05:37,970 --> 00:05:39,379
Then everything in your business

196
00:05:39,380 --> 00:05:40,709
should be your first.

197
00:05:40,710 --> 00:05:42,769
It's a bit of a question of agents

198
00:05:42,770 --> 00:05:44,119
which are going to be verticals

199
00:05:44,120 --> 00:05:45,349
right now for very specific

200
00:05:45,350 --> 00:05:47,389
functions as opposed to agents

201
00:05:47,390 --> 00:05:48,769
working together in tandem to

202
00:05:48,770 --> 00:05:50,509
actually have more of a net effect

203
00:05:50,510 --> 00:05:51,409
for an organization.

204
00:05:51,410 --> 00:05:53,029
And I think that's definitely out

205
00:05:53,030 --> 00:05:54,119
there in the future. I think we're

206
00:05:54,120 --> 00:05:56,239
going to get first is more siloed

207
00:05:56,240 --> 00:05:58,159
agents for Realm A

208
00:05:58,160 --> 00:05:59,479
versus Run B?

209
00:05:59,480 --> 00:06:00,709
Yeah, but like a lot of those agents

210
00:06:00,710 --> 00:06:01,999
are just siloed.

211
00:06:02,000 --> 00:06:03,769
Vertical agents are not interesting

212
00:06:03,770 --> 00:06:05,419
for me. My Nirvana.

213
00:06:05,420 --> 00:06:07,459
And I think at some point this year

214
00:06:07,460 --> 00:06:09,289
will be I

215
00:06:09,290 --> 00:06:11,269
have all of the CVS to sort

216
00:06:11,270 --> 00:06:12,469
through.

217
00:06:12,470 --> 00:06:14,239
I don't want to sort through them.

218
00:06:14,240 --> 00:06:16,369
I tell the agent, these

219
00:06:16,370 --> 00:06:18,049
are the topics that I'm looking for

220
00:06:18,050 --> 00:06:19,309
on the CVS.

221
00:06:19,310 --> 00:06:21,259
This is what good looks like.

222
00:06:21,260 --> 00:06:23,689
Go and read all of these and

223
00:06:23,690 --> 00:06:25,849
bring out the ones

224
00:06:25,850 --> 00:06:27,199
that I'm looking for rather than a

225
00:06:27,200 --> 00:06:28,249
rules based.

226
00:06:28,250 --> 00:06:29,419
And then you're like, No, no, you

227
00:06:29,420 --> 00:06:30,739
didn't quite get the CVS, right?

228
00:06:30,740 --> 00:06:32,269
Really? I want to have blah, blah,

229
00:06:32,270 --> 00:06:33,199
blah, blah, blah, blah. And, you

230
00:06:33,200 --> 00:06:35,749
know, like actually doing

231
00:06:35,750 --> 00:06:37,579
the boring work for you, but

232
00:06:37,580 --> 00:06:39,679
you're prompting them in English.

233
00:06:39,680 --> 00:06:40,699
But I would consider that to be a

234
00:06:40,700 --> 00:06:42,949
vertical sized agent for

235
00:06:42,950 --> 00:06:44,359
the people recruitment function as

236
00:06:44,360 --> 00:06:45,319
an example.

237
00:06:45,320 --> 00:06:47,119
I think that's definitely in sight

238
00:06:47,120 --> 00:06:48,169
right now. I think what I was

239
00:06:48,170 --> 00:06:49,279
thinking more in my head is more

240
00:06:49,280 --> 00:06:51,109
around broader activities

241
00:06:51,110 --> 00:06:52,579
where, you know, it's not just

242
00:06:52,580 --> 00:06:53,929
simply a recruiter and CV thing that

243
00:06:53,930 --> 00:06:55,369
they're doing, but in fact it's

244
00:06:55,370 --> 00:06:57,259
maybe the entire people function.

245
00:06:57,260 --> 00:06:58,429
Multiple agents need to work

246
00:06:58,430 --> 00:07:00,379
together and cross-pollinate to

247
00:07:00,380 --> 00:07:01,879
come up with better outcomes, I

248
00:07:01,880 --> 00:07:02,509
suppose.

249
00:07:02,510 --> 00:07:03,679
I think you're right in terms of

250
00:07:03,680 --> 00:07:05,059
exists now. But the problem is, is

251
00:07:05,060 --> 00:07:06,349
every single one of those things

252
00:07:06,350 --> 00:07:09,229
exists as a $10 subscription

253
00:07:09,230 --> 00:07:11,779
and I want to pay $130

254
00:07:11,780 --> 00:07:13,909
subscription and do it all

255
00:07:13,910 --> 00:07:15,049
where, you know, there's like one

256
00:07:15,050 --> 00:07:16,999
tool that the people can use

257
00:07:17,000 --> 00:07:19,429
to do their stuff and

258
00:07:19,430 --> 00:07:21,139
accounts payable can use to do

259
00:07:21,140 --> 00:07:23,239
accounts payable and sales

260
00:07:23,240 --> 00:07:25,249
can use to automate their stuff.

261
00:07:25,250 --> 00:07:27,229
And you have like a new

262
00:07:27,230 --> 00:07:29,089
automation layer that the

263
00:07:29,090 --> 00:07:30,919
entire business can

264
00:07:30,920 --> 00:07:32,029
use.

265
00:07:32,030 --> 00:07:33,229
I think from an operational

266
00:07:33,230 --> 00:07:34,669
standpoint, for a company, it seems

267
00:07:34,670 --> 00:07:36,499
like the low hanging fruit

268
00:07:36,500 --> 00:07:38,209
that is out there is simply your

269
00:07:38,210 --> 00:07:39,649
processes and policies.

270
00:07:39,650 --> 00:07:40,729
Anything that's been documented in

271
00:07:40,730 --> 00:07:42,769
that form, which is either

272
00:07:42,770 --> 00:07:44,789
internally creating a subset of GB

273
00:07:44,790 --> 00:07:46,939
TS that pass off different pieces

274
00:07:46,940 --> 00:07:48,709
of policy sections or process

275
00:07:48,710 --> 00:07:50,479
sections or functions or whatever.

276
00:07:50,480 --> 00:07:52,669
Or what I've also seen is

277
00:07:52,670 --> 00:07:54,499
using GB to your court or what have

278
00:07:54,500 --> 00:07:55,999
you. But what I've also seen is

279
00:07:56,000 --> 00:07:57,889
actual agents that

280
00:07:57,890 --> 00:07:59,929
are set up to ingest your

281
00:07:59,930 --> 00:08:02,269
policies or ingest your processes

282
00:08:02,270 --> 00:08:03,529
and do something more advanced than

283
00:08:03,530 --> 00:08:04,849
what you get with an internal chat

284
00:08:04,850 --> 00:08:06,649
and GPT that you create yourself

285
00:08:06,650 --> 00:08:07,489
basically.

286
00:08:07,490 --> 00:08:09,349
So one example of this was

287
00:08:09,350 --> 00:08:11,329
the internal knowledge base curator,

288
00:08:11,330 --> 00:08:12,259
GPT.

289
00:08:12,260 --> 00:08:14,299
So that product and it is a product,

290
00:08:14,300 --> 00:08:15,409
you basically take your

291
00:08:15,410 --> 00:08:17,239
documentation with your company,

292
00:08:17,240 --> 00:08:18,649
ingest it into the CBT.

293
00:08:18,650 --> 00:08:20,749
It's all set up to sort, categorize

294
00:08:20,750 --> 00:08:22,399
and provide a more advanced

295
00:08:22,400 --> 00:08:23,779
experience for the end user in terms

296
00:08:23,780 --> 00:08:24,799
of being able to search for things

297
00:08:24,800 --> 00:08:26,539
and so on with a taxonomy that's

298
00:08:26,540 --> 00:08:27,499
created and so on.

299
00:08:27,500 --> 00:08:28,729
And that seems very interesting to

300
00:08:28,730 --> 00:08:29,149
me.

301
00:08:29,150 --> 00:08:30,229
Yeah, I just don't want to have to

302
00:08:30,230 --> 00:08:31,489
pay 10 pounds for every single one

303
00:08:31,490 --> 00:08:32,569
of them or 20 pounds for every

304
00:08:32,570 --> 00:08:33,798
single one of them.

305
00:08:33,799 --> 00:08:34,639
I'm looking forward to the

306
00:08:34,640 --> 00:08:36,558
consolidation and it just at

307
00:08:36,559 --> 00:08:37,849
what level does that consolidation

308
00:08:37,850 --> 00:08:38,719
happen?

309
00:08:38,720 --> 00:08:40,548
So I was reading Charlie's

310
00:08:40,549 --> 00:08:42,229
LinkedIn posts and he created one

311
00:08:42,230 --> 00:08:44,329
post around hiring

312
00:08:44,330 --> 00:08:46,609
for an AI first organization,

313
00:08:46,610 --> 00:08:47,929
and his little description that he

314
00:08:47,930 --> 00:08:49,909
put in there was to create a task

315
00:08:49,910 --> 00:08:52,249
for a candidate where

316
00:08:52,250 --> 00:08:54,169
they need to write a prompt or

317
00:08:54,170 --> 00:08:56,739
create a. Claude Church project

318
00:08:56,740 --> 00:08:58,809
that can accelerate a key task

319
00:08:58,810 --> 00:08:59,859
in the role that they're being hired

320
00:08:59,860 --> 00:09:01,779
for and to demonstrate

321
00:09:01,780 --> 00:09:03,819
an example of how they would use it.

322
00:09:03,820 --> 00:09:04,899
And thus that's really interesting

323
00:09:04,900 --> 00:09:06,069
to me. Or maybe think at the very

324
00:09:06,070 --> 00:09:07,989
least in the sense of like, okay, if

325
00:09:07,990 --> 00:09:09,819
we want to hire people

326
00:09:09,820 --> 00:09:11,739
that are on a I using

327
00:09:11,740 --> 00:09:13,929
a, I have some level of competence

328
00:09:13,930 --> 00:09:16,359
and skill and interest in A.I..

329
00:09:16,360 --> 00:09:17,859
It seems like at the very least

330
00:09:17,860 --> 00:09:19,659
asking an interview question in some

331
00:09:19,660 --> 00:09:21,099
form, maybe not a task, but some

332
00:09:21,100 --> 00:09:22,449
kind of questioning around this

333
00:09:22,450 --> 00:09:23,619
seems to make tremendous sense.

334
00:09:23,620 --> 00:09:24,999
I'm just curious what you think.

335
00:09:25,000 --> 00:09:26,919
I agree. And I actually think being

336
00:09:26,920 --> 00:09:28,719
a task would be totally legitimate.

337
00:09:28,720 --> 00:09:29,559
And it shows interest.

338
00:09:29,560 --> 00:09:31,539
It shows aptitude and it

339
00:09:31,540 --> 00:09:33,639
shows curiosity.

340
00:09:33,640 --> 00:09:35,169
I think there's still a level of

341
00:09:35,170 --> 00:09:37,569
resistance to

342
00:09:37,570 --> 00:09:39,609
a AI because it's seen

343
00:09:39,610 --> 00:09:40,779
as cheating.

344
00:09:40,780 --> 00:09:43,029
And so it's actually showing people

345
00:09:43,030 --> 00:09:45,189
who have a AI first

346
00:09:45,190 --> 00:09:47,019
mindset or forward mindset

347
00:09:47,020 --> 00:09:48,009
where using A.I.

348
00:09:48,010 --> 00:09:50,079
isn't cheating, it's the new normal

349
00:09:50,080 --> 00:09:52,179
and it's the smart thing to do.

350
00:09:52,180 --> 00:09:54,279
And so both you're testing

351
00:09:54,280 --> 00:09:56,109
for people in who

352
00:09:56,110 --> 00:09:57,369
are coming into your organization,

353
00:09:57,370 --> 00:09:59,289
who are afterward, but

354
00:09:59,290 --> 00:10:02,259
you're also massively demonstrating

355
00:10:02,260 --> 00:10:04,629
expectations that in your

356
00:10:04,630 --> 00:10:06,489
world, using

357
00:10:06,490 --> 00:10:08,259
AI is smart, not cheating.

358
00:10:08,260 --> 00:10:10,269
So when you think about rolling

359
00:10:10,270 --> 00:10:11,319
out A.I.

360
00:10:11,320 --> 00:10:13,089
within the organization right now in

361
00:10:13,090 --> 00:10:15,069
2025 for a company, we

362
00:10:15,070 --> 00:10:16,989
have our standard tools, if you want

363
00:10:16,990 --> 00:10:19,089
to call it that, which is clod

364
00:10:19,090 --> 00:10:20,859
chatbot, Gemini.

365
00:10:20,860 --> 00:10:21,759
What makes sense here?

366
00:10:21,760 --> 00:10:23,169
Would you encourage companies right

367
00:10:23,170 --> 00:10:25,269
now to say to themselves, okay,

368
00:10:25,270 --> 00:10:26,859
look, let's create a policy around

369
00:10:26,860 --> 00:10:28,689
this. Let's license

370
00:10:28,690 --> 00:10:30,009
some of these tools for different

371
00:10:30,010 --> 00:10:31,059
functions within the company or

372
00:10:31,060 --> 00:10:33,159
within the entire organization and

373
00:10:33,160 --> 00:10:34,479
launch them into the company.

374
00:10:34,480 --> 00:10:35,829
Upskill people in terms of how to

375
00:10:35,830 --> 00:10:37,929
write prompts or create projects,

376
00:10:37,930 --> 00:10:39,549
create some level of support

377
00:10:39,550 --> 00:10:41,289
structure around it where people are

378
00:10:41,290 --> 00:10:42,519
being supported to use it

379
00:10:42,520 --> 00:10:44,769
effectively, and some kind of reward

380
00:10:44,770 --> 00:10:46,179
mechanism to reward those who are

381
00:10:46,180 --> 00:10:47,379
actually integrating it back into

382
00:10:47,380 --> 00:10:48,309
their jobs.

383
00:10:48,310 --> 00:10:49,479
Is that something that we should do

384
00:10:49,480 --> 00:10:50,139
now?

385
00:10:50,140 --> 00:10:52,089
All the technology is changing so

386
00:10:52,090 --> 00:10:54,099
rapidly that I would

387
00:10:54,100 --> 00:10:55,779
not commit to a single one.

388
00:10:55,780 --> 00:10:57,279
I would go even if it's a bit more

389
00:10:57,280 --> 00:10:58,959
expensive for rolling monthly

390
00:10:58,960 --> 00:11:00,909
contracts and continue to

391
00:11:00,910 --> 00:11:02,019
experiment.

