WEBVTT

1
00:00:01.760 --> 00:00:03.759
- Join us as we gather around the hedge,

2
00:00:03.759 --> 00:00:06.740
- where we dig into technology, business, and culture

3
00:00:06.799 --> 00:00:09.139
- with the finest minds in computer networking.

4
00:00:21.054 --> 00:00:22.675
- Hello, Tom. Hello, Audrey.

5
00:00:23.294 --> 00:00:23.794
- And

6
00:00:24.175 --> 00:00:26.094
- I'm glad to see there is stuff on

7
00:00:26.094 --> 00:00:27.314
- your whiteboard, Tom.

8
00:00:27.780 --> 00:00:31.060
- Yeah. And Yeah. The rocket is gone. Was

9
00:00:31.060 --> 00:00:32.520
- it fun? Did it go off?

10
00:00:33.060 --> 00:00:34.359
- It's just off screen.

11
00:00:34.899 --> 00:00:36.579
- It's still there in its box and its

12
00:00:36.579 --> 00:00:37.560
- potential energy.

13
00:00:39.620 --> 00:00:41.479
- You just moved it to fool us.

14
00:00:42.414 --> 00:00:44.435
- That's that's what you did. Alright.

15
00:00:45.215 --> 00:00:47.854
- And today, we have Doug Dawson from the

16
00:00:47.854 --> 00:00:48.914
- other side of the mountains.

17
00:00:50.655 --> 00:00:52.414
- He's on the he's on the right side

18
00:00:52.414 --> 00:00:53.774
- of the mountains. I'm on the left side

19
00:00:53.774 --> 00:00:54.515
- of the mountains.

20
00:00:56.890 --> 00:00:58.990
- Well, depends where you're standing. So, yeah,

21
00:01:00.250 --> 00:01:02.010
- Asheville pretty much thinks they're on the left

22
00:01:02.010 --> 00:01:03.370
- side of everything. So

23
00:01:05.049 --> 00:01:08.409
- They do. But... Yeah. So, yeah, Doug is

24
00:01:08.409 --> 00:01:10.954
- just over the mountains for me, and we

25
00:01:10.954 --> 00:01:12.494
- were gonna talk today about

26
00:01:13.594 --> 00:01:14.094
- AI.

27
00:01:14.954 --> 00:01:15.454
- And,

28
00:01:17.194 --> 00:01:19.275
- you know, it's an interesting... Not not AI.

29
00:01:19.275 --> 00:01:20.795
- We've been talking a lot about AI, but

30
00:01:20.795 --> 00:01:22.415
- everybody has. So whatever.

31
00:01:22.875 --> 00:01:23.375
- But,

32
00:01:24.189 --> 00:01:26.049
- about the potential for a crash,

33
00:01:26.350 --> 00:01:28.109
- and I may have to move this show

34
00:01:28.109 --> 00:01:29.870
- up in the schedule a little bit to

35
00:01:29.870 --> 00:01:30.689
- beat the crash

36
00:01:32.750 --> 00:01:33.730
- before it happens.

37
00:01:34.590 --> 00:01:36.750
- Yeah. This this show becomes worthless the week

38
00:01:36.750 --> 00:01:38.585
- after the crash. Yeah. Yeah.

39
00:01:41.125 --> 00:01:43.064
- You predicted it after the crash.

40
00:01:43.844 --> 00:01:44.344
- Impressive.

41
00:01:46.085 --> 00:01:48.265
- That was quite brave of you. Yeah. So...

42
00:01:48.405 --> 00:01:48.905
- Yeah.

43
00:01:50.840 --> 00:01:52.519
- So let's talk about it because you wrote

44
00:01:52.519 --> 00:01:54.280
- an article about this, I think, right, on

45
00:01:54.280 --> 00:01:56.439
- your pots and pans blog, which I did.

46
00:01:56.519 --> 00:01:58.040
- By the way, you're not following pots and

47
00:01:58.040 --> 00:01:59.099
- pans, you should.

48
00:01:59.719 --> 00:02:00.219
- So

49
00:02:00.680 --> 00:02:02.119
- talk to us a little about what you're

50
00:02:02.119 --> 00:02:03.420
- seeing or what you're thinking.

51
00:02:03.734 --> 00:02:05.415
- I think we can add more fuel to

52
00:02:05.415 --> 00:02:07.435
- the fire over the course of the discussion.

53
00:02:07.814 --> 00:02:09.175
- But what are you seeing or what are

54
00:02:09.175 --> 00:02:09.754
- you thinking?

55
00:02:10.294 --> 00:02:11.814
- Well, first of all, let's put it in

56
00:02:11.814 --> 00:02:13.895
- context what I'm talking about. I mean, AI

57
00:02:13.895 --> 00:02:15.974
- as a technology is gonna be just fine

58
00:02:15.974 --> 00:02:17.489
- in the long run. I mean, it's a

59
00:02:17.489 --> 00:02:19.650
- wonderful technology. It does all sorts of pretty

60
00:02:19.650 --> 00:02:20.629
- amazing things.

61
00:02:21.090 --> 00:02:22.549
- What we're talking about crashing

62
00:02:22.849 --> 00:02:24.709
- is the big data center model.

63
00:02:25.169 --> 00:02:27.329
- I mean, the big data center model looks

64
00:02:27.329 --> 00:02:28.150
- to be unsustainable.

65
00:02:28.530 --> 00:02:30.289
- So there's lots of other ways to do

66
00:02:30.289 --> 00:02:32.610
- AI that doesn't take these giant data centers.

67
00:02:32.610 --> 00:02:32.955
- So,

68
00:02:33.915 --> 00:02:37.115
- and so... And they're... Interestingly, you're... Before we

69
00:02:37.115 --> 00:02:38.555
- got on here, we... You said there was

70
00:02:38.555 --> 00:02:40.794
- a lot of articles coming out. There was

71
00:02:40.794 --> 00:02:43.275
- a... There's probably been... I probably read 40

72
00:02:43.275 --> 00:02:45.534
- reports from various Wall Street analysts.

73
00:02:46.300 --> 00:02:48.460
- One just came out on Tuesday from Bain,

74
00:02:48.460 --> 00:02:49.580
- if you know who they are, if you

75
00:02:49.580 --> 00:02:52.060
- remember the Bain guys. Yeah. Yeah. And that...

76
00:02:52.060 --> 00:02:54.539
- So that's the that's the newest one. It

77
00:02:54.539 --> 00:02:56.219
- very much says what the other one does.

78
00:02:56.219 --> 00:02:58.319
- What's interesting is Bain is very

79
00:02:59.115 --> 00:02:59.615
- big

80
00:03:00.155 --> 00:03:02.715
- pro on AI, but but they look at

81
00:03:02.715 --> 00:03:04.395
- the numbers and they just shake their head.

82
00:03:04.395 --> 00:03:05.775
- So here's what they said.

83
00:03:06.314 --> 00:03:08.414
- They said by 2031,

84
00:03:08.875 --> 00:03:11.435
- the current... The big data center industry would

85
00:03:11.435 --> 00:03:13.275
- need $6,000,000,000,000

86
00:03:13.275 --> 00:03:14.014
- of revenue

87
00:03:14.540 --> 00:03:15.360
- a year to

88
00:03:15.659 --> 00:03:16.560
- sustain itself.

89
00:03:17.420 --> 00:03:17.920
- And

90
00:03:18.460 --> 00:03:20.219
- so... And now to put that number into

91
00:03:20.219 --> 00:03:20.719
- context,

92
00:03:21.580 --> 00:03:22.319
- The US

93
00:03:22.620 --> 00:03:25.680
- whole federal budget right now is 7,400,000,000,000.

94
00:03:27.715 --> 00:03:29.715
- Wow. Now that... Now the 6,000,000,000,000 is a

95
00:03:29.715 --> 00:03:30.694
- worldwide number,

96
00:03:31.394 --> 00:03:33.794
- but that doesn't really matter. That's just... You

97
00:03:33.794 --> 00:03:35.634
- know, that... It's like, well, we need the

98
00:03:35.634 --> 00:03:37.954
- second biggest economy after The United States just

99
00:03:37.954 --> 00:03:39.474
- to run AI. I mean, it's like, this

100
00:03:39.474 --> 00:03:41.759
- is not gonna work. Yeah. And and what

101
00:03:41.759 --> 00:03:44.020
- Bain did, Bain was very optimistic

102
00:03:44.800 --> 00:03:46.560
- more than the other folks. So here's what

103
00:03:46.560 --> 00:03:47.300
- Bain estimated.

104
00:03:47.840 --> 00:03:49.300
- They estimated the consumer

105
00:03:49.759 --> 00:03:51.379
- AI by 2031

106
00:03:51.680 --> 00:03:54.159
- would generate 200 to 400,000,000,000

107
00:03:54.159 --> 00:03:54.819
- a year.

108
00:03:55.294 --> 00:03:56.735
- So from, you know, from people like me

109
00:03:56.735 --> 00:03:58.834
- and you buying the $200 subscriptions.

110
00:03:59.854 --> 00:04:01.715
- And then they said that corporate,

111
00:04:02.655 --> 00:04:06.974
- and government AI would generate 1 to $1,400,000,000,000

112
00:04:06.974 --> 00:04:09.750
- a year, meaning that even if... And those

113
00:04:09.750 --> 00:04:12.150
- numbers, which I'll talk about a second, sound

114
00:04:12.150 --> 00:04:15.110
- both ridiculously high to me. But even if

115
00:04:15.110 --> 00:04:18.329
- they achieve that, there's still 4,200,000,000,000

116
00:04:18.389 --> 00:04:18.889
- short

117
00:04:19.669 --> 00:04:20.329
- a year.

118
00:04:21.029 --> 00:04:23.414
- It's like, so not... So let's talk about

119
00:04:23.414 --> 00:04:25.034
- why they need that big number.

120
00:04:25.654 --> 00:04:28.055
- And and the answer is fairly easy. It's

121
00:04:28.055 --> 00:04:30.055
- not people. It's the cost of running those

122
00:04:30.055 --> 00:04:32.875
- data centers. Those things are expensive as hell.

123
00:04:33.094 --> 00:04:35.094
- And it's not just the operating cost. They

124
00:04:35.094 --> 00:04:36.534
- do suck a lot of power. They do

125
00:04:36.534 --> 00:04:37.709
- suck a lot of water.

126
00:04:38.269 --> 00:04:39.949
- But the worst thing is the cards in

127
00:04:39.949 --> 00:04:42.610
- the data centers have gotten ungodly expensive.

128
00:04:43.389 --> 00:04:45.229
- I mean, to do a rack... Just to

129
00:04:45.229 --> 00:04:46.990
- do a shelf on on one of their

130
00:04:46.990 --> 00:04:49.149
- data centers is, like, a $100,000

131
00:04:49.149 --> 00:04:49.649
- now.

132
00:04:49.985 --> 00:04:51.504
- And so by the time you put all

133
00:04:51.504 --> 00:04:53.345
- the components in there... And and these things

134
00:04:53.345 --> 00:04:54.785
- go on and on and on. I mean,

135
00:04:54.785 --> 00:04:57.024
- they're building these things, you know, the size

136
00:04:57.024 --> 00:04:59.024
- of three Super Bowlmarts. You know, these are

137
00:04:59.024 --> 00:05:00.805
- just monstrously large buildings.

138
00:05:01.264 --> 00:05:03.420
- In fact, one one of the ones that's

139
00:05:03.420 --> 00:05:05.180
- being built right now is the size of

140
00:05:05.180 --> 00:05:08.540
- of, you know, of Central Park in in

141
00:05:08.540 --> 00:05:10.220
- New York City. It's like, this is an

142
00:05:10.220 --> 00:05:11.199
- insane building.

143
00:05:11.660 --> 00:05:13.899
- Yeah. And so and so so they they

144
00:05:13.899 --> 00:05:16.080
- made billions of dollars worth of hardware

145
00:05:16.985 --> 00:05:18.904
- and thing... And those things are only lasting

146
00:05:18.904 --> 00:05:20.345
- three and four years, so they have to

147
00:05:20.345 --> 00:05:21.944
- keep buying. That's what I was gonna say.

148
00:05:21.944 --> 00:05:23.645
- Yeah. Then... Right. Exactly.

149
00:05:24.024 --> 00:05:25.384
- That's what I was gonna say is that

150
00:05:25.384 --> 00:05:27.305
- the life cycle is so short, and yet

151
00:05:27.305 --> 00:05:29.939
- they're trying to make it sound like it's

152
00:05:29.939 --> 00:05:32.259
- almost a permanent investment. Like, they go to

153
00:05:32.259 --> 00:05:34.259
- a community, and they say we're putting in

154
00:05:34.259 --> 00:05:36.819
- this data center. And we're gonna add a

155
00:05:36.819 --> 00:05:38.279
- 100 jobs to the community,

156
00:05:38.819 --> 00:05:40.899
- and they don't say how long. But the

157
00:05:40.899 --> 00:05:43.639
- fact is is that data center buildings themselves

158
00:05:44.285 --> 00:05:46.064
- are not eternal things.

159
00:05:46.444 --> 00:05:48.305
- Right? Yeah. They're they're...

