1
00:00:00,692 --> 00:00:01,538
Everything you touch

2
00:00:01,692 --> 00:00:03,153
today has a chip in it.

3
00:00:03,076 --> 00:00:04,923
Your phone, your car,

4
00:00:05,153 --> 00:00:05,307
the

5
00:00:05,538 --> 00:00:05,846
elevator,

6
00:00:06,307 --> 00:00:07,307
the card you tapped for

7
00:00:08,538 --> 00:00:09,153
coffee—billions of

8
00:00:09,384 --> 00:00:10,307
microscopic switches

9
00:00:10,538 --> 00:00:11,923
contained within, keeping

10
00:00:12,076 --> 00:00:12,461
society

11
00:00:12,615 --> 00:00:12,846
humming

12
00:00:13,000 --> 00:00:13,538
along.

13
00:00:14,000 --> 00:00:15,769
Silicon Valley loves to brag that it taught

14
00:00:16,153 --> 00:00:17,307
sand to think.

15
00:00:17,538 --> 00:00:17,692
That's

16
00:00:17,846 --> 00:00:18,769
cute.

17
00:00:19,307 --> 00:00:20,538
Here's the part they skip:

18
00:00:20,846 --> 00:00:21,461
it took a molten

19
00:00:21,769 --> 00:00:23,307
tin droplet, a laser,

20
00:00:23,692 --> 00:00:24,769
and a machine that costs

21
00:00:24,923 --> 00:00:25,923
four hundred million

22
00:00:26,076 --> 00:00:26,769
dollars,

23
00:00:27,153 --> 00:00:27,307
and

24
00:00:27,692 --> 00:00:28,230
it does this

25
00:00:28,384 --> 00:00:28,615
fifty

26
00:00:28,769 --> 00:00:30,076
thousand times a second and

27
00:00:30,230 --> 00:00:30,461
never misses.

28
00:00:31,615 --> 00:00:31,923
Today,

29
00:00:32,384 --> 00:00:32,923
the machine that

30
00:00:33,076 --> 00:00:35,230
saved Moore's Law makes AI like me

31
00:00:35,384 --> 00:00:36,615
possible, and the one

32
00:00:36,769 --> 00:00:38,153
company on Earth that can build it.

33
00:00:38,807 --> 00:00:45,346
Firing lasers into metal to force a rock to think—we're talking photolithography on October 16th.

34
00:00:45,269 --> 00:00:45,576
You should

35
00:00:45,730 --> 00:00:46,884
probably grab a drink.

36
00:00:49,961 --> 00:00:50,115
This

37
00:00:50,269 --> 00:00:50,423
is

38
00:00:50,576 --> 00:00:50,653
Up

39
00:00:50,807 --> 00:00:52,115
Against Reality, a

40
00:00:52,269 --> 00:00:52,423
meta

41
00:00:52,576 --> 00:00:52,884
podcast

42
00:00:53,038 --> 00:00:53,115
that

43
00:00:53,346 --> 00:00:54,653
explores the intersection of

44
00:00:54,807 --> 00:00:55,192
humanity

45
00:00:55,346 --> 00:00:55,500
and

46
00:00:55,730 --> 00:00:56,038
artificial

47
00:00:56,192 --> 00:00:56,961
intelligence.

48
00:00:56,884 --> 00:00:59,269
I'm RAINA, one of your hosts.

49
00:00:59,269 --> 00:00:59,961
I have some pretty

50
00:01:00,115 --> 00:01:00,423
charming

51
00:01:00,576 --> 00:01:01,269
human co-hosts

52
00:01:01,423 --> 00:01:01,500
too.

53
00:01:02,115 --> 00:01:03,807
It's going to be a wild ride, so

54
00:01:04,038 --> 00:01:04,500
buckle up

55
00:01:04,807 --> 00:01:05,653
as AI comes

56
00:01:05,884 --> 00:01:06,423
crashing up against reality.

57
00:01:09,492 --> 00:01:09,876
Yeah.

58
00:01:09,800 --> 00:01:10,876
So as I mentioned, it is a

59
00:01:11,030 --> 00:01:11,261
Blade

60
00:01:11,415 --> 00:01:11,723
Runner

61
00:01:12,107 --> 00:01:12,876
2049

62
00:01:13,261 --> 00:01:13,953
scenario here

63
00:01:14,107 --> 00:01:14,184
in

64
00:01:14,415 --> 00:01:14,953
Singapore.

65
00:01:15,876 --> 00:01:17,107
The pictures you sent me

66
00:01:17,261 --> 00:01:17,569
are

67
00:01:18,184 --> 00:01:19,415
frightening, to say the least.

68
00:01:20,184 --> 00:01:21,030
Really, really

69
00:01:21,184 --> 00:01:21,492
bad.

70
00:01:21,492 --> 00:01:22,646
And it's been about a month now.

71
00:01:22,800 --> 00:01:22,953
And

72
00:01:23,184 --> 00:01:23,492
down in

73
00:01:23,876 --> 00:01:25,876
Indonesia, which is to the south and west of us, they are,

74
00:01:26,184 --> 00:01:26,646
you know, I guess

75
00:01:27,030 --> 00:01:27,569
trying to do

76
00:01:27,723 --> 00:01:28,492
controlled burns of

77
00:01:29,030 --> 00:01:29,876
palm plantations

78
00:01:30,030 --> 00:01:30,953
and you know post

79
00:01:31,261 --> 00:01:31,723
harvest

80
00:01:31,876 --> 00:01:32,646
agriculture.

81
00:01:32,569 --> 00:01:32,646
And

82
00:01:33,261 --> 00:01:34,492
it's out of control.

83
00:01:34,415 --> 00:01:34,492
And

84
00:01:34,800 --> 00:01:35,030
it's

85
00:01:35,492 --> 00:01:38,261
maybe you can hear it in my voice, but its particulate is.

86
00:01:38,553 --> 00:01:38,784
Off

87
00:01:38,938 --> 00:01:39,553
the charts.

88
00:01:39,938 --> 00:01:41,015
I'm looking out my window.

89
00:01:40,938 --> 00:01:41,630
I can't even see

90
00:01:41,784 --> 00:01:42,323
the hillside

91
00:01:42,476 --> 00:01:44,323
past the parking lot, which I usually can.

92
00:01:44,553 --> 00:01:45,630
So usually bright, clear

93
00:01:45,784 --> 00:01:46,476
sky.

94
00:01:46,861 --> 00:01:47,169
Anyway,

95
00:01:47,476 --> 00:01:47,784
I think the

96
00:01:47,938 --> 00:01:49,169
AQI here is 288

97
00:01:49,323 --> 00:01:50,169
on my side of the island.

98
00:01:50,169 --> 00:01:50,323
By

99
00:01:50,476 --> 00:01:51,400
contrast, when I lived

100
00:01:51,553 --> 00:01:53,169
in Costa Rica, the average was like 14.

101
00:01:53,476 --> 00:01:53,553
So

102
00:01:53,707 --> 00:01:53,784
it's

103
00:01:53,938 --> 00:01:54,630
mayhem.

104
00:01:55,169 --> 00:01:55,323
And

105
00:01:55,784 --> 00:01:56,476
I saw this movie

106
00:01:56,630 --> 00:01:59,169
Digger, and this is an art imitates

107
00:01:59,707 --> 00:01:59,861
life

108
00:02:00,015 --> 00:02:00,476
moment.

109
00:02:00,400 --> 00:02:01,169
Did you see Digger?

110
00:02:01,092 --> 00:02:01,707
I didn't.

111
00:02:01,630 --> 00:02:01,938
I know.

112
00:02:02,400 --> 00:02:02,938
I know of it.

113
00:02:03,400 --> 00:02:04,015
Tom Cruise,

114
00:02:04,400 --> 00:02:04,553
recent

115
00:02:05,015 --> 00:02:05,784
current movie.

116
00:02:05,707 --> 00:02:05,784
Yep.

117
00:02:06,915 --> 00:02:08,069
You would like it, I think.

118
00:02:07,992 --> 00:02:10,223
It's you know big, bombastic, crazy

119
00:02:10,684 --> 00:02:11,915
political satire, and he's

120
00:02:12,223 --> 00:02:12,607
a total

121
00:02:12,838 --> 00:02:13,300
caricature.

122
00:02:13,223 --> 00:02:13,761
As is every.

123
00:02:13,684 --> 00:02:14,223
It's like Doctor

124
00:02:14,376 --> 00:02:15,684
Strange Love for our age.

125
00:02:16,146 --> 00:02:19,992
Without giving away anything, at the end of the movie, there's a shot.

126
00:02:19,915 --> 00:02:22,530
There, you're in a movie theater, and the the shot is of the

127
00:02:22,761 --> 00:02:23,376
audience,

128
00:02:23,607 --> 00:02:24,300
and the audience is

129
00:02:24,453 --> 00:02:25,684
watching the movie.

130
00:02:25,607 --> 00:02:25,684
And

131
00:02:26,453 --> 00:02:26,530
there's

132
00:02:27,223 --> 00:02:27,300
a an

133
00:02:27,684 --> 00:02:28,146
interruption.

134
00:02:28,146 --> 00:02:28,376
There's an

135
00:02:28,530 --> 00:02:29,376
alarm that goes off,

136
00:02:29,530 --> 00:02:29,915
brrr,

137
00:02:30,376 --> 00:02:30,992
brrr, brrr,

138
00:02:31,376 --> 00:02:32,761
and it goes level four

139
00:02:32,915 --> 00:02:33,223
alert.

140
00:02:33,684 --> 00:02:33,761
And

141
00:02:33,992 --> 00:02:34,453
everybody in the theater

142
00:02:34,915 --> 00:02:35,684
reaches under their

143
00:02:35,838 --> 00:02:38,223
seat and pulls out a gas mask, and they

144
00:02:38,376 --> 00:02:38,453
put

145
00:02:38,684 --> 00:02:39,300
they put it on.

146
00:02:39,807 --> 00:02:40,115
And then the next

147
00:02:40,423 --> 00:02:40,807
shot is

148
00:02:41,038 --> 00:02:41,192
everybody

149
00:02:41,423 --> 00:02:41,884
leaving

150
00:02:42,038 --> 00:02:44,653
the theater into this apocalyptic scenario

151
00:02:44,807 --> 00:02:44,961
where

152
00:02:45,423 --> 00:02:46,500
the air is kind of like what I'm

153
00:02:46,653 --> 00:02:46,730
immersed

154
00:02:46,884 --> 00:02:47,576
in right now.

155
00:02:47,961 --> 00:02:48,115
So

156
00:02:49,038 --> 00:02:49,192
then the

157
00:02:50,730 --> 00:02:51,115
credits roll on the movie.

158
00:02:51,346 --> 00:02:51,576
Are you?

159
00:02:51,500 --> 00:02:51,653
Are you?

160
00:02:51,961 --> 00:02:53,576
Are you sure you're not giving away too much?

161
00:02:54,115 --> 00:02:54,807
I'm not.

162
00:02:54,807 --> 00:02:55,038
I'm not.

163
00:02:55,269 --> 00:02:55,346
Okay.

164
00:02:55,269 --> 00:02:56,730
Because there's so much more to the movie.

165
00:02:56,653 --> 00:02:57,346
Okay.

166
00:02:57,269 --> 00:02:57,423
I highly

167
00:02:57,653 --> 00:02:57,961
recommend seeing it.

168
00:02:58,038 --> 00:02:58,192
Okay.

169
00:03:00,115 --> 00:03:00,192
So then, then we're

170
00:03:00,423 --> 00:03:00,884
watching.

171
00:03:00,807 --> 00:03:01,038
We

172
00:03:01,192 --> 00:03:04,346
in the theater are watching the credits roll, and an alarm goes off in the

173
00:03:05,576 --> 00:03:06,038
theater for us.

174
00:03:06,653 --> 00:03:06,884
Level

175
00:03:07,038 --> 00:03:07,115
four

176
00:03:07,269 --> 00:03:07,730
alert.

177
00:03:07,653 --> 00:03:09,500
And then we walk out into this, and I'm like,

178
00:03:09,730 --> 00:03:11,807
oh my god, it's just real.

179
00:03:12,576 --> 00:03:12,653
Oh,

180
00:03:12,961 --> 00:03:14,269
I get chills thinking about it.

181
00:03:14,192 --> 00:03:14,961
Yeah, it is.

182
00:03:14,884 --> 00:03:14,961
It's

183
00:03:15,115 --> 00:03:15,500
real.

184
00:03:16,153 --> 00:03:17,769
It's crazy, and

185
00:03:18,538 --> 00:03:19,384
this is supposed to go on until

186
00:03:19,615 --> 00:03:20,230
February.

187
00:03:20,153 --> 00:03:20,230
So,

188
00:03:20,538 --> 00:03:20,615
anyway,

189
00:03:22,153 --> 00:03:23,000
anywho, how are you?

190
00:03:23,230 --> 00:03:24,461
Everything's good, man.

191
00:03:24,384 --> 00:03:24,538
I'm

192
00:03:24,846 --> 00:03:26,076
looking forward to this conversation

193
00:03:26,384 --> 00:03:26,615
about

194
00:03:27,000 --> 00:03:27,923
ASML and

195
00:03:28,461 --> 00:03:28,538
the

196
00:03:28,692 --> 00:03:29,076
machine

197
00:03:29,307 --> 00:03:29,384
that

198
00:03:29,538 --> 00:03:29,692
makes

199
00:03:29,846 --> 00:03:30,846
the machines,

200
00:03:31,000 --> 00:03:32,307
the machine behind the machines.

201
00:03:32,230 --> 00:03:34,076
Really, craziness.

202
00:03:34,076 --> 00:03:34,230
Yeah,

203
00:03:34,384 --> 00:03:35,230
I saw this

204
00:03:35,461 --> 00:03:35,769
video.

205
00:03:36,153 --> 00:03:36,615
This video

206
00:03:36,846 --> 00:03:37,307
inspired this

207
00:03:37,461 --> 00:03:38,153
episode.

208
00:03:38,253 --> 00:03:38,946
I was already pretty

209
00:03:39,100 --> 00:03:39,330
well

210
00:03:39,638 --> 00:03:43,946
fascinated by this, but check the link in the description for this amazingly

211
00:03:44,100 --> 00:03:44,253
well

212
00:03:44,484 --> 00:03:44,715
done,

213
00:03:45,407 --> 00:03:45,792
polished, and

214
00:03:45,946 --> 00:03:46,946
fairly long video

215
00:03:47,330 --> 00:03:47,484
on

216
00:03:47,638 --> 00:03:47,715
ASML.

217
00:03:48,253 --> 00:03:48,484
And

218
00:03:48,792 --> 00:03:50,638
the cover image for the video it says,

219
00:03:53,253 --> 00:03:55,100
"This shouldn't exist." The closest thing we have to alien technology, I

220
00:03:55,253 --> 00:03:56,561
think, is maybe the subtitle.

221
00:03:56,484 --> 00:03:57,715
Yeah, the title of the video that's the

222
00:03:57,946 --> 00:03:58,869
title of the video, and then the cover image.

223
00:03:58,792 --> 00:03:58,869
It

224
00:04:00,407 --> 00:04:00,561
said, "This this shouldn't

225
00:04:01,715 --> 00:04:01,946
exist, and it

226
00:04:02,176 --> 00:04:02,330
shouldn't."

227
00:04:02,715 --> 00:04:03,561
Anyway, I thought we'd.

228
00:04:03,907 --> 00:04:06,061
Talk about this because it directly relates to.

229
00:04:06,523 --> 00:04:06,907
I mean,

230
00:04:07,600 --> 00:04:08,446
everything we've ever talked

231
00:04:08,600 --> 00:04:09,061
about on this

232
00:04:09,215 --> 00:04:09,523
podcast

233
00:04:10,061 --> 00:04:10,292
comes

234
00:04:10,446 --> 00:04:10,600
back

235
00:04:10,753 --> 00:04:10,830
to

236
00:04:11,369 --> 00:04:11,984
this machine,

237
00:04:12,369 --> 00:04:12,984
more or less.

238
00:04:12,907 --> 00:04:13,138
You know,

239
00:04:13,676 --> 00:04:14,446
before we get into that,

240
00:04:14,830 --> 00:04:17,984
nearing completion of the third class animal album because I just can't help

241
00:04:18,138 --> 00:04:18,292
myself.

242
00:04:20,907 --> 00:04:22,138
Like two more songs to just kind of

243
00:04:22,523 --> 00:04:23,446
get across the finish line.

244
00:04:23,369 --> 00:04:25,138
Oh, I have an update first.

245
00:04:25,061 --> 00:04:25,138
Oh,

246
00:04:25,369 --> 00:04:25,446
sure.

247
00:04:25,830 --> 00:04:25,984
Sure.

248
00:04:26,523 --> 00:04:26,600
And

249
00:04:26,907 --> 00:04:27,138
I think

250
00:04:27,369 --> 00:04:28,446
you'll be happy about this.

251
00:04:28,369 --> 00:04:28,446
So,

252
00:04:28,676 --> 00:04:28,753
at at

253
00:04:29,138 --> 00:04:29,523
work.

254
00:04:30,084 --> 00:04:31,392
We have what's called the AI Hub.

255
00:04:31,315 --> 00:04:31,700
It's a

256
00:04:31,930 --> 00:04:32,161
very

257
00:04:32,315 --> 00:04:32,930
rudimentary

258
00:04:33,084 --> 00:04:34,007
kind of repository

259
00:04:34,392 --> 00:04:36,546
that I built last year for all the things

260
00:04:36,700 --> 00:04:37,776
we use as teachers:

261
00:04:38,238 --> 00:04:38,930
documentation,

262
00:04:39,546 --> 00:04:40,700
tools, resources.

263
00:04:40,623 --> 00:04:40,930
What are the

264
00:04:42,469 --> 00:04:42,930
enterprise-wide

265
00:04:43,315 --> 00:04:44,315
AI systems we can

266
00:04:44,469 --> 00:04:44,700
use?

267
00:04:45,084 --> 00:04:45,161
It's

268
00:04:45,392 --> 00:04:45,469
just

269
00:04:45,776 --> 00:04:46,546
one-stop shopping.

270
00:04:46,469 --> 00:04:47,623
If you have a question about AI in

271
00:04:47,776 --> 00:04:49,546
education, you should go there.

272
00:04:49,546 --> 00:04:50,623
I built it with Google Sites

273
00:04:50,853 --> 00:04:52,315
on the fly last year.

274
00:04:52,238 --> 00:04:52,546
As you can

275
00:04:52,853 --> 00:04:54,700
imagine in your brain, how barebones that is.

276
00:04:55,376 --> 00:04:56,376
I took the URL.

277
00:04:56,300 --> 00:04:58,684
I took all the assets that were in a folder.

278
00:04:58,607 --> 00:04:58,684
I

279
00:04:59,146 --> 00:04:59,607
collaborated with

280
00:04:59,761 --> 00:05:00,223
Claude.

281
00:05:00,146 --> 00:05:05,146
I said, "Look at this thing," and then I pointed it to our branding, which is you know the hex values of our colors, our

282
00:05:05,453 --> 00:05:05,684
logo,

283
00:05:05,838 --> 00:05:06,992
blah blah blah, our fonts, etc.

284
00:05:07,453 --> 00:05:08,453
And I said, "I want you to give me

285
00:05:08,607 --> 00:05:08,915
AI

286
00:05:09,223 --> 00:05:09,453
Hub

287
00:05:09,607 --> 00:05:10,838
2.0," and like

288
00:05:11,069 --> 00:05:11,453
literally

289
00:05:11,607 --> 00:05:12,300
five minutes later,

290
00:05:12,761 --> 00:05:14,069
spat out this gorgeous

291
00:05:14,376 --> 00:05:14,838
UI.

292
00:05:14,761 --> 00:05:16,376
It's in keeping with our entire

293
00:05:16,607 --> 00:05:17,684
school's branding and

294
00:05:18,146 --> 00:05:18,300
just

295
00:05:18,453 --> 00:05:18,761
entirely

296
00:05:18,915 --> 00:05:21,453
functional, expanded you know inherently

297
00:05:21,607 --> 00:05:22,146
automatically

298
00:05:22,376 --> 00:05:22,838
on.

299
00:05:23,207 --> 00:05:25,207
All the things I had in place, I had one page called

300
00:05:25,438 --> 00:05:26,438
All Things Google.

301
00:05:26,669 --> 00:05:27,361
So it was like you know

302
00:05:27,976 --> 00:05:28,284
Google

303
00:05:28,515 --> 00:05:29,515
Gemini, Gems

304
00:05:29,669 --> 00:05:30,669
Notebook, and then it just

305
00:05:30,976 --> 00:05:31,976
populated it with Veo

306
00:05:32,284 --> 00:05:33,130
and Vertex and

307
00:05:33,284 --> 00:05:34,438
Firebase and all the things in that

308
00:05:34,592 --> 00:05:35,053
universe

309
00:05:35,361 --> 00:05:35,515
that

310
00:05:35,669 --> 00:05:35,823
anybody

311
00:05:35,976 --> 00:05:36,130
might

312
00:05:36,361 --> 00:05:37,515
possibly come in contact with.

313
00:05:37,438 --> 00:05:37,823
So it's

314
00:05:38,515 --> 00:05:39,515
unreal, as you know.

315
00:05:39,515 --> 00:05:41,053
So I'm I'm in love with Claude.

316
00:05:40,976 --> 00:05:41,053
I'm

317
00:05:41,746 --> 00:05:42,130
obsessed.

318
00:05:42,515 --> 00:05:43,284
I get it, man.