392
00:11:02,020 --> 00:11:04,569
It's too early

393
00:11:04,570 --> 00:11:06,549
to sign up for an annual

394
00:11:06,550 --> 00:11:08,409
commitment to anything right now.

395
00:11:08,410 --> 00:11:09,579
I don't even know if you can, but if

396
00:11:09,580 --> 00:11:11,109
you can, I would not suggest annual

397
00:11:11,110 --> 00:11:12,069
commitments.

398
00:11:12,070 --> 00:11:12,849
It's not worth it.

399
00:11:12,850 --> 00:11:14,349
Things are moving fast.

400
00:11:14,350 --> 00:11:15,879
Policies, Definitely.

401
00:11:15,880 --> 00:11:17,619
Because it makes you think through

402
00:11:17,620 --> 00:11:18,939
all of the gotchas and it makes it

403
00:11:18,940 --> 00:11:20,679
really clear for the organization

404
00:11:20,680 --> 00:11:22,599
just to push

405
00:11:22,600 --> 00:11:24,309
Charlie's LinkedIn a little bit

406
00:11:24,310 --> 00:11:26,349
more. He happened to have a LinkedIn

407
00:11:26,350 --> 00:11:28,179
post this morning that was a seven

408
00:11:28,180 --> 00:11:30,489
page, a AI policy

409
00:11:30,490 --> 00:11:32,469
template to use that looked pretty

410
00:11:32,470 --> 00:11:34,329
decent for what a lot of

411
00:11:34,330 --> 00:11:36,189
the main thoughts if you want, if

412
00:11:36,190 --> 00:11:37,959
you haven't put one in place yet or

413
00:11:37,960 --> 00:11:38,889
you want to make sure that you're

414
00:11:38,890 --> 00:11:40,659
covering the key areas, I suggest

415
00:11:40,660 --> 00:11:41,679
having a look at it.

416
00:11:41,680 --> 00:11:44,049
And then the reward mechanism or

417
00:11:44,050 --> 00:11:47,079
for us at peak,

418
00:11:47,080 --> 00:11:48,939
we have a few different ways that

419
00:11:48,940 --> 00:11:50,979
we are focusing on the team

420
00:11:50,980 --> 00:11:53,349
using and getting more experimental

421
00:11:53,350 --> 00:11:55,299
and interested in generally because

422
00:11:55,300 --> 00:11:56,709
we have pockets that are super

423
00:11:56,710 --> 00:11:58,569
interested in pockets that aren't.

424
00:11:58,570 --> 00:12:01,179
So in our weekly

425
00:12:01,180 --> 00:12:03,049
we have like a company all hands

426
00:12:03,050 --> 00:12:04,479
that's weekly.

427
00:12:04,480 --> 00:12:06,759
There is a five minute

428
00:12:06,760 --> 00:12:08,679
segment every single week where

429
00:12:08,680 --> 00:12:10,989
somebody does a show and tell on

430
00:12:10,990 --> 00:12:12,129
something they've done.

431
00:12:12,130 --> 00:12:14,379
The second thing

432
00:12:14,380 --> 00:12:16,209
that we've done is we

433
00:12:16,210 --> 00:12:17,529
have a Slack channel that's a

434
00:12:17,530 --> 00:12:18,669
Jenn-air channel. We actually used

435
00:12:18,670 --> 00:12:19,899
to have two and have just merged it

436
00:12:19,900 --> 00:12:21,459
into one because it was getting a

437
00:12:21,460 --> 00:12:23,199
bit messy as to which is which.

438
00:12:23,200 --> 00:12:25,329
And so all news, all

439
00:12:25,330 --> 00:12:27,369
thoughts, all new technology,

440
00:12:27,370 --> 00:12:29,499
all uses goes

441
00:12:29,500 --> 00:12:32,019
into that channel and

442
00:12:32,020 --> 00:12:33,309
it's a self-selecting channel.

443
00:12:33,310 --> 00:12:34,479
But I think most of the businesses

444
00:12:34,480 --> 00:12:35,259
in it.

445
00:12:35,260 --> 00:12:37,449
And then the third thing that we are

446
00:12:37,450 --> 00:12:39,999
rolling out this year is

447
00:12:40,000 --> 00:12:42,429
for our technical teams,

448
00:12:42,430 --> 00:12:45,009
a Jenn-air training course

449
00:12:45,010 --> 00:12:46,029
we're running internally.

450
00:12:46,030 --> 00:12:47,349
We've not found anything externally

451
00:12:47,350 --> 00:12:49,689
that matches our needs and

452
00:12:49,690 --> 00:12:51,189
that is an enablement course.

453
00:12:51,190 --> 00:12:52,689
It'll be running for the first half

454
00:12:52,690 --> 00:12:54,069
of the year and then we'll see what

455
00:12:54,070 --> 00:12:55,779
the second half brings us.

456
00:12:55,780 --> 00:12:58,269
So when it comes to budgeting

457
00:12:58,270 --> 00:13:00,279
and providing different

458
00:13:00,280 --> 00:13:01,749
functions that have different

459
00:13:01,750 --> 00:13:03,129
aspirations and different needs

460
00:13:03,130 --> 00:13:04,809
around some of these generic tools,

461
00:13:04,810 --> 00:13:05,859
what are you doing from a budgeting

462
00:13:05,860 --> 00:13:06,819
standpoint? You're you're creating

463
00:13:06,820 --> 00:13:08,409
like a budget for each of the

464
00:13:08,410 --> 00:13:10,869
functions or how does that work?

465
00:13:10,870 --> 00:13:12,879
So we're doing a bit of of that

466
00:13:12,880 --> 00:13:14,829
for experimentation and like what

467
00:13:14,830 --> 00:13:15,969
are the right tools, what are the

468
00:13:15,970 --> 00:13:16,989
best tools?

469
00:13:16,990 --> 00:13:19,209
And then also for all

470
00:13:19,210 --> 00:13:21,699
of our systems of records

471
00:13:21,700 --> 00:13:24,099
or standard SAS tools

472
00:13:24,100 --> 00:13:26,409
as they come up for renewal.

473
00:13:26,410 --> 00:13:28,839
We are investigating

474
00:13:28,840 --> 00:13:31,029
the market and looking for

475
00:13:31,030 --> 00:13:32,979
eye first alternatives

476
00:13:32,980 --> 00:13:34,509
and seeing whether or not they're

477
00:13:34,510 --> 00:13:36,759
too early stage to move to.

478
00:13:36,760 --> 00:13:38,829
But our general

479
00:13:38,830 --> 00:13:40,869
assumption of hypothesis is

480
00:13:40,870 --> 00:13:42,699
all of the SAS businesses

481
00:13:42,700 --> 00:13:44,679
that have a bit of AI stuck

482
00:13:44,680 --> 00:13:46,569
on the side are not going to

483
00:13:46,570 --> 00:13:48,699
be as good as AI first businesses.

484
00:13:48,700 --> 00:13:50,889
And so we're constantly

485
00:13:50,890 --> 00:13:53,139
searching in the market for what

486
00:13:53,140 --> 00:13:54,499
is the new system of record.

487
00:13:54,500 --> 00:13:56,029
Like in this new world.

488
00:13:56,030 --> 00:13:57,799
And for my vision, I think we have

489
00:13:57,800 --> 00:13:59,269
one system of record with lots of

490
00:13:59,270 --> 00:14:00,559
different elements to it.

491
00:14:00,560 --> 00:14:02,329
Because what's a system?

492
00:14:02,330 --> 00:14:04,129
A record, but a massive database.

493
00:14:04,130 --> 00:14:05,959
And so and there's always

494
00:14:05,960 --> 00:14:07,139
been that the reason why the systems

495
00:14:07,140 --> 00:14:08,899
record are different is because of

496
00:14:08,900 --> 00:14:11,089
the interface and the tooling

497
00:14:11,090 --> 00:14:12,049
required.

498
00:14:12,050 --> 00:14:14,689
But if the interface becomes

499
00:14:14,690 --> 00:14:16,669
a chat box

500
00:14:16,670 --> 00:14:18,799
for pretty much everybody,

501
00:14:18,800 --> 00:14:20,329
do we need different systems of

502
00:14:20,330 --> 00:14:21,439
record long term?

503
00:14:21,440 --> 00:14:22,369
Short term we do.

504
00:14:22,370 --> 00:14:23,869
Because I don't think anybody's

505
00:14:23,870 --> 00:14:25,079
reached that long term goal.

506
00:14:25,080 --> 00:14:26,959
But that's part of why we're not

507
00:14:26,960 --> 00:14:28,819
committing to

508
00:14:28,820 --> 00:14:30,919
long term contracts right now while

509
00:14:30,920 --> 00:14:32,029
we watch the space.

510
00:14:32,030 --> 00:14:33,919
And so for our existing systems of

511
00:14:33,920 --> 00:14:34,999
record, when they're coming up for

512
00:14:35,000 --> 00:14:36,199
renewal, we're not doing three year

513
00:14:36,200 --> 00:14:37,639
renewals for anything.

514
00:14:37,640 --> 00:14:39,169
We're doing one year renewals so

515
00:14:39,170 --> 00:14:41,029
that we have this choice, even if

516
00:14:41,030 --> 00:14:41,989
nothing's fit for purpose.

517
00:14:41,990 --> 00:14:43,879
Yet our our guest is something will

518
00:14:43,880 --> 00:14:45,259
be fit for purpose in a year.

519
00:14:45,260 --> 00:14:47,209
So when it comes to the policy

520
00:14:47,210 --> 00:14:49,069
that you now have and you also

521
00:14:49,070 --> 00:14:51,139
mentioned Charlie on his LinkedIn

522
00:14:51,140 --> 00:14:52,429
profile, putting together a

523
00:14:52,430 --> 00:14:54,349
template, a policy.

524
00:14:54,350 --> 00:14:55,189
He talked about a couple of

525
00:14:55,190 --> 00:14:57,049
interesting things. One was in the

526
00:14:57,050 --> 00:14:59,209
policy itself to provide

527
00:14:59,210 --> 00:15:01,039
a list of encouraged uses for

528
00:15:01,040 --> 00:15:02,869
each function that are endorsed by

529
00:15:02,870 --> 00:15:03,649
the leaders.

530
00:15:03,650 --> 00:15:04,939
And I think this is quite important

531
00:15:04,940 --> 00:15:06,439
and useful in the sense that the

532
00:15:06,440 --> 00:15:08,659
policy itself is not simply there

533
00:15:08,660 --> 00:15:10,579
to tell people what to not do, but

534
00:15:10,580 --> 00:15:11,839
in fact to do the opposite in this

535
00:15:11,840 --> 00:15:14,209
case, because what we want with AI

536
00:15:14,210 --> 00:15:16,039
is to truly embed in

537
00:15:16,040 --> 00:15:17,809
people's thoughts that we want the

538
00:15:17,810 --> 00:15:18,649
company to win.

539
00:15:18,650 --> 00:15:19,639
And the way the company is going to

540
00:15:19,640 --> 00:15:20,989
win of the future is using A.I.

541
00:15:20,990 --> 00:15:22,849
tools. And we need you to try them.

542
00:15:22,850 --> 00:15:24,349
We need you to experiment.

543
00:15:24,350 --> 00:15:25,849
And by providing a list of

544
00:15:25,850 --> 00:15:27,769
encouraged use cases and

545
00:15:27,770 --> 00:15:28,579
it's not to say that it's going to be

546
00:15:28,580 --> 00:15:30,469
exhaustive, but you need to put

547
00:15:30,470 --> 00:15:31,549
something out there where people

548
00:15:31,550 --> 00:15:32,929
like, okay, within my function of

549
00:15:32,930 --> 00:15:35,029
sales, here's like the ten things

550
00:15:35,030 --> 00:15:36,859
that the cells like right now has

551
00:15:36,860 --> 00:15:38,179
tacitly endorsed, basically.

552
00:15:38,180 --> 00:15:39,169
So maybe I should start doing

553
00:15:39,170 --> 00:15:40,459
something about it, I guess.

554
00:15:40,460 --> 00:15:42,139
What do you think of that first one?

555
00:15:42,140 --> 00:15:43,969
So we have that, but not

556
00:15:43,970 --> 00:15:45,139
in our policy.

557
00:15:45,140 --> 00:15:47,539
We have it in our strategy document.

558
00:15:47,540 --> 00:15:50,119
And so what we've done is

559
00:15:50,120 --> 00:15:51,979
for each department, we

560
00:15:51,980 --> 00:15:54,559
have like low hanging fruit,

561
00:15:54,560 --> 00:15:55,789
you know, what are like the immediate

562
00:15:55,790 --> 00:15:58,009
things that we can be doing.

563
00:15:58,010 --> 00:16:00,889
What are areas where we want to

564
00:16:00,890 --> 00:16:02,899
do more, investigate and experiment

565
00:16:02,900 --> 00:16:04,429
with, and what is our long term

566
00:16:04,430 --> 00:16:05,329
vision?

567
00:16:05,330 --> 00:16:06,919
So what is our ideal in each of

568
00:16:06,920 --> 00:16:08,119
those areas?

569
00:16:08,120 --> 00:16:09,499
What we have also done that I forgot

570
00:16:09,500 --> 00:16:11,509
to mention is we have an A.I.

571
00:16:11,510 --> 00:16:13,309
evangelist in every team.

572
00:16:13,310 --> 00:16:15,769
So that's the person

573
00:16:15,770 --> 00:16:17,929
who self-selected

574
00:16:17,930 --> 00:16:19,189
could be very different in each

575
00:16:19,190 --> 00:16:20,329
team, but it's the person who just

576
00:16:20,330 --> 00:16:21,649
like naturally is reading

577
00:16:21,650 --> 00:16:24,169
everything, naturally experimenting.

578
00:16:24,170 --> 00:16:26,059
They tend to be the person who's

579
00:16:26,060 --> 00:16:27,619
in the all hands showing

580
00:16:27,620 --> 00:16:29,149
information, but they're like the go

581
00:16:29,150 --> 00:16:31,039
to person of Ah,

582
00:16:31,040 --> 00:16:32,869
I've just realized that stars can

583
00:16:32,870 --> 00:16:33,499
do whatever.

584
00:16:33,500 --> 00:16:35,989
Or in the

585
00:16:35,990 --> 00:16:37,459
second one that he suggested should

586
00:16:37,460 --> 00:16:39,349
be in the policy is deciding

587
00:16:39,350 --> 00:16:41,209
on a list of company approved A.I.