160
00:05:48.685 --> 00:05:50.845
- You you you have a change in power

161
00:05:50.845 --> 00:05:53.485
- requirements or a change in cooling requirements, and

162
00:05:53.485 --> 00:05:55.485
- it's often cheaper to build a new building

163
00:05:55.485 --> 00:05:57.324
- than it is to retrofit the one you

164
00:05:57.324 --> 00:05:57.824
- have.

165
00:05:58.500 --> 00:06:00.500
- And Yeah. And that becomes... You end up

166
00:06:00.500 --> 00:06:02.600
- with abandoned stuff all over the place.

167
00:06:03.300 --> 00:06:05.860
- Now there are scientists and engineers working on

168
00:06:05.860 --> 00:06:08.100
- making the cards last better because the problem...

169
00:06:08.100 --> 00:06:10.439
- The reason they're burning out is they

170
00:06:10.745 --> 00:06:12.504
- they just let them run. And so when...

171
00:06:12.504 --> 00:06:14.345
- So those... These things are running twenty four

172
00:06:14.345 --> 00:06:16.425
- seven at full capacity, and they generate a

173
00:06:16.425 --> 00:06:17.245
- lot of heat.

174
00:06:17.625 --> 00:06:20.345
- You know, there's smart strategies to say run

175
00:06:20.345 --> 00:06:21.545
- them at 60%

176
00:06:21.545 --> 00:06:22.045
- load,

177
00:06:22.665 --> 00:06:23.485
- which takes...

178
00:06:24.199 --> 00:06:26.759
- It's a really complicated set of programming to

179
00:06:26.759 --> 00:06:28.379
- make every chip do that.

180
00:06:28.759 --> 00:06:30.120
- So it's just not like you can't just

181
00:06:30.120 --> 00:06:31.720
- go to the data center and go slow

182
00:06:31.720 --> 00:06:34.060
- my chips down. But but that would probably...

183
00:06:34.279 --> 00:06:36.279
- That might even double their shelf life, but

184
00:06:36.279 --> 00:06:38.654
- then it also cuts the computing power now.

185
00:06:38.714 --> 00:06:41.454
- Right. So... But, you know... And and

186
00:06:41.754 --> 00:06:42.334
- the... Now...

187
00:06:42.794 --> 00:06:44.875
- So so the money is just not making

188
00:06:44.875 --> 00:06:47.694
- any sense. This makes absolutely no sense. We...

189
00:06:47.834 --> 00:06:50.314
- Now let's let's talk about the current revenue.

190
00:06:50.314 --> 00:06:50.814
- So

191
00:06:51.539 --> 00:06:53.720
- this year in 2026,

192
00:06:54.819 --> 00:06:55.319
- Anthropic

193
00:06:55.779 --> 00:06:59.959
- and OpenAI together will make one point... 105,000,000,000

194
00:07:00.339 --> 00:07:01.800
- in revenues, total revenues.

195
00:07:02.579 --> 00:07:04.259
- So somehow they have to grow from a

196
00:07:04.259 --> 00:07:05.220
- 105...

197
00:07:05.220 --> 00:07:07.314
- And and then I thought I don't... I

198
00:07:07.314 --> 00:07:09.875
- didn't go look up the number for, for

199
00:07:09.875 --> 00:07:12.035
- Grok, but it's not very big. So when

200
00:07:12.035 --> 00:07:14.274
- you add all those together, there... It's a

201
00:07:14.274 --> 00:07:16.214
- really long way to get the 6,000,000,000,000.

202
00:07:18.675 --> 00:07:20.274
- It's like that... You know, it's just really

203
00:07:20.274 --> 00:07:20.935
- not there.

204
00:07:21.329 --> 00:07:24.220
- So so now we can just talk about

205
00:07:24.220 --> 00:07:27.111
- what's wrong even if they had a path

206
00:07:27.111 --> 00:07:30.001
- to get there, which they don't. Let's talk

207
00:07:30.001 --> 00:07:32.892
- about their big number. Their big number from

208
00:07:32.892 --> 00:07:35.794
- Bing was 1,000,000,000,000 to 1,400,000,000,000

209
00:07:36.175 --> 00:07:37.314
- in corporate revenues.

210
00:07:38.894 --> 00:07:40.654
- There's a whole lot of these studies I

211
00:07:40.654 --> 00:07:42.014
- read in the last few months that say

212
00:07:42.014 --> 00:07:44.334
- corporations have soured on the big AI model

213
00:07:44.334 --> 00:07:45.555
- because they're too expensive.

214
00:07:46.000 --> 00:07:48.160
- And they're... You know, to ask the major

215
00:07:48.160 --> 00:07:50.100
- corporations in the world to ship... To

216
00:07:50.480 --> 00:07:52.720
- to spend a trillion dollars a year on

217
00:07:52.720 --> 00:07:53.220
- AI,

218
00:07:53.600 --> 00:07:55.139
- it's like, you know, again,

219
00:07:56.079 --> 00:07:58.240
- most of them all added altogether don't even

220
00:07:58.240 --> 00:08:00.000
- have that much revenue. That's a lot of

221
00:08:00.000 --> 00:08:01.555
- money. That's a lot of money for me.

222
00:08:01.555 --> 00:08:03.714
- What's the ROI on that? Is that... That's

223
00:08:03.714 --> 00:08:05.555
- part of the problem is what's the ROI

224
00:08:05.555 --> 00:08:07.235
- on that? If I'm if I'm a big

225
00:08:07.235 --> 00:08:08.294
- company and I'm spending,

226
00:08:08.595 --> 00:08:11.394
- you know, 5,000,000 a year, 10,000,000 a year,

227
00:08:11.394 --> 00:08:13.949
- am I getting 5 or 10,000,000 back? And

228
00:08:13.949 --> 00:08:16.269
- then the other question is, just like when

229
00:08:16.269 --> 00:08:18.370
- you look at cloud versus on premises,

230
00:08:19.949 --> 00:08:22.529
- this whole thing might actually drive

231
00:08:22.990 --> 00:08:26.129
- a movement back into on premises networking

232
00:08:26.750 --> 00:08:29.089
- and on premise data centers because

233
00:08:29.764 --> 00:08:30.585
- the cost

234
00:08:30.964 --> 00:08:33.764
- of them running these huge data centers and

235
00:08:33.764 --> 00:08:34.825
- spreading it everywhere.

236
00:08:35.285 --> 00:08:37.365
- And then not only that, but, like, all

237
00:08:37.365 --> 00:08:39.065
- the data you're giving them

238
00:08:39.684 --> 00:08:42.245
- to do this Right. That Right. That's... I

239
00:08:42.245 --> 00:08:43.625
- think that's very problematic.

240
00:08:44.159 --> 00:08:46.639
- If I Well... And and you just hit

241
00:08:46.639 --> 00:08:48.259
- on you just hit on the problem.

242
00:08:49.039 --> 00:08:51.200
- The reason it's so expensive to run these

243
00:08:51.200 --> 00:08:53.200
- is they've tried to load all the data

244
00:08:53.200 --> 00:08:55.379
- of mankind has with these data centers.

245
00:08:55.759 --> 00:08:58.004
- So just for the just for the data

246
00:08:58.004 --> 00:08:59.845
- center to find what it is you ask

247
00:08:59.845 --> 00:09:00.345
- it,

248
00:09:00.804 --> 00:09:02.964
- they are now... You know, the public here,

249
00:09:02.964 --> 00:09:05.125
- as we talked today, unless the guys tune

250
00:09:05.125 --> 00:09:06.904
- and use AI, they don't get it. But

251
00:09:06.964 --> 00:09:10.559
- the current OpenAI model uses 400,000

252
00:09:10.559 --> 00:09:11.860
- tokens just to get

253
00:09:13.679 --> 00:09:14.419
- started, just

254
00:09:14.799 --> 00:09:16.100
- to find the data

255
00:09:16.399 --> 00:09:18.240
- to try to answer your question. And that

256
00:09:18.240 --> 00:09:20.899
- includes... If that's my mother going,

257
00:09:21.575 --> 00:09:23.414
- you know, could you tell me if the

258
00:09:23.414 --> 00:09:26.774
- bus schedule changed for tomorrow? It's gonna cost

259
00:09:26.774 --> 00:09:28.934
- 400,000 tokens because it looks through the whole

260
00:09:28.934 --> 00:09:30.615
- world of data. It doesn't know how to

261
00:09:30.615 --> 00:09:32.855
- go straight to that. And so, you know,

262
00:09:32.855 --> 00:09:35.254
- so that I... They... They've overtrained them. They...

263
00:09:35.254 --> 00:09:36.934
- This idea that it has to have all

264
00:09:36.934 --> 00:09:39.580
- the information in the world in there just

265
00:09:39.580 --> 00:09:41.120
- to do that stuff is crazy

266
00:09:41.660 --> 00:09:45.179
- because companies want stuff. Bank of America just

267
00:09:45.179 --> 00:09:47.179
- wants to run stuff on Bank of America

268
00:09:47.179 --> 00:09:49.259
- data. They don't care about, you know, they

269
00:09:49.259 --> 00:09:50.399
- don't care about Charles

270
00:09:50.985 --> 00:09:53.945
- Darwin's books and and, you know, statistics about

271
00:09:53.945 --> 00:09:56.904
- volcanoes and everything that's in these data centers.

272
00:09:56.904 --> 00:09:59.065
- Right? So so the the way they... They're

273
00:09:59.065 --> 00:10:00.745
- gonna have to they're gonna have to back

274
00:10:00.745 --> 00:10:01.804
- off of that model,

275
00:10:02.825 --> 00:10:03.964
- because it's just

276
00:10:04.360 --> 00:10:06.700
- incredibly inefficient, and it's getting more inefficient.

277
00:10:07.240 --> 00:10:09.080
- The smarter the models get, the worse they

278
00:10:09.080 --> 00:10:11.559
- are at at getting started. So the older

279
00:10:11.559 --> 00:10:13.879
- ones were actually way more efficient than the

280
00:10:13.879 --> 00:10:16.279
- new ones are. So So, Greg, I wanna

281
00:10:16.279 --> 00:10:18.440
- I wanna Yeah. The math is so bad.

282
00:10:18.440 --> 00:10:19.004
- Yes, sir.

283
00:10:19.485 --> 00:10:20.924
- I I just wanted to ask, have you

284
00:10:20.924 --> 00:10:22.465
- noticed any studies about,

285
00:10:23.085 --> 00:10:26.225
- impact on revenue for companies for large corporations

286
00:10:26.445 --> 00:10:28.924
- spending lots of money? Like Yeah. Is there

287
00:10:28.924 --> 00:10:30.945
- is there any... Yeah. What's the evidence there?

288
00:10:31.325 --> 00:10:33.825
- Well, here's a great story. So Uber

289
00:10:35.320 --> 00:10:37.399
- Uber had a budget for AI this year,

290
00:10:37.399 --> 00:10:38.299
- and in June,

291
00:10:38.759 --> 00:10:40.919
- they they had a $40,000,000

292
00:10:40.919 --> 00:10:41.419
- spend

293
00:10:41.879 --> 00:10:43.259
- plan for Uber headquarters.

294
00:10:43.960 --> 00:10:45.720
- They hit that number in June, and they

295
00:10:45.720 --> 00:10:47.240
- just turned it off. They go, we can't

296
00:10:47.240 --> 00:10:48.059
- afford this.

297
00:10:49.160 --> 00:10:51.284
- So they hit their annual spend in half

298
00:10:51.284 --> 00:10:53.125
- a year, and now they're not gonna use

299
00:10:53.125 --> 00:10:54.664
- the big open AI anymore.

300
00:10:55.445 --> 00:10:57.044
- Because the whole point of them using the

301
00:10:57.044 --> 00:10:57.865
- 40,000,000

302
00:10:58.164 --> 00:11:00.404
- was to replace people, and they weren't replacing

303
00:11:00.404 --> 00:11:02.404
- them. So not only did they spend the

304
00:11:02.404 --> 00:11:05.579
- 40,000,000, they they have nowhere near $40,000,000

305
00:11:05.579 --> 00:11:06.240
- of savings.

306
00:11:06.700 --> 00:11:08.700
- And so the the head of... And this

307
00:11:08.700 --> 00:11:10.639
- is a tech company, and they're just going,

308
00:11:11.100 --> 00:11:14.059
- you know, our analysts are just doing it

309
00:11:14.059 --> 00:11:15.820
- to to save them time on their work.

310
00:11:15.820 --> 00:11:17.259
- It's like, hey. That's... We pay them to

311
00:11:17.259 --> 00:11:18.644
- do that work. They can do the damn

312
00:11:18.884 --> 00:11:22.345
- analysis themselves without spending $40,000,000.

313
00:11:22.884 --> 00:11:24.644
- Right? And so so they so they so

314
00:11:24.644 --> 00:11:27.125
- they candid it. And I forget who did

315
00:11:27.125 --> 00:11:27.785
- the study,

316
00:11:28.165 --> 00:11:28.665
- but

317
00:11:29.045 --> 00:11:31.379
- someone recently did one that said something like

318
00:11:31.379 --> 00:11:34.100
- 80% of corporations are double thinking whether they

319
00:11:34.100 --> 00:11:35.539
- can use this at the corporate level, and

320
00:11:35.539 --> 00:11:37.779
- it's just not panning out at the bottom

321
00:11:37.779 --> 00:11:38.759
- line for them.