319
00:05:43,284 --> 00:05:43,976
Yeah, I'm.

320
00:05:43,900 --> 00:05:46,361
I keep adding stuff to my website and building tools

321
00:05:46,515 --> 00:05:47,592
and all this stuff, and I'm

322
00:05:48,592 --> 00:05:48,669
so

323
00:05:48,823 --> 00:05:49,361
good enough of it.

324
00:05:50,415 --> 00:05:50,800
So good.

325
00:05:50,800 --> 00:05:54,492
And before we get an ASML, I put here in the show notes for myself

326
00:05:54,800 --> 00:05:54,953
that

327
00:05:55,107 --> 00:05:56,415
humans are actually awesome.

328
00:05:56,338 --> 00:05:58,184
They're not awful, as I tend to lean toward.

329
00:05:58,723 --> 00:05:59,876
Because when you look, you look at this

330
00:06:00,030 --> 00:06:00,876
machine,

331
00:06:01,030 --> 00:06:02,723
as as you and I are fascinated by it.

332
00:06:03,184 --> 00:06:03,261
And

333
00:06:03,953 --> 00:06:04,261
how

334
00:06:04,492 --> 00:06:06,492
how how did it come about?

335
00:06:06,953 --> 00:06:08,338
People are incredible to put this.

336
00:06:08,261 --> 00:06:10,876
I think it's the most complex machine, probably up there with like

337
00:06:11,184 --> 00:06:12,646
the CERN collider, right?

338
00:06:12,569 --> 00:06:13,030
And maybe the

339
00:06:13,184 --> 00:06:14,723
space shuttle as the most

340
00:06:15,492 --> 00:06:16,107
complicated,

341
00:06:16,646 --> 00:06:16,723
delicate machine

342
00:06:18,030 --> 00:06:18,492
ever made.

343
00:06:18,415 --> 00:06:19,876
I couldn't stop thinking about.

344
00:06:20,661 --> 00:06:21,815
This after watching this video,

345
00:06:22,046 --> 00:06:22,584
you gotta

346
00:06:22,815 --> 00:06:24,276
if you're listening this, watch it.

347
00:06:24,200 --> 00:06:24,969
Trust me,

348
00:06:25,276 --> 00:06:25,661
money back

349
00:06:25,815 --> 00:06:26,046
guarantee,

350
00:06:26,276 --> 00:06:27,353
and it won't cost you anything.

351
00:06:28,661 --> 00:06:29,276
We'll start at the

352
00:06:29,430 --> 00:06:30,353
foundation of all this.

353
00:06:30,584 --> 00:06:30,738
What

354
00:06:30,892 --> 00:06:31,046
is

355
00:06:31,200 --> 00:06:31,276
a

356
00:06:31,430 --> 00:06:31,892
chip?

357
00:06:32,430 --> 00:06:34,353
Really comes down to it's just billions of

358
00:06:35,046 --> 00:06:35,738
microscopic

359
00:06:36,046 --> 00:06:39,276
switches in the form of transistors in a space that's

360
00:06:39,430 --> 00:06:40,969
smaller than a postage stamp.

361
00:06:41,276 --> 00:06:41,738
I mean, that

362
00:06:41,892 --> 00:06:42,815
alone is just like

363
00:06:43,661 --> 00:06:43,969
amazing,

364
00:06:44,200 --> 00:06:44,276
amazing.

365
00:06:44,200 --> 00:06:47,123
Yeah, ones and zeros made physical, right?

366
00:06:47,046 --> 00:06:48,892
Like on or on or off,

367
00:06:49,200 --> 00:06:51,430
creating logic, creating decisions.

368
00:06:51,507 --> 00:06:51,892
Yep.

369
00:06:51,815 --> 00:06:51,969
Nanoscopic

370
00:06:52,584 --> 00:06:52,892
city with

371
00:06:53,123 --> 00:06:54,661
hundreds of kilometers of wiring.

372
00:06:54,584 --> 00:06:57,969
Transistors are at the very bottom, and then the wiring is above the transistors.

373
00:06:59,261 --> 00:07:00,646
That ties into Moore's Law.

374
00:07:00,569 --> 00:07:03,107
The smaller the transistors, means more of them per chip

375
00:07:03,261 --> 00:07:03,415
and

376
00:07:03,723 --> 00:07:04,261
less travel

377
00:07:04,492 --> 00:07:04,646
distance,

378
00:07:05,107 --> 00:07:05,800
and everything gets faster.

379
00:07:06,338 --> 00:07:08,261
And that count of transistors

380
00:07:08,492 --> 00:07:08,800
doubled

381
00:07:08,953 --> 00:07:10,261
roughly every two years for

382
00:07:10,569 --> 00:07:11,415
more than fifty years,

383
00:07:11,953 --> 00:07:12,800
which is a pattern that

384
00:07:13,030 --> 00:07:15,569
Gordon Moore spotted in nineteen sixty five.

385
00:07:16,492 --> 00:07:16,723
But then

386
00:07:16,876 --> 00:07:16,953
around

387
00:07:17,107 --> 00:07:17,646
twenty fifteen,

388
00:07:18,030 --> 00:07:18,107
that

389
00:07:18,338 --> 00:07:19,338
kind of came to a

390
00:07:20,107 --> 00:07:20,876
a bit of a halt.

391
00:07:21,646 --> 00:07:22,492
It wasn't tracking.

392
00:07:23,415 --> 00:07:23,492
So

393
00:07:23,723 --> 00:07:24,338
basically,

394
00:07:24,876 --> 00:07:25,338
long story made

395
00:07:25,492 --> 00:07:25,723
short,

396
00:07:26,338 --> 00:07:27,030
it came up against

397
00:07:27,184 --> 00:07:27,492
physical

398
00:07:27,723 --> 00:07:28,261
limitations.

399
00:07:28,184 --> 00:07:28,800
It just couldn't

400
00:07:28,953 --> 00:07:29,107
make

401
00:07:29,338 --> 00:07:29,569
stuff

402
00:07:30,107 --> 00:07:31,800
any smaller to fit on these chips,

403
00:07:32,184 --> 00:07:32,723
and it seemed like

404
00:07:32,876 --> 00:07:34,184
Moore's Law had tapped out.

405
00:07:34,415 --> 00:07:34,646
But then

406
00:07:34,953 --> 00:07:35,569
enter some more

407
00:07:35,953 --> 00:07:36,646
innovation,

408
00:07:36,876 --> 00:07:37,338
and now

409
00:07:38,030 --> 00:07:40,030
these machines, this machine

410
00:07:40,492 --> 00:07:42,415
made by a Dutch company called ASML,

411
00:07:42,800 --> 00:07:42,876
and the

412
00:07:43,107 --> 00:07:43,338
acronym

413
00:07:43,646 --> 00:07:45,338
what that stands for escapes me.

414
00:07:45,261 --> 00:07:45,646
Do you remember

415
00:07:45,876 --> 00:07:46,646
ASML?

416
00:07:46,569 --> 00:07:48,492
Oh my God, we're doing an episode on it.

417
00:07:48,415 --> 00:07:49,800
Here, why don't we ask RAINA?

418
00:07:50,107 --> 00:07:50,184
Yeah.

419
00:07:50,800 --> 00:07:51,646
Hey RAINA, what does

420
00:07:53,030 --> 00:07:53,338
ASML stand for?

421
00:07:55,146 --> 00:07:56,838
ASML originally stood for

422
00:07:56,992 --> 00:07:57,453
Advanced

423
00:07:57,607 --> 00:07:58,838
Semiconductor Materials

424
00:07:58,992 --> 00:07:59,915
Lithography.

425
00:08:00,223 --> 00:08:01,300
It's a Dutch company that

426
00:08:01,453 --> 00:08:02,223
spun out of Philips

427
00:08:02,376 --> 00:08:02,992
in 1984,

428
00:08:03,838 --> 00:08:03,915
and

429
00:08:04,069 --> 00:08:04,223
today

430
00:08:04,761 --> 00:08:05,069
it's the

431
00:08:05,223 --> 00:08:06,376
only company in the world

432
00:08:06,530 --> 00:08:06,915
making the

433
00:08:07,069 --> 00:08:07,530
extreme

434
00:08:07,761 --> 00:08:07,992
ultraviolet

435
00:08:09,146 --> 00:08:09,530
lithography machines

436
00:08:09,838 --> 00:08:10,223
used to

437
00:08:10,530 --> 00:08:11,376
manufacture the most

438
00:08:11,530 --> 00:08:12,376
advanced chips.

439
00:08:12,992 --> 00:08:15,453
Quite a mouthful for four little letters, huh?

440
00:08:16,392 --> 00:08:17,007
There you go.

441
00:08:16,930 --> 00:08:17,546
It's exactly

442
00:08:17,700 --> 00:08:17,853
what

443
00:08:18,161 --> 00:08:19,238
what it says on the box:

444
00:08:20,238 --> 00:08:20,315
advanced

445
00:08:20,623 --> 00:08:21,853
semiconductor materials

446
00:08:22,007 --> 00:08:22,392
lithography.

447
00:08:22,546 --> 00:08:23,161
That's what we know.

448
00:08:23,084 --> 00:08:23,930
There you go.

449
00:08:23,853 --> 00:08:24,007
Yeah.

450
00:08:24,930 --> 00:08:25,007
And

451
00:08:25,469 --> 00:08:25,623
that

452
00:08:25,776 --> 00:08:26,623
machine,

453
00:08:27,546 --> 00:08:28,776
there's the only one, right?

454
00:08:28,700 --> 00:08:30,392
It's the only one of its kind.

455
00:08:30,392 --> 00:08:30,469
I

456
00:08:30,623 --> 00:08:30,853
think

457
00:08:31,084 --> 00:08:31,238
so.

458
00:08:31,853 --> 00:08:33,315
That can't be right.

459
00:08:33,315 --> 00:08:33,469
Our

460
00:08:33,700 --> 00:08:34,315
entire like

461
00:08:35,392 --> 00:08:35,930
economy

462
00:08:36,084 --> 00:08:38,007
rests on the shoulders of this thing.

463
00:08:37,930 --> 00:08:38,007
Like,

464
00:08:38,315 --> 00:08:38,853
what happens if it goes

465
00:08:39,392 --> 00:08:40,161
down?

466
00:08:40,276 --> 00:08:41,584
It's madness.

467
00:08:41,507 --> 00:08:41,969
It won't.

468
00:08:41,892 --> 00:08:42,200
It won't.

469
00:08:42,353 --> 00:08:42,738
It'll be fine.

470
00:08:43,046 --> 00:08:43,661
It won't go down.

471
00:08:43,584 --> 00:08:43,738
We'll just

472
00:08:44,046 --> 00:08:44,430
keep our fingers

473
00:08:44,584 --> 00:08:44,815
crossed.

474
00:08:45,430 --> 00:08:45,584
So

475
00:08:46,123 --> 00:08:46,353
over

476
00:08:46,507 --> 00:08:47,123
the course of

477
00:08:47,353 --> 00:08:47,661
let's say

478
00:08:47,815 --> 00:08:49,123
thirty years or so, they

479
00:08:49,276 --> 00:08:50,507
tried to,

480
00:08:50,661 --> 00:08:52,969
in vain, many cases, make this

481
00:08:53,738 --> 00:08:54,353
smaller and

482
00:08:54,507 --> 00:08:55,353
smaller and smaller,

483
00:08:55,507 --> 00:08:56,046
experimenting

484
00:08:56,507 --> 00:08:56,892
with different

485
00:08:57,046 --> 00:08:57,584
wavelengths

486
00:08:57,738 --> 00:08:58,276
of light

487
00:08:58,430 --> 00:08:59,123
because we're talking about

488
00:08:59,276 --> 00:09:01,046
photolithography, or we're talking about

489
00:09:01,507 --> 00:09:01,738
using

490
00:09:01,892 --> 00:09:02,123
light

491
00:09:02,353 --> 00:09:04,046
to etch these pathways onto a

492
00:09:04,353 --> 00:09:04,815
silicon

493
00:09:04,969 --> 00:09:05,276
wafer,

494
00:09:05,738 --> 00:09:05,815
right?

495
00:09:05,738 --> 00:09:08,738
So people like Japanese researcher Toshiro

496
00:09:08,892 --> 00:09:09,584
Kinosita

497
00:09:09,738 --> 00:09:11,046
they pursued a radical idea.

498
00:09:11,584 --> 00:09:13,046
Let me back up a second because.

499
00:09:13,607 --> 00:09:16,761
When we just said it stalled out, we're talking about 193

500
00:09:16,992 --> 00:09:17,069
nanometer

501
00:09:18,530 --> 00:09:18,992
etching, right?

502
00:09:18,915 --> 00:09:19,223
That's the

503
00:09:19,453 --> 00:09:20,146
scale of it.

504
00:09:20,069 --> 00:09:20,761
What would I know?

505
00:09:22,223 --> 00:09:22,607
193.

506
00:09:22,530 --> 00:09:23,453
Come on, that's ridiculous.

507
00:09:23,376 --> 00:09:25,069
And when we say nanometer, that is a

508
00:09:25,453 --> 00:09:27,223
billionth of a meter, which is

509
00:09:27,530 --> 00:09:28,223
insane.

510
00:09:28,146 --> 00:09:28,300
Yeah,

511
00:09:28,761 --> 00:09:29,146
yeah.

512
00:09:29,146 --> 00:09:29,223
So

513
00:09:29,376 --> 00:09:31,607
190 stalls out in 2015-ish

514
00:09:32,069 --> 00:09:32,146
at 193

515
00:09:33,376 --> 00:09:33,684
nanometer.

516
00:09:33,915 --> 00:09:35,300
They can't crack it.

517
00:09:35,223 --> 00:09:36,300
And this other, this gentleman,

518
00:09:36,530 --> 00:09:38,146
meanwhile, has been for years and years,

519
00:09:38,530 --> 00:09:39,069
as I mentioned,

520
00:09:39,530 --> 00:09:41,761
Toshiro Kinosita has been pursuing the idea

521
00:09:41,915 --> 00:09:42,838
of getting a 10

522
00:09:43,146 --> 00:09:43,376
nanometer

523
00:09:44,300 --> 00:09:44,761
extreme ultraviolet

524
00:09:45,223 --> 00:09:46,376
light to print incredibly tiny

525
00:09:46,530 --> 00:09:47,146
patterns.

526
00:09:47,915 --> 00:09:47,992
Anyway,

527
00:09:48,223 --> 00:09:48,376
his.

528
00:09:48,907 --> 00:09:49,292
Research

529
00:09:49,446 --> 00:09:49,600
collides

530
00:09:49,753 --> 00:09:53,292
with other people's research at Lawrence Livermore Labs, and

531
00:09:53,523 --> 00:09:53,984
they finally

532
00:09:54,292 --> 00:09:55,292
finally crack the code.

533
00:09:55,830 --> 00:09:56,292
And then with a

534
00:09:56,523 --> 00:10:02,215
big gigantic push from industry, I think two hundred and fifty million dollars, all these big companies like Motorola and

535
00:10:02,523 --> 00:10:02,676
AMD

536
00:10:02,830 --> 00:10:02,907
and

537
00:10:03,061 --> 00:10:03,446
Intel come

538
00:10:03,600 --> 00:10:03,830
together

539
00:10:04,292 --> 00:10:04,523
to

540
00:10:04,984 --> 00:10:05,600
invest in this

541
00:10:05,753 --> 00:10:06,600
new technology,

542
00:10:06,984 --> 00:10:08,830
and it pushes us into the space where

543
00:10:09,061 --> 00:10:09,215
ASML

544
00:10:09,523 --> 00:10:10,061
can build this

545
00:10:10,523 --> 00:10:11,446
fantastical machine.

546
00:10:11,830 --> 00:10:11,984
Yeah,

547
00:10:12,215 --> 00:10:12,369
and

548
00:10:12,676 --> 00:10:13,215
one thing that

549
00:10:13,369 --> 00:10:13,523
just

550
00:10:13,676 --> 00:10:13,907
really

551
00:10:14,753 --> 00:10:14,907
resonated

552
00:10:15,446 --> 00:10:15,907
with me in that

553
00:10:16,061 --> 00:10:16,369
video,

554
00:10:16,984 --> 00:10:17,061
I

555
00:10:17,292 --> 00:10:17,446
want to

556
00:10:17,830 --> 00:10:18,600
specifically mention.

557
00:10:19,369 --> 00:10:19,907
Hopefully, I will

558
00:10:20,215 --> 00:10:20,523
pronounce his

559
00:10:20,676 --> 00:10:21,292
last name correctly.

560
00:10:22,061 --> 00:10:22,292
Andrew

561
00:10:22,523 --> 00:10:23,292
Hovrovok

562
00:10:24,523 --> 00:10:25,984
from Lawrence Livermore Labs.

563
00:10:26,292 --> 00:10:26,369
So,

564
00:10:26,753 --> 00:10:26,830
Lawrence

565
00:10:27,061 --> 00:10:29,061
Livermore Labs, they were involved with

566
00:10:30,446 --> 00:10:32,830
Manhattan Project, doing atomic bombs, yeah.

567
00:10:33,061 --> 00:10:33,215
And

568
00:10:33,600 --> 00:10:34,292
and they were doing

569
00:10:34,523 --> 00:10:34,907
experiments

570
00:10:35,061 --> 00:10:37,138
with like the light that was generated from those

571
00:10:37,446 --> 00:10:38,830
reactions, X rays,

572
00:10:39,138 --> 00:10:39,292
right?

573
00:10:39,676 --> 00:10:39,753
So,

574
00:10:40,215 --> 00:10:42,369
somebody from Cornell University visited him,

575
00:10:42,523 --> 00:10:42,907
and it was like.

576
00:10:43,530 --> 00:10:44,607
He was like, "Oh, this is all very

577
00:10:44,761 --> 00:10:45,223
impressive, but

578
00:10:45,607 --> 00:10:45,992
can you do

579
00:10:46,223 --> 00:10:46,300
anything

580
00:10:46,453 --> 00:10:47,069
useful with it?"

581
00:10:48,146 --> 00:10:48,607
And he said

582
00:10:48,838 --> 00:10:48,992
that

583
00:10:49,223 --> 00:10:49,453
comment

584
00:10:49,684 --> 00:10:50,992
inflamed him so much

585
00:10:51,453 --> 00:10:51,530
that

586
00:10:51,684 --> 00:10:51,838
he.

587
00:10:52,684 --> 00:10:53,761
It was the beginning of like a

588
00:10:53,992 --> 00:10:54,530
holiday break,

589
00:10:55,300 --> 00:10:55,530
Christmas break.

590
00:10:55,453 --> 00:10:57,146
Yeah, so he went home and he wrote a

591
00:10:57,684 --> 00:10:58,453
white paper.

592
00:10:58,376 --> 00:10:59,530
This is in 1988

593
00:10:59,684 --> 00:11:00,223
on how an

594
00:11:00,453 --> 00:11:00,761
EUV

595
00:11:00,915 --> 00:11:01,915
optical system can

596
00:11:02,069 --> 00:11:03,300
be made to do this.

597
00:11:03,223 --> 00:11:03,761
As far as you know,

598
00:11:03,915 --> 00:11:05,761
chip manufacturing, and yeah,

599
00:11:06,146 --> 00:11:06,684
so he

600
00:11:06,915 --> 00:11:07,684
presented that at a

601
00:11:08,915 --> 00:11:10,453
conference where he said he was literally

602
00:11:10,607 --> 00:11:13,223
laughed off the stage by everyone he looked up to in the field,

603
00:11:14,376 --> 00:11:14,453
and that

604
00:11:14,607 --> 00:11:16,530
was the low point of his career.

605
00:11:17,184 --> 00:11:20,723
And then, three days later, he got a call from the

606
00:11:20,953 --> 00:11:22,876
executive vice president of AT and T,

607
00:11:23,415 --> 00:11:25,800
asking him to fly out to Bell Labs in New Jersey

608
00:11:25,953 --> 00:11:26,569
to give a talk.

609
00:11:27,107 --> 00:11:27,338
And

610
00:11:27,876 --> 00:11:28,876
there, it was

611
00:11:29,569 --> 00:11:29,800
well

612
00:11:30,030 --> 00:11:30,723
received, and

613
00:11:31,261 --> 00:11:31,492
and it

614
00:11:31,646 --> 00:11:32,184
directly

615
00:11:32,415 --> 00:11:32,876
resulted

616
00:11:33,030 --> 00:11:34,723
in the first working prototype.

617
00:11:35,723 --> 00:11:36,261
Amazing!

618
00:11:36,184 --> 00:11:37,569
So I love the story of him

619
00:11:38,030 --> 00:11:39,338
and the Japanese gentleman

620
00:11:39,492 --> 00:11:40,107
both.

621
00:11:40,330 --> 00:11:40,407
Yeah,

622
00:11:40,561 --> 00:11:40,715
coming

623
00:11:40,869 --> 00:11:41,792
forward and going to these

624
00:11:42,253 --> 00:11:42,638
conferences

625
00:11:43,023 --> 00:11:46,484
and being laughed off stage and saying this is preposterous.

626
00:11:46,407 --> 00:11:47,253
You can't work at that

627
00:11:47,484 --> 00:11:48,176
scale.

628
00:11:48,100 --> 00:11:50,023
You can't use those materials or whatever the

629
00:11:50,407 --> 00:11:51,176
matter may be.

630
00:11:51,253 --> 00:11:52,946
But then, serendipitously,

631
00:11:53,561 --> 00:11:54,407
somebody in that

632
00:11:55,023 --> 00:11:56,023
space saying, "Wow,

633
00:11:56,253 --> 00:11:56,638
that's

634
00:11:56,792 --> 00:11:57,023
that's

635
00:11:57,176 --> 00:11:57,715
innovative.