588
00:16:41,210 --> 00:16:43,549
tools in a particular a streamlined

589
00:16:43,550 --> 00:16:45,349
process to get tools and versions

590
00:16:45,350 --> 00:16:47,029
approved, i.e.

591
00:16:47,030 --> 00:16:48,499
somebody see something, they make a

592
00:16:48,500 --> 00:16:50,659
request, they get an approval within

593
00:16:50,660 --> 00:16:51,949
days as an SLA.

594
00:16:51,950 --> 00:16:53,239
And the default is to say yes,

595
00:16:53,240 --> 00:16:54,229
pretty much.

596
00:16:54,230 --> 00:16:56,029
What do you make of that?

597
00:16:56,030 --> 00:16:57,919
So we already have that in place,

598
00:16:57,920 --> 00:16:59,809
but not specifically for genitals,

599
00:16:59,810 --> 00:17:02,479
just in general for buying

600
00:17:02,480 --> 00:17:03,859
for our security reasons.

601
00:17:03,860 --> 00:17:05,118
So it has to go through a security

602
00:17:05,119 --> 00:17:06,769
audit, which means that it's not

603
00:17:06,770 --> 00:17:08,419
within a couple hours.

604
00:17:08,420 --> 00:17:10,429
It just depends on how much

605
00:17:10,430 --> 00:17:12,318
information we can get about these

606
00:17:12,319 --> 00:17:13,699
companies and how they're treating

607
00:17:13,700 --> 00:17:15,529
our data and whether or not they

608
00:17:15,530 --> 00:17:17,449
have ISO 27,001,

609
00:17:17,450 --> 00:17:18,559
whether or not they're GDPR

610
00:17:18,560 --> 00:17:20,179
compliant, etc.

611
00:17:20,180 --> 00:17:22,429
And sometimes that can take weeks.

612
00:17:22,430 --> 00:17:24,709
But I think it's worth

613
00:17:24,710 --> 00:17:26,719
it because there's so many new tiny

614
00:17:26,720 --> 00:17:29,029
companies that

615
00:17:29,030 --> 00:17:30,949
the chat you or

616
00:17:30,950 --> 00:17:33,079
the Open Eyes and Gemini

617
00:17:33,080 --> 00:17:35,089
like you can find out

618
00:17:35,090 --> 00:17:36,979
all of their security really quickly

619
00:17:36,980 --> 00:17:38,899
and make decisions and understand

620
00:17:38,900 --> 00:17:40,099
whether or not your data is training

621
00:17:40,100 --> 00:17:41,899
the models, etc..

622
00:17:41,900 --> 00:17:43,999
Tiny company with 15

623
00:17:44,000 --> 00:17:46,339
people, we end up contacting

624
00:17:46,340 --> 00:17:48,259
them, getting their certifications,

625
00:17:48,260 --> 00:17:49,759
understanding what their policies

626
00:17:49,760 --> 00:17:51,379
all are. If they don't have
certifications

627
00:17:51,380 --> 00:17:53,089
yet, and talking through whether or

628
00:17:53,090 --> 00:17:54,889
not we feel comfortable working with

629
00:17:54,890 --> 00:17:56,539
them. Quite often they choose small

630
00:17:56,540 --> 00:17:57,979
and haven't not mature enough and

631
00:17:57,980 --> 00:17:59,299
don't have it yet, or maybe haven't

632
00:17:59,300 --> 00:18:00,469
even been around enough for an

633
00:18:00,470 --> 00:18:01,639
audit.

634
00:18:01,640 --> 00:18:03,349
But we will have conversations with

635
00:18:03,350 --> 00:18:04,849
them and ask them how they're

636
00:18:04,850 --> 00:18:06,799
thinking about things and if

637
00:18:06,800 --> 00:18:08,539
they seem to understand what they're

638
00:18:08,540 --> 00:18:10,039
talking about and be quite forward

639
00:18:10,040 --> 00:18:11,929
thinking and have policies in

640
00:18:11,930 --> 00:18:13,669
place and a roadmap of where they're

641
00:18:13,670 --> 00:18:15,499
going to get to, depending on the

642
00:18:15,500 --> 00:18:16,789
sensitivity of the data that might

643
00:18:16,790 --> 00:18:18,859
be going there will still say

644
00:18:18,860 --> 00:18:19,639
yes.

645
00:18:19,640 --> 00:18:20,899
But if we talk to somebody and it's

646
00:18:20,900 --> 00:18:22,249
just like, okay, these people are

647
00:18:22,250 --> 00:18:25,099
very immature, they have no idea.

648
00:18:25,100 --> 00:18:26,689
We cannot trust them with any level

649
00:18:26,690 --> 00:18:28,279
of data, we'll say no.

650
00:18:28,280 --> 00:18:29,329
The other thing that occurs to me

651
00:18:29,330 --> 00:18:31,339
for companies right now,

652
00:18:31,340 --> 00:18:33,049
irrespective of any policy, are

653
00:18:33,050 --> 00:18:35,299
actually rolling on AI at all

654
00:18:35,300 --> 00:18:37,489
is just this obvious thing where

655
00:18:37,490 --> 00:18:39,169
right now today, if you don't have

656
00:18:39,170 --> 00:18:40,999
any AI policy, you've got a problem

657
00:18:41,000 --> 00:18:42,619
a little bit because people in your

658
00:18:42,620 --> 00:18:44,479
company are using personal accounts

659
00:18:44,480 --> 00:18:46,609
to use chatbot to write their emails

660
00:18:46,610 --> 00:18:48,559
or whatever that is happening.

661
00:18:48,560 --> 00:18:50,089
So having a very simple statement to

662
00:18:50,090 --> 00:18:51,559
the company, irrespective of

663
00:18:51,560 --> 00:18:53,149
anything else, just saying, look, if

664
00:18:53,150 --> 00:18:54,189
you're using a personal.

665
00:18:54,190 --> 00:18:56,079
Howard, please, please, please

666
00:18:56,080 --> 00:18:58,029
go into your settings and toggle

667
00:18:58,030 --> 00:19:00,339
the data learning off essentially,

668
00:19:00,340 --> 00:19:02,229
because if you don't, anything that

669
00:19:02,230 --> 00:19:03,549
you do will be sucked into the

670
00:19:03,550 --> 00:19:05,409
vortex of the the

671
00:19:05,410 --> 00:19:07,269
large language models and we

672
00:19:07,270 --> 00:19:08,559
cannot have that.

673
00:19:08,560 --> 00:19:10,479
And also then write an AI policy.

674
00:19:10,480 --> 00:19:11,709
The last bit I wanted to bring up

675
00:19:11,710 --> 00:19:12,969
was best practices.

676
00:19:12,970 --> 00:19:14,739
And you had a couple of nuggets here

677
00:19:14,740 --> 00:19:15,939
that I thought were useful.

678
00:19:15,940 --> 00:19:17,949
One was when it comes to

679
00:19:17,950 --> 00:19:19,839
clod projects and

680
00:19:19,840 --> 00:19:21,699
church EPD projects and projects

681
00:19:21,700 --> 00:19:23,619
are essentially a framework to

682
00:19:23,620 --> 00:19:25,359
allow you to have a consistent set

683
00:19:25,360 --> 00:19:27,339
of context for whatever

684
00:19:27,340 --> 00:19:28,989
it is that you're asking of the GPU

685
00:19:28,990 --> 00:19:31,329
in this case to respond to around

686
00:19:31,330 --> 00:19:32,919
the company or around your products

687
00:19:32,920 --> 00:19:34,599
or around your policies or whatever,

688
00:19:34,600 --> 00:19:36,489
whereby you don't have to enter

689
00:19:36,490 --> 00:19:38,469
a prompt every single time to

690
00:19:38,470 --> 00:19:39,609
put in that context.

691
00:19:39,610 --> 00:19:41,439
And his best

692
00:19:41,440 --> 00:19:43,569
practice was to ensure that

693
00:19:43,570 --> 00:19:45,999
those projects and those statements

694
00:19:46,000 --> 00:19:48,009
around the company were there

695
00:19:48,010 --> 00:19:50,109
to allow any user within the company

696
00:19:50,110 --> 00:19:51,579
to take that project, take that

697
00:19:51,580 --> 00:19:53,739
context and apply it to their

698
00:19:53,740 --> 00:19:56,049
their account very specifically

699
00:19:56,050 --> 00:19:57,489
to ensure that all the context was

700
00:19:57,490 --> 00:19:58,409
all pre set up.

701
00:19:58,410 --> 00:19:59,349
When you think about one.

702
00:19:59,350 --> 00:20:01,539
That is beyond what I've experienced

703
00:20:01,540 --> 00:20:02,709
so far. So I just be like, if

704
00:20:02,710 --> 00:20:04,179
Charlie says it's a good idea, it's

705
00:20:04,180 --> 00:20:05,289
a good idea.

706
00:20:05,290 --> 00:20:06,699
So what do we park it here and get

707
00:20:06,700 --> 00:20:08,199
on to our conversation with Mr.

708
00:20:08,200 --> 00:20:09,200
Charlie Cullen?

709
00:20:13,240 --> 00:20:14,289
Why don't you talk a little bit

710
00:20:14,290 --> 00:20:15,159
about Koala?

711
00:20:15,160 --> 00:20:16,779
Not so much for a product that

712
00:20:16,780 --> 00:20:18,309
people should go and buy, but the

713
00:20:18,310 --> 00:20:20,529
journey of creating it.

714
00:20:20,530 --> 00:20:22,629
So in August 2024,

715
00:20:22,630 --> 00:20:24,189
this is when I was in pink.

716
00:20:24,190 --> 00:20:26,439
My goodness. I'm going to set up a

717
00:20:26,440 --> 00:20:28,749
i boutique consultancy.

718
00:20:28,750 --> 00:20:30,609
And then suddenly, my

719
00:20:30,610 --> 00:20:32,379
goodness. Like, who am I to go and

720
00:20:32,380 --> 00:20:34,269
tell people how

721
00:20:34,270 --> 00:20:35,799
they should be implementing and

722
00:20:35,800 --> 00:20:37,389
integrating these technologies if

723
00:20:37,390 --> 00:20:39,369
I've not done it myself?

724
00:20:39,370 --> 00:20:41,079
That's just snake oil.

725
00:20:41,080 --> 00:20:41,949
This was August.

726
00:20:41,950 --> 00:20:44,019
I was on a ferry to a Greek

727
00:20:44,020 --> 00:20:46,119
island with my family,

728
00:20:46,120 --> 00:20:48,009
and the kids were sleeping

729
00:20:48,010 --> 00:20:50,229
and my wife and I were reading

730
00:20:50,230 --> 00:20:51,609
and I was reading a book called The

731
00:20:51,610 --> 00:20:52,449
Jolt Effect.

732
00:20:52,450 --> 00:20:54,429
And they did some research where

733
00:20:54,430 --> 00:20:56,559
they looked at why enterprise

734
00:20:56,560 --> 00:20:58,659
sales end up with the status

735
00:20:58,660 --> 00:20:59,709
quo.

736
00:20:59,710 --> 00:21:02,139
And the prevailing thought was that

737
00:21:02,140 --> 00:21:03,939
people stayed with the status quo

738
00:21:03,940 --> 00:21:05,859
because as a sales person,

739
00:21:05,860 --> 00:21:07,779
you had not been able to sell

740
00:21:07,780 --> 00:21:09,699
the difference from what they've

741
00:21:09,700 --> 00:21:10,659
got today.

742
00:21:10,660 --> 00:21:12,159
So I came back from this holiday and

743
00:21:12,160 --> 00:21:13,479
I was thinking, okay, well, I'm

744
00:21:13,480 --> 00:21:15,759
going to build an AI type solution

745
00:21:15,760 --> 00:21:17,319
for this and it will solve my

746
00:21:17,320 --> 00:21:19,329
problem of I've never built

747
00:21:19,330 --> 00:21:21,069
anything with these tools and then

748
00:21:21,070 --> 00:21:22,899
I'll be able to advise

749
00:21:22,900 --> 00:21:24,039
people.

750
00:21:24,040 --> 00:21:25,929
And even in September

751
00:21:25,930 --> 00:21:27,849
of 2024, I

752
00:21:27,850 --> 00:21:30,039
was WhatsApp paying friends.

753
00:21:30,040 --> 00:21:31,539
Do you know anyone that is a

754
00:21:31,540 --> 00:21:32,919
developer that could know how to

755
00:21:32,920 --> 00:21:34,299
write this thing? Because I'm a

756
00:21:34,300 --> 00:21:36,429
non-technical founder.

757
00:21:36,430 --> 00:21:37,929
Does anyone know anyone that could

758
00:21:37,930 --> 00:21:40,029
help me either as an agency

759
00:21:40,030 --> 00:21:41,709
or whatever? But just over the

760
00:21:41,710 --> 00:21:43,599
course of the summer, some amazing

761
00:21:43,600 --> 00:21:45,879
new tools had been coming out

762
00:21:45,880 --> 00:21:47,709
new versions of Code,

763
00:21:47,710 --> 00:21:49,539
which is one of my favorite albums,

764
00:21:49,540 --> 00:21:50,540
which is like a

765
00:21:51,430 --> 00:21:53,349
a new tool called Cosa came

766
00:21:53,350 --> 00:21:55,539
out, which is an AI development

767
00:21:55,540 --> 00:21:56,589
environment.

768
00:21:56,590 --> 00:21:58,629
A new tool called V Zero

769
00:21:58,630 --> 00:22:00,459
came out, which is a UI

770
00:22:00,460 --> 00:22:01,719
development tool.

771
00:22:01,720 --> 00:22:03,519
And I just thought one weekend, you

772
00:22:03,520 --> 00:22:05,679
know, I'll just see how

773
00:22:05,680 --> 00:22:08,259
far I can get as a non

774
00:22:08,260 --> 00:22:09,129
developer.

775
00:22:09,130 --> 00:22:11,139
And so I started off with

776
00:22:11,140 --> 00:22:12,139
Claude.

777
00:22:12,140 --> 00:22:13,959
So if you use EPG, then you

778
00:22:13,960 --> 00:22:15,369
know how to use Claude.

779
00:22:15,370 --> 00:22:17,529
One of my favorite features of Tord

780
00:22:17,530 --> 00:22:19,689
is that you can set up a project

781
00:22:19,690 --> 00:22:21,489
and a project you can think of as a

782
00:22:21,490 --> 00:22:23,859
wrapper around a specific subject

783
00:22:23,860 --> 00:22:24,969
or topic.