322
00:11:39.620 --> 00:11:41.720
- So, you know, where AI comes useful,

323
00:11:42.419 --> 00:11:43.079
- like, universities

324
00:11:43.539 --> 00:11:45.220
- love this. I mean, if you're doing drug

325
00:11:45.220 --> 00:11:46.875
- research or or something, as long as someone

326
00:11:46.875 --> 00:11:47.855
- will give you the money,

327
00:11:48.475 --> 00:11:50.815
- you know, for the research, it's really wonderful

328
00:11:50.955 --> 00:11:53.115
- for stuff like that. But for, you know,

329
00:11:53.115 --> 00:11:54.254
- for your for your

330
00:11:54.955 --> 00:11:57.534
- regular analyst who's doing budgets in a corporation,

331
00:11:57.595 --> 00:11:58.875
- you can let them use AI to do

332
00:11:58.875 --> 00:12:01.500
- that. It's, like, costing way more than the

333
00:12:01.500 --> 00:12:03.660
- cost of than the cost of... It's like,

334
00:12:03.660 --> 00:12:06.139
- what? Be cheaper to hire another analyst or

335
00:12:06.139 --> 00:12:07.420
- to or to go get one that actually

336
00:12:07.420 --> 00:12:08.559
- knows what they're doing.

337
00:12:09.660 --> 00:12:12.059
- So, you know, so now here's the here's

338
00:12:12.059 --> 00:12:13.360
- the other math because...

339
00:12:13.784 --> 00:12:15.784
- I mean, a friend of mine is an

340
00:12:15.784 --> 00:12:18.664
- AI scientist, and I've known him forever. And

341
00:12:18.664 --> 00:12:18.904
- and,

342
00:12:19.625 --> 00:12:21.544
- he's been he's been doing AI since the

343
00:12:21.544 --> 00:12:23.304
- old old days of the seventies when it

344
00:12:23.304 --> 00:12:25.704
- was something very different than now. And and,

345
00:12:27.120 --> 00:12:29.920
- he says that when you... That the average

346
00:12:29.920 --> 00:12:31.840
- person who gives them a $200 a month

347
00:12:31.840 --> 00:12:35.059
- for their home subscription is costing OpenAI between

348
00:12:35.120 --> 00:12:36.480
- 30 and $50,000

349
00:12:36.480 --> 00:12:38.160
- a month in cost to give you your

350
00:12:38.160 --> 00:12:38.660
- stuff.

351
00:12:39.475 --> 00:12:42.514
- I mean, that's how unsustainable this is. Obviously,

352
00:12:42.514 --> 00:12:44.855
- they're not gonna charge you 30 to 50,000,

353
00:12:45.154 --> 00:12:47.014
- but they're off... They can't charge the corporations

354
00:12:47.154 --> 00:12:48.294
- that either. So...

355
00:12:48.674 --> 00:12:50.274
- And that's that's the gap to get to

356
00:12:50.274 --> 00:12:51.174
- the 6,000,000,000,000.

357
00:12:51.315 --> 00:12:52.774
- It's just an... It's a ridiculous

358
00:12:53.154 --> 00:12:53.654
- mess.

359
00:12:54.059 --> 00:12:54.559
- Now

360
00:12:55.659 --> 00:12:56.720
- why are we there?

361
00:12:58.059 --> 00:13:00.080
- That... That's the second big problem.

362
00:13:00.460 --> 00:13:01.200
- We have,

363
00:13:01.580 --> 00:13:04.379
- you know, y'all... We... I... Russ is an

364
00:13:04.379 --> 00:13:06.539
- intro. Are you in the telecom industry, Tom?

365
00:13:06.539 --> 00:13:08.235
- I assume you are. Yes, sir. Yeah. Yeah.

366
00:13:08.235 --> 00:13:10.235
- Are you... We all remember what happened in

367
00:13:10.235 --> 00:13:10.894
- 2000.

368
00:13:11.434 --> 00:13:14.235
- Exactly the same thing. We had we had

369
00:13:14.235 --> 00:13:16.554
- we had Lucent and Nortel and a couple

370
00:13:16.554 --> 00:13:17.535
- other big companies

371
00:13:18.634 --> 00:13:19.134
- giving...

372
00:13:20.240 --> 00:13:22.500
- The vendors were giving money to the ISPs

373
00:13:22.720 --> 00:13:24.240
- to turn around and give back to them

374
00:13:24.240 --> 00:13:26.159
- to buy their equipment. And so they were

375
00:13:26.159 --> 00:13:27.940
- booking huge amounts of sales,

376
00:13:29.199 --> 00:13:31.279
- but but then they... You know? And it

377
00:13:31.279 --> 00:13:33.059
- just was a circular financing,

378
00:13:33.440 --> 00:13:35.324
- and that's really what broke. That was the

379
00:13:35.324 --> 00:13:37.564
- number one item that broke the the the

380
00:13:37.564 --> 00:13:39.264
- the burgeoning sea like world,

381
00:13:40.125 --> 00:13:41.725
- because they were just handing out money, and

382
00:13:41.725 --> 00:13:42.544
- it was ridiculously

383
00:13:43.245 --> 00:13:43.745
- spent

384
00:13:44.125 --> 00:13:46.284
- way too much money. People were told to

385
00:13:46.284 --> 00:13:47.964
- go fast, so they did, and that's very

386
00:13:47.964 --> 00:13:50.460
- wasteful. And so that's where we're at again.

387
00:13:50.620 --> 00:13:51.360
- I mean,

388
00:13:51.820 --> 00:13:54.160
- there's just this round robin. I

389
00:13:54.540 --> 00:13:56.779
- I should have saved it somewhere, but there's

390
00:13:56.779 --> 00:13:59.040
- this round robin diagram that shows

391
00:13:59.580 --> 00:14:01.679
- Microsoft going to OpenAI

392
00:14:01.980 --> 00:14:04.735
- going to NVIDIA, and the money just goes

393
00:14:04.735 --> 00:14:06.495
- around and around, and they're all booking it

394
00:14:06.495 --> 00:14:07.075
- to revenue.

395
00:14:07.774 --> 00:14:09.295
- So it gets booked three or four times

396
00:14:09.295 --> 00:14:10.434
- as industry revenue.

397
00:14:10.815 --> 00:14:12.754
- And it's like... And none of it's revenue

398
00:14:12.815 --> 00:14:13.315
- because

399
00:14:13.695 --> 00:14:15.615
- what happened was the basis of that money

400
00:14:15.615 --> 00:14:17.950
- was somebody borrowed it. It's like when you

401
00:14:17.950 --> 00:14:20.429
- borrow money, that's not revenue to anybody. So

402
00:14:20.429 --> 00:14:21.629
- even if you turn around and give it

403
00:14:21.629 --> 00:14:23.149
- to somebody else, you still have to pay

404
00:14:23.149 --> 00:14:25.549
- those loans off. So so that's the other

405
00:14:25.549 --> 00:14:27.230
- one of the... Yeah. Go ahead. There was

406
00:14:27.230 --> 00:14:29.710
- another story this morning about... I believe it

407
00:14:29.710 --> 00:14:30.450
- was Amazon

408
00:14:30.830 --> 00:14:34.415
- who was trying to raise $7,000,000,000,000

409
00:14:36.315 --> 00:14:38.235
- to... Maybe maybe it was 7,000,000,000.

410
00:14:38.235 --> 00:14:40.634
- I don't know. Whatever. Some stupid number. And

411
00:14:40.634 --> 00:14:41.855
- maybe it was 7,000,000,000.

412
00:14:41.915 --> 00:14:43.774
- They're trying to raise $7,000,000,000

413
00:14:43.995 --> 00:14:45.134
- to sell off,

414
00:14:46.190 --> 00:14:48.690
- a huge chunk of their GPUs

415
00:14:49.629 --> 00:14:51.950
- and then lease them back so they can

416
00:14:51.950 --> 00:14:54.450
- continue using them as a way to, like,

417
00:14:54.830 --> 00:14:57.549
- build their their their their bottom line, build

418
00:14:57.549 --> 00:15:01.149
- their their sheet. Right? Their financial sheet. And,

419
00:15:01.149 --> 00:15:01.605
- like,

420
00:15:02.164 --> 00:15:02.664
- guys,

421
00:15:03.845 --> 00:15:06.884
- who... First of all, the market is not

422
00:15:06.884 --> 00:15:08.824
- made of infinite amounts of money.

423
00:15:09.125 --> 00:15:09.945
- We cannot...

424
00:15:10.404 --> 00:15:12.824
- People cannot invest this kind of money

425
00:15:13.204 --> 00:15:16.105
- in these things forever. It's already dried up.

426
00:15:16.220 --> 00:15:18.299
- I was reading QTR the other day, and

427
00:15:18.299 --> 00:15:20.379
- they were saying... QTR was saying that... Quote

428
00:15:20.379 --> 00:15:22.240
- the Raven, and he was saying, look.

429
00:15:22.940 --> 00:15:25.580
- If you take the big seven out of

430
00:15:25.580 --> 00:15:27.500
- the S and P 500 right now, it's

431
00:15:27.500 --> 00:15:29.279
- been losing money for the last

432
00:15:30.215 --> 00:15:32.774
- however long... A long time. It's only the

433
00:15:32.774 --> 00:15:35.335
- big seven floating it. And part of that

434
00:15:35.335 --> 00:15:36.875
- is just simply people

435
00:15:37.335 --> 00:15:39.575
- not investing in other things so they can

436
00:15:39.575 --> 00:15:41.355
- invest in the big seven.

437
00:15:41.870 --> 00:15:44.190
- Yeah. Well, that money's gonna run out. I

438
00:15:44.190 --> 00:15:45.490
- don't know. Like,

439
00:15:46.350 --> 00:15:48.929
- who has another 7,000,000,000 in the bank?

440
00:15:50.829 --> 00:15:53.709
- Let let alone let let alone settle. Now

441
00:15:53.709 --> 00:15:55.949
- The US spending alone for data centers this

442
00:15:55.949 --> 00:15:57.490
- year is 600,000,000,000.

443
00:15:57.924 --> 00:16:00.245
- Yeah. The world... The worldwide spending on data

444
00:16:00.245 --> 00:16:01.605
- centers, just 2026

445
00:16:01.605 --> 00:16:03.945
- is over $10,000,000,000,001,000,000,000,000.

446
00:16:04.565 --> 00:16:06.345
- Next year, if you could look on paper,

447
00:16:06.965 --> 00:16:09.625
- the plans that they've all announced are 1,300,000,000,000.

448
00:16:09.684 --> 00:16:10.184
- Now

449
00:16:10.589 --> 00:16:12.509
- people are backing off of data centers now,

450
00:16:12.509 --> 00:16:13.949
- so it's starting to cool down, which is

451
00:16:13.949 --> 00:16:15.470
- why the crash is probably coming. So they're

452
00:16:15.470 --> 00:16:17.870
- never gonna spend the 1.3, but but these

453
00:16:17.870 --> 00:16:18.929
- are really unbelievably

454
00:16:19.470 --> 00:16:21.870
- large numbers. So... Yeah. Now what you described

455
00:16:21.870 --> 00:16:24.190
- a minute ago was called... I... On top

456
00:16:24.190 --> 00:16:27.184
- of being a tech guy, I also... And

457
00:16:27.365 --> 00:16:29.764
- what... Gosh. Now it's about fifty years ago,

458
00:16:29.764 --> 00:16:31.764
- I passed the CPA exam. So I'm also

459
00:16:31.764 --> 00:16:34.004
- an accountant even though I never I never

460
00:16:34.004 --> 00:16:36.164
- actually did that for a living, but but,

461
00:16:36.164 --> 00:16:38.084
- you know, while I have that background, what

462
00:16:38.084 --> 00:16:40.424
- you're talking about is called the backstop clause.

463
00:16:41.910 --> 00:16:44.950
- It's a funny money game that they're playing.

464
00:16:44.950 --> 00:16:46.169
- What they're doing now

465
00:16:46.629 --> 00:16:48.250
- is is they are

466
00:16:48.710 --> 00:16:49.210
- booking

467
00:16:50.149 --> 00:16:50.649
- capacity,

468
00:16:51.429 --> 00:16:53.029
- and they're putting it on the books as

469
00:16:53.029 --> 00:16:53.690
- an asset.

470
00:16:54.414 --> 00:16:56.735
- And and everyone in in the AICPA world

471
00:16:56.735 --> 00:16:58.035
- who are the guys who are

472
00:16:58.414 --> 00:16:59.695
- on the county and the audits of the

473
00:16:59.695 --> 00:17:02.035
- world are going, that's not an asset.

474
00:17:03.294 --> 00:17:05.454
- Capacity is not an asset. Money is an,

475
00:17:05.454 --> 00:17:07.615
- you know, is an asset. The cards you

476
00:17:07.615 --> 00:17:09.055
- buy are an asset. That's what you're talking

477
00:17:09.055 --> 00:17:10.829
- about Microsoft doing. They're trying

478
00:17:11.210 --> 00:17:13.129
- to invent an asset because they want to

479
00:17:13.129 --> 00:17:15.450
- make their balance sheet look better. Yeah. So

480
00:17:15.450 --> 00:17:17.149
- they're taking they're taking a cost,

481
00:17:18.009 --> 00:17:20.429
- not even a cost, a theoretical cost

482
00:17:21.049 --> 00:17:23.125
- that... From that big circular chain of stuff

483
00:17:23.125 --> 00:17:24.804
- we're talking about, and they're trying to book

484
00:17:24.804 --> 00:17:26.664
- that to make their books look better.