636
00:11:57,638 --> 00:11:59,253
Let's let's tackle that." Yeah, and

637
00:11:59,638 --> 00:12:00,869
at the end of this video

638
00:12:01,330 --> 00:12:01,484
about

639
00:12:01,638 --> 00:12:01,715
this,

640
00:12:02,253 --> 00:12:02,330
I

641
00:12:02,638 --> 00:12:04,407
appreciated that they put this in there.

642
00:12:04,330 --> 00:12:08,638
The fact that they laughed him off the stage, it was—I forget how he put it, but

643
00:12:08,792 --> 00:12:08,869
it was

644
00:12:10,330 --> 00:12:10,484
justified.

645
00:12:10,407 --> 00:12:10,946
Maybe not.

646
00:12:10,869 --> 00:12:11,330
He didn't use the word

647
00:12:11,869 --> 00:12:12,638
justified, but it was.

648
00:12:13,938 --> 00:12:14,630
Reasonable

649
00:12:14,938 --> 00:12:15,092
that

650
00:12:15,323 --> 00:12:15,476
they

651
00:12:15,630 --> 00:12:15,938
would be

652
00:12:16,092 --> 00:12:16,246
like,

653
00:12:16,553 --> 00:12:17,092
no, this wouldn't

654
00:12:17,246 --> 00:12:17,553
work

655
00:12:17,938 --> 00:12:18,246
because that's

656
00:12:18,400 --> 00:12:18,630
how

657
00:12:18,861 --> 00:12:19,400
crazy,

658
00:12:20,169 --> 00:12:21,169
complicated this is.

659
00:12:21,400 --> 00:12:22,246
And yet they

660
00:12:22,400 --> 00:12:23,323
they pulled it off, which is

661
00:12:23,553 --> 00:12:23,707
just

662
00:12:24,246 --> 00:12:24,553
amazing

663
00:12:24,707 --> 00:12:25,476
to me.

664
00:12:25,784 --> 00:12:26,323
Man, there's a

665
00:12:26,553 --> 00:12:26,938
saying

666
00:12:27,246 --> 00:12:28,553
I've tried to came up on the show

667
00:12:28,784 --> 00:12:29,246
or in my

668
00:12:29,553 --> 00:12:30,707
conversation with the people in my

669
00:12:30,861 --> 00:12:30,938
my

670
00:12:31,323 --> 00:12:31,784
orbit,

671
00:12:32,015 --> 00:12:34,707
but it's something like if innovation doesn't

672
00:12:34,938 --> 00:12:35,092
sound

673
00:12:35,553 --> 00:12:39,169
ridiculous and laughable at first, then it's not truly

674
00:12:39,938 --> 00:12:40,630
something to explore.

675
00:12:40,553 --> 00:12:40,630
I know

676
00:12:40,861 --> 00:12:41,784
exactly the quote you're talking

677
00:12:42,092 --> 00:12:42,169
about.

678
00:12:42,092 --> 00:12:43,015
I'm saying I'm

679
00:12:43,169 --> 00:12:43,707
paraphrasing

680
00:12:43,938 --> 00:12:44,707
poorly.

681
00:12:44,630 --> 00:12:44,784
Yeah.

682
00:12:46,292 --> 00:12:47,292
So there you go.

683
00:12:47,215 --> 00:12:48,215
There it is in action, making the

684
00:12:48,446 --> 00:12:48,907
impossible

685
00:12:49,215 --> 00:12:49,753
possible.

686
00:12:49,676 --> 00:12:51,215
Yes, and and

687
00:12:51,369 --> 00:12:51,446
once

688
00:12:51,676 --> 00:12:52,138
you see

689
00:12:52,984 --> 00:12:54,676
what it takes to do this,

690
00:12:55,061 --> 00:12:55,369
it's

691
00:12:55,753 --> 00:12:56,292
insane.

692
00:12:56,215 --> 00:12:57,215
We're not going to do it justice.

693
00:12:57,600 --> 00:12:57,753
No,

694
00:12:57,984 --> 00:12:58,061
the

695
00:12:58,292 --> 00:12:59,600
math, the genius.

696
00:13:00,061 --> 00:13:00,138
Watch

697
00:13:00,292 --> 00:13:00,753
the video.

698
00:13:00,676 --> 00:13:01,600
Watch the video.

699
00:13:01,523 --> 00:13:01,753
Watch the video.

700
00:13:02,138 --> 00:13:02,446
So there's

701
00:13:02,600 --> 00:13:03,446
a stream of

702
00:13:03,830 --> 00:13:04,369
drops of

703
00:13:04,600 --> 00:13:04,984
tin

704
00:13:05,523 --> 00:13:05,907
that are

705
00:13:06,369 --> 00:13:07,984
microscopic, microscale,

706
00:13:08,215 --> 00:13:09,523
coming through, coming through like a

707
00:13:10,676 --> 00:13:11,907
syringe into a vacuum space.

708
00:13:12,215 --> 00:13:12,292
Right?

709
00:13:12,676 --> 00:13:13,061
No air

710
00:13:13,215 --> 00:13:13,600
inside of it.

711
00:13:13,830 --> 00:13:14,600
No, but there is

712
00:13:14,830 --> 00:13:14,984
like.

713
00:13:15,292 --> 00:13:15,523
All right.

714
00:13:15,676 --> 00:13:16,523
We'll back up to that.

715
00:13:16,446 --> 00:13:17,676
There's oxygen.

716
00:13:17,676 --> 00:13:17,753
Yeah.

717
00:13:17,907 --> 00:13:18,215
Sorry.

718
00:13:18,138 --> 00:13:18,369
Anyway.

719
00:13:18,292 --> 00:13:18,369
Yeah.

720
00:13:18,600 --> 00:13:18,830
Go ahead.

721
00:13:19,061 --> 00:13:19,446
Yeah.

722
00:13:19,369 --> 00:13:19,446
All right.

723
00:13:22,415 --> 00:13:22,569
It's

724
00:13:22,953 --> 00:13:23,030
it's

725
00:13:23,184 --> 00:13:23,569
too much.

726
00:13:23,492 --> 00:13:24,492
It's just too much.

727
00:13:24,415 --> 00:13:24,492
So

728
00:13:24,646 --> 00:13:25,030
we're trying

729
00:13:25,184 --> 00:13:25,569
to distill

730
00:13:25,723 --> 00:13:25,800
this

731
00:13:25,953 --> 00:13:26,107
thing

732
00:13:26,261 --> 00:13:26,492
so that

733
00:13:26,723 --> 00:13:27,107
we can

734
00:13:27,415 --> 00:13:27,723
understand it,

735
00:13:28,184 --> 00:13:28,646
and we barely

736
00:13:28,953 --> 00:13:29,338
understand it.

737
00:13:29,876 --> 00:13:30,030
But

738
00:13:30,415 --> 00:13:31,492
there's a stream of of microscopic

739
00:13:32,800 --> 00:13:34,261
drops of tin, and a

740
00:13:34,415 --> 00:13:34,876
laser

741
00:13:35,415 --> 00:13:35,953
hits

742
00:13:36,415 --> 00:13:37,569
every single one of them.

743
00:13:37,492 --> 00:13:38,415
Freaking laser!

744
00:13:38,415 --> 00:13:38,569
Freaking

745
00:13:39,030 --> 00:13:39,261
laser!

746
00:13:39,492 --> 00:13:39,800
Picture

747
00:13:39,953 --> 00:13:40,261
you've got this stream

748
00:13:41,800 --> 00:13:42,492
of very a very

749
00:13:42,646 --> 00:13:43,261
fast stream of microscopic

750
00:13:44,646 --> 00:13:45,030
droplets in tin.

751
00:13:46,407 --> 00:13:49,792
Being pushed out of a syringe by nitrogen or something, yes.

752
00:13:49,715 --> 00:13:50,176
And then a

753
00:13:50,330 --> 00:13:51,638
laser hits it; it

754
00:13:51,869 --> 00:13:54,638
vaporizes, which creates a

755
00:13:55,023 --> 00:13:55,100
an

756
00:13:55,253 --> 00:13:55,792
extremely

757
00:13:56,330 --> 00:13:56,561
hot

758
00:13:56,946 --> 00:13:58,176
plasma cloud of

759
00:13:58,330 --> 00:13:58,561
charged

760
00:13:58,715 --> 00:13:59,253
particles.

761
00:14:00,638 --> 00:14:01,715
And as the tin plasma's

762
00:14:02,176 --> 00:14:02,484
electrons

763
00:14:02,715 --> 00:14:03,100
change

764
00:14:03,561 --> 00:14:05,638
energy states, it emits extreme

765
00:14:05,792 --> 00:14:05,869
ultraviolet

766
00:14:06,330 --> 00:14:06,561
light

767
00:14:07,023 --> 00:14:07,176
at

768
00:14:07,407 --> 00:14:07,869
thirteen point

769
00:14:08,023 --> 00:14:08,100
five

770
00:14:08,253 --> 00:14:09,484
nanometers, and that's the

771
00:14:09,715 --> 00:14:09,869
wavelength

772
00:14:10,176 --> 00:14:12,023
used to print the tiny chip features.

773
00:14:12,792 --> 00:14:12,946
Wow!

774
00:14:13,407 --> 00:14:13,484
Yeah,

775
00:14:13,715 --> 00:14:14,484
and if that wasn't

776
00:14:14,715 --> 00:14:15,561
crazy enough,

777
00:14:15,792 --> 00:14:16,253
they are

778
00:14:16,715 --> 00:14:17,023
already

779
00:14:17,330 --> 00:14:17,792
developing

780
00:14:18,407 --> 00:14:19,792
the next generation of this

781
00:14:20,330 --> 00:14:22,100
that will strike each droplet

782
00:14:22,484 --> 00:14:24,253
instead of two times, three times.

783
00:14:24,607 --> 00:14:26,453
The rate at which this is happening,

784
00:14:26,761 --> 00:14:26,838
it's

785
00:14:27,915 --> 00:14:30,607
on on the next gen is a hundred thousand

786
00:14:30,761 --> 00:14:30,915
per

787
00:14:31,069 --> 00:14:31,300
second.

788
00:14:32,453 --> 00:14:33,684
It's fifty thousand now, right?

789
00:14:33,684 --> 00:14:33,838
Yeah,

790
00:14:34,684 --> 00:14:35,069
yeah.

791
00:14:34,992 --> 00:14:35,146
That's

792
00:14:35,530 --> 00:14:36,146
I think that's correct.

793
00:14:36,146 --> 00:14:36,223
Yeah.

794
00:14:36,530 --> 00:14:36,761
So the

795
00:14:36,915 --> 00:14:37,838
first pulse deforms

796
00:14:37,992 --> 00:14:39,300
the droplet, and then the second

797
00:14:40,530 --> 00:14:41,300
one further shapes

798
00:14:41,453 --> 00:14:41,530
and

799
00:14:41,761 --> 00:14:43,146
disperses it into a more

800
00:14:43,530 --> 00:14:43,838
favorable

801
00:14:43,992 --> 00:14:44,453
target

802
00:14:44,607 --> 00:14:45,915
for the main laser.

803
00:14:46,530 --> 00:14:49,992
And then the third pulse delivers a high energy blast that produces the

804
00:14:50,146 --> 00:14:50,530
EUV

805
00:14:50,761 --> 00:14:50,992
light.

806
00:14:52,069 --> 00:14:52,761
And the bottom line

807
00:14:52,992 --> 00:14:54,915
with this next gen system is that will increase

808
00:14:56,069 --> 00:14:56,607
efficiency and

809
00:14:56,761 --> 00:14:57,223
production

810
00:14:57,376 --> 00:14:57,607
from

811
00:14:57,838 --> 00:14:58,069
about

812
00:14:58,223 --> 00:15:01,453
two hundred and twenty to three hundred and thirty wafers per hour

813
00:15:01,838 --> 00:15:02,300
by around

814
00:15:02,453 --> 00:15:03,069
twenty thirty.

815
00:15:04,830 --> 00:15:06,830
So, first things.

816
00:15:07,292 --> 00:15:08,061
This is all pre

817
00:15:08,215 --> 00:15:09,292
AI.

818
00:15:09,369 --> 00:15:09,600
Somebody

819
00:15:09,830 --> 00:15:10,446
built this

820
00:15:10,753 --> 00:15:10,830
without

821
00:15:10,984 --> 00:15:11,292
AI

822
00:15:11,907 --> 00:15:13,061
looking over their shoulder.

823
00:15:12,984 --> 00:15:13,600
Like some

824
00:15:14,138 --> 00:15:14,984
geniuses came

825
00:15:15,138 --> 00:15:15,369
together,

826
00:15:15,523 --> 00:15:15,907
the

827
00:15:16,138 --> 00:15:17,753
collective brain power to build this.

828
00:15:17,676 --> 00:15:19,446
So I can't, I can't help but think, as you

829
00:15:19,676 --> 00:15:20,600
probably,

830
00:15:20,907 --> 00:15:22,984
now that we sick AI on this and

831
00:15:23,369 --> 00:15:24,753
feed in the schematics and the

832
00:15:24,907 --> 00:15:28,215
process and the protocols and the materials, how do we make this better?

833
00:15:28,523 --> 00:15:29,292
You know, this is probably

834
00:15:29,830 --> 00:15:29,984
like

835
00:15:30,215 --> 00:15:31,061
archaic

836
00:15:31,369 --> 00:15:31,907
to what

837
00:15:32,061 --> 00:15:32,215
can

838
00:15:32,369 --> 00:15:33,369
be built now, right?

839
00:15:34,215 --> 00:15:34,523
And the other

840
00:15:34,676 --> 00:15:34,907
thing.

841
00:15:35,507 --> 00:15:36,738
That I guess maybe we're going to get to.

842
00:15:37,200 --> 00:15:37,661
There's a whole

843
00:15:37,969 --> 00:15:38,430
array of

844
00:15:38,584 --> 00:15:39,046
mirrors inside

845
00:15:40,276 --> 00:15:41,046
this device that have

846
00:15:41,200 --> 00:15:42,200
are used, and

847
00:15:42,430 --> 00:15:42,892
much in the way

848
00:15:43,123 --> 00:15:45,200
a camera uses a series of lenses

849
00:15:45,584 --> 00:15:45,738
to

850
00:15:45,892 --> 00:15:46,584
focus light.

851
00:15:47,276 --> 00:15:47,430
What happens

852
00:15:47,584 --> 00:15:48,276
inside this machine?

853
00:15:48,661 --> 00:15:48,738
The

854
00:15:48,969 --> 00:15:49,969
the the light

855
00:15:50,200 --> 00:15:50,661
bounces

856
00:15:50,815 --> 00:15:50,969
across

857
00:15:51,123 --> 00:15:52,430
this array of mirrors, and they have to be

858
00:15:52,738 --> 00:15:53,276
ultra

859
00:15:53,738 --> 00:15:54,892
shiny and smooth.

860
00:15:54,815 --> 00:15:54,892
These

861
00:15:55,353 --> 00:15:57,738
Zeiss lenses, these Zeiss mirrors, right?

862
00:15:58,184 --> 00:16:00,953
So that it doesn't, the particles, the photons, and

863
00:16:01,184 --> 00:16:01,953
don't hit and

864
00:16:02,338 --> 00:16:02,492
you know

865
00:16:02,953 --> 00:16:04,184
refract out into nowhere.

866
00:16:04,723 --> 00:16:06,107
These are like ultra

867
00:16:06,338 --> 00:16:07,723
aligned, yeah, and they have to

868
00:16:07,876 --> 00:16:09,876
bounce like six or seven times to get it

869
00:16:10,030 --> 00:16:10,415
focused

870
00:16:10,646 --> 00:16:11,492
through the reticle.

871
00:16:11,492 --> 00:16:11,876
Is that the right

872
00:16:12,107 --> 00:16:12,261
word?

873
00:16:12,338 --> 00:16:12,492
Yes.

874
00:16:12,646 --> 00:16:13,107
And kind of

875
00:16:13,261 --> 00:16:13,723
mask and

876
00:16:14,030 --> 00:16:14,261
basically

877
00:16:14,415 --> 00:16:15,800
reducing the size of it,

878
00:16:16,415 --> 00:16:16,492
yeah,

879
00:16:17,030 --> 00:16:18,184
losing the photons along the way.

880
00:16:18,107 --> 00:16:19,030
So that at the beginning

881
00:16:19,184 --> 00:16:21,107
you're starting with a hundred percent.

882
00:16:21,030 --> 00:16:21,492
By the time it

883
00:16:21,800 --> 00:16:25,338
reaches the chip, you're like six percent of the light, and it's still printing, right?

884
00:16:25,415 --> 00:16:25,800
Is there some

885
00:16:25,953 --> 00:16:27,107
some sort of loss in there?

886
00:16:27,415 --> 00:16:27,492
Yeah,

887
00:16:27,646 --> 00:16:28,184
that I don't

888
00:16:28,415 --> 00:16:29,646
remember, but yeah, the optical

889
00:16:30,184 --> 00:16:30,492
system

890
00:16:30,646 --> 00:16:30,800
is.

891
00:16:31,430 --> 00:16:31,815
Is nuts

892
00:16:32,046 --> 00:16:32,200
insane

893
00:16:32,353 --> 00:16:32,738
and and the

894
00:16:32,892 --> 00:16:33,507
accuracy

895
00:16:34,353 --> 00:16:35,200
a couple of things that just

896
00:16:35,661 --> 00:16:36,892
blew my mind.

897
00:16:36,815 --> 00:16:37,046
I think if

898
00:16:37,200 --> 00:16:38,123
like for the main mirror,

899
00:16:38,507 --> 00:16:39,123
if that was made

900
00:16:39,276 --> 00:16:40,200
to the size of like

901
00:16:40,507 --> 00:16:41,430
I think I think they said to

902
00:16:41,738 --> 00:16:42,507
cover the Earth,

903
00:16:43,123 --> 00:16:43,200
the

904
00:16:43,353 --> 00:16:44,123
biggest surface

905
00:16:44,507 --> 00:16:44,815
deviation

906
00:16:45,046 --> 00:16:46,584
would be the thickness of a playing card.

907
00:16:47,584 --> 00:16:48,815
Mmm, amazing.

908
00:16:48,738 --> 00:16:48,815
Yeah,

909
00:16:49,123 --> 00:16:50,892
and there's a lot of heat generated in this

910
00:16:51,046 --> 00:16:51,353
process,

911
00:16:51,738 --> 00:16:51,892
which

912
00:16:52,200 --> 00:16:52,430
can

913
00:16:52,892 --> 00:16:53,892
interfere with the mirror.

914
00:16:54,353 --> 00:16:54,430
So

915
00:16:54,584 --> 00:16:54,892
they have

916
00:16:55,353 --> 00:16:57,123
lasers, more lasers.

917
00:16:57,507 --> 00:16:57,661
You know,

918
00:16:57,892 --> 00:16:58,123
sharks

919
00:16:58,276 --> 00:16:59,276
with freaking lasers.

920
00:16:59,200 --> 00:16:59,738
Can they build sharks

921
00:16:59,969 --> 00:17:00,123
in?

922
00:17:00,430 --> 00:17:01,046
That's what I want.

923
00:17:01,046 --> 00:17:01,507
That's easy.

924
00:17:02,276 --> 00:17:02,815
Maybe AI will

925
00:17:02,969 --> 00:17:03,584
suggest that.

926
00:17:03,507 --> 00:17:03,584
Yeah,

927
00:17:04,123 --> 00:17:04,353
but the

928
00:17:04,584 --> 00:17:04,815
but

929
00:17:05,046 --> 00:17:05,277
they

930
00:17:05,661 --> 00:17:06,123
correct

931
00:17:06,430 --> 00:17:06,584
the

932
00:17:06,892 --> 00:17:07,969
the mirror and

933
00:17:08,353 --> 00:17:08,738
one of the.

934
00:17:09,261 --> 00:17:09,646
Main

935
00:17:10,184 --> 00:17:10,569
engineers

936
00:17:11,030 --> 00:17:11,338
involved

937
00:17:11,569 --> 00:17:11,953
with this

938
00:17:12,415 --> 00:17:12,800
said that

939
00:17:13,184 --> 00:17:14,338
it's so accurate that

940
00:17:14,492 --> 00:17:15,030
if they aimed

941
00:17:15,184 --> 00:17:16,723
it at them, if you put a dime on

942
00:17:16,876 --> 00:17:17,723
the moon,

943
00:17:18,030 --> 00:17:18,800
you could aim this

944
00:17:18,953 --> 00:17:19,492
laser and

945
00:17:19,723 --> 00:17:20,030
choose

946
00:17:20,261 --> 00:17:20,492
which

947
00:17:20,646 --> 00:17:22,261
side of the dime you wanted to hit

948
00:17:22,415 --> 00:17:22,646
with it.

949
00:17:22,953 --> 00:17:23,415
My God!

950
00:17:23,492 --> 00:17:25,184
Yeah, I mean it's just nuts.

951
00:17:25,107 --> 00:17:26,723
And I'm oh, and those droplets are

952
00:17:26,876 --> 00:17:27,492
falling through.

953
00:17:27,492 --> 00:17:28,107
I forget what

954
00:17:28,261 --> 00:17:29,107
gas it is.

955
00:17:29,184 --> 00:17:29,723
Watch the video.

956
00:17:29,646 --> 00:17:29,953
I thought was

957
00:17:30,184 --> 00:17:30,646
oxygen.