784
00:22:24,970 --> 00:22:26,799
So I set up a project

785
00:22:26,800 --> 00:22:28,749
called Koala, which is the

786
00:22:28,750 --> 00:22:30,069
name of the application that I

787
00:22:30,070 --> 00:22:31,989
built, and I gave it some custom

788
00:22:31,990 --> 00:22:34,149
instructions. You are my technical

789
00:22:34,150 --> 00:22:36,069
co-founder and sure

790
00:22:36,070 --> 00:22:37,629
I am. Got a clue what I'm doing.

791
00:22:37,630 --> 00:22:39,429
I've got an idea for an app, but I

792
00:22:39,430 --> 00:22:41,739
want you to be my co-founder

793
00:22:41,740 --> 00:22:43,689
and to help me to architect it, to

794
00:22:43,690 --> 00:22:45,729
help me structure the project,

795
00:22:45,730 --> 00:22:47,589
to help me to understand what

796
00:22:47,590 --> 00:22:48,909
tech stack I should use.

797
00:22:48,910 --> 00:22:50,109
And I'm going to come back to you

798
00:22:50,110 --> 00:22:51,969
with questions and you should

799
00:22:51,970 --> 00:22:53,829
push back on me if what I'm

800
00:22:53,830 --> 00:22:55,779
asking for is a stupid

801
00:22:55,780 --> 00:22:57,909
idea. And our goal here

802
00:22:57,910 --> 00:23:00,069
is just to get to an MVP

803
00:23:00,070 --> 00:23:01,479
where we can get some paying

804
00:23:01,480 --> 00:23:02,469
customers.

805
00:23:02,470 --> 00:23:04,299
So there's more that goes into it

806
00:23:04,300 --> 00:23:05,469
than that. But that was basically

807
00:23:05,470 --> 00:23:07,179
the custom instruction.

808
00:23:07,180 --> 00:23:09,219
And then as I started chatting

809
00:23:09,220 --> 00:23:11,319
with Claude, I would build out

810
00:23:11,320 --> 00:23:12,609
Google Docs. That might be the

811
00:23:12,610 --> 00:23:14,409
project overview, it might be the

812
00:23:14,410 --> 00:23:16,119
file structure, it might be the

813
00:23:16,120 --> 00:23:17,949
licensing model, it might

814
00:23:17,950 --> 00:23:19,869
be some of the go to market.

815
00:23:19,870 --> 00:23:21,699
And all of this becomes

816
00:23:21,700 --> 00:23:23,769
the context that is in this flawed

817
00:23:23,770 --> 00:23:24,819
project.

818
00:23:24,820 --> 00:23:26,649
And my goodness, I chat

819
00:23:26,650 --> 00:23:28,509
to this thing all day long

820
00:23:28,510 --> 00:23:29,739
at the weekends because this is a

821
00:23:29,740 --> 00:23:31,779
weekend project from eight

822
00:23:31,780 --> 00:23:33,249
in the morning I'm chatting to

823
00:23:33,250 --> 00:23:34,589
Claude, say, Hey, Claude, you know

824
00:23:34,590 --> 00:23:35,439
what you think about this. What do

825
00:23:35,440 --> 00:23:36,609
you think about that?

826
00:23:36,610 --> 00:23:38,739
And one of the great things

827
00:23:38,740 --> 00:23:40,869
about having Claude as a technical

828
00:23:40,870 --> 00:23:43,059
co-founder is you can ask

829
00:23:43,060 --> 00:23:44,769
stupid questions.

830
00:23:44,770 --> 00:23:45,969
You can ask the question that you

831
00:23:45,970 --> 00:23:47,799
asked Claude yesterday, and, you

832
00:23:47,800 --> 00:23:49,209
know, he doesn't complain.

833
00:23:49,210 --> 00:23:50,379
And because you give him the

834
00:23:50,380 --> 00:23:52,089
instructions, he'll push back and

835
00:23:52,090 --> 00:23:53,499
say, I wouldn't do it like that.

836
00:23:53,500 --> 00:23:54,879
Keep focused on the fact that you're

837
00:23:54,880 --> 00:23:55,929
building an MVP.

838
00:23:55,930 --> 00:23:57,639
So I wouldn't do that right now.

839
00:23:57,640 --> 00:23:59,499
What about might be so Claude

840
00:23:59,500 --> 00:24:01,119
was sort of the first, you know,

841
00:24:01,120 --> 00:24:02,799
build out the architecture, build

842
00:24:02,800 --> 00:24:04,659
out the project plan, build

843
00:24:04,660 --> 00:24:06,639
out the user stories and so

844
00:24:06,640 --> 00:24:08,679
on. Next step is going

845
00:24:08,680 --> 00:24:10,869
over to this tool called V zero.

846
00:24:10,870 --> 00:24:13,359
It's an AI powered generative

847
00:24:13,360 --> 00:24:14,859
UI tool.

848
00:24:14,860 --> 00:24:16,779
So here you set up a new project

849
00:24:16,780 --> 00:24:18,819
into which I upload

850
00:24:18,820 --> 00:24:20,199
the context which has come from

851
00:24:20,200 --> 00:24:21,039
code.

852
00:24:21,040 --> 00:24:23,409
So Claude's given me this whole page

853
00:24:23,410 --> 00:24:25,599
product requirements document

854
00:24:25,600 --> 00:24:27,549
that goes straight into V zero,

855
00:24:27,550 --> 00:24:29,709
and now I can start, you know, can

856
00:24:29,710 --> 00:24:31,599
you design the

857
00:24:31,600 --> 00:24:33,129
homescreen for me?

858
00:24:33,130 --> 00:24:35,379
I want these sections.

859
00:24:35,380 --> 00:24:37,539
Here is my brand colors.

860
00:24:37,540 --> 00:24:39,489
Here's I want a dark theme.

861
00:24:39,490 --> 00:24:40,539
And without me having to write a

862
00:24:40,540 --> 00:24:42,039
very long prompt because it's

863
00:24:42,040 --> 00:24:44,799
already got the entire payload,

864
00:24:44,800 --> 00:24:47,229
the outcomes, the sidebar

865
00:24:47,230 --> 00:24:48,879
outcomes, the project, upload

866
00:24:48,880 --> 00:24:50,829
document outcomes, the chat screen.

867
00:24:50,830 --> 00:24:52,749
I'm not were even better

868
00:24:52,750 --> 00:24:53,739
than that.

869
00:24:53,740 --> 00:24:55,659
What you can put into V zero is

870
00:24:55,660 --> 00:24:57,549
screenshots of other

871
00:24:57,550 --> 00:24:59,469
apps that you like and say

872
00:24:59,470 --> 00:25:01,449
I want that but in

873
00:25:01,450 --> 00:25:03,039
my project context.

874
00:25:03,040 --> 00:25:05,079
So why do I need to reinvent

875
00:25:05,080 --> 00:25:07,149
the wheel? Code and chat

876
00:25:07,150 --> 00:25:09,159
have got the kind of chat interface

877
00:25:09,160 --> 00:25:10,239
fairly well nailed.

878
00:25:10,240 --> 00:25:12,159
If my product did that, I'd be

879
00:25:12,160 --> 00:25:13,809
happy. Screenshot.

880
00:25:13,810 --> 00:25:15,519
Drop into V0.

881
00:25:15,520 --> 00:25:17,169
Give me like that.

882
00:25:17,170 --> 00:25:19,209
And so as a non developer

883
00:25:19,210 --> 00:25:21,099
or non designer, I should say

884
00:25:21,100 --> 00:25:23,169
at this point, you're now

885
00:25:23,170 --> 00:25:25,089
designing screens that are rapid,

886
00:25:25,090 --> 00:25:25,869
right?

887
00:25:25,870 --> 00:25:28,629
And not only does it do the design,

888
00:25:28,630 --> 00:25:30,759
but it then creates all of the

889
00:25:30,760 --> 00:25:32,799
code, whether that's TypeScript

890
00:25:32,800 --> 00:25:35,139
JavaScript in the back end.

891
00:25:35,140 --> 00:25:37,269
So how much did you spend

892
00:25:37,270 --> 00:25:39,789
on it to build

893
00:25:39,790 --> 00:25:41,079
koala?

894
00:25:41,080 --> 00:25:43,119
So when GPT

895
00:25:43,120 --> 00:25:45,159
launched and they pulled a number

896
00:25:45,160 --> 00:25:47,259
out of thin air of $20

897
00:25:47,260 --> 00:25:49,089
a month and as Sam Altman

898
00:25:49,090 --> 00:25:50,559
has said, we basically pulled a

899
00:25:50,560 --> 00:25:51,759
number of out of thin air.

900
00:25:51,760 --> 00:25:53,709
We're going to charge $20 a month.

901
00:25:53,710 --> 00:25:55,089
Well, that's what everyone else has

902
00:25:55,090 --> 00:25:57,399
had to follow. So clod is $20

903
00:25:57,400 --> 00:25:59,289
a month. I pay $20 a month

904
00:25:59,290 --> 00:26:00,579
for GPT.

905
00:26:00,580 --> 00:26:02,949
And I use them both interchangeably

906
00:26:02,950 --> 00:26:04,029
throughout the day.

907
00:26:04,030 --> 00:26:07,119
Rappler, I think, was $120

908
00:26:07,120 --> 00:26:09,339
for a year's license to call

909
00:26:09,340 --> 00:26:11,799
V0 $20 a month.

910
00:26:11,800 --> 00:26:13,829
Kazu I think I pay maybe 30

911
00:26:13,830 --> 00:26:15,759
or $40 a month and

912
00:26:15,760 --> 00:26:16,899
then I'm paying for the others

913
00:26:16,900 --> 00:26:17,929
because I'm testing them out.

914
00:26:17,930 --> 00:26:19,479
That loveable as well.

915
00:26:19,480 --> 00:26:21,489
But they're all in this order.

916
00:26:21,490 --> 00:26:22,569
A couple of things that are a little

917
00:26:22,570 --> 00:26:23,679
bit interesting.

918
00:26:23,680 --> 00:26:25,359
Just when people are thinking about

919
00:26:25,360 --> 00:26:27,429
pricing models for these things, two

920
00:26:27,430 --> 00:26:28,479
things that I've seen are quite

921
00:26:28,480 --> 00:26:30,939
interesting v zero

922
00:26:30,940 --> 00:26:32,619
and lovable.

923
00:26:32,620 --> 00:26:34,839
You can set up a free account.

924
00:26:34,840 --> 00:26:36,699
But when you have a free account,

925
00:26:36,700 --> 00:26:39,099
your designs are public

926
00:26:39,100 --> 00:26:41,319
and so you pay to go private.

927
00:26:41,320 --> 00:26:44,139
And I remember when I was first

928
00:26:44,140 --> 00:26:46,419
designing some of the UI screens

929
00:26:46,420 --> 00:26:48,219
for Koala, I'm sure no one would

930
00:26:48,220 --> 00:26:49,419
have been interested in what I was

931
00:26:49,420 --> 00:26:51,159
doing. But for me, I was.

932
00:26:51,160 --> 00:26:53,289
I felt compelled to make this thing

933
00:26:53,290 --> 00:26:54,339
private.

934
00:26:54,340 --> 00:26:55,539
I'm not seeing that before.

935
00:26:55,540 --> 00:26:56,529
You could imagine that for

936
00:26:56,530 --> 00:26:58,209
Salesforce or HubSpot, you can use

937
00:26:58,210 --> 00:26:59,829
it for free, but your contact

938
00:26:59,830 --> 00:27:00,859
records are public.

939
00:27:00,860 --> 00:27:02,049
Do you want to go private?

940
00:27:02,050 --> 00:27:02,919
So that was an interesting

941
00:27:02,920 --> 00:27:04,299
dimension.

942
00:27:04,300 --> 00:27:06,039
And so you did all of this from

943
00:27:06,040 --> 00:27:08,139
September to Wednesday launch.

944
00:27:08,140 --> 00:27:09,489
December.

945
00:27:09,490 --> 00:27:10,389
December, yeah.

946
00:27:10,390 --> 00:27:12,249
So it was 12 weeks from having

947
00:27:12,250 --> 00:27:14,319
never written a line of code

948
00:27:14,320 --> 00:27:16,659
to having a production app.

949
00:27:16,660 --> 00:27:18,589
I would say that and this

950
00:27:18,590 --> 00:27:20,499
was all done during the weekends and

951
00:27:20,500 --> 00:27:22,419
a few evenings.

952
00:27:22,420 --> 00:27:24,249
There were three weekends

953
00:27:24,250 --> 00:27:26,289
when I literally broke

954
00:27:26,290 --> 00:27:28,809
down in tears in front of my wife

955
00:27:28,810 --> 00:27:30,759
and I was like, Who

956
00:27:30,760 --> 00:27:31,719
am I kidding?

957
00:27:31,720 --> 00:27:33,349
I don't know what I'm talking about.

958
00:27:33,350 --> 00:27:34,869
I should stick to doing what I'm

959
00:27:34,870 --> 00:27:35,859
good at.

960
00:27:35,860 --> 00:27:37,069
You know, what was I thinking?

961
00:27:37,070 --> 00:27:38,079
Like, there's a reason why

962
00:27:38,080 --> 00:27:39,699
developers get paid a ton of money

963
00:27:39,700 --> 00:27:41,679
is because I this is complicated

964
00:27:41,680 --> 00:27:43,929
and I don't know what I'm doing.

965
00:27:43,930 --> 00:27:45,759
And this is one of the problems

966
00:27:45,760 --> 00:27:47,829
with AI development tools,

967
00:27:47,830 --> 00:27:49,539
is that, yes, it gives you this

968
00:27:49,540 --> 00:27:51,999
acceleration from not to

969
00:27:52,000 --> 00:27:53,859
amateur, but when something

970
00:27:53,860 --> 00:27:55,809
goes wrong, you don't know

971
00:27:55,810 --> 00:27:57,249
what's gone wrong because you don't

972
00:27:57,250 --> 00:27:59,349
actually understand at the start

973
00:27:59,350 --> 00:28:00,819
what you're looking at.

974
00:28:00,820 --> 00:28:02,979
And I was trying to deal with quite

975
00:28:02,980 --> 00:28:05,889
complex user authentication

976
00:28:05,890 --> 00:28:07,809
through a third party and I just

977
00:28:07,810 --> 00:28:09,819
couldn't understand why when certain

978
00:28:09,820 --> 00:28:10,779
people were logging in, they

979
00:28:10,780 --> 00:28:12,639
couldn't see certain things.

