1
00:00:00,000 --> 00:00:04,879
Brace yourself. I'm going to show you an

2
00:00:02,640 --> 00:00:07,599
aesthetic catastrophe.

3
00:00:04,879 --> 00:00:10,000
It's the worst PowerPoint slide in human

4
00:00:07,599 --> 00:00:12,080
history. It was produced by the US

5
00:00:10,000 --> 00:00:13,839
military in 2010 to illustrate the

6
00:00:12,080 --> 00:00:16,480
difficulties they were facing during the

7
00:00:13,839 --> 00:00:19,039
war in Afghanistan. When the public got

8
00:00:16,480 --> 00:00:22,679
a hold of it, it became a case study for

9
00:00:19,039 --> 00:00:24,480
how not to do PowerPoint. This is actual

10
00:00:22,679 --> 00:00:27,000
slide of the Afghanistan [music]

11
00:00:24,480 --> 00:00:28,519
strategy last year and it was prepared

12
00:00:27,000 --> 00:00:30,600
by the staff to General Stanley

13
00:00:28,519 --> 00:00:31,640
McChrystal. Listen here, Thomas.

14
00:00:30,600 --> 00:00:32,320
We have met the enemy and he is

15
00:00:31,640 --> 00:00:34,719
PowerPoint.

16
00:00:32,320 --> 00:00:36,200
>> It was this incredibly complicated

17
00:00:34,719 --> 00:00:38,839
PowerPoint chart.

18
00:00:36,200 --> 00:00:40,840
>> Even the head of US forces at the time,

19
00:00:38,840 --> 00:00:43,880
General Stanley McChrystal, knew

20
00:00:40,840 --> 00:00:45,280
immediately how bad the slide was. When

21
00:00:43,880 --> 00:00:47,359
the staff brought this to McChrystal,

22
00:00:45,280 --> 00:00:49,320
this PowerPoint slide, he said, "When we

23
00:00:47,359 --> 00:00:51,960
understand that slide, we will have won

24
00:00:49,320 --> 00:00:54,119
the war." At the time, the US military's

25
00:00:51,960 --> 00:00:56,000
overuse of PowerPoint was an easy

26
00:00:54,119 --> 00:00:57,439
punchline for commanders when they were

27
00:00:56,000 --> 00:00:58,759
giving public speeches.

28
00:00:57,439 --> 00:01:00,879
>> It says in there that it is the

29
00:00:58,759 --> 00:01:03,000
inalienable right of every four-star

30
00:01:00,880 --> 00:01:05,239
Army general to use PowerPoint slides

31
00:01:03,000 --> 00:01:08,040
when communicating relying on only

32
00:01:05,239 --> 00:01:10,760
commander's intent, not detailed orders

33
00:01:08,040 --> 00:01:12,840
in multicolored PowerPoint format.

34
00:01:10,760 --> 00:01:14,960
Marine Corps General James Mattis was

35
00:01:12,840 --> 00:01:17,240
fond of saying, "PowerPoint makes us

36
00:01:14,959 --> 00:01:18,759
dumb." I have a friend who runs

37
00:01:17,239 --> 00:01:21,879
Microsoft, by the way, and he doesn't

38
00:01:18,760 --> 00:01:24,200
like it when I say that.

39
00:01:21,879 --> 00:01:27,359
10 years later, things hadn't improved

40
00:01:24,200 --> 00:01:29,520
much. Another Pentagon PowerPoint slide

41
00:01:27,359 --> 00:01:31,439
that showed the way different parts of

42
00:01:29,519 --> 00:01:34,759
the military were interconnected went

43
00:01:31,439 --> 00:01:36,840
viral in 2020. It also was an

44
00:01:34,760 --> 00:01:38,640
abomination of communication. But,

45
00:01:36,840 --> 00:01:41,000
there's a reason the US military kept

46
00:01:38,640 --> 00:01:42,879
spitting out these awful diagrams. They

47
00:01:41,000 --> 00:01:45,519
were in the midst of a decades-long

48
00:01:42,879 --> 00:01:47,159
battle to figure out how all the

49
00:01:45,519 --> 00:01:49,799
different parts of the armed forces

50
00:01:47,159 --> 00:01:51,799
could work together more efficiently.

51
00:01:49,799 --> 00:01:53,039
The interesting thing is, in the last 3

52
00:01:51,799 --> 00:01:55,959
years or so,

53
00:01:53,040 --> 00:01:58,400
it appears that one company has finally

54
00:01:55,959 --> 00:02:00,519
figured it all out. To maintain US and

55
00:01:58,400 --> 00:02:01,960
Western military superiority requires

56
00:02:00,519 --> 00:02:03,959
deliberate action and [music]

57
00:02:01,959 --> 00:02:05,919
technological innovation.

58
00:02:03,959 --> 00:02:07,559
At Palantir, we deliver secure,

59
00:02:05,920 --> 00:02:09,599
innovative, and scalable software

60
00:02:07,560 --> 00:02:11,280
solutions at speed.

61
00:02:09,599 --> 00:02:13,840
You might have heard of this company,

62
00:02:11,280 --> 00:02:16,080
Palantir. It seems to be the evil

63
00:02:13,840 --> 00:02:19,000
multinational corporation de jure.

64
00:02:16,080 --> 00:02:21,760
Largely secretive, Palantir specializes

65
00:02:19,000 --> 00:02:23,800
in the shadowy practice of data mining.

66
00:02:21,759 --> 00:02:26,120
For most of its existence, Palantir flew

67
00:02:23,800 --> 00:02:28,719
under the radar. It was an extremely

68
00:02:26,120 --> 00:02:31,000
obscure US defense contractor that few

69
00:02:28,719 --> 00:02:32,120
people outside the industry had ever

70
00:02:31,000 --> 00:02:33,919
heard of.