485
00:17:27.525 --> 00:17:28.025
- Now

486
00:17:28.565 --> 00:17:30.804
- here's something. Some of the reports I read,

487
00:17:30.804 --> 00:17:32.025
- this is really interesting

488
00:17:32.404 --> 00:17:34.345
- because you talked about the big seven.

489
00:17:34.964 --> 00:17:37.464
- We think most of them have actually shielded

490
00:17:37.525 --> 00:17:39.629
- themselves from a whole lot of the loss

491
00:17:39.629 --> 00:17:41.329
- because they know that crash is coming.

492
00:17:42.029 --> 00:17:44.269
- So when the crash does come, they're gonna

493
00:17:44.269 --> 00:17:46.509
- get hurt. No doubt about it. Nvidia probably

494
00:17:46.509 --> 00:17:47.950
- is the one who takes it the worst

495
00:17:47.950 --> 00:17:48.769
- in the shorts.

496
00:17:49.549 --> 00:17:50.990
- But what they've done is they're not the

497
00:17:50.990 --> 00:17:53.515
- ones pay... We say there's 600,000,000,000

498
00:17:53.515 --> 00:17:56.154
- going into data centers. They're not paying for

499
00:17:56.154 --> 00:17:56.654
- that.

500
00:17:57.194 --> 00:17:58.634
- Do you wanna know where that money's coming

501
00:17:58.634 --> 00:18:00.815
- from? You're you're not... You're gonna hate this.

502
00:18:01.275 --> 00:18:02.335
- Oh, I'm saying private

503
00:18:03.275 --> 00:18:04.335
- equity and,

504
00:18:05.194 --> 00:18:07.930
- It's not it's not even come... It's retirement

505
00:18:07.930 --> 00:18:10.330
- accounts. It's not even private equity. It's coming

506
00:18:10.330 --> 00:18:12.809
- from retirement funds, like state retirement funds. This

507
00:18:12.809 --> 00:18:14.509
- is Enron all over again.

508
00:18:14.970 --> 00:18:16.730
- The little guy is gonna get killed in

509
00:18:16.730 --> 00:18:20.054
- his 04/2001 k when this stuff crashes. Now

510
00:18:20.054 --> 00:18:21.355
- these... The big seven,

511
00:18:21.734 --> 00:18:23.815
- Microsoft, NVIDIA, you know, all those guys, they're

512
00:18:23.815 --> 00:18:25.894
- gonna get hurt, but they're not gonna take

513
00:18:25.894 --> 00:18:27.734
- the brunt of it except that none of

514
00:18:27.734 --> 00:18:29.375
- us will ever trust them again

515
00:18:29.815 --> 00:18:31.654
- Yeah. For ten years. You know? We'll have

516
00:18:31.654 --> 00:18:33.734
- an... After AI crashes, there'll be a ten

517
00:18:33.734 --> 00:18:35.839
- year AI desert where people like, I won't

518
00:18:35.839 --> 00:18:37.680
- touch that stuff with a 10 foot pole.

519
00:18:37.680 --> 00:18:39.059
- Just like happened to telecom.

520
00:18:39.599 --> 00:18:41.359
- I mean, boy, there was no new investment

521
00:18:41.359 --> 00:18:44.160
- after 2000 for quite a while. So Yeah.

522
00:18:44.160 --> 00:18:46.400
- So we're gonna see that again. So so

523
00:18:46.400 --> 00:18:48.480
- these these guys have had... They're... They pegged

524
00:18:48.480 --> 00:18:50.160
- their money back, so they're not even gonna

525
00:18:50.160 --> 00:18:52.125
- get hurt. A lot of these companies are

526
00:18:52.125 --> 00:18:52.625
- actually...

527
00:18:53.485 --> 00:18:55.164
- Or they're being accused. We don't know that

528
00:18:55.164 --> 00:18:57.005
- they're being accused of buying shorts of their

529
00:18:57.005 --> 00:18:58.924
- own stock so they're gonna crash as it

530
00:18:58.924 --> 00:18:59.904
- evens out.

531
00:19:01.725 --> 00:19:02.545
- That's crazy.

532
00:19:03.085 --> 00:19:03.985
- That is crazy.

533
00:19:04.429 --> 00:19:05.970
- So so from an individual

534
00:19:06.349 --> 00:19:09.169
- networking... For the individual engineer's perspective,

535
00:19:09.950 --> 00:19:11.869
- the the first thing that that you might

536
00:19:11.869 --> 00:19:13.490
- think is, number one,

537
00:19:14.349 --> 00:19:15.950
- I think there's a lot of things that

538
00:19:15.950 --> 00:19:17.630
- go through my head, you know, if I'm

539
00:19:17.630 --> 00:19:20.014
- if I'm working in this space, which by

540
00:19:20.014 --> 00:19:21.575
- the way, I do work in this space.

541
00:19:21.575 --> 00:19:23.274
- So I do think about this.

542
00:19:24.615 --> 00:19:26.134
- One of the things I think that runs

543
00:19:26.134 --> 00:19:27.274
- through my head is,

544
00:19:27.734 --> 00:19:29.974
- oh, no. People won't be building these big

545
00:19:29.974 --> 00:19:32.154
- networks. The prices are gonna go down,

546
00:19:32.529 --> 00:19:34.930
- and the demand is gonna go down, and

547
00:19:34.930 --> 00:19:37.589
- therefore, there will just be less jobs

548
00:19:38.369 --> 00:19:40.150
- in the world of engineering.

549
00:19:41.250 --> 00:19:42.930
- What what do you think? Is that right,

550
00:19:42.930 --> 00:19:44.849
- or do you think that's that's kind of

551
00:19:44.849 --> 00:19:45.589
- off base?

552
00:19:46.130 --> 00:19:48.049
- I don't think so because almost all this

553
00:19:48.049 --> 00:19:50.345
- much money... I mean, here's how the money

554
00:19:50.345 --> 00:19:51.325
- is being spent.

555
00:19:51.944 --> 00:19:54.345
- There's certainly jobs created when they build these

556
00:19:54.345 --> 00:19:56.105
- data centers. You know, the guys who do

557
00:19:56.105 --> 00:19:58.744
- construction and bulldozers and all that stuff are

558
00:19:58.744 --> 00:20:00.585
- making money, but that's really not a big

559
00:20:00.585 --> 00:20:02.360
- pile of people to do. It's it's not

560
00:20:02.360 --> 00:20:04.299
- a whole lot different than building a giant,

561
00:20:04.519 --> 00:20:07.080
- you know, UPS warehouse. I mean, it's just

562
00:20:07.080 --> 00:20:08.619
- it's just a giant warehouse.

563
00:20:09.559 --> 00:20:11.320
- The real cost is in the card. So

564
00:20:11.320 --> 00:20:13.320
- the real the real cost is in the

565
00:20:13.320 --> 00:20:14.700
- electronics on the inside,

566
00:20:15.274 --> 00:20:17.595
- and then the real operating cost going on

567
00:20:17.595 --> 00:20:18.095
- is

568
00:20:18.554 --> 00:20:20.414
- for electricity and and water.

569
00:20:20.954 --> 00:20:21.694
- And so

570
00:20:21.994 --> 00:20:24.474
- when we're not spending that money anymore for

571
00:20:24.474 --> 00:20:26.634
- that, I don't see how that hurts the

572
00:20:26.634 --> 00:20:28.174
- other jobs. Because right now,

573
00:20:29.194 --> 00:20:30.255
- one of my friends

574
00:20:30.740 --> 00:20:32.680
- owns a phone company in a county

575
00:20:33.059 --> 00:20:35.220
- where the county next door, they built one

576
00:20:35.220 --> 00:20:36.840
- of these data centers in Arkansas.

577
00:20:37.859 --> 00:20:39.400
- He can't find an electrician

578
00:20:40.420 --> 00:20:42.019
- because... Because they didn't have that many to

579
00:20:42.019 --> 00:20:43.940
- start with, and all the electricians went to

580
00:20:43.940 --> 00:20:45.724
- work as a full time job at the

581
00:20:45.724 --> 00:20:47.644
- data center. That's one of the two kinds

582
00:20:47.644 --> 00:20:49.404
- of jobs they create. They take care of

583
00:20:49.404 --> 00:20:49.904
- the

584
00:20:50.205 --> 00:20:51.965
- they take care of the generators and all,

585
00:20:51.965 --> 00:20:53.644
- you know, all the stuff and and the

586
00:20:53.644 --> 00:20:56.065
- wiring and all that stuff. So the... So

587
00:20:56.205 --> 00:20:58.125
- so, you know, those folks will go back

588
00:20:58.125 --> 00:21:00.190
- to the economy, and we'll actually have we'll

589
00:21:00.190 --> 00:21:02.769
- actually have electricians who could build houses again.

590
00:21:03.430 --> 00:21:03.930
- Yeah.

591
00:21:04.590 --> 00:21:06.590
- Yeah. So I I don't think these jobs

592
00:21:06.590 --> 00:21:07.410
- will be lost.

593
00:21:08.029 --> 00:21:09.970
- AI is not generating jobs.

594
00:21:10.670 --> 00:21:12.750
- The actual AI industry is not generating jobs

595
00:21:12.750 --> 00:21:14.615
- at all. The typical data center has between

596
00:21:14.615 --> 00:21:16.695
- twenty and fifty employees. It's not a big

597
00:21:16.695 --> 00:21:18.234
- number. So Yeah.

598
00:21:18.934 --> 00:21:21.494
- Yeah. And so the other thing that as

599
00:21:21.494 --> 00:21:23.494
- a network engineer or as an engineer in

600
00:21:23.494 --> 00:21:26.394
- general, I might think or worry about is

601
00:21:27.015 --> 00:21:29.815
- all this automation and all these systems I've

602
00:21:29.815 --> 00:21:30.089
- built,

603
00:21:30.970 --> 00:21:31.470
- they

604
00:21:31.929 --> 00:21:33.869
- run on all of these things,

605
00:21:34.730 --> 00:21:36.509
- all these public models.

606
00:21:37.609 --> 00:21:39.929
- Like, what am I gonna do when all

607
00:21:39.929 --> 00:21:41.789
- of that... What's the disruption

608
00:21:42.904 --> 00:21:44.984
- when all of that starts to either go

609
00:21:44.984 --> 00:21:47.144
- away or just... It won't go away. Let's

610
00:21:47.144 --> 00:21:49.224
- be honest. It won't go away. It will

611
00:21:49.224 --> 00:21:50.684
- stop being maintained.

612
00:21:51.544 --> 00:21:54.184
- That's that's what will happen. There will be

613
00:21:54.184 --> 00:21:56.365
- no new features added essentially

614
00:21:57.049 --> 00:21:59.849
- over time, and the quality of the answers

615
00:21:59.849 --> 00:22:01.470
- will become worse and worse

616
00:22:02.410 --> 00:22:05.849
- as, you know, as it degrades. Because these

617
00:22:05.849 --> 00:22:07.929
- systems, they degrade over time. They... They're like

618
00:22:07.929 --> 00:22:10.890
- any other software. They require constant maintenance to

619
00:22:10.890 --> 00:22:12.625
- keep them up. And when the money goes

620
00:22:12.625 --> 00:22:14.244
- away, the maintenance goes away.

621
00:22:14.625 --> 00:22:16.944
- And people, like, no big vendor is just

622
00:22:16.944 --> 00:22:19.345
- gonna come along and say, okay. Well, we're

623
00:22:19.345 --> 00:22:22.244
- sorry you bought our AI network management product.

624
00:22:22.625 --> 00:22:24.384
- It's going to be shut down. That'll never

625
00:22:24.384 --> 00:22:27.090
- happen. They want the revenue from those rentals

626
00:22:27.090 --> 00:22:29.490
- and stuff. But I do think that over

627
00:22:29.490 --> 00:22:31.029
- time, the quality will reduce.

628
00:22:31.410 --> 00:22:31.910
- So

629
00:22:32.210 --> 00:22:34.470
- from that perspective, do you see any,

630
00:22:35.009 --> 00:22:35.509
- like,

631
00:22:35.809 --> 00:22:37.970
- may I actually wanna dis... I wanna disagree

632
00:22:37.970 --> 00:22:40.335
- with you. Okay. If if we have an

633
00:22:40.335 --> 00:22:41.234
- AI crash,

634
00:22:42.015 --> 00:22:44.095
- software that runs directly in the big data

635
00:22:44.095 --> 00:22:46.815
- centers is gonna stop overnight. It's just gonna

636
00:22:46.815 --> 00:22:47.715
- be shut down

637
00:22:48.734 --> 00:22:51.055
- because because the AI companies are gonna shut

638
00:22:51.055 --> 00:22:52.734
- down a whole bunch of the data centers.

639
00:22:52.734 --> 00:22:54.434
- And and if you have a

640
00:22:54.809 --> 00:22:55.309
- software

641
00:22:55.690 --> 00:22:57.869
- that actually works in those data centers,

642
00:22:58.649 --> 00:23:01.230
- you're you're out of business. And now

643
00:23:02.089 --> 00:23:05.929
- this is a crazy number. There's four... 40,000,000

644
00:23:05.929 --> 00:23:07.769
- apps have been written for AI. That that

645
00:23:07.769 --> 00:23:09.130
- was, like, a month ago, so now it's

646
00:23:09.130 --> 00:23:11.115
- probably 45.