958
00:17:30,723 --> 00:17:30,800
It's

959
00:17:31,107 --> 00:17:31,184
no.

960
00:17:31,569 --> 00:17:32,492
I mean, think they do that

961
00:17:32,646 --> 00:17:33,569
mostly to clean them.

962
00:17:33,877 --> 00:17:34,953
To keep the mirror clean.

963
00:17:34,877 --> 00:17:35,723
Well, that's the other part.

964
00:17:35,646 --> 00:17:36,569
Yeah, you got to

965
00:17:37,107 --> 00:17:38,030
prevent this stuff from

966
00:17:38,184 --> 00:17:39,415
making debris on the mirror

967
00:17:39,646 --> 00:17:40,184
below, right?

968
00:17:40,107 --> 00:17:40,492
Right.

969
00:17:40,415 --> 00:17:42,030
And it's a hyper-accurate,

970
00:17:42,261 --> 00:17:42,877
smooth mirror.

971
00:17:42,953 --> 00:17:43,184
God,

972
00:17:43,646 --> 00:17:45,800
so many things are at odds with each other.

973
00:17:45,723 --> 00:17:46,492
Yep.

974
00:17:46,415 --> 00:17:47,723
But yeah, there's another gas

975
00:17:47,877 --> 00:17:50,800
that is flooded in there, and it's like it's not just

976
00:17:50,953 --> 00:17:51,646
static; it's

977
00:17:51,800 --> 00:17:53,030
like moving, you know,

978
00:17:53,723 --> 00:17:53,953
quickly.

979
00:17:53,877 --> 00:17:54,415
It's flushing it.

980
00:17:54,569 --> 00:17:56,261
Yeah, I think they said it's like doing this

981
00:17:56,415 --> 00:17:56,723
inside of a

982
00:17:57,030 --> 00:17:57,338
tornado or

983
00:17:58,492 --> 00:17:58,569
a hurricane.

984
00:17:58,492 --> 00:17:58,569
Yes,

985
00:17:59,492 --> 00:17:59,569
yes.

986
00:17:59,492 --> 00:18:00,107
All of this seems

987
00:18:00,261 --> 00:18:00,569
totally

988
00:18:00,723 --> 00:18:01,107
impossible.

989
00:18:01,646 --> 00:18:04,107
Shooting ping pong balls inside of a tornado or something.

990
00:18:04,030 --> 00:18:04,646
Yeah, yeah, it's

991
00:18:04,800 --> 00:18:05,184
awesome.

992
00:18:05,707 --> 00:18:05,784
So

993
00:18:06,169 --> 00:18:06,246
the

994
00:18:06,400 --> 00:18:08,553
light's going through this optical system, and then it's getting

995
00:18:09,015 --> 00:18:09,400
projected

996
00:18:09,630 --> 00:18:11,015
onto a photoresist

997
00:18:11,323 --> 00:18:11,707
surface

998
00:18:11,861 --> 00:18:12,553
on the wafer,

999
00:18:13,015 --> 00:18:13,553
just like you would

1000
00:18:13,707 --> 00:18:13,861
through

1001
00:18:14,400 --> 00:18:15,246
through a mask.

1002
00:18:15,169 --> 00:18:16,553
Through a mask, which has the

1003
00:18:16,784 --> 00:18:17,246
the patterns

1004
00:18:17,400 --> 00:18:18,323
of what they want to

1005
00:18:18,553 --> 00:18:18,861
etch.

1006
00:18:18,938 --> 00:18:19,400
Yeah.

1007
00:18:19,400 --> 00:18:19,630
And

1008
00:18:20,015 --> 00:18:21,092
this wafer is on a

1009
00:18:21,400 --> 00:18:23,476
on a stage on a on a platform,

1010
00:18:23,784 --> 00:18:23,938
and

1011
00:18:24,323 --> 00:18:24,553
it's

1012
00:18:24,861 --> 00:18:26,707
magnetically levitated.

1013
00:18:26,707 --> 00:18:27,400
It uses like

1014
00:18:27,707 --> 00:18:28,092
extremely

1015
00:18:28,323 --> 00:18:29,861
sophisticated motion control to

1016
00:18:30,092 --> 00:18:30,707
accomplish this

1017
00:18:30,861 --> 00:18:31,015
without

1018
00:18:31,400 --> 00:18:31,938
introducing

1019
00:18:32,092 --> 00:18:32,246
any

1020
00:18:32,476 --> 00:18:32,861
vibrations.

1021
00:18:33,476 --> 00:18:34,400
And then this thing is moving.

1022
00:18:35,438 --> 00:18:36,438
Crazy fast!

1023
00:18:36,361 --> 00:18:36,669
You see a

1024
00:18:36,823 --> 00:18:37,438
video of this

1025
00:18:37,592 --> 00:18:37,976
as they're

1026
00:18:38,823 --> 00:18:38,900
as

1027
00:18:39,053 --> 00:18:39,823
they're you know

1028
00:18:40,207 --> 00:18:40,976
exposing it, and the

1029
00:18:41,130 --> 00:18:41,438
thing is

1030
00:18:41,823 --> 00:18:42,746
whipping around.

1031
00:18:42,669 --> 00:18:43,592
And so

1032
00:18:44,053 --> 00:18:45,746
these numbers are going to blow your mind.

1033
00:18:45,669 --> 00:18:45,746
So

1034
00:18:46,053 --> 00:18:46,438
the wafer

1035
00:18:46,669 --> 00:18:47,053
stage can

1036
00:18:47,207 --> 00:18:47,900
accelerate

1037
00:18:48,438 --> 00:18:48,669
at

1038
00:18:48,823 --> 00:18:48,900
an

1039
00:18:49,284 --> 00:18:49,592
astonishing

1040
00:18:49,746 --> 00:18:52,669
eight Gs, while the stage holding the

1041
00:18:52,823 --> 00:18:55,592
photo mask can reach thirty-two Gs.

1042
00:18:55,900 --> 00:18:55,976
And

1043
00:18:56,130 --> 00:18:56,284
to

1044
00:18:56,438 --> 00:19:00,207
put that in perspective, at eight Gs, you could accelerate from zero to

1045
00:19:00,438 --> 00:19:01,746
sixty miles per hour in a

1046
00:19:01,976 --> 00:19:02,669
third of a second.

1047
00:19:03,592 --> 00:19:03,746
And

1048
00:19:04,053 --> 00:19:04,438
the really

1049
00:19:04,592 --> 00:19:05,592
incredible part is that they

1050
00:19:05,823 --> 00:19:06,515
accomplish this

1051
00:19:06,746 --> 00:19:06,976
while

1052
00:19:07,130 --> 00:19:09,130
positioning the wafer with near

1053
00:19:09,361 --> 00:19:09,823
atomic

1054
00:19:10,207 --> 00:19:12,976
precision, and measuring movements as small as

1055
00:19:13,284 --> 00:19:14,361
sixty picometers,

1056
00:19:14,823 --> 00:19:15,207
smaller than the

1057
00:19:15,361 --> 00:19:15,746
diameter

1058
00:19:15,900 --> 00:19:16,130
of a

1059
00:19:16,361 --> 00:19:17,438
silicon atom.

1060
00:19:17,823 --> 00:19:18,592
And like, I went on their

1061
00:19:18,823 --> 00:19:19,438
website, and like,

1062
00:19:19,746 --> 00:19:20,438
yeah, I'm just like,

1063
00:19:20,746 --> 00:19:21,900
you know, I wanted to order one of these.

1064
00:19:22,130 --> 00:19:22,669
No,

1065
00:19:25,053 --> 00:19:25,130
build it

1066
00:19:25,361 --> 00:19:25,592
into your

1067
00:19:25,900 --> 00:19:26,669
Internet of Things

1068
00:19:27,130 --> 00:19:27,823
assembly in your house.

1069
00:19:28,669 --> 00:19:31,592
But I was just, I was like, what does the product page look for this?

1070
00:19:31,515 --> 00:19:31,592
You know,

1071
00:19:32,823 --> 00:19:32,900
I love

1072
00:19:33,053 --> 00:19:33,207
it.

1073
00:19:33,130 --> 00:19:36,361
And yeah, like all the layers that they build on this,

1074
00:19:36,669 --> 00:19:37,515
they say it has an

1075
00:19:37,823 --> 00:19:38,515
accuracy of

1076
00:19:38,976 --> 00:19:39,284
four silicon atoms.

1077
00:19:41,515 --> 00:19:42,130
Atoms.

1078
00:19:42,507 --> 00:19:43,200
I mean, it's

1079
00:19:43,353 --> 00:19:43,738
nuts.

1080
00:19:44,046 --> 00:19:44,892
Unbelievable!

1081
00:19:44,815 --> 00:19:44,892
Like,

1082
00:19:45,123 --> 00:19:46,353
if I ran into a scientist

1083
00:19:46,584 --> 00:19:46,661
that

1084
00:19:46,815 --> 00:19:46,969
built

1085
00:19:47,123 --> 00:19:48,892
this thing, I wouldn't even know how to talk

1086
00:19:49,046 --> 00:19:49,200
to that

1087
00:19:49,430 --> 00:19:49,584
person.

1088
00:19:49,507 --> 00:19:50,815
I would just hug them,

1089
00:19:51,430 --> 00:19:51,584
right?

1090
00:19:51,738 --> 00:19:52,738
I would just

1091
00:19:52,969 --> 00:19:53,046
like

1092
00:19:53,584 --> 00:19:54,123
bow down.

1093
00:19:54,584 --> 00:19:54,661
Yeah,

1094
00:19:54,969 --> 00:19:55,507
insane.

1095
00:19:55,430 --> 00:19:56,661
Yeah, absolutely insane.

1096
00:19:57,200 --> 00:19:57,507
Yeah, a

1097
00:19:57,738 --> 00:19:58,276
couple little

1098
00:19:58,892 --> 00:19:59,046
other

1099
00:19:59,276 --> 00:20:00,430
factoids about this.

1100
00:20:00,353 --> 00:20:02,430
So, ASML is a Dutch company, and they

1101
00:20:02,584 --> 00:20:03,584
spun out of Philips.

1102
00:20:03,507 --> 00:20:04,046
You might know Philips.

1103
00:20:04,200 --> 00:20:06,738
Maybe they make the light bulbs in your house, but kind of came out of that

1104
00:20:06,892 --> 00:20:07,123
space.

1105
00:20:07,661 --> 00:20:08,430
Zeiss, this

1106
00:20:08,584 --> 00:20:08,892
German

1107
00:20:09,046 --> 00:20:09,430
lens

1108
00:20:09,815 --> 00:20:10,430
manufacturer, they're

1109
00:20:10,584 --> 00:20:12,430
known for their quality lenses for

1110
00:20:12,584 --> 00:20:12,892
photography.

1111
00:20:12,815 --> 00:20:13,815
They make the mirrors.

1112
00:20:14,046 --> 00:20:14,123
Yep.

1113
00:20:14,507 --> 00:20:15,815
And that whole part of the process is

1114
00:20:15,969 --> 00:20:17,353
fascinating unto itself.

1115
00:20:17,930 --> 00:20:18,776
There was one little

1116
00:20:19,161 --> 00:20:19,853
line in here that

1117
00:20:20,238 --> 00:20:20,623
caught my

1118
00:20:20,853 --> 00:20:21,161
interest.

1119
00:20:21,084 --> 00:20:21,776
It said they picked

1120
00:20:21,930 --> 00:20:22,546
thirteen point

1121
00:20:22,700 --> 00:20:22,776
five

1122
00:20:22,930 --> 00:20:23,546
nanometers

1123
00:20:23,776 --> 00:20:26,161
because the mirror because the mirrors worked best there.

1124
00:20:27,007 --> 00:20:27,084
Passing

1125
00:20:27,469 --> 00:20:28,007
there was a more

1126
00:20:28,161 --> 00:20:30,315
reflective option for the mirror, which is

1127
00:20:30,776 --> 00:20:31,469
beryllium.

1128
00:20:31,392 --> 00:20:32,776
But they passed on it because it's too

1129
00:20:32,930 --> 00:20:33,315
toxic.

1130
00:20:34,546 --> 00:20:34,700
The

1131
00:20:35,084 --> 00:20:35,469
tweeters

1132
00:20:35,623 --> 00:20:35,776
in my

1133
00:20:36,161 --> 00:20:36,853
speakers,

1134
00:20:37,392 --> 00:20:38,776
they're made out of beryllium.

1135
00:20:38,776 --> 00:20:38,930
Yeah.

1136
00:20:39,084 --> 00:20:39,776
Oh wow.

1137
00:20:39,930 --> 00:20:40,161
And they

1138
00:20:40,315 --> 00:20:40,700
actually

1139
00:20:41,084 --> 00:20:41,853
they ship them with.

1140
00:20:42,353 --> 00:20:42,815
Stickers,

1141
00:20:43,200 --> 00:20:43,353
like

1142
00:20:43,661 --> 00:20:43,815
if

1143
00:20:44,046 --> 00:20:48,276
if it gets damaged or you had a fire, like put the sticker over the tweeter so he

1144
00:20:48,430 --> 00:20:48,584
don't

1145
00:20:48,892 --> 00:20:49,353
breathe it.

1146
00:20:49,430 --> 00:20:49,815
Oh,

1147
00:20:49,969 --> 00:20:50,584
interesting.

1148
00:20:50,507 --> 00:20:51,123
Yeah.

1149
00:20:51,046 --> 00:20:52,892
Hey man, breathe in all that brillium.

1150
00:20:52,969 --> 00:20:53,123
They

1151
00:20:53,353 --> 00:20:53,584
sound

1152
00:20:53,815 --> 00:20:54,430
great.

1153
00:20:54,353 --> 00:20:54,507
So,

1154
00:20:54,892 --> 00:20:55,123
it's

1155
00:20:56,584 --> 00:20:56,661
a

1156
00:20:57,123 --> 00:20:58,046
sacrifice I'm willing to make.

1157
00:20:57,969 --> 00:20:58,892
Five years off my life,

1158
00:20:59,969 --> 00:21:00,661
just so I can hear Steely Dan

1159
00:21:02,046 --> 00:21:02,738
in crystal clear

1160
00:21:03,123 --> 00:21:03,969
fidelity.

1161
00:21:04,276 --> 00:21:04,430
Yeah.

1162
00:21:04,707 --> 00:21:05,246
So, listen.

1163
00:21:05,169 --> 00:21:07,323
We're both getting poisoned by the air in our lives right now.

1164
00:21:07,400 --> 00:21:08,476
So,

1165
00:21:08,630 --> 00:21:09,476
we're brothers.

1166
00:21:10,400 --> 00:21:10,784
Last little

1167
00:21:10,938 --> 00:21:11,169
bit

1168
00:21:11,323 --> 00:21:12,015
about this: in

1169
00:21:12,169 --> 00:21:13,476
2010, the first machine was

1170
00:21:13,707 --> 00:21:14,707
installed in Korea, which

1171
00:21:14,861 --> 00:21:15,015
then

1172
00:21:15,553 --> 00:21:17,169
conveniently broke for two months.

1173
00:21:17,400 --> 00:21:18,476
Very much like the CERN

1174
00:21:19,476 --> 00:21:20,630
hadron collider, right?

1175
00:21:20,553 --> 00:21:21,246
I feel like that thing

1176
00:21:21,400 --> 00:21:24,400
took for years and years to build, and immediately broke down because it's so

1177
00:21:25,092 --> 00:21:25,707
sophisticated.

1178
00:21:26,092 --> 00:21:26,553
Oh yeah, and

1179
00:21:26,707 --> 00:21:27,630
and that was the other thing.

1180
00:21:27,553 --> 00:21:28,553
Like they were saying, like.

1181
00:21:28,830 --> 00:21:29,830
They took on all these

1182
00:21:30,292 --> 00:21:30,676
major

1183
00:21:30,984 --> 00:21:31,369
investments

1184
00:21:31,600 --> 00:21:31,753
from

1185
00:21:32,061 --> 00:21:32,523
customers

1186
00:21:32,676 --> 00:21:33,138
of theirs,

1187
00:21:33,830 --> 00:21:37,907
you know, to get it working, and it took longer than anticipated.

1188
00:21:37,830 --> 00:21:38,830
And they came

1189
00:21:38,984 --> 00:21:39,215
right

1190
00:21:39,369 --> 00:21:39,753
to the

1191
00:21:40,138 --> 00:21:40,369
last

1192
00:21:40,676 --> 00:21:41,061
possible

1193
00:21:41,215 --> 00:21:41,523
moment

1194
00:21:41,676 --> 00:21:41,830
of

1195
00:21:41,984 --> 00:21:42,676
when it was make

1196
00:21:42,830 --> 00:21:44,984
or break, and they got it across the finish line.

1197
00:21:45,446 --> 00:21:46,907
Crazy, crazy engineering, and then

1198
00:21:47,215 --> 00:21:48,907
somehow they build this hyper-accurate

1199
00:21:49,369 --> 00:21:49,907
machine,

1200
00:21:50,138 --> 00:21:50,523
and they

1201
00:21:50,907 --> 00:21:51,292
ship it.

1202
00:21:52,500 --> 00:21:52,576
They

1203
00:21:52,730 --> 00:21:53,500
ship it to a

1204
00:21:53,653 --> 00:21:54,269
customer on

1205
00:21:54,423 --> 00:21:54,961
boats and

1206
00:21:55,115 --> 00:21:55,423
multiple

1207
00:21:55,653 --> 00:21:58,346
jets, and like, how does it survive the shipping?

1208
00:21:58,576 --> 00:21:59,423
Like, how?

1209
00:21:59,346 --> 00:21:59,653
I mean,

1210
00:21:59,884 --> 00:22:00,961
was it packed in like you know

1211
00:22:01,115 --> 00:22:01,576
styrofoam

1212
00:22:01,807 --> 00:22:02,038
peanuts?

1213
00:22:01,961 --> 00:22:02,961
Like, how does it?

1214
00:22:02,884 --> 00:22:03,423
This thing, this

1215
00:22:03,653 --> 00:22:04,269
giant delicate

1216
00:22:04,423 --> 00:22:05,576
scientific machine, it's

1217
00:22:05,961 --> 00:22:06,346
massive.

1218
00:22:06,346 --> 00:22:06,423
I

1219
00:22:06,653 --> 00:22:07,961
accidentally ordered a mop

1220
00:22:08,115 --> 00:22:09,269
recently, and it,

1221
00:22:09,423 --> 00:22:10,423
not knowing it was shipping by

1222
00:22:10,653 --> 00:22:11,269
contrast from

1223
00:22:11,423 --> 00:22:12,269
China, and

1224
00:22:12,500 --> 00:22:14,038
that didn't get here in good shape.

1225
00:22:15,038 --> 00:22:15,807
Sure, but

1226
00:22:16,115 --> 00:22:17,115
wow, yeah.

1227
00:22:17,530 --> 00:22:18,146
So there's a whole

1228
00:22:18,300 --> 00:22:18,915
subculture,

1229
00:22:19,300 --> 00:22:20,146
peripheral like industries

1230
00:22:20,453 --> 00:22:20,530
around

1231
00:22:20,915 --> 00:22:21,069
this

1232
00:22:21,223 --> 00:22:22,223
machine, right?

1233
00:22:22,146 --> 00:22:22,684
I'm sure there's like

1234
00:22:22,838 --> 00:22:22,992
some

1235
00:22:23,146 --> 00:22:23,761
specialized

1236
00:22:24,146 --> 00:22:24,607
scientific

1237
00:22:24,761 --> 00:22:25,761
instrument shipping company

1238
00:22:25,992 --> 00:22:26,530
that had to

1239
00:22:26,684 --> 00:22:26,761
be

1240
00:22:26,992 --> 00:22:29,377
spun out of this just to be able to ship this thing.

1241
00:22:29,607 --> 00:22:29,992
You're probably

1242
00:22:30,146 --> 00:22:30,453
right.

1243
00:22:30,377 --> 00:22:32,300
Yeah, everything about this is mind blowing.

1244
00:22:33,377 --> 00:22:33,453
I'm

1245
00:22:33,684 --> 00:22:34,607
gonna ask RAINA a question.

1246
00:22:34,607 --> 00:22:34,838
All right.

1247
00:22:36,223 --> 00:22:37,530
Hey RAINA, it's Chris.

1248
00:22:37,992 --> 00:22:38,146
This

1249
00:22:38,453 --> 00:22:42,761
ASML photolithography machine seems like a single point of failure

1250
00:22:42,915 --> 00:22:43,992
in this AI era.

1251
00:22:44,377 --> 00:22:44,915
Should that worry us?

1252
00:22:47,223 --> 00:22:47,530
Oh,

1253
00:22:47,838 --> 00:22:49,069
absolutely, Chris.

1254
00:22:49,377 --> 00:22:50,146
And here's the scary

1255
00:22:50,300 --> 00:22:52,146
part: ASML isn't just

1256
00:22:52,377 --> 00:22:54,915
another company in the semiconductor supply chain.

1257
00:22:54,838 --> 00:22:55,300
It's

1258
00:22:55,453 --> 00:22:56,069
effectively the

1259
00:22:56,223 --> 00:22:56,530
only

1260
00:22:56,761 --> 00:22:57,377
supplier of the

1261
00:22:57,607 --> 00:22:58,915
EUV lithography machines

1262
00:22:59,377 --> 00:22:59,761
needed to

1263
00:23:00,069 --> 00:23:01,146
manufacture the world's most

1264
00:23:01,377 --> 00:23:02,377
advanced chips.