980
00:28:12,640 --> 00:28:14,719
And I literally broke down in tears

981
00:28:14,720 --> 00:28:15,849
with my wife going, I don't know

982
00:28:15,850 --> 00:28:17,799
what I'm doing. And she'd say, Go

983
00:28:17,800 --> 00:28:19,329
for a walk or go for a run.

984
00:28:19,330 --> 00:28:20,889
That always helps you listen to a

985
00:28:20,890 --> 00:28:21,999
podcast.

986
00:28:22,000 --> 00:28:24,129
And so I'd put on a Lenny Chatzky

987
00:28:24,130 --> 00:28:26,619
podcast and I would come back

988
00:28:26,620 --> 00:28:28,539
inspired and I got I'll

989
00:28:28,540 --> 00:28:30,369
just try this one thing,

990
00:28:30,370 --> 00:28:32,499
and that one thing always solved

991
00:28:32,500 --> 00:28:34,179
it. And I was like, Yes, I'm back in

992
00:28:34,180 --> 00:28:36,229
the game and then would carry on.

993
00:28:36,230 --> 00:28:38,109
So I finished each weekend on

994
00:28:38,110 --> 00:28:38,949
a high.

995
00:28:38,950 --> 00:28:41,019
But Saturday afternoons were my

996
00:28:41,020 --> 00:28:42,639
my low points.

997
00:28:42,640 --> 00:28:44,169
There's some really interesting

998
00:28:44,170 --> 00:28:46,359
conversations happening on

999
00:28:46,360 --> 00:28:48,789
X or on LinkedIn

1000
00:28:48,790 --> 00:28:51,009
where you've got real

1001
00:28:51,010 --> 00:28:52,839
developers that actually

1002
00:28:52,840 --> 00:28:54,369
know what it takes to build a

1003
00:28:54,370 --> 00:28:56,589
production enterprise application

1004
00:28:56,590 --> 00:28:58,149
and backend infrastructure and

1005
00:28:58,150 --> 00:29:00,339
scaling versus

1006
00:29:00,340 --> 00:29:02,229
this whole new breed of

1007
00:29:02,230 --> 00:29:04,119
indie hackers or

1008
00:29:04,120 --> 00:29:05,739
non-technical founders that are now

1009
00:29:05,740 --> 00:29:07,719
spinning up these PCs

1010
00:29:07,720 --> 00:29:08,979
and the people that have been doing

1011
00:29:08,980 --> 00:29:10,449
this for a long time and actually

1012
00:29:10,450 --> 00:29:11,599
know what they're doing.

1013
00:29:11,600 --> 00:29:13,299
They're like, You think that's

1014
00:29:13,300 --> 00:29:14,769
development, That's not development,

1015
00:29:14,770 --> 00:29:15,999
You haven't got a clue.

1016
00:29:16,000 --> 00:29:18,159
And the same thing happened

1017
00:29:18,160 --> 00:29:20,169
when Canva came out

1018
00:29:20,170 --> 00:29:22,389
and real designers who are using

1019
00:29:22,390 --> 00:29:24,519
Photoshop or whatever are looking

1020
00:29:24,520 --> 00:29:26,019
at people using camera going, that's

1021
00:29:26,020 --> 00:29:27,399
not real design.

1022
00:29:27,400 --> 00:29:29,649
And I know this trend

1023
00:29:29,650 --> 00:29:31,599
will continue, but

1024
00:29:31,600 --> 00:29:33,219
the tools will improve.

1025
00:29:33,220 --> 00:29:34,749
Like I said, I listen to this Lenny

1026
00:29:34,750 --> 00:29:37,059
Rich Jet Ski podcast and

1027
00:29:37,060 --> 00:29:39,069
always out of my runs and there's

1028
00:29:39,070 --> 00:29:41,139
two episodes or two sort of sound

1029
00:29:41,140 --> 00:29:42,999
bites that really stuck with

1030
00:29:43,000 --> 00:29:43,989
me when I was going through this

1031
00:29:43,990 --> 00:29:44,949
process.

1032
00:29:44,950 --> 00:29:46,779
The first and I'll

1033
00:29:46,780 --> 00:29:48,279
have to remember his name, he was

1034
00:29:48,280 --> 00:29:50,259
the founder of Right, which

1035
00:29:50,260 --> 00:29:52,269
became Google Docs, and he's

1036
00:29:52,270 --> 00:29:54,399
telling Lenny about the process

1037
00:29:54,400 --> 00:29:56,289
of writing original Google

1038
00:29:56,290 --> 00:29:57,099
Docs.

1039
00:29:57,100 --> 00:29:58,149
And he said, We didn't know what we

1040
00:29:58,150 --> 00:29:59,979
were doing, but throughout my

1041
00:29:59,980 --> 00:30:01,719
entire career, I've just got to the

1042
00:30:01,720 --> 00:30:03,609
edge of what I know and fucked

1043
00:30:03,610 --> 00:30:05,559
around by getting to the edge

1044
00:30:05,560 --> 00:30:07,869
of what you know and fucking around.

1045
00:30:07,870 --> 00:30:09,369
You end up learning stuff.

1046
00:30:09,370 --> 00:30:11,199
If you stay within what you're

1047
00:30:11,200 --> 00:30:12,809
doing, then you don't learn.

1048
00:30:12,810 --> 00:30:14,669
So that was one thing on those

1049
00:30:14,670 --> 00:30:16,019
days when I was like, I don't know

1050
00:30:16,020 --> 00:30:17,099
what I'm doing, I don't know what

1051
00:30:17,100 --> 00:30:18,989
I'm doing. I had his phrase in my

1052
00:30:18,990 --> 00:30:20,159
mind like, No, but this is where

1053
00:30:20,160 --> 00:30:21,239
learning is happening.

1054
00:30:21,240 --> 00:30:23,339
And the second episode that really

1055
00:30:23,340 --> 00:30:25,049
stuck with me was a guy called

1056
00:30:25,050 --> 00:30:26,999
Nikita Baer, who as

1057
00:30:27,000 --> 00:30:29,099
well a number of B2C apps

1058
00:30:29,100 --> 00:30:31,139
that have gone viral and be bought

1059
00:30:31,140 --> 00:30:32,549
by Facebook matter.

1060
00:30:32,550 --> 00:30:34,859
I think 1 or 2 of them,

1061
00:30:34,860 --> 00:30:36,749
he said this phrase stuck with me is

1062
00:30:36,750 --> 00:30:38,039
like, you know, one of the things no

1063
00:30:38,040 --> 00:30:39,899
one ever tells you is the moment you

1064
00:30:39,900 --> 00:30:41,939
hit virality, you have to build

1065
00:30:41,940 --> 00:30:44,279
the whole thing again because

1066
00:30:44,280 --> 00:30:46,139
it isn't built for the scale you

1067
00:30:46,140 --> 00:30:48,329
need. And I took the positive

1068
00:30:48,330 --> 00:30:49,439
in that. I was like, So what you're

1069
00:30:49,440 --> 00:30:51,719
telling me is I don't need to build

1070
00:30:51,720 --> 00:30:53,189
an enterprise solution.

1071
00:30:53,190 --> 00:30:54,749
I just need to build something that

1072
00:30:54,750 --> 00:30:56,579
gets to the first 5 or 10

1073
00:30:56,580 --> 00:30:58,949
users and solves that problem.

1074
00:30:58,950 --> 00:31:00,929
And if I can get that far

1075
00:31:00,930 --> 00:31:02,759
and then prove the value in the use

1076
00:31:02,760 --> 00:31:04,679
case, then I can get proper

1077
00:31:04,680 --> 00:31:05,759
people that actually know what

1078
00:31:05,760 --> 00:31:07,589
they're doing and then build for the

1079
00:31:07,590 --> 00:31:08,609
next stage.

1080
00:31:08,610 --> 00:31:10,529
So that gave me this kind of license

1081
00:31:10,530 --> 00:31:12,389
to like, I'm not purporting to be

1082
00:31:12,390 --> 00:31:13,889
an enterprise developer that's going

1083
00:31:13,890 --> 00:31:15,779
to build a SOC two compliant,

1084
00:31:15,780 --> 00:31:17,609
scalable thing, but I don't need

1085
00:31:17,610 --> 00:31:18,689
to build that yet.

1086
00:31:18,690 --> 00:31:20,549
I just need to prove that I

1087
00:31:20,550 --> 00:31:22,409
can solve a problem for someone and

1088
00:31:22,410 --> 00:31:23,699
that they would come and log in.

1089
00:31:23,700 --> 00:31:25,799
And if I can prove that, then solve

1090
00:31:25,800 --> 00:31:27,809
the next stage later on.

1091
00:31:27,810 --> 00:31:29,579
So this is me pushing your LinkedIn

1092
00:31:29,580 --> 00:31:31,529
content, but there's probably

1093
00:31:31,530 --> 00:31:33,419
a piece of content

1094
00:31:33,420 --> 00:31:35,279
a week that you write that

1095
00:31:35,280 --> 00:31:37,049
I end up sharing internally.

1096
00:31:37,050 --> 00:31:38,819
And this week's piece of content was

1097
00:31:38,820 --> 00:31:40,679
around public versus

1098
00:31:40,680 --> 00:31:42,629
private data and information

1099
00:31:42,630 --> 00:31:44,429
with chat shaped.

1100
00:31:44,430 --> 00:31:46,469
So you want to talk about that?

1101
00:31:46,470 --> 00:31:48,569
One of the things that I'm

1102
00:31:48,570 --> 00:31:50,399
always having to remind myself

1103
00:31:50,400 --> 00:31:52,289
as someone that would now consider

1104
00:31:52,290 --> 00:31:54,839
myself to be in the AI bubble

1105
00:31:54,840 --> 00:31:57,059
is that the majority of people

1106
00:31:57,060 --> 00:31:59,009
are not in the AI bubble.

1107
00:31:59,010 --> 00:32:00,389
They're just, you know, going to

1108
00:32:00,390 --> 00:32:02,279
work doing the work the same way

1109
00:32:02,280 --> 00:32:04,289
they have. Maybe

1110
00:32:04,290 --> 00:32:05,819
they've asked a couple of questions,

1111
00:32:05,820 --> 00:32:07,349
you know, what is or how should I do

1112
00:32:07,350 --> 00:32:08,350
that?

1113
00:32:09,030 --> 00:32:10,199
But they've not really embedded it

1114
00:32:10,200 --> 00:32:12,089
in their work.

1115
00:32:12,090 --> 00:32:13,919
And to that extent,

1116
00:32:13,920 --> 00:32:15,779
a lot of the companies I speak to

1117
00:32:15,780 --> 00:32:18,239
either have not got an AI policy

1118
00:32:18,240 --> 00:32:20,099
or they've got quite a

1119
00:32:20,100 --> 00:32:22,049
pessimistic and restrictive

1120
00:32:22,050 --> 00:32:22,899
AI policy.

1121
00:32:22,900 --> 00:32:24,929
You know, you must not use GPT

1122
00:32:24,930 --> 00:32:26,309
on your company laptop.

1123
00:32:26,310 --> 00:32:28,619
You must now upload company data

1124
00:32:28,620 --> 00:32:30,419
to GBG.

1125
00:32:30,420 --> 00:32:32,369
And definitely because

1126
00:32:32,370 --> 00:32:34,559
I've come across almost no companies

1127
00:32:34,560 --> 00:32:36,719
that have this, they have not got a

1128
00:32:36,720 --> 00:32:38,849
team or enterprise account for

1129
00:32:38,850 --> 00:32:40,859
GPT or Clod.

1130
00:32:40,860 --> 00:32:42,119
So if you're asking people how are

1131
00:32:42,120 --> 00:32:43,649
you using it, I say either I've got

1132
00:32:43,650 --> 00:32:45,539
a free account or I pay for the pro

1133
00:32:45,540 --> 00:32:47,429
account on my own personal

1134
00:32:47,430 --> 00:32:48,209
card.

1135
00:32:48,210 --> 00:32:50,069
Now, this is where the data

1136
00:32:50,070 --> 00:32:52,289
privacy issues come in.

1137
00:32:52,290 --> 00:32:54,869
So if you're on a free

1138
00:32:54,870 --> 00:32:56,969
or a paid personal

1139
00:32:56,970 --> 00:32:59,129
account for GPT

1140
00:32:59,130 --> 00:33:01,049
by default chat

1141
00:33:01,050 --> 00:33:03,119
GPT open, I can

1142
00:33:03,120 --> 00:33:05,039
use what you upload in

1143
00:33:05,040 --> 00:33:06,119
terms of your content.

1144
00:33:06,120 --> 00:33:07,589
So that can either be what you write

1145
00:33:07,590 --> 00:33:09,539
in the chat or any files that you

1146
00:33:09,540 --> 00:33:12,209
upload that is defined as content.

1147
00:33:12,210 --> 00:33:14,159
They can use that to train the model

1148
00:33:14,160 --> 00:33:15,779
by default.

1149
00:33:15,780 --> 00:33:17,759
You can go into the settings

1150
00:33:17,760 --> 00:33:19,739
and that is a

1151
00:33:19,740 --> 00:33:21,159
I love the way they phrased it.

1152
00:33:21,160 --> 00:33:22,889
The setting does not say train our

1153
00:33:22,890 --> 00:33:23,669
models.

1154
00:33:23,670 --> 00:33:25,619
The setting is would you like to

1155
00:33:25,620 --> 00:33:27,839
help improve the model for everyone?

1156
00:33:27,840 --> 00:33:29,189
It's a really nice one.

1157
00:33:29,190 --> 00:33:30,629
I would like to help improve the

1158
00:33:30,630 --> 00:33:31,859
model. It doesn't say what that

1159
00:33:31,860 --> 00:33:33,089
actually means.

1160
00:33:33,090 --> 00:33:35,069
Your data then goes into

1161
00:33:35,070 --> 00:33:36,359
training the model by name.

1162
00:33:36,360 --> 00:33:38,129
You can turn that off.

1163
00:33:38,130 --> 00:33:39,989
But that is you have to opt out

1164
00:33:39,990 --> 00:33:41,939
of that and out of the people

1165
00:33:41,940 --> 00:33:43,829
that I speak to, No one does

1166
00:33:43,830 --> 00:33:44,999
that because I haven't thought about

1167
00:33:45,000 --> 00:33:47,009
it. Most people are low tech users

1168
00:33:47,010 --> 00:33:48,299
of it. They just start using it and

1169
00:33:48,300 --> 00:33:50,879
uploading things on the teams.