647
00:23:11.734 --> 00:23:13.575
- None of them make any money, but the

648
00:23:13.575 --> 00:23:15.434
- fact is there is all these programs.

649
00:23:15.974 --> 00:23:17.755
- You know? Talk... You know?

650
00:23:18.454 --> 00:23:21.434
- Medical research, talk to your vet, you know,

651
00:23:21.494 --> 00:23:23.815
- schedule your bus rides. All these people have

652
00:23:23.815 --> 00:23:26.359
- these functions running on data centers. And if

653
00:23:26.359 --> 00:23:28.920
- they're running into made data centers and they

654
00:23:28.920 --> 00:23:30.839
- crash, the next morning they go to use

655
00:23:30.839 --> 00:23:32.680
- those, they're not there anymore. They're gonna go

656
00:23:32.680 --> 00:23:34.680
- completely dead because they're gonna turn off the

657
00:23:34.680 --> 00:23:37.000
- power at the data centers. What's gonna crash

658
00:23:37.000 --> 00:23:39.044
- is they can't afford to pay the bills

659
00:23:39.044 --> 00:23:41.924
- with data centers. Again, anthropically, these guys don't

660
00:23:41.924 --> 00:23:43.765
- own these data centers. The data center guys

661
00:23:43.765 --> 00:23:44.984
- are gonna go no revenue,

662
00:23:45.525 --> 00:23:46.585
- no no computing.

663
00:23:47.765 --> 00:23:50.424
- The other funny thing about all this is

664
00:23:50.869 --> 00:23:52.809
- we keep talking about all these data centers.

665
00:23:53.830 --> 00:23:55.750
- Over half of them don't have the chips

666
00:23:55.750 --> 00:23:57.509
- in them yet. They're still waiting. There's a

667
00:23:57.509 --> 00:23:59.910
- two year backlog on these chips. We have

668
00:23:59.910 --> 00:24:01.590
- a whole bunch of empty data centers that

669
00:24:01.590 --> 00:24:04.090
- don't even have chips in them yet. But

670
00:24:04.150 --> 00:24:05.529
- but... No. I think... Now

671
00:24:06.545 --> 00:24:08.224
- what you're talking about that we keep going,

672
00:24:08.224 --> 00:24:10.785
- I have three friends just in this area

673
00:24:10.785 --> 00:24:12.865
- where we live who have built their own

674
00:24:12.865 --> 00:24:15.345
- AI data centers, which just means one rack

675
00:24:15.345 --> 00:24:17.505
- of of the cards that they run for

676
00:24:17.505 --> 00:24:18.404
- their own business.

677
00:24:18.990 --> 00:24:21.490
- And so those guys will keep running forever,

678
00:24:22.029 --> 00:24:23.490
- but the problem will be

679
00:24:23.950 --> 00:24:26.509
- if if the anthropic truly crashes, when those

680
00:24:26.509 --> 00:24:28.430
- cards wear out, they may never find any

681
00:24:28.430 --> 00:24:28.930
- replacements

682
00:24:29.789 --> 00:24:32.035
- because that's not the same cards. There's gonna

683
00:24:32.035 --> 00:24:33.394
- be a flood of the market of the

684
00:24:33.394 --> 00:24:36.214
- big cards from the data centers, but that's...

685
00:24:36.434 --> 00:24:38.755
- Those are really not practical for me and

686
00:24:38.755 --> 00:24:40.674
- you to put into our own little rack

687
00:24:40.674 --> 00:24:42.775
- of equipment because they suck too much power.

688
00:24:44.755 --> 00:24:46.375
- So these guys are sort of running,

689
00:24:46.940 --> 00:24:49.440
- you know, three or four generations of NVIDIA

690
00:24:49.500 --> 00:24:51.259
- old. And so they're... And they're not even

691
00:24:51.259 --> 00:24:53.579
- making those now anymore. So so, you know,

692
00:24:53.579 --> 00:24:56.140
- the little data centers... The new industry, if

693
00:24:56.140 --> 00:24:58.619
- the big IA crashes, little AI, somebody has

694
00:24:58.619 --> 00:25:00.480
- to step in and make those cards again.

695
00:25:00.779 --> 00:25:01.759
- In the long term

696
00:25:02.394 --> 00:25:04.315
- model is there will be hundreds of thousands

697
00:25:04.315 --> 00:25:06.075
- of these little data centers in the world,

698
00:25:06.075 --> 00:25:08.474
- and and whoever supplies them will make good

699
00:25:08.474 --> 00:25:08.974
- money.

700
00:25:09.674 --> 00:25:11.454
- Kinda like when NVIDIA made

701
00:25:12.075 --> 00:25:13.214
- cards for game

702
00:25:14.075 --> 00:25:16.154
- game. Yeah. But that's what... But now they're

703
00:25:16.154 --> 00:25:18.819
- they're gonna go from big big Nvidia back

704
00:25:18.819 --> 00:25:21.480
- to little Nvidia, and it's a permanent business

705
00:25:21.700 --> 00:25:23.139
- if they don't if they don't die in

706
00:25:23.139 --> 00:25:25.619
- this crash. Right? There's a good chance they

707
00:25:25.619 --> 00:25:26.440
- die. So

708
00:25:26.819 --> 00:25:28.899
- now in our industry, here's an interesting set

709
00:25:28.899 --> 00:25:30.599
- of people who have a lot of risk.

710
00:25:32.434 --> 00:25:32.934
- Zayo

711
00:25:33.795 --> 00:25:34.295
- and

712
00:25:34.755 --> 00:25:37.555
- and Lumen are... And, and and the old

713
00:25:37.555 --> 00:25:38.055
- CenturyLink

714
00:25:38.914 --> 00:25:41.714
- are are building... They're building massive amounts of

715
00:25:41.714 --> 00:25:43.815
- fiber in an... On the interstate networks

716
00:25:44.275 --> 00:25:45.575
- to carry all this traffic.

717
00:25:45.950 --> 00:25:47.390
- Because when you make a request to a

718
00:25:47.390 --> 00:25:49.309
- data center, it literally... You might use seven

719
00:25:49.309 --> 00:25:50.589
- or eight or 10 or 20 of these

720
00:25:50.589 --> 00:25:52.589
- data centers. So the amount of traffic bouncing

721
00:25:52.589 --> 00:25:55.230
- between them is gigantic. So they're spending a

722
00:25:55.230 --> 00:25:55.730
- fortune

723
00:25:56.190 --> 00:25:57.410
- on these new routes.

724
00:25:57.789 --> 00:25:59.710
- And if if there is an AI crash,

725
00:25:59.710 --> 00:26:02.325
- their traffic's just gonna go away and and,

726
00:26:02.325 --> 00:26:04.484
- you know, to some much lower level. And

727
00:26:04.484 --> 00:26:06.724
- they're and they're... I mean, that's all... That

728
00:26:06.724 --> 00:26:08.565
- might be all that it takes to to

729
00:26:08.565 --> 00:26:10.265
- finally put CenturyLink under.

730
00:26:10.884 --> 00:26:12.184
- They're they're on the edge anyway.

731
00:26:12.644 --> 00:26:14.984
- So... Yeah. I... Yeah. I remember,

732
00:26:16.230 --> 00:26:17.509
- at a place that I worked a long

733
00:26:17.509 --> 00:26:19.829
- time ago, there was, there was a data

734
00:26:19.829 --> 00:26:21.609
- center shell that was built out,

735
00:26:22.230 --> 00:26:25.109
- in... It was like '99 or something. Somebody

736
00:26:25.109 --> 00:26:27.029
- took an old coke bottling plant, and they're

737
00:26:27.029 --> 00:26:29.109
- like, everybody's gonna be building, you know, they...

738
00:26:29.109 --> 00:26:30.549
- So they... So they built the shell and

739
00:26:30.549 --> 00:26:32.335
- then everything crashed, and it was just a

740
00:26:32.335 --> 00:26:34.734
- shell. And then the place I worked went

741
00:26:34.734 --> 00:26:36.414
- and bought it at a nice discount, and

742
00:26:36.414 --> 00:26:38.575
- it was like the property was already developed.

743
00:26:38.575 --> 00:26:39.954
- It was awesome. It was perfect.

744
00:26:40.255 --> 00:26:40.755
- And,

745
00:26:41.294 --> 00:26:43.054
- and they were able to take that asset,

746
00:26:43.054 --> 00:26:44.734
- and and they would never would have been

747
00:26:44.734 --> 00:26:45.794
- able to build it themselves.

748
00:26:46.420 --> 00:26:48.839
- But because of, you know, the the aftereffects

749
00:26:48.900 --> 00:26:50.099
- of the crash, right,

750
00:26:50.740 --> 00:26:52.820
- they they found a lot of utility. Do

751
00:26:52.820 --> 00:26:55.619
- you see anything similar to that happening here

752
00:26:55.619 --> 00:26:57.380
- with these assets? I mean, yeah, we can't

753
00:26:57.380 --> 00:26:59.080
- run them in our garage or whatever.

754
00:26:59.384 --> 00:27:00.984
- But is there, do you think there's a

755
00:27:00.984 --> 00:27:02.904
- secondary effect where that same sort of thing

756
00:27:02.904 --> 00:27:03.404
- happens?

757
00:27:06.025 --> 00:27:07.625
- I don't know that the buildings are ever

758
00:27:07.625 --> 00:27:09.224
- gonna be useful. Have you been to one

759
00:27:09.224 --> 00:27:11.305
- of the day centers, either of you? Just

760
00:27:11.305 --> 00:27:13.065
- on just on the just on the outside

761
00:27:13.065 --> 00:27:15.109
- of the building, they have, forty and fifty

762
00:27:15.250 --> 00:27:16.950
- tractor trailer size generators.

763
00:27:19.329 --> 00:27:20.849
- You have to sit... You have to see

764
00:27:20.849 --> 00:27:23.349
- one of these to believe them. And so

765
00:27:23.410 --> 00:27:25.430
- when they don't need those generators anymore,

766
00:27:26.049 --> 00:27:28.130
- GE Capital and whoever holds the lease on

767
00:27:28.130 --> 00:27:29.845
- those generators is gonna come and drag them

768
00:27:29.845 --> 00:27:32.085
- away because they will be reusable somewhere in

769
00:27:32.085 --> 00:27:32.984
- the world. Right?

770
00:27:34.005 --> 00:27:34.744
- And so

771
00:27:35.125 --> 00:27:37.284
- so that... It's a very weird building. These

772
00:27:37.284 --> 00:27:39.444
- were... These are being built, and they have

773
00:27:39.444 --> 00:27:42.519
- this gigantic water circulation system and all. They're

774
00:27:42.519 --> 00:27:43.740
- really not friendly

775
00:27:44.039 --> 00:27:45.579
- towards a normal use.

776
00:27:45.960 --> 00:27:48.039
- I I I kind of... I'm at the

777
00:27:48.039 --> 00:27:50.380
- Northern Virginia at the two thousand crash.

778
00:27:50.680 --> 00:27:53.160
- I remember the row of dead sea legs,

779
00:27:53.160 --> 00:27:55.480
- and it took ten to fifteen years before

780
00:27:55.480 --> 00:27:57.019
- people moved into those buildings.

781
00:27:57.335 --> 00:27:59.734
- I kinda see the same thing. So, I

782
00:27:59.734 --> 00:28:01.734
- mean, some of them will get reused, but

783
00:28:01.734 --> 00:28:02.934
- but, you know, when you have one of

784
00:28:02.934 --> 00:28:05.174
- these giant data centers in rural Arkansas, there's

785
00:28:05.174 --> 00:28:06.634
- no use for it. It's

786
00:28:07.494 --> 00:28:09.894
- never gonna get used again. So some of

787
00:28:09.894 --> 00:28:12.210
- them will get used, but they're they're not...

788
00:28:12.210 --> 00:28:14.609
- They can't easily... It'd be... Cost more money

789
00:28:14.609 --> 00:28:16.549
- to convert one of those into a Walmart,

790
00:28:17.089 --> 00:28:18.609
- you know, shipping center than it would be

791
00:28:18.609 --> 00:28:20.130
- to build a new one, I think, because

792
00:28:20.130 --> 00:28:21.650
- they're not they're not laid out for any

793
00:28:21.650 --> 00:28:22.549
- of those functions.

794
00:28:23.490 --> 00:28:24.069
- So... Yeah.

795
00:28:25.035 --> 00:28:27.115
- I I mean, it's a very wasteful, weird

796
00:28:27.115 --> 00:28:29.535
- thing that we're doing here. So... Yeah. Yeah.

797
00:28:30.394 --> 00:28:30.894
- Interesting.

798
00:28:31.195 --> 00:28:33.115
- Yeah. So I know I saw some statistics

799
00:28:33.115 --> 00:28:34.715
- recently, and I was looking for them, and

800
00:28:34.715 --> 00:28:36.715
- I can't find them. But something like a

801
00:28:36.715 --> 00:28:38.559
- 119,000,000,

802
00:28:38.559 --> 00:28:39.440
- a 190,000,000

803
00:28:39.440 --> 00:28:42.319
- miles, something crazy of new fiber needed in

804
00:28:42.319 --> 00:28:43.220
- The United States

805
00:28:43.599 --> 00:28:44.419
- in the next

806
00:28:44.799 --> 00:28:46.419
- ten years or something.