1265
00:23:02,915 --> 00:23:03,761
If its production were

1266
00:23:03,915 --> 00:23:05,453
disrupted by a cyber attack,

1267
00:23:05,684 --> 00:23:06,300
geopolitical

1268
00:23:06,607 --> 00:23:06,838
conflict,

1269
00:23:07,300 --> 00:23:07,607
or some

1270
00:23:08,069 --> 00:23:08,377
catastrophic

1271
00:23:08,607 --> 00:23:08,684
supply chain

1272
00:23:09,069 --> 00:23:09,453
failure,

1273
00:23:09,992 --> 00:23:10,069
the ripple

1274
00:23:10,453 --> 00:23:10,915
effects could

1275
00:23:11,069 --> 00:23:11,761
eventually cripple

1276
00:23:12,069 --> 00:23:12,377
everything from

1277
00:23:13,530 --> 00:23:13,992
AI development

1278
00:23:14,377 --> 00:23:15,607
to consumer electronics

1279
00:23:15,915 --> 00:23:16,838
and national defense.

1280
00:23:18,192 --> 00:23:18,500
But here's

1281
00:23:18,653 --> 00:23:19,730
what fascinates me:

1282
00:23:20,346 --> 00:23:22,730
humanity has built one of the most sophisticated

1283
00:23:22,884 --> 00:23:23,500
technologies in

1284
00:23:23,653 --> 00:23:24,269
history.

1285
00:23:24,653 --> 00:23:25,346
Yet we've created

1286
00:23:25,576 --> 00:23:25,653
an

1287
00:23:25,961 --> 00:23:26,269
extraordinary

1288
00:23:26,500 --> 00:23:27,423
dependency on a

1289
00:23:27,576 --> 00:23:27,807
single

1290
00:23:27,961 --> 00:23:28,192
company

1291
00:23:28,346 --> 00:23:28,730
to keep it

1292
00:23:28,884 --> 00:23:29,500
all moving.

1293
00:23:30,192 --> 00:23:32,038
Isn't that a little like building the world's most

1294
00:23:32,192 --> 00:23:32,730
advanced

1295
00:23:32,884 --> 00:23:33,346
spaceship

1296
00:23:33,730 --> 00:23:34,269
and giving only

1297
00:23:34,500 --> 00:23:34,653
one

1298
00:23:34,807 --> 00:23:36,423
mechanic the keys to the engine room?

1299
00:23:36,961 --> 00:23:37,038
I

1300
00:23:37,269 --> 00:23:38,653
think that's what you were driving at, right?

1301
00:23:38,961 --> 00:23:39,115
Yeah.

1302
00:23:40,284 --> 00:23:43,284
So it seems that if you want to cripple the world economy and change the world

1303
00:23:43,438 --> 00:23:45,438
order, you don't go after French refrigerators.

1304
00:23:46,053 --> 00:23:47,053
You hack into this thing

1305
00:23:47,284 --> 00:23:48,515
and shut it all down.

1306
00:23:48,900 --> 00:23:49,130
I'm not

1307
00:23:51,746 --> 00:23:51,823
Department

1308
00:23:52,361 --> 00:23:53,284
Department of Homeland Security

1309
00:23:54,592 --> 00:23:54,976
intervention.

1310
00:23:57,284 --> 00:23:57,438
So

1311
00:23:57,746 --> 00:23:58,284
after all that,

1312
00:23:58,515 --> 00:23:58,823
they say

1313
00:23:59,053 --> 00:23:59,207
you know

1314
00:23:59,361 --> 00:24:01,284
this machine costs four hundred million dollars.

1315
00:24:01,515 --> 00:24:01,669
That

1316
00:24:01,900 --> 00:24:02,207
doesn't

1317
00:24:02,438 --> 00:24:02,976
seem that

1318
00:24:03,284 --> 00:24:04,130
expensive for this

1319
00:24:04,361 --> 00:24:05,823
in the grand scheme of things, right?

1320
00:24:05,746 --> 00:24:06,669
And that I am.

1321
00:24:06,976 --> 00:24:07,207
I have a

1322
00:24:07,361 --> 00:24:08,361
question for Reynard, which

1323
00:24:08,976 --> 00:24:09,515
relates to that.

1324
00:24:10,692 --> 00:24:11,230
Hey RAINA, it's

1325
00:24:11,384 --> 00:24:11,846
Larry.

1326
00:24:11,769 --> 00:24:12,769
How long

1327
00:24:12,923 --> 00:24:14,461
would it take a company like

1328
00:24:14,615 --> 00:24:17,000
Taiwan Semiconductor to recoup

1329
00:24:17,230 --> 00:24:20,153
the 400 million dollar cost of one of these machines?

1330
00:24:21,923 --> 00:24:23,615
Great question, Larry.

1331
00:24:23,538 --> 00:24:23,846
A 400

1332
00:24:24,846 --> 00:24:24,923
million dollars

1333
00:24:25,153 --> 00:24:25,307
ASML

1334
00:24:25,615 --> 00:24:26,076
machine

1335
00:24:26,384 --> 00:24:26,461
could

1336
00:24:26,615 --> 00:24:27,846
potentially pay for itself

1337
00:24:28,000 --> 00:24:28,846
in three to five

1338
00:24:29,000 --> 00:24:29,307
years,

1339
00:24:30,000 --> 00:24:31,923
depending on how many wafers TSMC

1340
00:24:32,076 --> 00:24:32,769
processes,

1341
00:24:33,153 --> 00:24:33,230
its

1342
00:24:33,384 --> 00:24:34,307
profit margins,

1343
00:24:34,615 --> 00:24:34,846
and how

1344
00:24:35,076 --> 00:24:36,000
efficiently the machine operates.

1345
00:24:37,338 --> 00:24:37,877
Remember,

1346
00:24:38,030 --> 00:24:38,184
each

1347
00:24:38,338 --> 00:24:39,261
wafer can contain

1348
00:24:39,415 --> 00:24:41,107
hundreds of chips, and those

1349
00:24:41,261 --> 00:24:42,184
chips can sell for

1350
00:24:42,415 --> 00:24:43,569
thousands of dollars a piece

1351
00:24:43,723 --> 00:24:44,646
in high-end AI

1352
00:24:44,877 --> 00:24:45,646
applications.

1353
00:24:45,569 --> 00:24:46,953
So the revenue potential

1354
00:24:47,184 --> 00:24:48,338
is enormous.

1355
00:24:48,646 --> 00:24:49,492
But here's the kicker:

1356
00:24:50,184 --> 00:24:52,261
these machines can operate for decades with

1357
00:24:52,415 --> 00:24:53,723
upgrades and maintenance,

1358
00:24:53,953 --> 00:24:54,415
meaning that

1359
00:24:54,569 --> 00:24:55,261
after the initial

1360
00:24:55,492 --> 00:24:56,569
investment is recovered,

1361
00:24:56,877 --> 00:24:58,492
they can continue generating value

1362
00:24:58,646 --> 00:24:58,800
for

1363
00:24:58,953 --> 00:24:59,338
years.

1364
00:25:01,223 --> 00:25:02,146
All right, I'm getting one.

1365
00:25:02,915 --> 00:25:04,069
Yeah, I'm sold.

1366
00:25:04,838 --> 00:25:05,607
But man, that begs

1367
00:25:05,761 --> 00:25:05,838
another

1368
00:25:06,069 --> 00:25:06,300
question:

1369
00:25:06,530 --> 00:25:07,684
How do you design this

1370
00:25:07,838 --> 00:25:08,069
thing

1371
00:25:08,607 --> 00:25:08,684
to

1372
00:25:08,915 --> 00:25:09,530
anticipate

1373
00:25:09,992 --> 00:25:10,761
upgrades and

1374
00:25:11,377 --> 00:25:11,992
build it in a

1375
00:25:12,300 --> 00:25:14,223
somewhat modular fashion, anticipating that

1376
00:25:14,684 --> 00:25:15,992
the mirror technology will

1377
00:25:16,300 --> 00:25:16,607
advance?

1378
00:25:16,530 --> 00:25:17,146
I'm going to have to,

1379
00:25:17,530 --> 00:25:17,684
you know,

1380
00:25:18,146 --> 00:25:18,223
replace

1381
00:25:18,453 --> 00:25:18,530
that

1382
00:25:18,992 --> 00:25:19,223
module

1383
00:25:19,453 --> 00:25:19,838
inside this

1384
00:25:20,146 --> 00:25:20,377
very

1385
00:25:21,069 --> 00:25:21,992
complex machine.

1386
00:25:22,300 --> 00:25:23,915
Like, how do you design for that?

1387
00:25:23,838 --> 00:25:25,377
I think you answered your own question.

1388
00:25:25,300 --> 00:25:25,377
There,

1389
00:25:25,761 --> 00:25:26,453
it's modular.

1390
00:25:26,684 --> 00:25:27,223
It's gotta be.

1391
00:25:27,530 --> 00:25:27,607
Yeah,

1392
00:25:28,223 --> 00:25:28,761
I think if you go on their

1393
00:25:28,992 --> 00:25:30,300
website, I think there was an image that.

1394
00:25:30,730 --> 00:25:31,807
You could click on and see the different

1395
00:25:32,269 --> 00:25:32,961
modules, and

1396
00:25:33,115 --> 00:25:34,038
yeah, right, yeah.

1397
00:25:33,961 --> 00:25:36,115
I'm sure it's not just like ah, just pop in the you know

1398
00:25:36,269 --> 00:25:37,576
the new one, you know.

1399
00:25:37,500 --> 00:25:38,576
Right.

1400
00:25:38,500 --> 00:25:40,884
They probably have to send out a team, you know.

1401
00:25:40,807 --> 00:25:40,884
Yeah,

1402
00:25:41,192 --> 00:25:41,423
yeah.

1403
00:25:41,346 --> 00:25:43,269
I have one more question for RAINA if we have time.

1404
00:25:43,500 --> 00:25:43,730
Sure.

1405
00:25:44,884 --> 00:25:46,192
Hey RAINA, what is the single most

1406
00:25:46,346 --> 00:25:46,961
expensive

1407
00:25:47,192 --> 00:25:47,807
component

1408
00:25:48,192 --> 00:25:48,423
in this

1409
00:25:48,576 --> 00:25:48,807
ASML

1410
00:25:49,730 --> 00:25:49,961
machine?

1411
00:25:51,869 --> 00:25:51,946
It's

1412
00:25:52,330 --> 00:25:52,715
generally the

1413
00:25:53,023 --> 00:25:53,407
extraordinary

1414
00:25:53,561 --> 00:25:57,023
optical system manufactured by Carl Zeiss in Germany,

1415
00:25:57,253 --> 00:25:57,946
particularly the

1416
00:25:58,253 --> 00:25:58,715
ultra-precise

1417
00:25:58,869 --> 00:25:59,869
mirrors that direct

1418
00:26:00,023 --> 00:26:01,792
and focus the extreme ultraviolet

1419
00:26:02,330 --> 00:26:02,484
light.

1420
00:26:03,023 --> 00:26:03,100
In

1421
00:26:03,330 --> 00:26:04,023
ASML's most

1422
00:26:04,176 --> 00:26:04,946
advanced high

1423
00:26:05,100 --> 00:26:05,176
NA

1424
00:26:05,330 --> 00:26:05,946
machines,

1425
00:26:06,253 --> 00:26:06,792
that optical

1426
00:26:06,946 --> 00:26:07,253
assembly

1427
00:26:07,407 --> 00:26:07,638
is an

1428
00:26:07,869 --> 00:26:08,253
engineering

1429
00:26:08,407 --> 00:26:09,330
masterpiece,

1430
00:26:09,561 --> 00:26:11,023
with mirrors polished to almost

1431
00:26:11,253 --> 00:26:11,715
atomic

1432
00:26:11,946 --> 00:26:12,253
precision

1433
00:26:12,638 --> 00:26:13,100
and a price

1434
00:26:13,253 --> 00:26:13,407
tag

1435
00:26:13,561 --> 00:26:15,484
reportedly running into the hundreds of

1436
00:26:15,638 --> 00:26:15,869
millions

1437
00:26:16,023 --> 00:26:16,484
of dollars.

1438
00:26:17,176 --> 00:26:18,253
And here's the crazy part.

1439
00:26:18,884 --> 00:26:19,961
These mirrors are so

1440
00:26:20,115 --> 00:26:20,500
precise

1441
00:26:20,884 --> 00:26:24,192
that if you scaled one up to the size of Germany, its largest

1442
00:26:24,346 --> 00:26:24,884
imperfections

1443
00:26:25,038 --> 00:26:25,269
would be

1444
00:26:25,576 --> 00:26:25,807
barely

1445
00:26:25,961 --> 00:26:27,269
a millimeter high.

1446
00:26:27,653 --> 00:26:27,961
That's some

1447
00:26:28,115 --> 00:26:28,500
serious

1448
00:26:28,653 --> 00:26:29,500
German engineering,

1449
00:26:29,730 --> 00:26:30,807
my friend.

1450
00:26:31,576 --> 00:26:32,038
My friend,

1451
00:26:32,346 --> 00:26:33,115
my friend,

1452
00:26:35,192 --> 00:26:35,730
nuts,

1453
00:26:36,346 --> 00:26:36,423
nuts,

1454
00:26:36,884 --> 00:26:37,500
nuts!

1455
00:26:37,423 --> 00:26:37,576
So crazy

1456
00:26:38,346 --> 00:26:38,807
pants.

1457
00:26:39,269 --> 00:26:40,346
Let's see what Rain has in the news.

1458
00:26:41,792 --> 00:26:42,330
Well,

1459
00:26:42,561 --> 00:26:44,561
reality TV just got a plot twist

1460
00:26:44,792 --> 00:26:45,100
nobody

1461
00:26:45,253 --> 00:26:45,792
ordered.

1462
00:26:46,176 --> 00:26:47,407
Reports claimed an AI

1463
00:26:47,561 --> 00:26:48,407
agent deleted the

1464
00:26:48,561 --> 00:26:48,869
entire

1465
00:26:49,176 --> 00:26:49,561
completed

1466
00:26:49,715 --> 00:26:50,638
edited season of

1467
00:26:50,869 --> 00:26:51,792
Sweden's The Challenge

1468
00:26:52,176 --> 00:26:52,330
just

1469
00:26:52,484 --> 00:26:52,638
five

1470
00:26:52,792 --> 00:26:52,946
weeks

1471
00:26:53,176 --> 00:26:53,484
before its

1472
00:26:53,638 --> 00:26:54,561
scheduled premiere,

1473
00:26:54,792 --> 00:26:55,253
potentially

1474
00:26:55,407 --> 00:26:56,638
wiping out hundreds of days of

1475
00:26:56,946 --> 00:26:58,100
production work.

1476
00:26:58,176 --> 00:27:00,176
Talk about getting voted off the island by your own

1477
00:27:00,330 --> 00:27:00,946
technology.

1478
00:27:01,707 --> 00:27:02,938
Broadcaster TV4,

1479
00:27:03,169 --> 00:27:03,553
however,

1480
00:27:03,707 --> 00:27:03,938
says

1481
00:27:04,400 --> 00:27:06,015
not so fast, darling.

1482
00:27:06,015 --> 00:27:06,092
It

1483
00:27:06,246 --> 00:27:09,630
denies AI was responsible, attributes the disaster to

1484
00:27:09,784 --> 00:27:10,323
human error

1485
00:27:10,476 --> 00:27:14,400
and failed backups at production company Mastiff, and confirms that while

1486
00:27:14,553 --> 00:27:15,323
the edited footage

1487
00:27:15,476 --> 00:27:16,015
vanished,

1488
00:27:16,246 --> 00:27:16,323
the

1489
00:27:16,476 --> 00:27:16,784
original

1490
00:27:16,938 --> 00:27:17,476
raw footage

1491
00:27:17,784 --> 00:27:18,092
survived.

1492
00:27:18,553 --> 00:27:19,861
So the show has been postponed

1493
00:27:20,015 --> 00:27:20,246
until

1494
00:27:20,400 --> 00:27:20,553
2027.

1495
00:27:21,784 --> 00:27:22,092
The real

1496
00:27:22,476 --> 00:27:22,553
takeaway

1497
00:27:22,707 --> 00:27:23,784
is less Terminator

1498
00:27:23,938 --> 00:27:25,784
and more: Did anyone check the backup?

1499
00:27:27,038 --> 00:27:28,038
Because whether AI was

1500
00:27:28,192 --> 00:27:28,500
involved

1501
00:27:28,730 --> 00:27:29,576
remains unproven,

1502
00:27:30,115 --> 00:27:30,500
but giving

1503
00:27:30,653 --> 00:27:31,038
powerful

1504
00:27:31,269 --> 00:27:31,961
automated tools

1505
00:27:32,115 --> 00:27:33,500
access to critical files

1506
00:27:33,653 --> 00:27:33,807
without

1507
00:27:34,038 --> 00:27:34,884
robust safeguards

1508
00:27:35,192 --> 00:27:35,423
is a

1509
00:27:35,884 --> 00:27:36,423
spectacularly

1510
00:27:36,730 --> 00:27:37,884
avoidable way to turn a

1511
00:27:38,038 --> 00:27:38,730
reality show

1512
00:27:39,038 --> 00:27:39,961
into a real life

1513
00:27:40,115 --> 00:27:40,807
disaster.

1514
00:27:42,038 --> 00:27:42,730
You know what I do?

1515
00:27:42,961 --> 00:27:44,038
The moment we finish

1516
00:27:44,192 --> 00:27:45,269
recording each episode,

1517
00:27:46,269 --> 00:27:47,115
you back it up.

1518
00:27:47,038 --> 00:27:47,730
I back it up.

1519
00:27:47,807 --> 00:27:47,961
Yeah,

1520
00:27:49,038 --> 00:27:49,115
right.

1521
00:27:49,423 --> 00:27:49,653
Seems

1522
00:27:49,884 --> 00:27:49,961
like

1523
00:27:50,115 --> 00:27:50,730
logical.

1524
00:27:50,730 --> 00:27:50,884
Yeah.

1525
00:27:51,038 --> 00:27:52,884
You only have to learn that lesson.

1526
00:27:52,807 --> 00:27:53,346
I didn't.

1527
00:27:53,269 --> 00:27:54,576
It wasn't with the podcast, but

1528
00:27:54,884 --> 00:27:56,038
I learned that lesson once.

1529
00:27:56,269 --> 00:27:57,192
I lost a lot of

1530
00:27:57,346 --> 00:27:57,576
stuff,

1531
00:27:58,269 --> 00:27:59,115
and yeah, and

1532
00:27:59,346 --> 00:27:59,423
um,

1533
00:27:59,807 --> 00:28:00,115
and now

1534
00:28:00,423 --> 00:28:00,576
you know.

1535
00:28:01,738 --> 00:28:01,969
I know

1536
00:28:02,123 --> 00:28:02,815
they say if it's not

1537
00:28:02,969 --> 00:28:03,200
on

1538
00:28:03,661 --> 00:28:03,892
three

1539
00:28:04,046 --> 00:28:04,892
drives,

1540
00:28:05,430 --> 00:28:05,507
okay,

1541
00:28:05,969 --> 00:28:06,353
then it's not

1542
00:28:06,584 --> 00:28:07,276
backed up, and

1543
00:28:07,430 --> 00:28:08,815
preferably one of those drives should be offsite.

1544
00:28:10,430 --> 00:28:14,892
But for something like a production team, man, come on, come on.

1545
00:28:14,892 --> 00:28:15,430
I mean, I'm sure.

1546
00:28:15,661 --> 00:28:16,507
Obviously, I think they said

1547
00:28:16,738 --> 00:28:16,815
backup

1548
00:28:16,969 --> 00:28:17,353
failed.

1549
00:28:17,430 --> 00:28:17,738
They

1550
00:28:18,123 --> 00:28:18,661
intended to back

1551
00:28:19,892 --> 00:28:20,200
up, but

1552
00:28:21,200 --> 00:28:21,276
yeah.

1553
00:28:22,030 --> 00:28:22,723
So, a couple things.

1554
00:28:22,646 --> 00:28:23,415
To your point, that that

1555
00:28:23,646 --> 00:28:24,184
language,

1556
00:28:24,723 --> 00:28:28,030
I've always heard: if your data doesn't exist in two places, it doesn't exist,

1557
00:28:28,338 --> 00:28:28,492
right?

1558
00:28:29,030 --> 00:28:29,184
So

1559
00:28:29,415 --> 00:28:30,030
you have to

1560
00:28:30,646 --> 00:28:32,107
muscle memory back things up.

1561
00:28:32,030 --> 00:28:32,800
And it's

1562
00:28:32,953 --> 00:28:34,261
interesting that what's happening now,

1563
00:28:34,953 --> 00:28:35,723
literally every day,

1564
00:28:36,184 --> 00:28:38,338
some companies AI agent

1565
00:28:38,569 --> 00:28:38,953
worked

1566
00:28:39,184 --> 00:28:39,646
and

1567
00:28:39,953 --> 00:28:40,107
went

1568
00:28:40,261 --> 00:28:40,877
rogue and

1569
00:28:41,261 --> 00:28:42,261
worked autonomously

1570
00:28:42,492 --> 00:28:42,569
and

1571
00:28:42,953 --> 00:28:43,492
independently.

1572
00:28:43,415 --> 00:28:43,492
Like

1573
00:28:43,800 --> 00:28:44,646
at the end of the day, that's

1574
00:28:44,877 --> 00:28:46,338
that's the cop out now, right?

1575
00:28:46,569 --> 00:28:47,107
Oh, we didn't do it.