1170
00:33:50,880 --> 00:33:52,799
This is GPT specifically when

1171
00:33:52,800 --> 00:33:54,599
I say teams or enterprise, but the

1172
00:33:54,600 --> 00:33:56,519
same thing on code as well.

1173
00:33:56,520 --> 00:33:58,349
If you are on one of those accounts

1174
00:33:58,350 --> 00:34:00,449
and by default, your

1175
00:34:00,450 --> 00:34:02,519
content is not used

1176
00:34:02,520 --> 00:34:04,469
to train the model unless you

1177
00:34:04,470 --> 00:34:06,389
explicitly opt in as part of

1178
00:34:06,390 --> 00:34:08,069
their feedback model.

1179
00:34:08,070 --> 00:34:09,959
So where I find this is quite

1180
00:34:09,960 --> 00:34:11,819
amusing in a way that the

1181
00:34:11,820 --> 00:34:14,009
companies that are most risk averse

1182
00:34:14,010 --> 00:34:17,069
that say, you know, you must not use

1183
00:34:17,070 --> 00:34:18,718
AI and we're certainly not going to

1184
00:34:18,719 --> 00:34:20,759
pay for an account for you to have

1185
00:34:20,760 --> 00:34:22,678
actually push everything

1186
00:34:22,679 --> 00:34:24,599
under the radar and under the carpet

1187
00:34:24,600 --> 00:34:26,069
to people that are still doing the

1188
00:34:26,070 --> 00:34:26,939
work.

1189
00:34:26,940 --> 00:34:28,379
But they're using a personal account

1190
00:34:28,380 --> 00:34:30,299
which by default gets all of your

1191
00:34:30,300 --> 00:34:32,218
company data up into

1192
00:34:32,219 --> 00:34:33,089
these models.

1193
00:34:33,090 --> 00:34:34,919
And so this is no

1194
00:34:34,920 --> 00:34:36,809
time to be an ostrich with

1195
00:34:36,810 --> 00:34:38,189
your head in the sand.

1196
00:34:38,190 --> 00:34:39,119
La la la la la.

1197
00:34:39,120 --> 00:34:39,988
Nothing's happening like it's

1198
00:34:39,989 --> 00:34:41,999
happening. People are

1199
00:34:42,000 --> 00:34:43,948
using this and you

1200
00:34:43,949 --> 00:34:45,658
can choose to ignore it, in which

1201
00:34:45,659 --> 00:34:47,638
case you create a data risk

1202
00:34:47,639 --> 00:34:49,289
or you can choose to embrace it.

1203
00:34:49,290 --> 00:34:51,269
And actually the default

1204
00:34:51,270 --> 00:34:53,099
is that your data is not going to

1205
00:34:53,100 --> 00:34:55,138
be used to train the models.

1206
00:34:55,139 --> 00:34:57,089
So the question of that

1207
00:34:57,090 --> 00:34:59,279
and this idea that, hey,

1208
00:34:59,280 --> 00:35:01,139
I'm a CEO and I'm

1209
00:35:01,140 --> 00:35:02,999
coming into this organization,

1210
00:35:03,000 --> 00:35:05,399
venture capital, back to 250 people,

1211
00:35:05,400 --> 00:35:07,229
and I want to have a real I

1212
00:35:07,230 --> 00:35:08,339
push in this organization

1213
00:35:08,340 --> 00:35:10,259
holistically across the company.

1214
00:35:10,260 --> 00:35:12,179
So the question to you as a CIO,

1215
00:35:12,180 --> 00:35:13,759
is it? What is it that I should do

1216
00:35:13,760 --> 00:35:15,319
or what should I think about if I

1217
00:35:15,320 --> 00:35:16,459
really want to transform this

1218
00:35:16,460 --> 00:35:17,719
organization?

1219
00:35:17,720 --> 00:35:19,909
Some of the most common first

1220
00:35:19,910 --> 00:35:21,829
use cases when you talk

1221
00:35:21,830 --> 00:35:23,959
to companies about using AI.

1222
00:35:23,960 --> 00:35:24,889
It comes down to sort of

1223
00:35:24,890 --> 00:35:27,619
productivity efficiency.

1224
00:35:27,620 --> 00:35:29,449
And there are two ways that you

1225
00:35:29,450 --> 00:35:30,859
can approach that.

1226
00:35:30,860 --> 00:35:33,379
So if you imagine you've got a 5000

1227
00:35:33,380 --> 00:35:35,359
person company and you can say, if

1228
00:35:35,360 --> 00:35:37,279
we use AI, we could be way

1229
00:35:37,280 --> 00:35:38,209
more efficient.

1230
00:35:38,210 --> 00:35:39,469
And therefore, instead of having

1231
00:35:39,470 --> 00:35:41,389
5000 people, we could

1232
00:35:41,390 --> 00:35:43,049
have 2000 people.

1233
00:35:43,050 --> 00:35:45,379
And so that's quite a defensive

1234
00:35:45,380 --> 00:35:47,149
cost saving increase.

1235
00:35:47,150 --> 00:35:49,639
Our margins way of looking at it.

1236
00:35:49,640 --> 00:35:52,039
The other is to say we're 5000

1237
00:35:52,040 --> 00:35:54,139
people. But actually, if we used

1238
00:35:54,140 --> 00:35:56,209
AI, we can immediately act

1239
00:35:56,210 --> 00:35:58,069
like a company of 50,000

1240
00:35:58,070 --> 00:36:00,019
people by giving everyone that we've

1241
00:36:00,020 --> 00:36:02,089
already got just way

1242
00:36:02,090 --> 00:36:03,619
more capability.

1243
00:36:03,620 --> 00:36:05,299
Benioff was talking about there not

1244
00:36:05,300 --> 00:36:07,309
hiring any more software

1245
00:36:07,310 --> 00:36:09,169
engineers this year

1246
00:36:09,170 --> 00:36:10,399
because everyone has a sort of wry

1247
00:36:10,400 --> 00:36:12,559
smile. Benioff The great marketer,

1248
00:36:12,560 --> 00:36:14,539
by his point is we don't need

1249
00:36:14,540 --> 00:36:16,639
to hire any more engineers

1250
00:36:16,640 --> 00:36:18,049
because we can get the engineers.

1251
00:36:18,050 --> 00:36:20,239
We've got to be way more productive.

1252
00:36:20,240 --> 00:36:21,619
And you're seeing that across so

1253
00:36:21,620 --> 00:36:23,239
many companies.

1254
00:36:23,240 --> 00:36:25,009
So the first thing that I'd be

1255
00:36:25,010 --> 00:36:27,139
canceling any senior executive

1256
00:36:27,140 --> 00:36:29,329
is to say, right, just imagine

1257
00:36:29,330 --> 00:36:31,219
that you could supercharge the

1258
00:36:31,220 --> 00:36:33,469
people that you've got and

1259
00:36:33,470 --> 00:36:35,539
get them to be way more capable

1260
00:36:35,540 --> 00:36:37,189
to be able to work in different

1261
00:36:37,190 --> 00:36:38,899
countries in different domains, to

1262
00:36:38,900 --> 00:36:40,669
know what way more than they do

1263
00:36:40,670 --> 00:36:41,689
today.

1264
00:36:41,690 --> 00:36:43,849
How would you approach that?

1265
00:36:43,850 --> 00:36:46,699
And a lot of this comes back to

1266
00:36:46,700 --> 00:36:48,289
not trying to do anything overly

1267
00:36:48,290 --> 00:36:50,179
complex. So even a couple

1268
00:36:50,180 --> 00:36:51,709
of years ago, when you talk about

1269
00:36:51,710 --> 00:36:53,449
AI, people are, you know, we're

1270
00:36:53,450 --> 00:36:55,159
going to use AI to completely

1271
00:36:55,160 --> 00:36:57,289
disrupt our supply chain

1272
00:36:57,290 --> 00:36:59,179
or we're going to use AI

1273
00:36:59,180 --> 00:37:00,529
to completely rebuild our

1274
00:37:00,530 --> 00:37:02,449
forecasting of footfall in a retail

1275
00:37:02,450 --> 00:37:03,439
environment or something.

1276
00:37:03,440 --> 00:37:04,699
It's quite a complex thing.

1277
00:37:04,700 --> 00:37:06,469
Machine learning, data science.

1278
00:37:06,470 --> 00:37:08,539
The big win right now is just get

1279
00:37:08,540 --> 00:37:10,459
every single person that works

1280
00:37:10,460 --> 00:37:11,719
in h.R.

1281
00:37:11,720 --> 00:37:13,009
Is a financial analyst.

1282
00:37:13,010 --> 00:37:14,809
That is a marketer that works in

1283
00:37:14,810 --> 00:37:16,279
customer support.

1284
00:37:16,280 --> 00:37:18,469
They have work that they have to do

1285
00:37:18,470 --> 00:37:19,789
as a human being.

1286
00:37:19,790 --> 00:37:21,469
And i sort of split this down into

1287
00:37:21,470 --> 00:37:22,279
three buckets.

1288
00:37:22,280 --> 00:37:23,809
There's there's work that you

1289
00:37:23,810 --> 00:37:25,279
receive from other people, so

1290
00:37:25,280 --> 00:37:27,049
someone gives you some requirements.

1291
00:37:27,050 --> 00:37:28,879
A customer gives you a request for

1292
00:37:28,880 --> 00:37:29,929
proposal.

1293
00:37:29,930 --> 00:37:31,969
Someone gives you a quote

1294
00:37:31,970 --> 00:37:33,289
or a contract.

1295
00:37:33,290 --> 00:37:35,299
Maybe there's a new regulation or

1296
00:37:35,300 --> 00:37:37,249
new compliance rules.

1297
00:37:37,250 --> 00:37:38,239
There's something that someone else

1298
00:37:38,240 --> 00:37:39,379
has given you.

1299
00:37:39,380 --> 00:37:41,269
And as a human being, you have to

1300
00:37:41,270 --> 00:37:43,099
understand that, analyze it,

1301
00:37:43,100 --> 00:37:45,049
summarize it, work out, what

1302
00:37:45,050 --> 00:37:46,279
are the most important bits and the

1303
00:37:46,280 --> 00:37:47,219
risks.

1304
00:37:47,220 --> 00:37:48,949
That's work that's given to you.

1305
00:37:48,950 --> 00:37:50,389
You've then got the work that you

1306
00:37:50,390 --> 00:37:52,309
actually do as a person.

1307
00:37:52,310 --> 00:37:54,139
So, you know, I need to

1308
00:37:54,140 --> 00:37:55,429
research accounts.

1309
00:37:55,430 --> 00:37:57,169
I need to prepare a proposal.

1310
00:37:57,170 --> 00:37:59,929
I need to manage a direct report.

1311
00:37:59,930 --> 00:38:01,069
Those are the things that you have

1312
00:38:01,070 --> 00:38:02,689
to do yourself.

1313
00:38:02,690 --> 00:38:04,339
And then there's work that you pass

1314
00:38:04,340 --> 00:38:05,929
on to other people.

1315
00:38:05,930 --> 00:38:07,579
So I need to create a proposal.

1316
00:38:07,580 --> 00:38:09,259
I need to create some requirements.

1317
00:38:09,260 --> 00:38:11,239
I need to create a summarization

1318
00:38:11,240 --> 00:38:12,739
for my manager.

1319
00:38:12,740 --> 00:38:14,359
I need to create an investor

1320
00:38:14,360 --> 00:38:16,369
briefing, whatever that might be.

1321
00:38:16,370 --> 00:38:18,169
And so these are very human things

1322
00:38:18,170 --> 00:38:19,519
that you have to do.

1323
00:38:19,520 --> 00:38:21,619
And how can you bolt on

1324
00:38:21,620 --> 00:38:23,509
the skills where they are to be able

1325
00:38:23,510 --> 00:38:25,819
to accelerate that much

1326
00:38:25,820 --> 00:38:28,429
more capable at ingesting

1327
00:38:28,430 --> 00:38:30,109
information and new requirements,

1328
00:38:30,110 --> 00:38:32,149
much more effective at doing the job

1329
00:38:32,150 --> 00:38:33,889
that I've got to do and a much

1330
00:38:33,890 --> 00:38:35,839
higher quality of output and

1331
00:38:35,840 --> 00:38:37,459
speed of output that I'm delivering

1332
00:38:37,460 --> 00:38:38,809
to other people.

1333
00:38:38,810 --> 00:38:40,759
And in it, you talk

1334
00:38:40,760 --> 00:38:42,799
about the ten x developer or the ten

1335
00:38:42,800 --> 00:38:44,539
x engineer, you know, what is the

1336
00:38:44,540 --> 00:38:45,949
ten x h.r.

1337
00:38:45,950 --> 00:38:46,909
Admin. What's that?

1338
00:38:46,910 --> 00:38:48,409
Ten x customer support

1339
00:38:48,410 --> 00:38:49,639
representative.

1340
00:38:49,640 --> 00:38:51,109
There's people that are kind of

1341
00:38:51,110 --> 00:38:52,969
biochemically enhancing themselves

1342
00:38:52,970 --> 00:38:54,859
with these tools suddenly

1343
00:38:54,860 --> 00:38:57,019
way more capable than the person

1344
00:38:57,020 --> 00:38:58,909
that sat on the desk right next to

1345
00:38:58,910 --> 00:38:59,689
them.

1346
00:38:59,690 --> 00:39:01,279
So I think that's a great vision.

1347
00:39:01,280 --> 00:39:02,269
There are a lot of people where all

1348
00:39:02,270 --> 00:39:03,709
CEOs, we're like, Yeah, okay,

1349
00:39:03,710 --> 00:39:05,509
whatever vision, how do we do it?

1350
00:39:05,510 --> 00:39:07,399
And dropping it down a level, I can

1351
00:39:07,400 --> 00:39:10,009
explain some of the ways that I use

1352
00:39:10,010 --> 00:39:11,959
cloud or chat type of ended

1353
00:39:11,960 --> 00:39:14,599
up buying Claud instead.

1354
00:39:14,600 --> 00:39:16,879
Because of the ability to

1355
00:39:16,880 --> 00:39:18,649
mimic your writing, I find that

1356
00:39:18,650 --> 00:39:20,779
Claude creates my writing better

1357
00:39:20,780 --> 00:39:22,459
than chatbot, creates my writing out

1358
00:39:22,460 --> 00:39:23,389
of the box, so that's why I

1359
00:39:23,390 --> 00:39:24,679
subscribe to it.