807
00:28:46.799 --> 00:28:49.279
- Like... So the fiber companies that are building

808
00:28:49.279 --> 00:28:51.619
- up for this are going to be hit

809
00:28:52.134 --> 00:28:53.894
- if this goes down. That's going to be

810
00:28:53.894 --> 00:28:55.674
- something that's going to happen. What?

811
00:28:56.054 --> 00:28:57.994
- Yes. Well, let's talk about Corning.

812
00:29:00.535 --> 00:29:02.375
- So why... I was talking about, you know,

813
00:29:02.375 --> 00:29:04.615
- Zayo and Interstate Fiber. That's the small piece

814
00:29:04.615 --> 00:29:05.434
- of the fiber.

815
00:29:05.779 --> 00:29:07.639
- The data center that they're building

816
00:29:08.259 --> 00:29:11.700
- at Louisiana that's almost finished has 1,000,000 feet

817
00:29:11.700 --> 00:29:13.159
- of fiber in it.

818
00:29:15.220 --> 00:29:17.139
- That... That's where all the fiber's going. It's

819
00:29:17.139 --> 00:29:18.359
- going between racks.

820
00:29:18.740 --> 00:29:20.359
- All these racks are not...

821
00:29:21.265 --> 00:29:22.304
- You know, in the old days, we would...

822
00:29:22.304 --> 00:29:23.744
- Inside the data center, we would have had

823
00:29:23.744 --> 00:29:25.664
- a lot of CAT six cables. Right? These

824
00:29:25.664 --> 00:29:26.644
- are all fiber.

825
00:29:27.025 --> 00:29:28.865
- Card to card to card to card, there's

826
00:29:28.865 --> 00:29:31.105
- shelf to shelf, rack to rack. It's all

827
00:29:31.105 --> 00:29:32.005
- fiber everywhere.

828
00:29:32.464 --> 00:29:33.125
- And so

829
00:29:33.825 --> 00:29:35.825
- so that... They're... They are sucking all of

830
00:29:35.825 --> 00:29:36.325
- the

831
00:29:36.970 --> 00:29:37.870
- fiber capability

832
00:29:38.570 --> 00:29:40.970
- from people like Corning because they're paying a

833
00:29:40.970 --> 00:29:43.370
- premium price. And so now there's a shortage

834
00:29:43.370 --> 00:29:45.450
- of... If you're an ISP wanting to build

835
00:29:45.450 --> 00:29:47.850
- fiber now, there's... We just went from... We

836
00:29:47.850 --> 00:29:50.250
- finally got over the we finally got over

837
00:29:50.250 --> 00:29:51.230
- the whole COVID

838
00:29:51.964 --> 00:29:53.884
- supply chain stuff. We got back to normal,

839
00:29:53.884 --> 00:29:56.285
- and now the now the the wait times

840
00:29:56.285 --> 00:29:57.964
- are back up again. We're up to six

841
00:29:57.964 --> 00:30:00.285
- to nine months to get fiber, and they're

842
00:30:00.285 --> 00:30:02.285
- warning it's gonna get worse because all the

843
00:30:02.285 --> 00:30:04.865
- fiber is going to the data centers. Now

844
00:30:05.005 --> 00:30:07.250
- Corning might take it in the shorts. They're

845
00:30:07.250 --> 00:30:10.630
- building three brand new factories to manufacture fiber.

846
00:30:11.170 --> 00:30:13.650
- At those quantities, if AI crashes, no one's

847
00:30:13.650 --> 00:30:14.789
- gonna want that fiber.

848
00:30:15.730 --> 00:30:17.650
- Yeah. Yeah. That was that was my other...

849
00:30:17.650 --> 00:30:19.570
- Like, that is another area that I worry

850
00:30:19.570 --> 00:30:21.970
- about. Now Yeah. On the on the other

851
00:30:21.970 --> 00:30:22.470
- side,

852
00:30:22.875 --> 00:30:25.355
- right, if I'm if I'm an individual engineer

853
00:30:25.355 --> 00:30:26.335
- looking at this,

854
00:30:26.954 --> 00:30:28.795
- it might... Like I... Like you said, you

855
00:30:28.795 --> 00:30:31.295
- know, there are all these... The small AI,

856
00:30:31.434 --> 00:30:33.835
- which I do have one one problem or

857
00:30:33.835 --> 00:30:36.234
- one thing I don't necessarily... I'm I'm a

858
00:30:36.234 --> 00:30:37.295
- little bit. But

859
00:30:40.519 --> 00:30:43.099
- the the average engineer also has to think

860
00:30:43.240 --> 00:30:44.779
- on the positive side

861
00:30:45.080 --> 00:30:46.920
- that if people are putting in, again, if

862
00:30:46.920 --> 00:30:47.820
- there's a repatriation

863
00:30:48.599 --> 00:30:51.080
- or if there's a little... We're gonna start

864
00:30:51.080 --> 00:30:53.684
- building little data centers for each company that's

865
00:30:53.684 --> 00:30:56.325
- gonna run AI or each of these access

866
00:30:56.325 --> 00:30:58.884
- providers. Right? Like, WISP and stuff like that

867
00:30:58.884 --> 00:31:01.204
- start saying, well, I wanna sell AI to

868
00:31:01.204 --> 00:31:03.924
- my customers. Right? I wanna sell bus schedule

869
00:31:03.924 --> 00:31:06.005
- AI to my customer or whatever it is

870
00:31:06.005 --> 00:31:08.424
- because that adds value, and I can charge

871
00:31:08.700 --> 00:31:10.940
- a little bit more. It adds stickiness because

872
00:31:10.940 --> 00:31:11.519
- I know,

873
00:31:12.059 --> 00:31:13.579
- like, a lot of the problem with access

874
00:31:13.579 --> 00:31:15.279
- providers is there's no stickiness.

875
00:31:15.659 --> 00:31:16.880
- People roll constantly.

876
00:31:17.259 --> 00:31:20.299
- There's a constant churn. And so if if

877
00:31:20.299 --> 00:31:21.759
- if you had somebody like...

878
00:31:22.095 --> 00:31:22.914
- I don't know,

879
00:31:23.454 --> 00:31:25.634
- big name ISPs used to

880
00:31:26.414 --> 00:31:28.974
- provide voice mail because it was a sticky

881
00:31:28.974 --> 00:31:30.815
- thing. Once my voice mail and email and

882
00:31:30.815 --> 00:31:31.714
- everything else,

883
00:31:32.095 --> 00:31:34.335
- maybe this becomes a service like that. So

884
00:31:34.335 --> 00:31:35.794
- on the positive side,

885
00:31:36.640 --> 00:31:38.099
- engineers have to be thinking

886
00:31:38.799 --> 00:31:40.900
- maybe the big stuff goes away,

887
00:31:41.359 --> 00:31:43.359
- but there's a lot of little stuff to

888
00:31:43.359 --> 00:31:44.099
- be built

889
00:31:44.720 --> 00:31:46.559
- if I'm in the right market. And if

890
00:31:46.559 --> 00:31:48.019
- I do the right thing,

891
00:31:48.559 --> 00:31:50.900
- maybe I can get into that market

892
00:31:51.704 --> 00:31:53.484
- and build something out of that.

893
00:31:54.025 --> 00:31:54.505
- Is that

894
00:31:55.224 --> 00:31:56.904
- And and and that's what... I... I'm thinking

895
00:31:56.904 --> 00:31:58.984
- that takes a decade because right now, again,

896
00:31:58.984 --> 00:32:00.345
- I said I know three guys who have

897
00:32:00.345 --> 00:32:01.325
- built their own.

898
00:32:01.625 --> 00:32:03.625
- The learning curve just to get it working

899
00:32:03.625 --> 00:32:05.464
- for a guy who really knows how AI

900
00:32:05.464 --> 00:32:06.744
- works is, like, a year and a half

901
00:32:06.744 --> 00:32:07.440
- or two years.

902
00:32:08.000 --> 00:32:09.759
- And and there's no off the shelf anybody

903
00:32:09.759 --> 00:32:10.960
- to help you do that. So the... And

904
00:32:10.960 --> 00:32:11.940
- industry has

905
00:32:12.240 --> 00:32:14.240
- to grow to say, I can come and

906
00:32:14.240 --> 00:32:16.659
- just give you a preloaded little

907
00:32:17.039 --> 00:32:19.119
- data center that works. These guys have to

908
00:32:19.119 --> 00:32:21.119
- figure out everything. They gotta figure out where

909
00:32:21.119 --> 00:32:23.045
- to get the cards, how to train the

910
00:32:23.045 --> 00:32:25.065
- model, how to get rid of the hallucinations.

911
00:32:25.125 --> 00:32:26.085
- I mean, they have a lot of work

912
00:32:26.085 --> 00:32:27.845
- to do before they can actually do anything

913
00:32:27.845 --> 00:32:28.825
- useful with it.

914
00:32:29.125 --> 00:32:31.365
- So so I I believe that is the

915
00:32:31.365 --> 00:32:33.365
- future model. You just... I believe that's the

916
00:32:33.365 --> 00:32:35.144
- ongoing survivable model.

917
00:32:35.470 --> 00:32:37.390
- Now the other thing that's happening we don't

918
00:32:37.390 --> 00:32:38.370
- really know about.

919
00:32:39.549 --> 00:32:41.950
- I... We know Bank of America has their

920
00:32:41.950 --> 00:32:43.789
- own AI. We, you know, we know Ford

921
00:32:43.789 --> 00:32:45.950
- Motor Company probably has their own AI, and

922
00:32:45.950 --> 00:32:47.390
- we don't know how they're doing it because

923
00:32:47.390 --> 00:32:48.370
- they're not talking.

924
00:32:48.724 --> 00:32:50.565
- So so they're doing this on a bigger

925
00:32:50.565 --> 00:32:52.105
- scale than my three buddies,

926
00:32:52.404 --> 00:32:53.765
- but they're not doing it on the scale

927
00:32:53.765 --> 00:32:55.845
- of the big data centers. And the reason

928
00:32:55.845 --> 00:32:58.805
- they're all doing it is one, cost. But

929
00:32:58.805 --> 00:33:00.644
- two, they're not about to get their data

930
00:33:00.644 --> 00:33:03.400
- to open AI. Like, no. You're you're not

931
00:33:03.400 --> 00:33:05.799
- gonna get my stuff. Because once you put

932
00:33:05.799 --> 00:33:07.720
- it in the cloud, it's there for everybody.

933
00:33:07.720 --> 00:33:09.820
- So nobody wants their stuff in the cloud.

934
00:33:10.279 --> 00:33:11.480
- You know, why I think they're all gonna

935
00:33:11.480 --> 00:33:13.080
- be surprised that one of these days is

936
00:33:13.080 --> 00:33:15.400
- find out that when all their analysts went

937
00:33:15.400 --> 00:33:17.019
- home in the evening and did their spreadsheets

938
00:33:17.080 --> 00:33:18.515
- on on their home OpenAI,

939
00:33:18.815 --> 00:33:20.674
- all their stuff is in the cloud anyway.

940
00:33:21.775 --> 00:33:23.795
- Yeah. Oh, yeah. Yeah. Definitely.

941
00:33:24.575 --> 00:33:25.775
- But but, you know, but I... But, you

942
00:33:25.775 --> 00:33:28.174
- know, so these... There is there is a...

943
00:33:28.174 --> 00:33:30.335
- That industry of people building their own who

944
00:33:30.335 --> 00:33:32.654
- are not mom and pops. They're bigger now.

945
00:33:32.654 --> 00:33:34.109
- And and so we don't know how many

946
00:33:34.109 --> 00:33:35.890
- of those there are. There's no statistics,

947
00:33:36.589 --> 00:33:38.349
- and none of them are talking about it.

948
00:33:38.349 --> 00:33:40.289
- I mean, you... I can't find any

949
00:33:40.910 --> 00:33:43.869
- literature articles, mench... People don't even talk about

950
00:33:43.869 --> 00:33:45.089
- it in their annual reports.

951
00:33:45.565 --> 00:33:47.085
- They have an AI department, and you don't

952
00:33:47.085 --> 00:33:48.525
- know if they're building their own or they're

953
00:33:48.525 --> 00:33:50.684
- buying it from Entropic. We just don't know.

954
00:33:50.684 --> 00:33:52.285
- So... But I I think there's a lot

955
00:33:52.285 --> 00:33:53.025
- of those.

956
00:33:53.644 --> 00:33:54.144
- I

957
00:33:54.684 --> 00:33:56.845
- mean, my butt... One buddy who's... How... Always

958
00:33:56.845 --> 00:33:58.445
- been in that business, he was pretty sure

959
00:33:58.445 --> 00:34:00.710
- that that, the Bank of America, as an

960
00:34:00.710 --> 00:34:02.630
- example, jumped into a great way to do

961
00:34:02.630 --> 00:34:04.390
- their own. But but but we don't have

962
00:34:04.390 --> 00:34:06.549
- a group. So there will be survivors of

963
00:34:06.549 --> 00:34:08.409
- the current model. But, again,

964
00:34:10.070 --> 00:34:12.150
- Bank of America probably can run their whole

965
00:34:12.150 --> 00:34:14.765
- business on three racks of these cards. I

966
00:34:14.765 --> 00:34:17.005
- mean, we don't need this giant data center

967
00:34:17.005 --> 00:34:18.925
- when it's just... Let's track all the money

968
00:34:18.925 --> 00:34:20.625
- in and out of my customers' accounts.