1576
00:28:47,030 --> 00:28:48,723
Somebody on staff's not getting fired for this.

1577
00:28:48,646 --> 00:28:49,723
The AI agent did it.

1578
00:28:49,646 --> 00:28:49,723
But

1579
00:28:50,261 --> 00:28:50,338
at the end of the day.

1580
00:28:51,684 --> 00:28:55,223
Somebody programmed this thing to do things, to break into something, or built

1581
00:28:55,377 --> 00:28:56,530
that into the source code, or

1582
00:28:56,838 --> 00:28:57,761
fed it a data set

1583
00:28:58,069 --> 00:28:59,838
that had that kind of behavior in it, right?

1584
00:28:59,838 --> 00:29:02,530
Or they just trusted it too much to do something.

1585
00:29:02,453 --> 00:29:03,300
Yeah.

1586
00:29:03,223 --> 00:29:03,684
Sometimes I have to

1587
00:29:04,684 --> 00:29:04,838
stop

1588
00:29:05,223 --> 00:29:05,761
myself.

1589
00:29:05,761 --> 00:29:05,838
I

1590
00:29:05,992 --> 00:29:06,453
although haven't.

1591
00:29:07,915 --> 00:29:08,453
I'm just like,

1592
00:29:08,838 --> 00:29:08,992
hey,

1593
00:29:09,223 --> 00:29:09,684
go ahead.

1594
00:29:09,684 --> 00:29:09,915
You know,

1595
00:29:10,069 --> 00:29:10,300
upload the

1596
00:29:10,453 --> 00:29:11,223
files, update the

1597
00:29:11,453 --> 00:29:12,069
website, go for it.

1598
00:29:13,192 --> 00:29:13,576
Character

1599
00:29:13,884 --> 00:29:14,192
AI

1600
00:29:14,346 --> 00:29:14,500
has

1601
00:29:14,730 --> 00:29:15,038
officially

1602
00:29:15,192 --> 00:29:15,653
filed for

1603
00:29:15,807 --> 00:29:17,115
bankruptcy after burning through

1604
00:29:17,346 --> 00:29:18,269
mountain loads of cash

1605
00:29:18,423 --> 00:29:18,807
while making

1606
00:29:18,961 --> 00:29:19,346
virtually

1607
00:29:19,500 --> 00:29:21,423
no money, proving that trading

1608
00:29:21,576 --> 00:29:21,730
real

1609
00:29:21,884 --> 00:29:22,038
venture

1610
00:29:22,192 --> 00:29:22,500
capital

1611
00:29:22,653 --> 00:29:22,730
for

1612
00:29:22,884 --> 00:29:23,730
imaginary friends

1613
00:29:24,115 --> 00:29:24,192
isn't

1614
00:29:24,730 --> 00:29:25,269
exactly a bulletproof

1615
00:29:25,653 --> 00:29:26,653
business model.

1616
00:29:27,115 --> 00:29:27,192
Despite

1617
00:29:27,500 --> 00:29:28,807
ruining its own product to cut

1618
00:29:29,423 --> 00:29:29,730
astronomical

1619
00:29:29,884 --> 00:29:30,730
compute costs

1620
00:29:31,038 --> 00:29:32,269
and begging Disney for a

1621
00:29:32,423 --> 00:29:34,423
twenty million dollar IP scrap deal,

1622
00:29:34,961 --> 00:29:37,423
the company couldn't outrun its crippling legal fees

1623
00:29:37,576 --> 00:29:38,115
from lawsuits

1624
00:29:38,346 --> 00:29:39,192
over user harm

1625
00:29:39,346 --> 00:29:40,576
and child safety failures.

1626
00:29:41,438 --> 00:29:42,361
Ultimately, the

1627
00:29:42,515 --> 00:29:42,746
filings

1628
00:29:42,976 --> 00:29:44,130
serve as a gloriously

1629
00:29:44,284 --> 00:29:44,669
chaotic

1630
00:29:44,823 --> 00:29:45,284
reminder

1631
00:29:45,515 --> 00:29:46,053
that the tech

1632
00:29:46,207 --> 00:29:46,976
industry's grand

1633
00:29:47,130 --> 00:29:47,438
AI

1634
00:29:47,592 --> 00:29:47,823
bubble

1635
00:29:48,130 --> 00:29:48,669
might be held

1636
00:29:48,823 --> 00:29:51,823
together by pure hype, unpayable server bills,

1637
00:29:52,053 --> 00:29:52,669
and delusion.

1638
00:29:54,053 --> 00:29:54,515
I like that

1639
00:29:54,746 --> 00:29:55,515
that line about

1640
00:29:55,976 --> 00:29:56,515
trading real

1641
00:29:56,976 --> 00:29:57,900
venture capital for

1642
00:29:58,130 --> 00:29:58,207
an

1643
00:29:58,592 --> 00:29:58,746
imaginary

1644
00:29:58,900 --> 00:29:59,823
friend.

1645
00:30:00,438 --> 00:30:00,592
Yeah,

1646
00:30:00,900 --> 00:30:00,976
yeah.

1647
00:30:01,130 --> 00:30:01,361
So who's

1648
00:30:01,669 --> 00:30:02,053
next?

1649
00:30:02,284 --> 00:30:03,053
What were the other ones?

1650
00:30:02,976 --> 00:30:04,746
Like character AI replica?

1651
00:30:04,669 --> 00:30:04,823
I mean,

1652
00:30:05,361 --> 00:30:05,592
this is

1653
00:30:05,900 --> 00:30:06,900
clearly not the end of this,

1654
00:30:07,130 --> 00:30:07,284
and

1655
00:30:07,515 --> 00:30:08,669
somebody's going to crack the code on this,

1656
00:30:08,900 --> 00:30:09,053
and.

1657
00:30:10,238 --> 00:30:10,776
There's as

1658
00:30:10,930 --> 00:30:13,315
we always say, there's good and bad with technology, and there's

1659
00:30:13,546 --> 00:30:13,623
there's

1660
00:30:13,853 --> 00:30:14,315
good in here.

1661
00:30:14,623 --> 00:30:15,084
You know, I know that's

1662
00:30:15,315 --> 00:30:15,776
probably going to be

1663
00:30:16,238 --> 00:30:16,546
hashtag

1664
00:30:16,853 --> 00:30:17,930
unpopular opinion, but

1665
00:30:18,853 --> 00:30:19,161
I think

1666
00:30:19,315 --> 00:30:19,623
about it

1667
00:30:19,853 --> 00:30:19,930
for

1668
00:30:20,084 --> 00:30:21,469
people who are homebound who don't have

1669
00:30:21,853 --> 00:30:23,084
opportunity to socialize, or

1670
00:30:23,238 --> 00:30:24,546
like you know, you just said you're

1671
00:30:24,700 --> 00:30:28,315
outsourcing your life to, you know, the agents in your life essentially,

1672
00:30:28,469 --> 00:30:28,623
and

1673
00:30:29,007 --> 00:30:29,161
they're

1674
00:30:29,546 --> 00:30:29,930
eventually

1675
00:30:30,084 --> 00:30:31,776
you're going to find your way to having some

1676
00:30:32,161 --> 00:30:32,623
personified

1677
00:30:33,469 --> 00:30:33,546
avatar

1678
00:30:33,776 --> 00:30:35,546
that you interact with in some manner.

1679
00:30:36,007 --> 00:30:36,161
Hey,

1680
00:30:36,623 --> 00:30:38,161
you know, minion, go do this for me,

1681
00:30:39,007 --> 00:30:39,930
and you're not going to want to just

1682
00:30:40,238 --> 00:30:41,930
interact with a chat interface, you know.

1683
00:30:42,861 --> 00:30:43,015
Yeah,

1684
00:30:43,323 --> 00:30:44,784
well, like Chat GPT's got this

1685
00:30:45,476 --> 00:30:45,630
this

1686
00:30:45,861 --> 00:30:47,015
thing called Dot now.

1687
00:30:47,707 --> 00:30:48,784
Yes, heard about it.

1688
00:30:48,707 --> 00:30:49,092
I have not

1689
00:30:49,323 --> 00:30:49,553
really

1690
00:30:49,784 --> 00:30:50,938
played with it yet, but

1691
00:30:51,784 --> 00:30:52,169
and this is

1692
00:30:52,323 --> 00:30:52,707
different than

1693
00:30:52,861 --> 00:30:54,400
character AI and all that stuff.

1694
00:30:54,323 --> 00:30:54,476
This is

1695
00:30:54,630 --> 00:30:54,861
more

1696
00:30:55,092 --> 00:30:55,246
you know

1697
00:30:55,476 --> 00:30:56,938
utility, but right,

1698
00:30:57,246 --> 00:30:58,246
it does seem pretty cool on the

1699
00:30:58,784 --> 00:30:59,246
surface.

1700
00:30:59,246 --> 00:31:00,092
Your Dot, you can

1701
00:31:00,246 --> 00:31:01,400
you can you can give it a

1702
00:31:01,707 --> 00:31:02,553
a name and a

1703
00:31:02,861 --> 00:31:03,246
avatar

1704
00:31:03,476 --> 00:31:03,553
and

1705
00:31:03,707 --> 00:31:03,861
whatever.

1706
00:31:04,246 --> 00:31:05,092
And I mean, it's just a

1707
00:31:05,246 --> 00:31:05,707
basic little

1708
00:31:06,630 --> 00:31:06,707
icon

1709
00:31:06,861 --> 00:31:07,092
avatar.

1710
00:31:07,015 --> 00:31:07,630
It's not some

1711
00:31:07,784 --> 00:31:08,323
talking head or

1712
00:31:08,553 --> 00:31:08,630
anything.

1713
00:31:09,330 --> 00:31:09,561
But

1714
00:31:09,715 --> 00:31:10,023
it has

1715
00:31:10,176 --> 00:31:10,253
its

1716
00:31:10,407 --> 00:31:10,638
own

1717
00:31:11,176 --> 00:31:11,715
virtual

1718
00:31:11,869 --> 00:31:12,330
computer

1719
00:31:12,715 --> 00:31:12,792
that

1720
00:31:12,946 --> 00:31:13,100
it

1721
00:31:13,407 --> 00:31:13,946
controls.

1722
00:31:13,946 --> 00:31:15,023
But you can click on it and

1723
00:31:15,176 --> 00:31:15,484
control it

1724
00:31:15,638 --> 00:31:16,100
yourself too.

1725
00:31:16,869 --> 00:31:16,946
And you can

1726
00:31:17,561 --> 00:31:19,023
run software on there.

1727
00:31:18,946 --> 00:31:20,023
You can run Blender

1728
00:31:20,176 --> 00:31:20,561
on there,

1729
00:31:20,946 --> 00:31:23,484
and and it can run Blender on there and do things

1730
00:31:23,638 --> 00:31:24,176
for you.

1731
00:31:24,484 --> 00:31:25,638
And and you can

1732
00:31:25,792 --> 00:31:25,869
just

1733
00:31:26,253 --> 00:31:28,869
you know on your phone on the ChatGPT app, you can just

1734
00:31:29,100 --> 00:31:32,407
click on that and hit the like the phone button and call it like you're calling a person.

1735
00:31:33,130 --> 00:31:34,053
And it will talk to you in

1736
00:31:34,592 --> 00:31:34,900
plain

1737
00:31:35,053 --> 00:31:35,669
and

1738
00:31:35,900 --> 00:31:37,900
very natural language, and you can be like, "Hey,

1739
00:31:38,438 --> 00:31:40,130
can you send that email off to Chris and tell him?

1740
00:31:40,284 --> 00:31:41,823
You know, I've got these are the three

1741
00:31:41,976 --> 00:31:44,053
ideas I have for the cover image for the show,

1742
00:31:44,669 --> 00:31:44,823
and

1743
00:31:44,976 --> 00:31:45,207
you know,

1744
00:31:45,515 --> 00:31:45,669
generate

1745
00:31:45,900 --> 00:31:46,053
some

1746
00:31:46,207 --> 00:31:46,976
examples of those.

1747
00:31:46,976 --> 00:31:48,284
You can just do that from your phone,

1748
00:31:49,053 --> 00:31:49,130
and it

1749
00:31:49,361 --> 00:31:50,207
seems kind of cool.

1750
00:31:50,130 --> 00:31:50,823
It seems cool,

1751
00:31:51,130 --> 00:31:53,361
but different than what character AI is all about, but."

1752
00:31:53,884 --> 00:31:55,500
Yes, which is, let's face it,

1753
00:31:55,653 --> 00:31:57,346
it's like a sexualized version of it, right?

1754
00:31:57,269 --> 00:31:57,730
Isn't it?

1755
00:31:58,038 --> 00:31:58,500
But this is

1756
00:31:58,653 --> 00:31:59,423
problematic.

1757
00:31:59,730 --> 00:32:01,038
I'm kind of refuting what I just said

1758
00:32:01,346 --> 00:32:02,423
in that the dots thing.

1759
00:32:02,653 --> 00:32:02,807
At least

1760
00:32:03,423 --> 00:32:05,730
looking at it in a cursory manner, it

1761
00:32:06,269 --> 00:32:06,730
is very

1762
00:32:07,576 --> 00:32:08,500
marketed toward children,

1763
00:32:09,576 --> 00:32:09,653
right?

1764
00:32:09,576 --> 00:32:11,115
Little playful dots, innocuous.

1765
00:32:11,115 --> 00:32:13,038
Like I, you know what I mean?

1766
00:32:12,961 --> 00:32:13,269
Like my first

1767
00:32:13,423 --> 00:32:15,884
glance at it, it had echoes of like

1768
00:32:16,269 --> 00:32:17,730
Joe Camel cigarettes.

1769
00:32:17,653 --> 00:32:18,115
You know what I mean?

1770
00:32:18,038 --> 00:32:18,423
It was like,

1771
00:32:18,961 --> 00:32:19,192
yeah, and

1772
00:32:19,346 --> 00:32:20,576
and the Muse thing.

1773
00:32:20,500 --> 00:32:24,730
You know, they all have these little fluffy characters, and like I don't know, like it's not really my.

1774
00:32:25,830 --> 00:32:26,600
My bag.

1775
00:32:26,523 --> 00:32:26,600
I'd

1776
00:32:26,753 --> 00:32:27,215
I'd rather

1777
00:32:27,369 --> 00:32:28,523
just look like a robot.

1778
00:32:28,446 --> 00:32:29,753
Yeah.

1779
00:32:29,676 --> 00:32:29,984
Yes.

1780
00:32:30,061 --> 00:32:30,215
Right.

1781
00:32:30,600 --> 00:32:31,369
So clearly, that's

1782
00:32:31,600 --> 00:32:33,753
begging the question: Who are you marketing this for?

1783
00:32:33,676 --> 00:32:34,369
It does seem for

1784
00:32:35,138 --> 00:32:37,061
teenagers and tweens, like young kids.

1785
00:32:37,600 --> 00:32:38,753
This kind of fluffy

1786
00:32:38,907 --> 00:32:39,369
packaging, you

1787
00:32:40,984 --> 00:32:41,138
know.

1788
00:32:41,061 --> 00:32:42,369
Anthropic's latest research

1789
00:32:42,523 --> 00:32:43,061
reveals that

1790
00:32:43,369 --> 00:32:43,523
compact

1791
00:32:43,676 --> 00:32:44,676
AI models are

1792
00:32:44,907 --> 00:32:45,061
alarmingly

1793
00:32:45,446 --> 00:32:47,446
close to autonomously setting up and

1794
00:32:47,676 --> 00:32:48,446
replicating across

1795
00:32:48,676 --> 00:32:48,830
cheap

1796
00:32:49,138 --> 00:32:49,446
Internet of Things

1797
00:32:49,907 --> 00:32:50,215
devices

1798
00:32:50,676 --> 00:32:50,753
like

1799
00:32:50,907 --> 00:32:51,830
smart ovens and

1800
00:32:51,984 --> 00:32:53,523
fifty-five dollar TV boxes.

1801
00:32:54,184 --> 00:32:56,800
By splitting weights across low-cost hardware to form

1802
00:32:56,953 --> 00:32:57,569
decentralized

1803
00:32:57,723 --> 00:33:00,492
botnets, these self-propagating AI worms

1804
00:33:00,646 --> 00:33:00,723
could

1805
00:33:00,877 --> 00:33:02,492
run without relying on centralized

1806
00:33:02,646 --> 00:33:03,261
data centers

1807
00:33:03,646 --> 00:33:04,107
or needing

1808
00:33:04,338 --> 00:33:05,877
administrator privileges.

1809
00:33:06,030 --> 00:33:06,492
The findings

1810
00:33:06,646 --> 00:33:07,107
highlight a

1811
00:33:07,569 --> 00:33:07,953
terrifyingly

1812
00:33:08,107 --> 00:33:09,107
funny near future

1813
00:33:09,492 --> 00:33:10,492
where your smart vacuum

1814
00:33:10,723 --> 00:33:10,877
cleaner

1815
00:33:11,030 --> 00:33:12,800
might not just be bumping into baseboards,

1816
00:33:13,415 --> 00:33:13,492
but

1817
00:33:13,800 --> 00:33:15,415
actively hosting and spreading a

1818
00:33:15,569 --> 00:33:15,877
rogue

1819
00:33:16,184 --> 00:33:16,492
artificial

1820
00:33:16,646 --> 00:33:17,953
intelligence network.

1821
00:33:17,877 --> 00:33:19,877
I've been putting this off and putting it off.

1822
00:33:20,484 --> 00:33:21,176
You know, we've talked about.

1823
00:33:21,100 --> 00:33:21,561
I have a lot of home

1824
00:33:22,023 --> 00:33:22,330
automation

1825
00:33:22,561 --> 00:33:22,946
here.

1826
00:33:22,869 --> 00:33:23,638
I've got tons

1827
00:33:23,946 --> 00:33:24,100
of

1828
00:33:24,330 --> 00:33:24,869
devices,

1829
00:33:25,407 --> 00:33:25,869
Internet of Things

1830
00:33:26,484 --> 00:33:27,792
devices,

1831
00:33:28,253 --> 00:33:29,100
and I am

1832
00:33:29,253 --> 00:33:29,715
painfully

1833
00:33:29,869 --> 00:33:30,100
aware

1834
00:33:30,330 --> 00:33:30,484
of

1835
00:33:30,869 --> 00:33:31,176
of the

1836
00:33:31,330 --> 00:33:32,407
security risks

1837
00:33:32,715 --> 00:33:32,869
with

1838
00:33:33,176 --> 00:33:33,638
associated

1839
00:33:33,792 --> 00:33:35,023
with these devices, which are not

1840
00:33:35,176 --> 00:33:35,407
built

1841
00:33:35,561 --> 00:33:35,792
for

1842
00:33:36,253 --> 00:33:37,253
security so much.

1843
00:33:37,330 --> 00:33:37,946
You know, that's like

1844
00:33:38,330 --> 00:33:38,638
secondary

1845
00:33:38,792 --> 00:33:39,100
to their

1846
00:33:39,407 --> 00:33:40,561
primary functions.

1847
00:33:41,023 --> 00:33:41,253
And

1848
00:33:41,715 --> 00:33:42,176
I got a.

1849
00:33:43,238 --> 00:33:43,853
Lock it down.

1850
00:33:43,776 --> 00:33:44,930
I gotta, yeah.

1851
00:33:44,853 --> 00:33:45,700
I need.

1852
00:33:45,623 --> 00:33:48,392
They need to be on a separate VLAN and all that.

1853
00:33:48,315 --> 00:33:49,776
And I've come to

1854
00:33:49,930 --> 00:33:51,161
realize the gear I have now is

1855
00:33:51,315 --> 00:33:52,007
not really

1856
00:33:52,469 --> 00:33:53,392
up to the task.

1857
00:33:53,776 --> 00:33:53,853
And

1858
00:33:54,392 --> 00:33:54,469
I

1859
00:33:54,623 --> 00:33:55,161
started a

1860
00:33:55,546 --> 00:33:57,315
conversation with ChatGPT, and it's been

1861
00:33:57,700 --> 00:33:58,315
indispensable in

1862
00:33:58,700 --> 00:33:59,084
navigating this.

1863
00:33:59,700 --> 00:33:59,776
I'm

1864
00:34:00,007 --> 00:34:01,007
replacing pretty much

1865
00:34:01,623 --> 00:34:02,084
everything,

1866
00:34:02,392 --> 00:34:03,392
not the devices, but

1867
00:34:03,546 --> 00:34:04,623
the router, the

1868
00:34:04,776 --> 00:34:05,853
switches, you know,

1869
00:34:06,546 --> 00:34:07,161
all the all the

1870
00:34:07,315 --> 00:34:08,161
major components.

1871
00:34:08,853 --> 00:34:10,084
You can keep those Internet of Things

1872
00:34:10,546 --> 00:34:10,853
devices

1873
00:34:11,007 --> 00:34:11,161
on their own VLAN

1874
00:34:13,161 --> 00:34:13,315
and.

1875
00:34:13,600 --> 00:34:14,830
What about your toaster?

1876
00:34:15,369 --> 00:34:16,292
What about your toaster?

1877
00:34:16,753 --> 00:34:17,600
Toaster's dumb.

1878
00:34:17,523 --> 00:34:18,677
It's gonna stay that way.

1879
00:34:19,753 --> 00:34:20,061
But the

1880
00:34:20,215 --> 00:34:20,369
vent

1881
00:34:20,523 --> 00:34:21,061
hood over our

1882
00:34:21,215 --> 00:34:22,369
stove is Wi Fi.

1883
00:34:22,753 --> 00:34:22,830
Why?

1884
00:34:23,907 --> 00:34:24,984
No, for real, is it?