1360
00:39:24,680 --> 00:39:26,509
I use it for things

1361
00:39:26,510 --> 00:39:27,859
when I have to be creative and I'm

1362
00:39:27,860 --> 00:39:29,359
not being very creative.

1363
00:39:29,360 --> 00:39:32,359
So I want an image

1364
00:39:32,360 --> 00:39:34,369
for a presentation that evokes

1365
00:39:34,370 --> 00:39:35,929
a certain feeling.

1366
00:39:35,930 --> 00:39:37,969
And in the old world,

1367
00:39:37,970 --> 00:39:39,229
I used to have an image and I

1368
00:39:39,230 --> 00:39:40,519
couldn't find it. And I talked to

1369
00:39:40,520 --> 00:39:42,169
our designer and he has the entire

1370
00:39:42,170 --> 00:39:44,029
Adobe back catalog and he

1371
00:39:44,030 --> 00:39:46,249
gets me some photo and it's done.

1372
00:39:46,250 --> 00:39:48,289
But now I can

1373
00:39:48,290 --> 00:39:49,729
either describe an image and have it

1374
00:39:49,730 --> 00:39:51,649
generated or I can just say

1375
00:39:51,650 --> 00:39:53,119
I don't really know what image I

1376
00:39:53,120 --> 00:39:55,069
want for

1377
00:39:55,070 --> 00:39:57,289
exceeding our target this quarter.

1378
00:39:57,290 --> 00:39:58,549
You know, like I can only come up

1379
00:39:58,550 --> 00:40:00,649
with like some really generic ones

1380
00:40:00,650 --> 00:40:01,939
and I'll say, give me ideas of

1381
00:40:01,940 --> 00:40:03,409
images that'll work and I'll come up

1382
00:40:03,410 --> 00:40:04,459
with loads of images. And I'm like,

1383
00:40:04,460 --> 00:40:05,299
No, I don't like that one.

1384
00:40:05,300 --> 00:40:06,679
I do like this one.

1385
00:40:06,680 --> 00:40:08,959
Okay, yeah, that sounds good.

1386
00:40:08,960 --> 00:40:10,279
Edit it this way.

1387
00:40:10,280 --> 00:40:11,059
And now create.

1388
00:40:11,060 --> 00:40:12,199
And then Claude doesn't I?

1389
00:40:12,200 --> 00:40:14,359
She creates images, but Chatty

1390
00:40:14,360 --> 00:40:15,649
Betty does.

1391
00:40:15,650 --> 00:40:16,939
Or Dolly does.

1392
00:40:16,940 --> 00:40:18,499
And so then I'll say. Write a prompt

1393
00:40:18,500 --> 00:40:19,789
for Dolly that will give me the

1394
00:40:19,790 --> 00:40:22,399
image. And then I'll put it in,

1395
00:40:22,400 --> 00:40:24,259
create the image, and then edit

1396
00:40:24,260 --> 00:40:25,159
between the two.

1397
00:40:25,160 --> 00:40:26,809
And so and and all of that can be

1398
00:40:26,810 --> 00:40:28,489
way faster than it used to take me

1399
00:40:28,490 --> 00:40:30,379
to trawl through images to try

1400
00:40:30,380 --> 00:40:32,509
and evoke the sensation

1401
00:40:32,510 --> 00:40:35,059
that I want to evoke for customers.

1402
00:40:35,060 --> 00:40:37,159
Or I need to write the weekly

1403
00:40:37,160 --> 00:40:39,079
report or the weekly customer

1404
00:40:39,080 --> 00:40:40,249
update. And there's bits of it that

1405
00:40:40,250 --> 00:40:41,689
are written for me in bits that are

1406
00:40:41,690 --> 00:40:43,789
end, and I have writer's

1407
00:40:43,790 --> 00:40:45,619
block. And I'll just say I

1408
00:40:45,620 --> 00:40:47,119
want to talk at the end of the year,

1409
00:40:47,120 --> 00:40:48,199
end of the week, I've actually

1410
00:40:48,200 --> 00:40:49,489
created a project for this.

1411
00:40:49,490 --> 00:40:50,809
It's the end of the week.

1412
00:40:50,810 --> 00:40:52,579
I need to make people feel this way.

1413
00:40:52,580 --> 00:40:53,809
These are some topics that I'd like

1414
00:40:53,810 --> 00:40:54,949
to talk about. Everything I'm

1415
00:40:54,950 --> 00:40:56,509
writing is really stupid.

1416
00:40:56,510 --> 00:40:57,439
Write me a paragraph.

1417
00:40:57,440 --> 00:40:58,519
That's not shit.

1418
00:40:58,520 --> 00:40:59,659
Obviously it's a better prompt than

1419
00:40:59,660 --> 00:41:00,829
that, although sometimes that is the

1420
00:41:00,830 --> 00:41:02,419
prompt and it'll at least give me

1421
00:41:02,420 --> 00:41:04,129
something I rarely cut and paste and

1422
00:41:04,130 --> 00:41:05,239
use that itself.

1423
00:41:05,240 --> 00:41:07,279
But it's a way of like

1424
00:41:07,280 --> 00:41:09,199
whenever I'm stuck with

1425
00:41:09,200 --> 00:41:11,329
the too hard project, that is me

1426
00:41:11,330 --> 00:41:12,799
having to think.

1427
00:41:12,800 --> 00:41:15,139
I turn to Claude to help me think.

1428
00:41:15,140 --> 00:41:17,839
And oftentimes it just releases

1429
00:41:17,840 --> 00:41:19,939
the writer's block in me

1430
00:41:19,940 --> 00:41:21,379
to produce something.

1431
00:41:21,380 --> 00:41:22,759
And that's not an efficiency thing.

1432
00:41:22,760 --> 00:41:23,989
That's actually like a thinking

1433
00:41:23,990 --> 00:41:25,639
thing. I mean, it is efficiency, but

1434
00:41:25,640 --> 00:41:28,369
it's not like summarizing my emails.

1435
00:41:28,370 --> 00:41:30,859
I'll give you a specific example,

1436
00:41:30,860 --> 00:41:33,289
which for SEO

1437
00:41:33,290 --> 00:41:35,539
is a really sort of tangible

1438
00:41:35,540 --> 00:41:37,639
way of using this.

1439
00:41:37,640 --> 00:41:39,379
Anyone that's in a sea level role

1440
00:41:39,380 --> 00:41:41,239
has got direct reports

1441
00:41:41,240 --> 00:41:43,429
managing or leading those direct

1442
00:41:43,430 --> 00:41:45,139
reports. Probably when you became a

1443
00:41:45,140 --> 00:41:46,249
leader, you had these great

1444
00:41:46,250 --> 00:41:48,109
aspirations of how you wanted to

1445
00:41:48,110 --> 00:41:49,939
be, you know, a great leader and

1446
00:41:49,940 --> 00:41:51,229
a coach and advisor.

1447
00:41:51,230 --> 00:41:53,179
And too often we get dragged down

1448
00:41:53,180 --> 00:41:55,099
to just being a manager,

1449
00:41:55,100 --> 00:41:56,569
a boss, and you're just going

1450
00:41:56,570 --> 00:41:58,849
through the day to day

1451
00:41:58,850 --> 00:42:00,919
exercise, take a code

1452
00:42:00,920 --> 00:42:03,409
project or

1453
00:42:03,410 --> 00:42:04,819
just launched projects just before

1454
00:42:04,820 --> 00:42:05,659
Christmas.

1455
00:42:05,660 --> 00:42:07,579
Create a project one for

1456
00:42:07,580 --> 00:42:09,709
each of your direct reports.

1457
00:42:09,710 --> 00:42:11,689
So I create a project for

1458
00:42:11,690 --> 00:42:13,519
Bethany As and

1459
00:42:13,520 --> 00:42:15,679
in that project I can upload

1460
00:42:15,680 --> 00:42:16,909
certain documents.

1461
00:42:16,910 --> 00:42:19,039
Now, I would add in

1462
00:42:19,040 --> 00:42:21,289
my regular 1 to 1 notes,

1463
00:42:21,290 --> 00:42:23,239
I would add in maybe the performance

1464
00:42:23,240 --> 00:42:25,159
review that came out of Workday last

1465
00:42:25,160 --> 00:42:27,229
year. I might add in some

1466
00:42:27,230 --> 00:42:28,969
chaos that have been set.

1467
00:42:28,970 --> 00:42:30,769
Maybe there was a desk review

1468
00:42:30,770 --> 00:42:32,599
assessment or something like that

1469
00:42:32,600 --> 00:42:33,919
and maybe just do my chat.

1470
00:42:33,920 --> 00:42:35,659
So my knowledge of working with you,

1471
00:42:35,660 --> 00:42:37,489
but I understand a little bit about

1472
00:42:37,490 --> 00:42:39,409
your sort of personal career

1473
00:42:39,410 --> 00:42:41,329
aspirations or maybe personal

1474
00:42:41,330 --> 00:42:43,189
goals around houses

1475
00:42:43,190 --> 00:42:45,619
or family or whatever that might be.

1476
00:42:45,620 --> 00:42:47,899
And having done that,

1477
00:42:47,900 --> 00:42:49,939
you can then start to ask

1478
00:42:49,940 --> 00:42:52,369
this project about

1479
00:42:52,370 --> 00:42:53,449
those kind of creative things.

1480
00:42:53,450 --> 00:42:54,719
You know, I'm preparing for my 1 to

1481
00:42:54,720 --> 00:42:55,729
1 with Beth.

1482
00:42:55,730 --> 00:42:57,739
Can you help me think through some

1483
00:42:57,740 --> 00:42:59,629
stretch goals that would support

1484
00:42:59,630 --> 00:43:01,819
what best career aspirations are?

1485
00:43:01,820 --> 00:43:03,799
And I need to provide Beth

1486
00:43:03,800 --> 00:43:05,839
as some challenging feedback around

1487
00:43:05,840 --> 00:43:07,219
the way that she handles a recent

1488
00:43:07,220 --> 00:43:07,999
project.

1489
00:43:08,000 --> 00:43:09,679
Can you help role play with me?

1490
00:43:09,680 --> 00:43:11,389
How I could deliver that feedback to

1491
00:43:11,390 --> 00:43:13,369
Beth? Understanding Desk and

1492
00:43:13,370 --> 00:43:14,689
the way that she is going to respond

1493
00:43:14,690 --> 00:43:15,499
to that.

1494
00:43:15,500 --> 00:43:17,689
Suddenly you've got ten

1495
00:43:17,690 --> 00:43:20,329
individual leadership co-chairs

1496
00:43:20,330 --> 00:43:22,999
for each of your direct reports,

1497
00:43:23,000 --> 00:43:24,829
and you know, that's useful for

1498
00:43:24,830 --> 00:43:26,659
managers, but it's about

1499
00:43:26,660 --> 00:43:28,459
thinking about what's the work that

1500
00:43:28,460 --> 00:43:30,289
someone is already doing.

1501
00:43:30,290 --> 00:43:31,489
They have to do it.

1502
00:43:31,490 --> 00:43:33,439
And how do we use AI

1503
00:43:33,440 --> 00:43:35,449
to support me and being better

1504
00:43:35,450 --> 00:43:37,129
at that specific role?

1505
00:43:37,130 --> 00:43:38,899
It might be managing someone.

1506
00:43:38,900 --> 00:43:40,159
You might have a project for a

1507
00:43:40,160 --> 00:43:42,079
customer that you're selling or

1508
00:43:42,080 --> 00:43:43,399
supporting into.

1509
00:43:43,400 --> 00:43:45,589
It might be a project for a country

1510
00:43:45,590 --> 00:43:47,539
that you're looking to open up.

1511
00:43:47,540 --> 00:43:49,369
It might be a project, in my

1512
00:43:49,370 --> 00:43:51,319
case, with a product that

1513
00:43:51,320 --> 00:43:52,699
I'm trying to build.

1514
00:43:52,700 --> 00:43:53,929
So it's thinking about breaking

1515
00:43:53,930 --> 00:43:55,549
down. What are the bits of work that

1516
00:43:55,550 --> 00:43:57,409
you're already doing and how do

1517
00:43:57,410 --> 00:43:59,539
we give enough context

1518
00:43:59,540 --> 00:44:01,669
around that to be able to give me

1519
00:44:01,670 --> 00:44:02,689
what I want?

1520
00:44:02,690 --> 00:44:04,129
Can you give us one example?

1521
00:44:04,130 --> 00:44:05,989
I need to ingest a wide swath

1522
00:44:05,990 --> 00:44:08,149
of information, either summarize

1523
00:44:08,150 --> 00:44:09,979
it or pull out some takeaways in

1524
00:44:09,980 --> 00:44:10,939
terms of something that I'm trying

1525
00:44:10,940 --> 00:44:12,979
to analyze or what have you.

1526
00:44:12,980 --> 00:44:14,599
I'm going to mention a tool called

1527
00:44:14,600 --> 00:44:16,819
Notebook Alarm, which

1528
00:44:16,820 --> 00:44:18,739
went pretty viral in the autumn of

1529
00:44:18,740 --> 00:44:20,569
2024. It comes from Google, so you

1530
00:44:20,570 --> 00:44:22,339
can just go to notebook alarm dot

1531
00:44:22,340 --> 00:44:23,899
Google.com.

1532
00:44:23,900 --> 00:44:26,389
And what they've done is take a very

1533
00:44:26,390 --> 00:44:28,369
sort of intuitive approach to using

1534
00:44:28,370 --> 00:44:30,259
AI. So you can upload

1535
00:44:30,260 --> 00:44:32,239
a source and a source

1536
00:44:32,240 --> 00:44:33,769
could be a document, it could be a

1537
00:44:33,770 --> 00:44:35,599
YouTube video, it could be

1538
00:44:35,600 --> 00:44:36,859
a website.

1539
00:44:36,860 --> 00:44:38,869
And what notebook will do

1540
00:44:38,870 --> 00:44:41,029
is ingest that source

1541
00:44:41,030 --> 00:44:43,759
and use that and only that

1542
00:44:43,760 --> 00:44:45,799
to drive a set of AI

1543
00:44:45,800 --> 00:44:47,189
outputs and responses.

1544
00:44:47,190 --> 00:44:49,309
So it's not accessing the wider

1545
00:44:49,310 --> 00:44:50,569
Internet.