969
00:34:21.244 --> 00:34:23.585
- These these chips are amazingly fast,

970
00:34:24.045 --> 00:34:25.405
- so they can do a lot of work

971
00:34:25.405 --> 00:34:27.849
- with a fairly small setup. So Yeah. So

972
00:34:27.849 --> 00:34:29.369
- they might have a million doll... They might

973
00:34:29.369 --> 00:34:31.289
- have a million dollar data center, which is

974
00:34:31.289 --> 00:34:33.449
- all they need. So Yeah. I was gonna

975
00:34:33.449 --> 00:34:34.989
- say it, but but still, it's

976
00:34:35.449 --> 00:34:37.630
- a set of racks sitting in

977
00:34:38.010 --> 00:34:39.630
- an existing data center

978
00:34:40.074 --> 00:34:41.215
- that has to be connected,

979
00:34:41.914 --> 00:34:43.535
- managed, dealt with,

980
00:34:44.155 --> 00:34:44.974
- etcetera, etcetera.

981
00:34:45.675 --> 00:34:46.494
- And and

982
00:34:46.954 --> 00:34:49.515
- there will be some value or a lot

983
00:34:49.515 --> 00:34:50.255
- of value

984
00:34:50.635 --> 00:34:52.255
- in engineers who learn

985
00:34:52.635 --> 00:34:53.454
- how to

986
00:34:54.074 --> 00:34:57.039
- manage and maintain such systems at that scale.

987
00:34:57.039 --> 00:34:58.179
- It seems to be Absolutely.

988
00:34:58.880 --> 00:35:01.519
- That's... Because there's not that many AI experts

989
00:35:01.519 --> 00:35:03.760
- today that actually know that kind of stuff.

990
00:35:03.760 --> 00:35:05.679
- So there... That's a real growth area, I

991
00:35:05.679 --> 00:35:06.179
- think.

992
00:35:06.639 --> 00:35:07.139
- Absolutely.

993
00:35:07.440 --> 00:35:08.494
- So, yep.

994
00:35:09.275 --> 00:35:10.414
- So... Yeah. Interesting.

995
00:35:10.954 --> 00:35:12.394
- Yeah. So I don't know. I mean, I

996
00:35:12.394 --> 00:35:14.954
- I see the I see the QTR and

997
00:35:14.954 --> 00:35:17.275
- the Hedgie, and I follow Hedgie on x

998
00:35:17.275 --> 00:35:19.134
- and all these other people.

999
00:35:19.514 --> 00:35:20.815
- And they're all, like,

1000
00:35:21.119 --> 00:35:21.940
- just explaining

1001
00:35:22.239 --> 00:35:24.500
- the financing and what's going on

1002
00:35:25.199 --> 00:35:27.779
- and what it's... How it's warping the market.

1003
00:35:28.400 --> 00:35:30.319
- And and to some degree, this worries me

1004
00:35:30.319 --> 00:35:31.940
- as well as the market warpage

1005
00:35:32.319 --> 00:35:34.099
- and the the negative fallout

1006
00:35:35.054 --> 00:35:35.554
- of...

1007
00:35:35.855 --> 00:35:38.094
- What was the .com? It was, like, 1%

1008
00:35:38.094 --> 00:35:39.474
- of the American economy

1009
00:35:39.775 --> 00:35:41.934
- or 2%. So it wasn't... That wasn't that

1010
00:35:41.934 --> 00:35:42.434
- much.

1011
00:35:42.815 --> 00:35:43.315
- AI

1012
00:35:43.934 --> 00:35:45.855
- right now is, like, three to 5% of

1013
00:35:45.855 --> 00:35:48.815
- the American economy. Yeah. Somewhere three to somewhere

1014
00:35:48.815 --> 00:35:51.190
- three to five. That's what everyone says. And

1015
00:35:51.190 --> 00:35:54.329
- that's a lot. Yeah. That's direct. That's not

1016
00:35:54.710 --> 00:35:55.210
- electricians,

1017
00:35:55.829 --> 00:35:56.809
- and that's not

1018
00:35:57.349 --> 00:35:57.849
- concrete,

1019
00:35:58.230 --> 00:36:01.369
- and that's not... You know, that's direct spend.

1020
00:36:01.989 --> 00:36:03.050
- And so that's

1021
00:36:03.594 --> 00:36:04.094
- crazy.

1022
00:36:05.675 --> 00:36:07.614
- That's a crazy amount of

1023
00:36:08.795 --> 00:36:11.135
- of stuff that's that's going down,

1024
00:36:11.594 --> 00:36:12.735
- that could go down

1025
00:36:13.114 --> 00:36:15.215
- if there's if there's a a financial

1026
00:36:15.940 --> 00:36:16.440
- fallout

1027
00:36:16.980 --> 00:36:17.960
- in the economy.

1028
00:36:18.739 --> 00:36:21.079
- Well, now now, you know...

1029
00:36:21.539 --> 00:36:23.219
- So now I ask you the question. When's

1030
00:36:23.219 --> 00:36:24.199
- it gonna crash?

1031
00:36:27.954 --> 00:36:30.594
- I'm Nobody nobody knows. None of the none

1032
00:36:30.594 --> 00:36:32.434
- of the pros know. None of the analysts

1033
00:36:32.434 --> 00:36:34.054
- know. They're all saying

1034
00:36:34.514 --> 00:36:36.994
- between this weekend and 2028.

1035
00:36:36.994 --> 00:36:38.614
- That's, like, that's the best prediction

1036
00:36:39.074 --> 00:36:40.214
- I I see it.

1037
00:36:40.760 --> 00:36:42.280
- Because right now, it's this house of cards,

1038
00:36:42.280 --> 00:36:44.280
- and they're putting a lot of money in

1039
00:36:44.280 --> 00:36:46.359
- to keep the house from sinking. So, you

1040
00:36:46.359 --> 00:36:48.199
- know, they they... It's a matter of how

1041
00:36:48.199 --> 00:36:49.960
- long they can keep the game going before

1042
00:36:49.960 --> 00:36:51.639
- they just can't keep the game. Yeah. I

1043
00:36:51.639 --> 00:36:53.420
- mean, we're watching we're watching

1044
00:36:53.960 --> 00:36:54.460
- billionaires

1045
00:36:55.319 --> 00:36:55.819
- who

1046
00:36:56.364 --> 00:36:57.184
- who are

1047
00:36:57.885 --> 00:37:00.945
- pledging their their their entire fortune

1048
00:37:02.204 --> 00:37:03.905
- to keep this game going.

1049
00:37:04.364 --> 00:37:05.505
- And you have to Couldn't happen.

1050
00:37:06.684 --> 00:37:08.465
- It couldn't happen to a nice aesthetic.

1051
00:37:12.820 --> 00:37:14.119
- Yeah. I mean,

1052
00:37:14.500 --> 00:37:15.239
- it's just...

1053
00:37:15.780 --> 00:37:17.699
- It just can't go forever. I mean, I

1054
00:37:17.699 --> 00:37:20.420
- saw... I'm like, oh, this company is seeking

1055
00:37:20.420 --> 00:37:22.980
- another 7,000,000,000. This company is seeking another 10,000,000,000.

1056
00:37:22.980 --> 00:37:25.005
- I know. It's like, where The numbers are

1057
00:37:25.005 --> 00:37:27.804
- the numbers are mind boggling. Yeah. Yeah. Who

1058
00:37:27.804 --> 00:37:30.444
- has the money? Who is going to invest

1059
00:37:30.444 --> 00:37:32.444
- this kind of money? And how are you

1060
00:37:32.444 --> 00:37:34.065
- not draining investment

1061
00:37:34.684 --> 00:37:37.164
- in every other sector of the economy when

1062
00:37:37.164 --> 00:37:37.984
- you do this?

1063
00:37:38.460 --> 00:37:41.119
- And, like, really, is the return on investment

1064
00:37:41.179 --> 00:37:41.920
- for AI

1065
00:37:42.699 --> 00:37:43.760
- that much better

1066
00:37:44.139 --> 00:37:46.619
- than the return on investment for, like, building

1067
00:37:46.619 --> 00:37:47.519
- rural broadband

1068
00:37:48.059 --> 00:37:48.800
- or building

1069
00:37:49.340 --> 00:37:51.519
- SATCOM networks that can do broadband

1070
00:37:52.380 --> 00:37:55.965
- or, like, something that... Even car car plants.

1071
00:37:55.965 --> 00:37:58.625
- Like, is it... Like, these things produce physical

1072
00:37:58.684 --> 00:38:00.304
- stuff that people can use.

1073
00:38:00.764 --> 00:38:03.184
- And here we have this ephemeral thing.

1074
00:38:03.885 --> 00:38:05.744
- I can go ask the AI questions.

1075
00:38:06.630 --> 00:38:07.130
- Okay.

1076
00:38:07.510 --> 00:38:09.589
- Really? That's more of a return on value

1077
00:38:09.589 --> 00:38:10.089
- than...

1078
00:38:10.469 --> 00:38:12.869
- Like, I Right. I... Our return on investment,

1079
00:38:12.869 --> 00:38:14.309
- I don't I don't... I'm not seeing Well,

1080
00:38:14.309 --> 00:38:16.069
- right now, the return on investment is so

1081
00:38:16.069 --> 00:38:18.069
- gigantically negative. They have a long way to

1082
00:38:18.069 --> 00:38:19.529
- go. They even get to zero.

1083
00:38:20.469 --> 00:38:23.324
- Yeah. So, yeah, building building building anything else

1084
00:38:23.324 --> 00:38:25.244
- that people will actually buy is a way

1085
00:38:25.244 --> 00:38:28.605
- better investment. So... Yeah. Yeah. So Well, these

1086
00:38:28.605 --> 00:38:31.264
- guys... Let... Let's be honest. These guys

1087
00:38:32.364 --> 00:38:34.545
- are benefiting from the gigantic

1088
00:38:34.844 --> 00:38:35.344
- stock

1089
00:38:35.885 --> 00:38:36.704
- price increases

1090
00:38:37.789 --> 00:38:39.329
- of the of the big companies.

1091
00:38:40.110 --> 00:38:42.190
- And, again, I think the smart ones have

1092
00:38:42.190 --> 00:38:43.949
- hedged their bet. They're not gonna get killed

1093
00:38:43.949 --> 00:38:46.190
- when it crashes. They're already ready for that.

1094
00:38:46.190 --> 00:38:47.329
- So... Yeah.

1095
00:38:48.030 --> 00:38:49.884
- Some of them will because not all not

1096
00:38:49.884 --> 00:38:51.724
- all rich guys are smart guys. You know?

1097
00:38:51.724 --> 00:38:52.224
- Yeah.

1098
00:38:52.844 --> 00:38:54.844
- Well, and not all not all not all

1099
00:38:54.844 --> 00:38:55.904
- hedges work

1100
00:38:56.844 --> 00:38:59.244
- in the long run. You know? The big

1101
00:38:59.244 --> 00:39:00.924
- big fans of mice and Well, if this

1102
00:39:00.924 --> 00:39:02.444
- is a really ugly crash, none of the

1103
00:39:02.444 --> 00:39:04.125
- hedges may work. This may be just so

1104
00:39:04.125 --> 00:39:06.380
- ugly that this blows past the hedges. Right?

1105
00:39:06.380 --> 00:39:08.539
- Yeah. Exactly. And that's kind of part of

1106
00:39:08.539 --> 00:39:10.780
- the impossible. Part part of the question or

1107
00:39:10.780 --> 00:39:11.680
- part of the problem.

1108
00:39:12.140 --> 00:39:14.780
- I don't know the answer to. Alright. So

1109
00:39:14.780 --> 00:39:15.599
- some negatives,

1110
00:39:16.619 --> 00:39:19.464
- some thoughts about, like, you know, I didn't

1111
00:39:19.464 --> 00:39:21.545
- wanna just go into this and talk about,

1112
00:39:21.545 --> 00:39:23.944
- oh, everything's bad, doom and gloom. I was

1113
00:39:23.944 --> 00:39:25.864
- trying to figure out, like, ways of what

1114
00:39:25.864 --> 00:39:28.684
- should I as an individual engineer think about,

1115
00:39:29.464 --> 00:39:31.864
- like, how do I position myself? What do

1116
00:39:31.864 --> 00:39:32.764
- I think about?

1117
00:39:33.570 --> 00:39:35.510
- What's what's the likely?

1118
00:39:36.930 --> 00:39:39.570
- And there's probably gonna be a ten year

1119
00:39:39.570 --> 00:39:41.510
- drought, like you said, and then

1120
00:39:41.809 --> 00:39:43.730
- things will pick back up. I I... That's

1121
00:39:43.730 --> 00:39:45.489
- what I think. I think they'll They absolutely

1122
00:39:45.489 --> 00:39:47.329
- will. I mean, if I'm a young engineer

1123
00:39:47.329 --> 00:39:47.829
- today,

1124
00:39:48.605 --> 00:39:49.585
- and I want it

1125
00:39:50.525 --> 00:39:52.525
- to really be part of that coming back.