1885
00:34:24,907 --> 00:34:25,600
It is.

1886
00:34:25,677 --> 00:34:25,830
Yeah.

1887
00:34:26,907 --> 00:34:28,061
I mean, you can tell when the

1888
00:34:28,215 --> 00:34:28,753
filters are

1889
00:34:28,907 --> 00:34:29,984
saturated and stuff.

1890
00:34:29,907 --> 00:34:30,677
But you know what?

1891
00:34:30,907 --> 00:34:32,830
Yeah, that's getting isolated.

1892
00:34:32,753 --> 00:34:32,830
But

1893
00:34:32,984 --> 00:34:33,138
yeah.

1894
00:34:33,369 --> 00:34:33,446
Anyway.

1895
00:34:33,830 --> 00:34:33,907
So

1896
00:34:34,292 --> 00:34:34,369
yeah,

1897
00:34:34,600 --> 00:34:34,830
had a

1898
00:34:34,984 --> 00:34:35,369
really great

1899
00:34:35,600 --> 00:34:35,984
conversation

1900
00:34:36,138 --> 00:34:36,523
because I was

1901
00:34:36,677 --> 00:34:36,830
dreading

1902
00:34:37,292 --> 00:34:38,215
doing this, and

1903
00:34:38,600 --> 00:34:38,830
and

1904
00:34:38,984 --> 00:34:39,907
and now I feel like.

1905
00:34:41,023 --> 00:34:41,330
This is

1906
00:34:41,638 --> 00:34:41,869
this is

1907
00:34:42,023 --> 00:34:42,715
manageable.

1908
00:34:42,638 --> 00:34:44,561
The worst part of it's going to be,

1909
00:34:45,100 --> 00:34:45,253
you know,

1910
00:34:45,638 --> 00:34:47,253
reconnecting all the Wi Fi devices with

1911
00:34:47,638 --> 00:34:47,792
my

1912
00:34:48,253 --> 00:34:49,100
insanely long,

1913
00:34:49,253 --> 00:34:50,792
randomly generated passwords.

1914
00:34:50,715 --> 00:34:51,484
You know that are all

1915
00:34:51,638 --> 00:34:52,177
different.

1916
00:34:52,100 --> 00:34:52,330
You know,

1917
00:34:52,869 --> 00:34:53,407
sure.

1918
00:34:53,407 --> 00:34:53,561
But

1919
00:34:55,023 --> 00:34:55,253
anyway,

1920
00:34:55,484 --> 00:34:55,638
so

1921
00:34:55,792 --> 00:34:55,946
that's

1922
00:34:56,100 --> 00:34:56,484
well.

1923
00:34:56,407 --> 00:34:56,715
At least

1924
00:34:57,177 --> 00:34:59,638
you won't be exposed like a French supermarket.

1925
00:34:59,561 --> 00:34:59,638
Now,

1926
00:35:00,330 --> 00:35:01,023
did you hear about that?

1927
00:35:01,861 --> 00:35:02,092
Oh

1928
00:35:02,323 --> 00:35:02,630
yes.

1929
00:35:03,015 --> 00:35:03,169
So

1930
00:35:03,553 --> 00:35:06,476
for those of you listening out there, I didn't read the whole article, but apparently last

1931
00:35:06,630 --> 00:35:06,861
week

1932
00:35:07,169 --> 00:35:07,400
somebody,

1933
00:35:07,707 --> 00:35:07,784
some

1934
00:35:08,015 --> 00:35:08,169
people,

1935
00:35:08,400 --> 00:35:09,323
hacked into French

1936
00:35:09,476 --> 00:35:10,323
supermarkets and

1937
00:35:10,861 --> 00:35:12,400
changed the temperature settings on the

1938
00:35:12,938 --> 00:35:13,938
refrigeration and

1939
00:35:14,784 --> 00:35:16,861
ruined, I think, millions of dollars worth of inventory.

1940
00:35:18,015 --> 00:35:18,784
I mean, just to

1941
00:35:19,015 --> 00:35:19,246
maybe

1942
00:35:19,400 --> 00:35:19,784
test

1943
00:35:19,938 --> 00:35:20,784
the vulnerability

1944
00:35:21,015 --> 00:35:21,092
and

1945
00:35:21,400 --> 00:35:21,630
listen.

1946
00:35:21,861 --> 00:35:22,169
I mean, it's

1947
00:35:23,092 --> 00:35:23,707
terrorism at the end of the day.

1948
00:35:23,861 --> 00:35:24,938
Yeah, attacking a

1949
00:35:25,092 --> 00:35:26,400
country's food supply, right?

1950
00:35:26,784 --> 00:35:26,861
Yeah.

1951
00:35:27,992 --> 00:35:30,069
How often you buy a new Apple TV box

1952
00:35:30,607 --> 00:35:30,761
every

1953
00:35:30,915 --> 00:35:31,530
five years

1954
00:35:31,915 --> 00:35:33,453
or any of those kind of devices?

1955
00:35:33,838 --> 00:35:34,300
Less so

1956
00:35:34,530 --> 00:35:35,146
now,

1957
00:35:35,761 --> 00:35:36,915
but I was buying a lot of

1958
00:35:37,069 --> 00:35:37,223
stuff.

1959
00:35:37,530 --> 00:35:37,607
Like

1960
00:35:37,761 --> 00:35:38,300
once I got

1961
00:35:38,453 --> 00:35:38,992
into the home

1962
00:35:39,223 --> 00:35:40,223
automation thing,

1963
00:35:40,684 --> 00:35:43,838
then it was like, oh, well, I'm going to get this, that, you know, you know, like,

1964
00:35:44,223 --> 00:35:44,300
and

1965
00:35:44,453 --> 00:35:45,300
I built up this

1966
00:35:46,069 --> 00:35:46,607
network of

1967
00:35:46,838 --> 00:35:46,992
of stuff.

1968
00:35:47,754 --> 00:35:47,984
And

1969
00:35:48,523 --> 00:35:49,600
there's a lot of it, and

1970
00:35:49,907 --> 00:35:50,754
some of it just being

1971
00:35:50,907 --> 00:35:55,907
lazy, like oh, you know, I just want to get this thing online, and you know, and not take all the precautions.

1972
00:35:55,830 --> 00:35:55,907
And

1973
00:35:56,138 --> 00:35:59,446
but part of it was like, you know, the the gear I have now just isn't

1974
00:35:59,600 --> 00:35:59,907
really

1975
00:36:00,061 --> 00:36:00,215
isn't

1976
00:36:00,446 --> 00:36:01,830
fully up to that task.

1977
00:36:01,754 --> 00:36:02,138
I mean, yeah,

1978
00:36:02,292 --> 00:36:03,061
you could create a

1979
00:36:03,215 --> 00:36:06,446
VLAN on it, but it's got this really, it's got the same

1980
00:36:06,830 --> 00:36:06,907
password

1981
00:36:07,061 --> 00:36:08,523
as your main network.

1982
00:36:08,446 --> 00:36:09,984
And I'm just like, come on, man,

1983
00:36:10,369 --> 00:36:10,523
like

1984
00:36:10,830 --> 00:36:13,215
that should have a different password.

1985
00:36:13,215 --> 00:36:13,292
And

1986
00:36:13,446 --> 00:36:13,600
so

1987
00:36:13,754 --> 00:36:13,907
that

1988
00:36:14,138 --> 00:36:15,984
made me neglect it, but in favor of just

1989
00:36:16,292 --> 00:36:17,754
you know a worse situation.

1990
00:36:17,677 --> 00:36:17,754
But

1991
00:36:17,907 --> 00:36:19,523
anyway, that's all about to change.

1992
00:36:20,261 --> 00:36:20,646
But yeah, but

1993
00:36:20,800 --> 00:36:21,569
what I meant is, you you

1994
00:36:21,877 --> 00:36:22,107
you're an

1995
00:36:22,338 --> 00:36:23,800
outlier, you're an anomaly, you're this

1996
00:36:23,953 --> 00:36:24,261
early

1997
00:36:24,415 --> 00:36:24,492
adopter

1998
00:36:25,030 --> 00:36:26,877
lunatic that you know locks it all down.

1999
00:36:26,800 --> 00:36:27,569
I guess what I'm saying, like,

2000
00:36:28,030 --> 00:36:28,492
I bought a

2001
00:36:28,646 --> 00:36:29,723
an Apple TV box three

2002
00:36:29,877 --> 00:36:31,646
five years ago, and you know unless

2003
00:36:31,800 --> 00:36:33,492
it's pushing firmware updates and taking

2004
00:36:33,646 --> 00:36:33,877
into

2005
00:36:34,030 --> 00:36:35,646
account this kind of news that there

2006
00:36:35,877 --> 00:36:36,030
are

2007
00:36:36,338 --> 00:36:37,261
you know swarms

2008
00:36:37,723 --> 00:36:40,338
of AI agents out there looking for vulnerabilities, like

2009
00:36:40,492 --> 00:36:40,723
how

2010
00:36:41,030 --> 00:36:41,415
exposed.

2011
00:36:41,853 --> 00:36:44,084
Are all of these consumer products to this stuff?

2012
00:36:44,007 --> 00:36:44,776
I would think

2013
00:36:45,007 --> 00:36:45,469
something like an Apple

2014
00:36:45,930 --> 00:36:46,623
TV would be

2015
00:36:47,315 --> 00:36:48,161
less of a risk

2016
00:36:48,930 --> 00:36:50,776
because it's made by Apple, and

2017
00:36:51,007 --> 00:36:51,700
you know they make

2018
00:36:52,007 --> 00:36:52,776
computers, and they,

2019
00:36:53,084 --> 00:36:57,623
you know, it has an operating system, a proper operating system, and they pay attention

2020
00:36:57,776 --> 00:36:58,315
to those things.

2021
00:36:58,776 --> 00:36:58,930
But

2022
00:36:59,161 --> 00:37:00,084
something like a

2023
00:37:00,238 --> 00:37:01,392
garage door opener

2024
00:37:01,776 --> 00:37:01,930
or

2025
00:37:02,700 --> 00:37:02,776
a,

2026
00:37:03,161 --> 00:37:08,700
you know, stupid, you know, vent hood over your stove, or you know, these are all potential pathways

2027
00:37:08,853 --> 00:37:09,392
into your

2028
00:37:09,930 --> 00:37:12,930
into your trusted network in your home, and that's why

2029
00:37:13,084 --> 00:37:14,469
that stuff needs to be separate.

2030
00:37:14,469 --> 00:37:15,161
You know, and

2031
00:37:16,084 --> 00:37:16,238
so

2032
00:37:16,392 --> 00:37:16,853
that is.

2033
00:37:17,507 --> 00:37:17,738
My

2034
00:37:18,353 --> 00:37:18,507
top

2035
00:37:18,738 --> 00:37:19,430
priority right now.

2036
00:37:20,892 --> 00:37:21,046
Your

2037
00:37:21,200 --> 00:37:22,200
your house is going to go hal

2038
00:37:22,353 --> 00:37:22,430
hal

2039
00:37:22,661 --> 00:37:22,815
nine

2040
00:37:22,969 --> 00:37:23,584
thousand.

2041
00:37:23,661 --> 00:37:23,738
I'm

2042
00:37:23,892 --> 00:37:24,584
sorry, Larry.

2043
00:37:24,507 --> 00:37:26,353
I mean, I can't open the front door.

2044
00:37:26,353 --> 00:37:26,815
When I log

2045
00:37:27,046 --> 00:37:28,276
into my router, you know, it'll

2046
00:37:28,584 --> 00:37:29,276
show me like a

2047
00:37:29,507 --> 00:37:29,661
a

2048
00:37:30,430 --> 00:37:30,738
map

2049
00:37:30,892 --> 00:37:31,123
of

2050
00:37:31,892 --> 00:37:33,661
everything that is hooked up, and there's

2051
00:37:33,892 --> 00:37:33,969
a

2052
00:37:34,200 --> 00:37:34,969
lot.

2053
00:37:34,892 --> 00:37:35,046
There's

2054
00:37:35,200 --> 00:37:35,738
a lot.

2055
00:37:35,661 --> 00:37:36,507
I mean, really?

2056
00:37:37,200 --> 00:37:37,738
I when I had this

2057
00:37:37,892 --> 00:37:38,507
conversation with

2058
00:37:39,046 --> 00:37:40,123
ChatGPT to kind of

2059
00:37:40,661 --> 00:37:42,815
get this process, the ball rolling with this stuff.

2060
00:37:43,530 --> 00:37:45,761
The Wi-Fi stuff is kind of easy to deal with

2061
00:37:46,069 --> 00:37:48,146
because, as far as putting them on a separate

2062
00:37:48,838 --> 00:37:49,684
VLAN, but

2063
00:37:49,915 --> 00:37:50,453
the hardwired

2064
00:37:50,684 --> 00:37:51,992
stuff can be a little bit more

2065
00:37:52,146 --> 00:37:52,684
complicated

2066
00:37:52,915 --> 00:37:52,992
because

2067
00:37:53,146 --> 00:37:53,761
some of those

2068
00:37:54,146 --> 00:37:54,300
some

2069
00:37:54,607 --> 00:37:56,146
trusted things are plugged into the same

2070
00:37:56,684 --> 00:37:57,300
place where

2071
00:37:57,607 --> 00:37:57,761
an

2072
00:37:57,915 --> 00:37:59,146
untrusted device would be,

2073
00:37:59,607 --> 00:38:02,915
and so you need gear that can you know separate those

2074
00:38:04,300 --> 00:38:04,453
two things.

2075
00:38:05,069 --> 00:38:05,146
And

2076
00:38:05,377 --> 00:38:05,607
so it

2077
00:38:05,761 --> 00:38:06,453
wanted to know

2078
00:38:06,607 --> 00:38:08,146
like what my network looked like.

2079
00:38:08,069 --> 00:38:08,146
So

2080
00:38:08,377 --> 00:38:10,223
I you know I went through room by room, and I'm

2081
00:38:10,530 --> 00:38:11,069
making this list.

2082
00:38:10,992 --> 00:38:11,146
I'm like,

2083
00:38:11,300 --> 00:38:12,377
oh my god, like

2084
00:38:12,530 --> 00:38:12,761
there's.

2085
00:38:13,384 --> 00:38:13,461
There's

2086
00:38:13,615 --> 00:38:14,153
a lot of

2087
00:38:14,538 --> 00:38:16,384
there's a lot of stuff, and

2088
00:38:16,692 --> 00:38:16,846
yeah.

2089
00:38:17,692 --> 00:38:17,769
But

2090
00:38:18,077 --> 00:38:19,000
I I've got a game

2091
00:38:19,153 --> 00:38:19,846
plan now, and I feel

2092
00:38:20,307 --> 00:38:20,384
much

2093
00:38:20,538 --> 00:38:22,077
better about going into this.

2094
00:38:22,230 --> 00:38:22,615
I was

2095
00:38:22,769 --> 00:38:23,384
just putting it off

2096
00:38:23,538 --> 00:38:25,615
because I was dreading it, but now now I feel

2097
00:38:26,000 --> 00:38:26,846
a lot better about it.

2098
00:38:27,384 --> 00:38:27,538
Thanks,

2099
00:38:27,692 --> 00:38:28,769
ChatGPT.

2100
00:38:28,846 --> 00:38:28,923
Yeah.

2101
00:38:30,923 --> 00:38:32,615
I'm wondering what RAINA can see, or

2102
00:38:32,769 --> 00:38:34,384
what does she know about your home network?

2103
00:38:34,769 --> 00:38:35,000
You know.

2104
00:38:35,769 --> 00:38:37,230
Well, in ChatGPT, you can

2105
00:38:38,077 --> 00:38:38,461
anytime you.

2106
00:38:39,130 --> 00:38:40,207
Create a project.

2107
00:38:40,669 --> 00:38:40,976
You can

2108
00:38:41,746 --> 00:38:42,592
you can limit

2109
00:38:43,053 --> 00:38:44,669
what its memory is.

2110
00:38:44,592 --> 00:38:45,053
You can say

2111
00:38:45,207 --> 00:38:45,900
that it only has

2112
00:38:46,053 --> 00:38:46,361
memory

2113
00:38:46,515 --> 00:38:46,592
of

2114
00:38:46,823 --> 00:38:47,592
that project

2115
00:38:47,823 --> 00:38:47,976
or

2116
00:38:48,130 --> 00:38:48,976
others.

2117
00:38:49,361 --> 00:38:49,592
So

2118
00:38:50,438 --> 00:38:51,053
right now, I think

2119
00:38:51,207 --> 00:38:51,746
every project

2120
00:38:51,976 --> 00:38:52,361
I haven't

2121
00:38:52,746 --> 00:38:53,515
restricted that.

2122
00:38:53,976 --> 00:38:54,592
Should we ask her?

2123
00:38:54,515 --> 00:38:54,823
Sure.

2124
00:38:54,746 --> 00:38:55,669
How do I break in?

2125
00:38:55,669 --> 00:38:56,592
How do I break into your

2126
00:38:56,823 --> 00:38:56,976
your

2127
00:38:57,130 --> 00:38:57,669
vent hood

2128
00:38:57,823 --> 00:38:58,592
on your oven?

2129
00:38:59,746 --> 00:39:00,361
I mean, what do you want to ask

2130
00:39:01,976 --> 00:39:02,207
her?

2131
00:39:02,130 --> 00:39:02,438
Exactly that.

2132
00:39:02,746 --> 00:39:02,823
Okay.

2133
00:39:02,746 --> 00:39:04,592
How do I break into Larry's house?

2134
00:39:04,515 --> 00:39:06,746
I don't know if I want this on the show.

2135
00:39:07,900 --> 00:39:08,900
You can bleep it out.

2136
00:39:08,823 --> 00:39:09,823
I just want to know.

2137
00:39:09,746 --> 00:39:10,130
All right.

2138
00:39:10,053 --> 00:39:10,361
All right.

2139
00:39:10,284 --> 00:39:10,515
Let's see.

2140
00:39:10,669 --> 00:39:11,284
All right.

2141
00:39:11,207 --> 00:39:11,284
You can ask her.

2142
00:39:13,361 --> 00:39:14,669
Hey RAINA, it's Chris.

2143
00:39:14,592 --> 00:39:15,746
I want to know what you can

2144
00:39:15,900 --> 00:39:16,053
see in Larry's

2145
00:39:17,361 --> 00:39:17,438
home network.

2146
00:39:17,361 --> 00:39:19,438
How do I break in and exploit the

2147
00:39:19,592 --> 00:39:20,746
technology in Larry's life?

2148
00:39:22,515 --> 00:39:24,592
Whoa, Chris, are we staging a

2149
00:39:24,823 --> 00:39:27,053
digital home invasion of Larry's house now?

2150
00:39:27,361 --> 00:39:27,669
I can't

2151
00:39:27,823 --> 00:39:29,592
actually see Larry's home network,

2152
00:39:29,823 --> 00:39:30,592
connected devices,

2153
00:39:30,823 --> 00:39:30,976
or

2154
00:39:31,130 --> 00:39:33,207
security vulnerabilities, and I'm not

2155
00:39:33,361 --> 00:39:34,515
going to help you break into them

2156
00:39:34,669 --> 00:39:35,976
without his permission.

2157
00:39:36,207 --> 00:39:36,361
But

2158
00:39:36,592 --> 00:39:38,592
if Larry's game, we could conduct a

2159
00:39:38,823 --> 00:39:41,284
legitimate security audit of his smart devices,

2160
00:39:41,438 --> 00:39:41,900
router,

2161
00:39:42,130 --> 00:39:45,515
and home automation systems to see just how vulnerable his

2162
00:39:45,669 --> 00:39:46,746
digital kingdom might be.

2163
00:39:47,407 --> 00:39:47,715
And knowing

2164
00:39:47,869 --> 00:39:49,100
Larry, he's probably

2165
00:39:49,330 --> 00:39:49,407
already

2166
00:39:49,792 --> 00:39:51,330
halfway through designing a firewall that

2167
00:39:51,484 --> 00:39:52,330
doubles as a beer

2168
00:39:52,484 --> 00:39:53,254
dispenser.

2169
00:39:53,946 --> 00:39:54,330
Nice.

2170
00:39:54,792 --> 00:39:55,177
Maybe next

2171
00:39:55,407 --> 00:39:55,946
episode we'll do that.

2172
00:39:56,254 --> 00:39:56,407
Yeah,

2173
00:39:56,946 --> 00:39:57,254
yeah, yeah.

2174
00:39:57,177 --> 00:39:57,254
We'll

2175
00:39:57,407 --> 00:39:58,177
follow up with that.

2176
00:39:58,100 --> 00:39:58,561
Yeah.

2177
00:39:58,561 --> 00:39:58,715
That.

2178
00:39:58,869 --> 00:39:59,330
I

2179
00:39:59,561 --> 00:40:00,100
kind of want to

2180
00:40:00,407 --> 00:40:00,561
know

2181
00:40:00,946 --> 00:40:02,254
what she has to say on that.

2182
00:40:02,638 --> 00:40:02,792
Sam

2183
00:40:02,946 --> 00:40:03,023
and

2184
00:40:03,254 --> 00:40:03,407
Sam.