1546
00:44:50,570 --> 00:44:52,309
And a couple of things that it

1547
00:44:52,310 --> 00:44:53,809
creates off the back of that

1548
00:44:53,810 --> 00:44:55,759
document is a briefing

1549
00:44:55,760 --> 00:44:56,569
document.

1550
00:44:56,570 --> 00:44:58,669
It creates some fake news

1551
00:44:58,670 --> 00:45:00,979
and the to resistance,

1552
00:45:00,980 --> 00:45:03,259
it creates a podcast

1553
00:45:03,260 --> 00:45:05,629
interview between two

1554
00:45:05,630 --> 00:45:07,789
AI generated hosts

1555
00:45:07,790 --> 00:45:09,769
and an example

1556
00:45:09,770 --> 00:45:11,809
that a provider may.

1557
00:45:11,810 --> 00:45:13,399
A little link to some of the

1558
00:45:13,400 --> 00:45:14,539
recording of this.

1559
00:45:14,540 --> 00:45:16,789
But is the the new EU

1560
00:45:16,790 --> 00:45:18,799
Air Act, which

1561
00:45:18,800 --> 00:45:20,419
starts to come into force in

1562
00:45:20,420 --> 00:45:22,339
February 2025.

1563
00:45:22,340 --> 00:45:24,439
So in just a month or so from now?

1564
00:45:24,440 --> 00:45:27,289
Now the EU Air Act is 144

1565
00:45:27,290 --> 00:45:29,659
page document, very dry

1566
00:45:29,660 --> 00:45:31,609
as you'd expect, and difficult for

1567
00:45:31,610 --> 00:45:33,049
most people, if not legal, to

1568
00:45:33,050 --> 00:45:33,929
quickly figure out.

1569
00:45:33,930 --> 00:45:35,089
Does this relate to me?

1570
00:45:35,090 --> 00:45:36,739
How must I consider it?

1571
00:45:36,740 --> 00:45:37,909
So what you're able to do in this

1572
00:45:37,910 --> 00:45:39,739
example is upload that

1573
00:45:39,740 --> 00:45:42,049
document and then suddenly

1574
00:45:42,050 --> 00:45:44,479
it generated for me a 31 minute

1575
00:45:44,480 --> 00:45:46,369
podcast interview with these two

1576
00:45:46,370 --> 00:45:47,869
hosts discussing it.

1577
00:45:47,870 --> 00:45:49,969
But even better than that, you

1578
00:45:49,970 --> 00:45:52,189
can join that podcast.

1579
00:45:52,190 --> 00:45:53,959
So as you listen to it, you can say,

1580
00:45:53,960 --> 00:45:55,519
Sorry, I don't understand that, or

1581
00:45:55,520 --> 00:45:56,959
just help me to understand that.

1582
00:45:56,960 --> 00:45:58,159
So suddenly you're having a

1583
00:45:58,160 --> 00:46:00,049
conversation about

1584
00:46:00,050 --> 00:46:01,969
this complex subject.

1585
00:46:01,970 --> 00:46:03,089
Right, Right. So it's it's right.

1586
00:46:03,090 --> 00:46:05,029
You can interrupt the podcast hosts

1587
00:46:05,030 --> 00:46:06,469
and basically ask a question

1588
00:46:06,470 --> 00:46:08,119
midstream. You're like, Hey, podcast

1589
00:46:08,120 --> 00:46:09,350
host, I have a question

1590
00:46:10,490 --> 00:46:11,209
100%.

1591
00:46:11,210 --> 00:46:12,569
And they go, Hey, what's up?

1592
00:46:12,570 --> 00:46:13,669
You know, what's your question?

1593
00:46:13,670 --> 00:46:14,929
And then you also you say, you know,

1594
00:46:14,930 --> 00:46:17,059
you talked about the

1595
00:46:17,060 --> 00:46:19,069
air coming in, but like, what's

1596
00:46:19,070 --> 00:46:21,169
the first deadline that

1597
00:46:21,170 --> 00:46:23,179
I have to be considering?

1598
00:46:23,180 --> 00:46:24,739
And they go, you know, well,

1599
00:46:24,740 --> 00:46:26,809
February the 2nd, 2025.

1600
00:46:26,810 --> 00:46:27,649
And so you're having this

1601
00:46:27,650 --> 00:46:28,759
conversation now.

1602
00:46:28,760 --> 00:46:30,709
This is about the Air

1603
00:46:30,710 --> 00:46:32,569
Act. But think about all of the

1604
00:46:32,570 --> 00:46:34,729
things that you have to consume.

1605
00:46:34,730 --> 00:46:36,739
Maybe it's the Mssa that's been sent

1606
00:46:36,740 --> 00:46:38,929
to you from a supplier.

1607
00:46:38,930 --> 00:46:40,399
Can I give you an example of what

1608
00:46:40,400 --> 00:46:41,519
we're looking to do?

1609
00:46:41,520 --> 00:46:43,459
So we're also it's

1610
00:46:43,460 --> 00:46:44,299
interesting that you've mentioned it

1611
00:46:44,300 --> 00:46:46,609
because my first experience

1612
00:46:46,610 --> 00:46:48,649
with it was yesterday when one

1613
00:46:48,650 --> 00:46:50,719
of our Gen I people pulled

1614
00:46:50,720 --> 00:46:52,579
together information

1615
00:46:52,580 --> 00:46:54,409
on ideas of what we're going

1616
00:46:54,410 --> 00:46:57,079
to do as a podcast.

1617
00:46:57,080 --> 00:46:58,519
Not only does it do it as a podcast,

1618
00:46:58,520 --> 00:47:00,499
but it can also play on your voice.

1619
00:47:00,500 --> 00:47:02,599
So it's his voice doing

1620
00:47:02,600 --> 00:47:03,649
the podcast.

1621
00:47:03,650 --> 00:47:05,479
And what we're going to experiment

1622
00:47:05,480 --> 00:47:07,549
is, do you remember years ago,

1623
00:47:07,550 --> 00:47:08,719
I can't remember which company it

1624
00:47:08,720 --> 00:47:09,859
was. It was a company that I worked

1625
00:47:09,860 --> 00:47:11,599
with that would give all of their

1626
00:47:11,600 --> 00:47:14,029
sales team CD's

1627
00:47:14,030 --> 00:47:15,889
to listen to for sales

1628
00:47:15,890 --> 00:47:17,119
enablement while they were on the

1629
00:47:17,120 --> 00:47:18,979
road. It was either Microsoft or

1630
00:47:18,980 --> 00:47:20,659
IBM, I don't remember.

1631
00:47:20,660 --> 00:47:21,589
And so what we're going to

1632
00:47:21,590 --> 00:47:23,419
experiment with is taking all

1633
00:47:23,420 --> 00:47:25,429
of our sales enablement content

1634
00:47:25,430 --> 00:47:27,559
and turning that into

1635
00:47:27,560 --> 00:47:29,419
an internal podcast for our sales

1636
00:47:29,420 --> 00:47:31,399
team so that while they're on

1637
00:47:31,400 --> 00:47:33,259
the road, not that people are

1638
00:47:33,260 --> 00:47:34,849
on the road as much as possible can

1639
00:47:34,850 --> 00:47:36,289
consume that content.

1640
00:47:36,290 --> 00:47:37,279
And we're looking at a couple of

1641
00:47:37,280 --> 00:47:39,499
tools to basically create

1642
00:47:39,500 --> 00:47:41,419
an internal playlist where you

1643
00:47:41,420 --> 00:47:43,159
can find all the content as a

1644
00:47:43,160 --> 00:47:43,969
podcast.

1645
00:47:43,970 --> 00:47:44,779
But I didn't realize about the

1646
00:47:44,780 --> 00:47:46,999
interactivity, so that's super cool.

1647
00:47:47,000 --> 00:47:48,889
So yeah, it's only just released

1648
00:47:48,890 --> 00:47:50,809
maybe a week or so ago that they put

1649
00:47:50,810 --> 00:47:52,279
the interactivity in.

1650
00:47:52,280 --> 00:47:53,239
It's amazing.

1651
00:47:53,240 --> 00:47:55,219
There's a maybe a two second

1652
00:47:55,220 --> 00:47:57,109
delay after your question

1653
00:47:57,110 --> 00:47:58,969
wallet reconfigures,

1654
00:47:58,970 --> 00:48:01,069
but it's amazing.

1655
00:48:01,070 --> 00:48:03,259
So my one question is

1656
00:48:03,260 --> 00:48:04,609
Gemini.

1657
00:48:04,610 --> 00:48:07,339
It was shit have heard that Gemini

1658
00:48:07,340 --> 00:48:08,929
two is not shit.

1659
00:48:08,930 --> 00:48:10,579
Have not played with it yet.

1660
00:48:10,580 --> 00:48:12,499
We're going to experiment.

1661
00:48:12,500 --> 00:48:14,239
Worth it or not.

1662
00:48:14,240 --> 00:48:16,069
So when I'm talking to clients,

1663
00:48:16,070 --> 00:48:17,419
I talk about three things.

1664
00:48:17,420 --> 00:48:20,089
I talk about Gemini chat

1665
00:48:20,090 --> 00:48:21,919
and Cod as being those two

1666
00:48:21,920 --> 00:48:23,299
providers that people should be

1667
00:48:23,300 --> 00:48:25,219
thinking about and definitely

1668
00:48:25,220 --> 00:48:26,839
want to be watching, especially if

1669
00:48:26,840 --> 00:48:28,519
you're a if you're a Google

1670
00:48:28,520 --> 00:48:30,469
workspace customer, then

1671
00:48:30,470 --> 00:48:31,859
that has to make sense.

1672
00:48:31,860 --> 00:48:33,229
I just say on that, it's one of

1673
00:48:33,230 --> 00:48:35,569
these things that is so different

1674
00:48:35,570 --> 00:48:38,029
to SAS five years ago.

1675
00:48:38,030 --> 00:48:39,829
If you're buying a CRM system or

1676
00:48:39,830 --> 00:48:41,779
you're buying an ERP system, you

1677
00:48:41,780 --> 00:48:43,459
go out to market, you pick which one

1678
00:48:43,460 --> 00:48:45,049
and you pick one.

1679
00:48:45,050 --> 00:48:46,819
That is not the way that it works

1680
00:48:46,820 --> 00:48:48,229
here. People are not going out to

1681
00:48:48,230 --> 00:48:50,659
market to pick one Al-Alam.

1682
00:48:50,660 --> 00:48:51,889
You will have this composable

1683
00:48:51,890 --> 00:48:52,849
architecture.

1684
00:48:52,850 --> 00:48:54,739
You will have multiple,

1685
00:48:54,740 --> 00:48:56,389
maybe some closed source, maybe some

1686
00:48:56,390 --> 00:48:58,309
open source, and you'll use

1687
00:48:58,310 --> 00:49:00,469
different ones even within the same

1688
00:49:00,470 --> 00:49:01,549
workflow.

1689
00:49:01,550 --> 00:49:02,989
You know, this is doing the research

1690
00:49:02,990 --> 00:49:04,549
and this is doing the writing.

1691
00:49:04,550 --> 00:49:06,829
And so from a CEO perspective,

1692
00:49:06,830 --> 00:49:09,529
from a CIO or CTO perspective,

1693
00:49:09,530 --> 00:49:11,749
I think about Lego blocks.

1694
00:49:11,750 --> 00:49:13,579
What you're building is a box of

1695
00:49:13,580 --> 00:49:15,439
Lego that people can build

1696
00:49:15,440 --> 00:49:16,939
what they want at the time.

1697
00:49:16,940 --> 00:49:18,379
This isn't like a traditional

1698
00:49:18,380 --> 00:49:20,599
procurement, so you would have

1699
00:49:20,600 --> 00:49:22,459
Gemini plus the others in

1700
00:49:22,460 --> 00:49:23,689
your box of Lego.

1701
00:49:23,690 --> 00:49:25,549
If our listeners can

1702
00:49:25,550 --> 00:49:27,529
only take away one thing from

1703
00:49:27,530 --> 00:49:29,359
today's conversation, what is

1704
00:49:29,360 --> 00:49:30,229
it?

1705
00:49:30,230 --> 00:49:32,449
It is to plant

1706
00:49:32,450 --> 00:49:34,729
the seed of

1707
00:49:34,730 --> 00:49:36,739
use and every

1708
00:49:36,740 --> 00:49:39,079
single one of your business

1709
00:49:39,080 --> 00:49:39,949
teams.

1710
00:49:39,950 --> 00:49:41,419
I say business to me.

1711
00:49:41,420 --> 00:49:43,309
This is not an IT project.

1712
00:49:43,310 --> 00:49:45,229
So for every one of your

1713
00:49:45,230 --> 00:49:47,209
teams, how can you give them just

1714
00:49:47,210 --> 00:49:49,789
one way of using

1715
00:49:49,790 --> 00:49:51,230
publicly available tools

1716
00:49:52,370 --> 00:49:54,589
called Gemini to improve

1717
00:49:54,590 --> 00:49:56,599
one aspect of their work?

1718
00:49:56,600 --> 00:49:59,239
Because it is those people

1719
00:49:59,240 --> 00:50:01,009
that are closest to the business

1720
00:50:01,010 --> 00:50:03,949
problem and the business process.

1721
00:50:03,950 --> 00:50:05,959
And if you can show them one

1722
00:50:05,960 --> 00:50:08,029
way of using this, that is where

1723
00:50:08,030 --> 00:50:09,919
you will suddenly find all of

1724
00:50:09,920 --> 00:50:12,009
these new ideas come from.

1725
00:50:12,010 --> 00:50:13,599
So figure out how to get this

1726
00:50:13,600 --> 00:50:15,819
groundswell going at the

1727
00:50:15,820 --> 00:50:17,859
core of your business and bring

1728
00:50:17,860 --> 00:50:19,809
that back up to the top rather than

1729
00:50:19,810 --> 00:50:21,849
going top down.

1730
00:50:21,850 --> 00:50:23,679
Lovely. Thank you, Charlie Cohen,

1731
00:50:23,680 --> 00:50:25,089
for joining us on the operations

1732
00:50:25,090 --> 00:50:25,959
room.

1733
00:50:25,960 --> 00:50:27,129
If you like what you hear, please

1734
00:50:27,130 --> 00:50:28,779
subscribe or leave us a comment.

1735
00:50:28,780 --> 00:50:30,249
And we will see you next week.