1126
00:39:52.525 --> 00:39:54.125
- Because the coming back, even during the ten

1127
00:39:54.125 --> 00:39:55.644
- years, it's a small level. I think there's

1128
00:39:55.644 --> 00:39:57.565
- gonna be a ton of work. And the

1129
00:39:57.565 --> 00:39:59.164
- and the the only way for you to

1130
00:39:59.164 --> 00:40:01.324
- do that today is you gotta scrape together

1131
00:40:01.324 --> 00:40:02.924
- money... Enough money to go out and buy

1132
00:40:02.924 --> 00:40:04.144
- a couple of these cards,

1133
00:40:04.849 --> 00:40:06.690
- put it in your garage, run it, and

1134
00:40:06.690 --> 00:40:08.849
- learn it because that's... You... The only way

1135
00:40:08.849 --> 00:40:10.849
- to learn small AI is to do small

1136
00:40:10.849 --> 00:40:12.849
- AI because there's nobody to tell you or

1137
00:40:12.849 --> 00:40:14.769
- teach you. So you have to just do

1138
00:40:14.769 --> 00:40:16.849
- it yourself. And if you do that, you've

1139
00:40:16.849 --> 00:40:18.849
- just built a business. I believe that you'll

1140
00:40:18.849 --> 00:40:20.635
- do fine. I mean, I think you could

1141
00:40:20.635 --> 00:40:22.315
- go to a 100 other businesses in the

1142
00:40:22.315 --> 00:40:24.075
- next five years and then put their data

1143
00:40:24.075 --> 00:40:26.155
- centers in for them and and make a

1144
00:40:26.155 --> 00:40:27.914
- really good living at that. So I think

1145
00:40:27.914 --> 00:40:29.454
- it's a massive growth area.

1146
00:40:30.795 --> 00:40:32.489
- The other thing we didn't mention, the one

1147
00:40:32.570 --> 00:40:34.969
- I... Everyone I know still scratches their head

1148
00:40:34.969 --> 00:40:37.230
- about why these guys gave all their software

1149
00:40:37.289 --> 00:40:38.269
- away for free.

1150
00:40:39.769 --> 00:40:41.929
- Yeah. It doesn't make any sense. If this

1151
00:40:41.929 --> 00:40:43.610
- is such a valuable thing, why are you

1152
00:40:43.610 --> 00:40:45.864
- publishing the code for the software? Because all

1153
00:40:45.864 --> 00:40:47.785
- my friends who have these small data centers,

1154
00:40:47.785 --> 00:40:49.385
- they went and got the software for free.

1155
00:40:49.385 --> 00:40:51.085
- They didn't even have to pay for it.

1156
00:40:51.465 --> 00:40:52.765
- It's like, well, thank you.

1157
00:40:53.465 --> 00:40:53.965
- Yeah.

1158
00:40:54.664 --> 00:40:56.985
- And not just open weight models, but actually,

1159
00:40:56.985 --> 00:40:58.744
- like, the actual source code that you can

1160
00:40:58.905 --> 00:41:01.405
- The actual source code. Oh, yeah. It's there.

1161
00:41:01.440 --> 00:41:03.699
- It's all out there, and it's... You know?

1162
00:41:04.239 --> 00:41:06.159
- So it's it's all setting out on Hugging

1163
00:41:06.159 --> 00:41:07.679
- Face right now. You can just get whatever

1164
00:41:07.679 --> 00:41:10.480
- you want out there. I wonder. I wonder.

1165
00:41:10.480 --> 00:41:12.079
- That's hard because there's a mill... There's a

1166
00:41:12.079 --> 00:41:14.239
- million different versions now of the AI software

1167
00:41:14.239 --> 00:41:15.164
- on Hugging Face.

1168
00:41:16.045 --> 00:41:17.565
- Yeah. I was gonna say... And that that

1169
00:41:17.565 --> 00:41:20.125
- makes me also wonder, like, this whole panic

1170
00:41:20.125 --> 00:41:22.525
- over AI doomsters and where it's gonna kill

1171
00:41:22.525 --> 00:41:23.585
- everybody or whatever

1172
00:41:24.204 --> 00:41:25.824
- might not be just a

1173
00:41:26.204 --> 00:41:26.704
- regulatory

1174
00:41:27.164 --> 00:41:27.664
- capture

1175
00:41:28.309 --> 00:41:31.110
- game that we're playing right now. Right? Yeah.

1176
00:41:31.110 --> 00:41:33.289
- Like, oh, please regulate us.

1177
00:41:34.070 --> 00:41:36.550
- I... Anytime a company goes to to any

1178
00:41:36.550 --> 00:41:37.610
- government organization,

1179
00:41:38.230 --> 00:41:40.309
- any place in the Earth, I don't mean

1180
00:41:40.309 --> 00:41:42.789
- just The United States. I mean, any place

1181
00:41:42.789 --> 00:41:45.005
- in the Earth, if If a company goes

1182
00:41:45.005 --> 00:41:46.545
- to a government and says,

1183
00:41:46.925 --> 00:41:48.625
- please regulate me, that

1184
00:41:49.965 --> 00:41:51.344
- to me just means

1185
00:41:51.885 --> 00:41:53.905
- that just an open admission

1186
00:41:54.445 --> 00:41:56.224
- that you can't take the competitive

1187
00:41:56.684 --> 00:41:57.184
- heat,

1188
00:41:57.769 --> 00:42:00.109
- and you want the government to come in

1189
00:42:00.329 --> 00:42:03.230
- and give you a lane to play in

1190
00:42:03.369 --> 00:42:04.750
- where you can make money.

1191
00:42:05.210 --> 00:42:08.489
- That's that's just my... Maybe it's... Maybe I'm

1192
00:42:08.489 --> 00:42:09.710
- wrong on that sometimes.

1193
00:42:10.250 --> 00:42:12.489
- But immediately when I see a company do

1194
00:42:12.489 --> 00:42:14.454
- that, that is my initial...

1195
00:42:15.394 --> 00:42:15.894
- Like,

1196
00:42:16.515 --> 00:42:18.454
- this is... They are just trying

1197
00:42:18.914 --> 00:42:20.934
- to build a moat around their business

1198
00:42:21.315 --> 00:42:22.695
- using government regulation.

1199
00:42:23.155 --> 00:42:25.554
- That's my immediate response every time I see

1200
00:42:25.554 --> 00:42:26.054
- that.

1201
00:42:26.390 --> 00:42:28.150
- I completely agree with you. That's exactly what

1202
00:42:28.150 --> 00:42:29.670
- I think is going on. I I think

1203
00:42:29.670 --> 00:42:31.849
- they're also looking for someone to blame. Well,

1204
00:42:32.150 --> 00:42:33.590
- well, yeah, we were doing good till they

1205
00:42:33.590 --> 00:42:34.570
- regulated us.

1206
00:42:34.869 --> 00:42:37.110
- Yeah. Right. It could be too. Yeah. It

1207
00:42:37.110 --> 00:42:39.190
- could be too. It... It's so hard to

1208
00:42:39.190 --> 00:42:41.175
- say. This stuff is, honestly, when you think

1209
00:42:41.175 --> 00:42:42.535
- about any of this stuff, it makes your

1210
00:42:42.535 --> 00:42:44.695
- head hurt because this doesn't... This makes no

1211
00:42:44.695 --> 00:42:46.855
- sense. It's like, you know, this stuff doesn't

1212
00:42:46.855 --> 00:42:47.835
- make any sense.

1213
00:42:48.295 --> 00:42:50.535
- Yeah. I don't... Yeah. I I honestly, I

1214
00:42:50.535 --> 00:42:51.255
- don't know how to

1215
00:42:52.135 --> 00:42:55.469
- No? Yeah. Anyway... Alright. Cool. Well, I think

1216
00:42:55.469 --> 00:42:57.389
- we beat that horse to the point where

1217
00:42:57.469 --> 00:42:58.690
- I think we did. Dismount.

1218
00:42:59.550 --> 00:43:01.630
- Yeah. I think so. It's been a good

1219
00:43:01.630 --> 00:43:04.510
- conversation now. So Yeah. Yeah. It's cool. And

1220
00:43:04.510 --> 00:43:06.590
- Doug Yeah. And I and I and I

1221
00:43:06.590 --> 00:43:08.349
- really liked how you didn't predict when it

1222
00:43:08.349 --> 00:43:09.744
- would crash. So, you know

1223
00:43:10.684 --> 00:43:11.184
- Very

1224
00:43:13.485 --> 00:43:15.344
- adroit at changing the subject.

1225
00:43:15.885 --> 00:43:16.545
- Yeah. You

1226
00:43:17.005 --> 00:43:18.704
- you did well. You did well.

1227
00:43:19.565 --> 00:43:21.085
- So, Doug, we need to have you more

1228
00:43:21.085 --> 00:43:21.985
- on more often.

1229
00:43:22.640 --> 00:43:23.140
- But,

1230
00:43:23.519 --> 00:43:25.360
- where can we... Where can people find you?

1231
00:43:25.360 --> 00:43:26.660
- I've said pots and pans.

1232
00:43:27.039 --> 00:43:29.119
- Pots and pans by CCG is the actual

1233
00:43:29.119 --> 00:43:31.539
- name. Pots and pans by CCG. I've written

1234
00:43:31.599 --> 00:43:34.019
- I've written a telecom blog for the industry

1235
00:43:34.400 --> 00:43:36.260
- every day for fourteen years. So

1236
00:43:37.835 --> 00:43:39.434
- which I added it all up in that

1237
00:43:39.434 --> 00:43:41.775
- equivalent of writing 40 books or some ridiculous

1238
00:43:42.074 --> 00:43:42.815
- amount of words.

1239
00:43:45.355 --> 00:43:47.994
- Nice. So Yeah. Yeah. Cool. So... Yeah. So

1240
00:43:47.994 --> 00:43:49.980
- I I cover every topic there. Is it

1241
00:43:50.059 --> 00:43:52.239
- it's either tech, regulatory, or

1242
00:43:52.539 --> 00:43:53.039
- operational,

1243
00:43:53.579 --> 00:43:55.980
- all all related to ISP in the telecom

1244
00:43:55.980 --> 00:43:57.980
- world. So Awesome. Yeah. You should go you

1245
00:43:57.980 --> 00:43:59.359
- should go read that. I follow

1246
00:43:59.739 --> 00:44:02.139
- I follow Doug's blog all the time. Alright.

1247
00:44:02.139 --> 00:44:02.639
- Tom.

1248
00:44:04.764 --> 00:44:06.065
- You can find me on LinkedIn.

1249
00:44:06.684 --> 00:44:09.085
- What? It was more than one word. I

1250
00:44:09.085 --> 00:44:10.465
- can't I can't hear that.

1251
00:44:11.885 --> 00:44:13.885
- Well, I'm on LinkedIn as well. But... Yes.

1252
00:44:13.885 --> 00:44:15.164
- So... But if you got... If you don't...

1253
00:44:15.164 --> 00:44:16.925
- Yes. But if you go to my blog,

1254
00:44:16.925 --> 00:44:19.289
- you'll find the LinkedIn. So... Yeah. Cool. Yeah.

1255
00:44:19.289 --> 00:44:20.489
- I'll try to put a link to the

1256
00:44:20.489 --> 00:44:22.730
- blog in the show notes here at some

1257
00:44:22.730 --> 00:44:24.329
- point. Well, I was I was thinking, Doug,

1258
00:44:24.329 --> 00:44:26.349
- with all of your content that you've produced,

1259
00:44:26.570 --> 00:44:28.650
- people can just go ask Chad GPT what

1260
00:44:28.650 --> 00:44:29.550
- you think. Right?

1261
00:44:29.930 --> 00:44:33.284
- Well, that's funny because I noticed yesterday that

1262
00:44:33.444 --> 00:44:34.744
- somebody scraped my

1263
00:44:35.924 --> 00:44:38.244
- site. So I'll... I had, like, 20,000

1264
00:44:38.244 --> 00:44:40.664
- reads yesterday since somebody scraped me.

1265
00:44:41.605 --> 00:44:42.505
- Oh, my goodness.

1266
00:44:42.964 --> 00:44:44.724
- So I'm I'm sure I've been into... I'm

1267
00:44:44.724 --> 00:44:46.344
- sure I know. I met all the models

1268
00:44:46.405 --> 00:44:48.590
- multiple times. Yeah. Because they're always looking they're

1269
00:44:48.590 --> 00:44:51.469
- always looking for words. Right? So Yeah. It's

1270
00:44:51.469 --> 00:44:51.969
- crazy.

1271
00:44:52.349 --> 00:44:54.030
- Alright. Well, I'm Russ White. You can find

1272
00:44:54.030 --> 00:44:55.390
- me here at the hedge at rule eleven

1273
00:44:55.390 --> 00:44:58.910
- dot tech, occasionally on LinkedIn, occasionally on x

1274
00:44:58.910 --> 00:45:00.105
- route at routing geek.

1275
00:45:01.465 --> 00:45:03.065
- We know that we live in an attention

1276
00:45:03.065 --> 00:45:05.224
- driven economy. So thank you for listening all

1277
00:45:05.224 --> 00:45:06.744
- the way to the bitter end of this

1278
00:45:06.744 --> 00:45:08.824
- episode of The Hedge, and we will catch

1279
00:45:08.824 --> 00:45:09.565
- you next.

1280
00:45:23.480 --> 00:45:23.768
- Time.