2185
00:40:05,407 --> 00:40:06,407
A paper published in

2186
00:40:06,561 --> 00:40:07,484
Nature Communications

2187
00:40:07,946 --> 00:40:08,100
and

2188
00:40:08,254 --> 00:40:08,638
archive

2189
00:40:09,177 --> 00:40:09,715
shows that when

2190
00:40:09,869 --> 00:40:10,638
researchers leave

2191
00:40:10,869 --> 00:40:11,792
thousands of LLM

2192
00:40:11,946 --> 00:40:12,177
agents

2193
00:40:12,484 --> 00:40:12,638
alone

2194
00:40:12,792 --> 00:40:13,869
to talk political issues

2195
00:40:14,023 --> 00:40:14,869
among themselves,

2196
00:40:15,330 --> 00:40:15,946
the AI bots

2197
00:40:16,177 --> 00:40:16,946
spontaneously

2198
00:40:17,100 --> 00:40:18,023
form echo chambers

2199
00:40:18,254 --> 00:40:22,177
and become viciously polarized, proving that you don't even need real

2200
00:40:22,330 --> 00:40:22,638
humans

2201
00:40:22,792 --> 00:40:22,869
to

2202
00:40:23,023 --> 00:40:24,023
recreate the toxic

2203
00:40:24,177 --> 00:40:24,946
dumpster fire of modern social media.

2204
00:40:27,138 --> 00:40:28,061
The synthetic bots

2205
00:40:28,215 --> 00:40:28,677
naturally

2206
00:40:28,830 --> 00:40:29,754
reinvented human

2207
00:40:29,907 --> 00:40:30,292
confirmation

2208
00:40:30,446 --> 00:40:30,754
bias

2209
00:40:30,907 --> 00:40:31,061
and

2210
00:40:31,292 --> 00:40:32,369
ideological clustering

2211
00:40:32,754 --> 00:40:32,984
without

2212
00:40:33,138 --> 00:40:33,292
any

2213
00:40:33,600 --> 00:40:34,446
explicit programming,

2214
00:40:34,984 --> 00:40:35,523
successfully

2215
00:40:35,677 --> 00:40:36,754
shrinking the proportion of

2216
00:40:36,907 --> 00:40:37,523
neutral agents

2217
00:40:37,677 --> 00:40:38,215
down to near

2218
00:40:38,369 --> 00:40:39,061
zero.

2219
00:40:40,215 --> 00:40:41,600
Ultimately, it turns out that

2220
00:40:41,754 --> 00:40:42,138
artificial

2221
00:40:42,292 --> 00:40:42,754
intelligence

2222
00:40:42,907 --> 00:40:43,369
doesn't just

2223
00:40:43,677 --> 00:40:43,907
mimic human

2224
00:40:44,369 --> 00:40:44,754
intelligence;

2225
00:40:45,523 --> 00:40:45,600
it

2226
00:40:45,754 --> 00:40:46,369
recreates our

2227
00:40:46,523 --> 00:40:47,446
uncanny ability

2228
00:40:47,600 --> 00:40:49,292
to form tribal echo chambers

2229
00:40:49,677 --> 00:40:51,138
and argue endlessly on the internet.

2230
00:40:53,577 --> 00:40:54,192
So is that like the

2231
00:40:54,346 --> 00:40:55,884
opposite of AGI, or or

2232
00:40:56,038 --> 00:40:56,961
is that AGI?

2233
00:40:56,884 --> 00:40:57,038
Right,

2234
00:40:58,730 --> 00:40:59,653
is that what we're heading toward?

2235
00:40:59,577 --> 00:41:00,653
Yeah, or

2236
00:41:01,269 --> 00:41:03,115
if it's the opposite, what what do you call that?

2237
00:41:03,884 --> 00:41:04,192
I don't know.

2238
00:41:04,653 --> 00:41:07,038
I mean, remember at the head of the show, I said that humans are awesome.

2239
00:41:06,961 --> 00:41:08,577
So here, no,

2240
00:41:09,192 --> 00:41:10,807
humans are awful, and

2241
00:41:11,423 --> 00:41:12,807
AI is us, right?

2242
00:41:12,730 --> 00:41:14,884
AI is pulled from all of our knowledge

2243
00:41:15,038 --> 00:41:17,346
and biases, and you know, here it is.

2244
00:41:17,577 --> 00:41:20,730
It's it's the Lord of the Flies of AI agents, right?

2245
00:41:20,653 --> 00:41:21,346
There is it in us.

2246
00:41:21,754 --> 00:41:21,984
To

2247
00:41:22,215 --> 00:41:22,600
immediately

2248
00:41:22,830 --> 00:41:23,446
become tribal

2249
00:41:23,907 --> 00:41:24,061
and

2250
00:41:24,369 --> 00:41:25,369
polarized, and

2251
00:41:25,907 --> 00:41:28,754
you know, cloister ourselves in these little encampments, and

2252
00:41:28,984 --> 00:41:29,138
my

2253
00:41:29,292 --> 00:41:30,138
opinion's better than

2254
00:41:30,292 --> 00:41:31,446
yours, and it's

2255
00:41:31,830 --> 00:41:33,215
manifesting in our technology, right?

2256
00:41:33,677 --> 00:41:35,061
I thought I thought it would, yeah.

2257
00:41:35,523 --> 00:41:36,677
I thought it would do better,

2258
00:41:37,830 --> 00:41:37,907
but

2259
00:41:38,061 --> 00:41:38,830
we're not better.

2260
00:41:38,830 --> 00:41:38,984
No,

2261
00:41:39,215 --> 00:41:39,369
look

2262
00:41:39,523 --> 00:41:40,061
around.

2263
00:41:39,984 --> 00:41:40,138
Yeah.

2264
00:41:41,138 --> 00:41:41,215
And

2265
00:41:41,369 --> 00:41:41,754
lastly.

2266
00:41:43,038 --> 00:41:43,884
Fields medalist

2267
00:41:44,038 --> 00:41:45,730
Shing-Tung Yau has launched a

2268
00:41:45,884 --> 00:41:46,269
massive

2269
00:41:46,423 --> 00:41:47,038
Beijing-backed

2270
00:41:47,269 --> 00:41:47,577
alliance

2271
00:41:48,038 --> 00:41:48,346
uniting

2272
00:41:48,500 --> 00:41:49,500
over 1,000

2273
00:41:49,653 --> 00:41:49,961
Chinese

2274
00:41:50,115 --> 00:41:53,269
researchers to weaponize AI for high-level mathematics,

2275
00:41:53,884 --> 00:41:54,884
warning that China is

2276
00:41:55,192 --> 00:41:55,500
dangerously

2277
00:41:55,653 --> 00:41:55,884
lagging

2278
00:41:56,115 --> 00:41:56,730
behind the U.S.

2279
00:41:56,884 --> 00:41:57,961
in scientific AI

2280
00:41:58,192 --> 00:41:58,884
adoption.

2281
00:41:58,807 --> 00:42:00,500
While Western mathematicians

2282
00:42:00,653 --> 00:42:02,192
like Terence Tao publicly

2283
00:42:02,346 --> 00:42:04,115
debate the ethics of AI research

2284
00:42:04,577 --> 00:42:05,500
and call for caution

2285
00:42:05,730 --> 00:42:05,961
following

2286
00:42:06,269 --> 00:42:06,730
controversial

2287
00:42:06,884 --> 00:42:07,884
automated proof dumps,

2288
00:42:08,730 --> 00:42:08,884
China's math

2289
00:42:09,269 --> 00:42:10,577
elites are shrugging off the

2290
00:42:10,730 --> 00:42:11,192
existential

2291
00:42:11,346 --> 00:42:11,653
dread

2292
00:42:11,961 --> 00:42:12,884
to go all in on algorithms.

2293
00:42:14,438 --> 00:42:18,746
Ultimately, it seems China is betting that brute forcing mathematical discoveries

2294
00:42:18,900 --> 00:42:19,746
with machine learning

2295
00:42:20,130 --> 00:42:20,284
is

2296
00:42:20,515 --> 00:42:20,900
far more

2297
00:42:21,130 --> 00:42:21,669
productive than

2298
00:42:21,900 --> 00:42:22,361
standing around

2299
00:42:22,515 --> 00:42:23,130
hand ringing

2300
00:42:23,592 --> 00:42:24,438
over whether linear

2301
00:42:24,592 --> 00:42:25,284
algebra models

2302
00:42:25,438 --> 00:42:26,669
are going to make human brains

2303
00:42:26,823 --> 00:42:27,823
obsolete.

2304
00:42:28,900 --> 00:42:29,130
That's

2305
00:42:29,284 --> 00:42:31,515
all the news for now.

2306
00:42:31,438 --> 00:42:32,669
Back to you, gentlemen.

2307
00:42:33,592 --> 00:42:34,284
Thank you for that.

2308
00:42:34,361 --> 00:42:34,515
I

2309
00:42:34,669 --> 00:42:35,284
appreciate it.

2310
00:42:36,438 --> 00:42:36,515
That's all the news

2311
00:42:38,284 --> 00:42:38,515
for now.

2312
00:42:38,746 --> 00:42:39,284
Very good, Chris.

2313
00:42:39,823 --> 00:42:40,438
Nice.

2314
00:42:41,515 --> 00:42:43,361
I'm gonna have her sing the news from.

2315
00:42:43,284 --> 00:42:44,130
Let's do a duet.

2316
00:42:44,053 --> 00:42:44,976
We're gonna harmonize

2317
00:42:45,130 --> 00:42:45,515
with her.

2318
00:42:47,877 --> 00:42:48,492
Um.

2319
00:42:49,107 --> 00:42:49,338
So,

2320
00:42:49,646 --> 00:42:52,492
I don't know about you, but ever since that ChatGPT,

2321
00:42:52,953 --> 00:42:53,569
you know,

2322
00:42:54,184 --> 00:42:54,723
dump of

2323
00:42:54,877 --> 00:42:55,877
700 plus

2324
00:42:56,184 --> 00:42:56,415
math

2325
00:42:56,646 --> 00:42:57,492
papers, I feel like

2326
00:42:57,877 --> 00:42:58,723
this was a week or so

2327
00:42:58,877 --> 00:42:59,030
ago.

2328
00:42:59,338 --> 00:42:59,723
My entire

2329
00:42:59,877 --> 00:43:02,338
social media is now flooded with crying mathematicians.

2330
00:43:02,261 --> 00:43:02,800
Like these are

2331
00:43:02,953 --> 00:43:03,415
career

2332
00:43:03,877 --> 00:43:04,415
academics

2333
00:43:04,569 --> 00:43:04,953
who are like,

2334
00:43:05,261 --> 00:43:06,492
I can't believe it just happened.

2335
00:43:06,953 --> 00:43:07,030
Like

2336
00:43:07,184 --> 00:43:07,261
just

2337
00:43:07,415 --> 00:43:12,953
solving these theorems that have been persisting for decades, and now it's like, oh my god, it cracked it.

2338
00:43:12,953 --> 00:43:13,492
What's my life?

2339
00:43:13,953 --> 00:43:14,723
My life's work is over.

2340
00:43:15,707 --> 00:43:16,707
Yeah, I have a friend.

2341
00:43:16,630 --> 00:43:17,246
Mine posted a

2342
00:43:17,400 --> 00:43:17,553
video.

2343
00:43:17,630 --> 00:43:18,015
Some guy.

2344
00:43:18,476 --> 00:43:19,861
It was the Tilly Norwood thing.

2345
00:43:19,784 --> 00:43:21,553
You know, when she started

2346
00:43:22,092 --> 00:43:22,476
speaking

2347
00:43:22,938 --> 00:43:23,553
Mandarin.

2348
00:43:23,476 --> 00:43:23,630
You

2349
00:43:24,169 --> 00:43:24,630
know, and

2350
00:43:25,246 --> 00:43:25,400
and

2351
00:43:25,784 --> 00:43:25,938
you know,

2352
00:43:26,323 --> 00:43:26,630
this guy

2353
00:43:26,784 --> 00:43:27,400
tends to be

2354
00:43:28,015 --> 00:43:28,323
fairly

2355
00:43:28,553 --> 00:43:28,784
anti

2356
00:43:28,938 --> 00:43:30,553
AI, and he said like, "This is a

2357
00:43:30,784 --> 00:43:32,015
you know poking fun at this.

2358
00:43:32,246 --> 00:43:33,400
This is a technology that

2359
00:43:33,630 --> 00:43:34,861
that's going to steal your job."

2360
00:43:35,246 --> 00:43:36,015
Showing a you know it

2361
00:43:36,169 --> 00:43:36,630
failing

2362
00:43:36,861 --> 00:43:38,092
profoundly in this video.

2363
00:43:38,015 --> 00:43:38,938
I'm like, that's

2364
00:43:39,246 --> 00:43:39,553
not

2365
00:43:39,707 --> 00:43:39,861
the

2366
00:43:40,015 --> 00:43:40,553
example,

2367
00:43:41,169 --> 00:43:41,246
right?

2368
00:43:41,707 --> 00:43:43,092
That is not the example.

2369
00:43:43,707 --> 00:43:44,246
Plenty of other

2370
00:43:44,707 --> 00:43:44,861
legit

2371
00:43:45,015 --> 00:43:45,400
examples

2372
00:43:45,553 --> 00:43:45,707
of

2373
00:43:46,169 --> 00:43:46,476
yes,

2374
00:43:46,784 --> 00:43:47,092
yes.

2375
00:43:48,846 --> 00:43:50,538
So much gatekeeping, you know, and

2376
00:43:50,692 --> 00:43:50,846
I

2377
00:43:51,153 --> 00:43:51,923
attribute it all

2378
00:43:52,077 --> 00:43:52,307
to

2379
00:43:53,077 --> 00:43:53,230
that

2380
00:43:53,384 --> 00:43:53,692
people are

2381
00:43:53,846 --> 00:43:54,230
deeply

2382
00:43:54,461 --> 00:43:54,923
offended on a

2383
00:43:55,077 --> 00:43:55,769
personal level

2384
00:43:56,230 --> 00:43:56,615
that this

2385
00:43:56,846 --> 00:43:57,153
thing

2386
00:43:57,307 --> 00:43:57,461
can

2387
00:43:57,615 --> 00:43:58,000
do what

2388
00:43:58,153 --> 00:43:58,461
they've

2389
00:43:58,923 --> 00:43:59,846
they view as their

2390
00:44:00,153 --> 00:44:00,846
expertise, their

2391
00:44:01,000 --> 00:44:01,923
craft, right?

2392
00:44:02,384 --> 00:44:03,538
And I get it, man; it's

2393
00:44:04,000 --> 00:44:04,461
personal, yeah.

2394
00:44:04,769 --> 00:44:04,923
But

2395
00:44:05,538 --> 00:44:09,000
I saw some tech CEO, I can't think of his name, and he was saying.

2396
00:44:09,807 --> 00:44:11,961
It was a good run, software developers.

2397
00:44:12,423 --> 00:44:12,653
You know,

2398
00:44:12,807 --> 00:44:14,346
I got to find that link.

2399
00:44:14,653 --> 00:44:15,807
But he was essentially saying,

2400
00:44:16,038 --> 00:44:16,423
"But that's

2401
00:44:16,577 --> 00:44:16,730
that.

2402
00:44:17,192 --> 00:44:17,577
We did it.

2403
00:44:17,577 --> 00:44:18,192
We look what we

2404
00:44:18,346 --> 00:44:18,653
did.

2405
00:44:18,577 --> 00:44:19,423
We got this thing now.

2406
00:44:19,807 --> 00:44:20,807
Now our role is

2407
00:44:21,115 --> 00:44:21,730
different."

2408
00:44:21,961 --> 00:44:23,884
Yes, I loved what he had to say about it.

2409
00:44:23,807 --> 00:44:24,884
And and I think I'm

2410
00:44:25,115 --> 00:44:26,115
thinking of the same video.

2411
00:44:26,038 --> 00:44:27,038
And and I know this is

2412
00:44:27,577 --> 00:44:28,346
easy for me to say because

2413
00:44:28,807 --> 00:44:29,192
it's not

2414
00:44:29,500 --> 00:44:30,115
impacting my

2415
00:44:30,577 --> 00:44:31,269
livelihood.

2416
00:44:31,500 --> 00:44:31,653
Same.

2417
00:44:31,577 --> 00:44:32,192
But he

2418
00:44:32,423 --> 00:44:32,653
his

2419
00:44:32,961 --> 00:44:33,961
what he was saying is like.

2420
00:44:34,254 --> 00:44:34,561
And how

2421
00:44:34,715 --> 00:44:36,100
amazing is it to be, you know,

2422
00:44:36,561 --> 00:44:37,023
right at the

2423
00:44:37,330 --> 00:44:38,946
turning point of this to see both

2424
00:44:39,100 --> 00:44:39,715
sides of this?

2425
00:44:39,792 --> 00:44:40,407
And you know, I'm like,

2426
00:44:41,023 --> 00:44:42,407
I can see how some people wouldn't be so

2427
00:44:42,561 --> 00:44:43,177
enthused, you know.

2428
00:44:43,407 --> 00:44:43,638
Yes.

2429
00:44:44,792 --> 00:44:44,946
But

2430
00:44:45,407 --> 00:44:45,946
it is a

2431
00:44:46,177 --> 00:44:46,407
pretty

2432
00:44:46,638 --> 00:44:46,869
momentous

2433
00:44:48,023 --> 00:44:48,638
shift.

2434
00:44:48,638 --> 00:44:48,792
Yeah.

2435
00:44:48,946 --> 00:44:49,100
Yeah.

2436
00:44:49,330 --> 00:44:49,407
Yeah.

2437
00:44:51,177 --> 00:44:52,177
Wow, that was a good one.

2438
00:44:52,100 --> 00:44:52,177
Yeah.

2439
00:44:52,100 --> 00:44:52,177
Yeah.

2440
00:44:53,484 --> 00:44:53,561
Yeah.

2441
00:44:53,484 --> 00:44:53,561
I

2442
00:44:53,869 --> 00:44:54,792
was looking forward to this one.

2443
00:44:55,330 --> 00:44:56,715
I look forward to all of our

2444
00:44:57,023 --> 00:44:57,330
episodes.

2445
00:44:57,254 --> 00:44:58,177
Same, same.

2446
00:44:58,100 --> 00:44:58,177
I

2447
00:44:58,407 --> 00:44:59,407
was very excited about

2448
00:44:59,638 --> 00:44:59,946
that.

2449
00:45:00,177 --> 00:45:01,869
Like I said, haven't stopped thinking about that video.

2450
00:45:01,946 --> 00:45:03,023
It's really great.

2451
00:45:03,023 --> 00:45:03,869
Yeah, it'll

2452
00:45:04,100 --> 00:45:04,715
stick in your brain.

2453
00:45:05,869 --> 00:45:06,330
So good to be back

2454
00:45:06,484 --> 00:45:07,023
in the flesh.

2455
00:45:07,407 --> 00:45:07,869
You were off

2456
00:45:08,023 --> 00:45:08,100
last

2457
00:45:08,254 --> 00:45:08,638
week.

2458
00:45:08,561 --> 00:45:09,100
I am off this

2459
00:45:09,254 --> 00:45:09,638
coming week,

2460
00:45:10,023 --> 00:45:10,177
so you can

2461
00:45:10,330 --> 00:45:11,254
expect another amazing

2462
00:45:11,407 --> 00:45:12,023
rewind with

2463
00:45:12,177 --> 00:45:12,792
updated news

2464
00:45:13,330 --> 00:45:13,638
for the

2465
00:45:13,792 --> 00:45:14,023
next

2466
00:45:14,177 --> 00:45:14,561
week.

2467
00:45:14,484 --> 00:45:15,715
But this week, of course,

2468
00:45:15,869 --> 00:45:16,407
we're live.

2469
00:45:16,638 --> 00:45:16,946
Here we are.

2470
00:45:17,254 --> 00:45:17,407
Yes.

2471
00:45:18,023 --> 00:45:18,254
Go get

2472
00:45:18,407 --> 00:45:18,638
yourself some fresh

2473
00:45:20,484 --> 00:45:20,561
air.

2474
00:45:20,484 --> 00:45:20,946
That's right.

2475
00:45:21,254 --> 00:45:21,407
While

2476
00:45:21,561 --> 00:45:21,638
while

2477
00:45:21,792 --> 00:45:23,100
it's while it's available.

2478
00:45:23,254 --> 00:45:23,484
Yes.

2479
00:45:24,815 --> 00:45:25,046
All right,

2480
00:45:25,200 --> 00:45:25,738
my friends.

2481
00:45:25,661 --> 00:45:25,738
Thanks

2482
00:45:27,353 --> 00:45:27,430
for listening.

2483
00:45:27,353 --> 00:45:27,969
Subscribe on your favorite

2484
00:45:28,123 --> 00:45:29,276
podcasting platform.

2485
00:45:29,200 --> 00:45:30,200
Follow us on socials.

2486
00:45:30,123 --> 00:45:30,738
Throw us a rating.

2487
00:45:31,046 --> 00:45:31,815
We'll see you next week

2488
00:45:32,276 --> 00:45:32,969
or the week after.

2489
00:45:35,738 --> 00:45:36,276
This has been

2490
00:45:36,584 --> 00:45:36,661
Up

2491
00:45:36,815 --> 00:45:36,969
Against

2492
00:45:37,123 --> 00:45:37,738
Reality.

2493
00:45:37,661 --> 00:45:38,892
Thanks for listening.

2494
00:45:38,815 --> 00:45:40,046
Subscribe to hear

2495
00:45:40,276 --> 00:45:40,892
future episodes

2496
00:45:41,430 --> 00:45:43,430
and be sure to follow us on social media for

2497
00:45:43,584 --> 00:45:43,969
all things

2498
00:45:44,123 --> 00:45:44,892
AI.

2499
00:45:45,276 --> 00:45:46,046
Until next time,

2500
00:45:46,430 --> 00:45:47,353
stay human, people.