71
00:02:32,120 --> 00:02:36,000
But in the last few years, it's kind of

72
00:02:33,919 --> 00:02:39,000
been everywhere. So, what do you think

73
00:02:36,000 --> 00:02:41,680
your local Coles has in common with CIA?

74
00:02:39,000 --> 00:02:43,599
Well, they both use Palantir. Palantir

75
00:02:41,680 --> 00:02:47,000
has been the top-performing stock on the

76
00:02:43,599 --> 00:02:49,479
entire US stock market in 2025. Special

77
00:02:47,000 --> 00:02:52,159
thanks to our sponsor, Palantir. From

78
00:02:49,479 --> 00:02:54,359
the global war on terror to space domain

79
00:02:52,159 --> 00:02:56,479
awareness, Palantir continues to be a

80
00:02:54,360 --> 00:02:58,640
mission partner for the US and allied

81
00:02:56,479 --> 00:03:00,639
nations. These days, Palantir is

82
00:02:58,639 --> 00:03:02,479
involved in everything from retail to

83
00:03:00,639 --> 00:03:05,238
military operations to immigration

84
00:03:02,479 --> 00:03:06,479
enforcement. Up until recently, the CEO

85
00:03:05,239 --> 00:03:08,560
has tried to remain out of the

86
00:03:06,479 --> 00:03:11,000
spotlight, and perhaps that was for a

87
00:03:08,560 --> 00:03:12,680
good reason. Because since becoming more

88
00:03:11,000 --> 00:03:15,959
of a public figure, he's been saying

89
00:03:12,680 --> 00:03:19,777
things like this. We are at the core

90
00:03:15,959 --> 00:03:22,479
of making the obvious superiority of the

91
00:03:19,776 --> 00:03:25,479
[music] West. Our product is used on

92
00:03:22,479 --> 00:03:28,560
occasion to kill people. Okay-dokey.

93
00:03:25,479 --> 00:03:30,799
Despite its exponential growth in size

94
00:03:28,560 --> 00:03:33,360
and strategic significance, Palantir is

95
00:03:30,800 --> 00:03:35,200
still a bit of a mystery to most people.

96
00:03:33,360 --> 00:03:37,000
We're going to spend the next 2 weeks on

97
00:03:35,199 --> 00:03:39,639
this company, taking a look at

98
00:03:37,000 --> 00:03:41,840
Palantir's origin story, and how it's

99
00:03:39,639 --> 00:03:44,679
grown to be one of the biggest companies

100
00:03:41,840 --> 00:03:47,759
in the world. I'm Matt Bevan,

101
00:03:44,680 --> 00:03:47,760
and this is If You Listen.

102
00:03:49,560 --> 00:03:53,000
This story begins with a terrorist

103
00:03:51,039 --> 00:03:53,836
attack on the World Trade Center in New

104
00:03:53,000 --> 00:03:53,960
York City,

105
00:03:53,836 --> 00:03:56,799
>> [laughter]

106
00:03:53,960 --> 00:04:00,480
>> but not the one you're thinking of. This

107
00:03:56,800 --> 00:04:02,520
one happened in 1993. As New Yorkers

108
00:04:00,479 --> 00:04:05,239
came to terms with a possible act of

109
00:04:02,520 --> 00:04:06,800
terrorism on US soil, investigators

110
00:04:05,240 --> 00:04:09,240
released the first pictures of the

111
00:04:06,800 --> 00:04:11,760
explosion site, leaving little doubt

112
00:04:09,240 --> 00:04:12,959
that a powerful bomb was the most likely

113
00:04:11,759 --> 00:04:15,560
cause.

114
00:04:12,959 --> 00:04:17,920
A van packed with 550 kilos of

115
00:04:15,560 --> 00:04:19,759
explosives was driven into an

116
00:04:17,920 --> 00:04:21,920
underground car park of the World Trade

117
00:04:19,759 --> 00:04:23,480
Center. Now, the crater is 100 ft beyond

118
00:04:21,920 --> 00:04:25,720
where you're looking right now. It

119
00:04:23,480 --> 00:04:27,919
blasted a hole through three floors,

120
00:04:25,720 --> 00:04:30,040
killing six people instantly, and

121
00:04:27,918 --> 00:04:32,639
injuring hundreds more.

122
00:04:30,040 --> 00:04:35,600
For some time, it wasn't clear who was

123
00:04:32,639 --> 00:04:37,360
behind the attack. As many as 19 groups

124
00:04:35,600 --> 00:04:39,760
have claimed responsibility for the

125
00:04:37,360 --> 00:04:42,319
explosion, but the mystery of who really

126
00:04:39,759 --> 00:04:44,079
did it remains unsolved.

127
00:04:42,319 --> 00:04:46,680
One of the FBI agents involved in the

128
00:04:44,079 --> 00:04:49,159
investigation was John O'Neill. He'd

129
00:04:46,680 --> 00:04:50,959
always wanted to be in the FBI ever

130
00:04:49,160 --> 00:04:54,000
since he watched this imaginatively

131
00:04:50,959 --> 00:04:57,000
titled TV show back in the 1960s. The

132
00:04:54,000 --> 00:04:57,000
FBI

133
00:04:57,879 --> 00:05:03,079
A QM production.

134
00:05:00,399 --> 00:05:05,879
The people behind the 1993 attack were

135
00:05:03,079 --> 00:05:08,519
religious extremists from Pakistan, and

136
00:05:05,879 --> 00:05:10,360
John was instrumental in their capture.

137
00:05:08,519 --> 00:05:12,839
He became obsessed with researching

138
00:05:10,360 --> 00:05:14,600
Islamic extremism, and quickly became an

139
00:05:12,839 --> 00:05:17,759
expert on the growing threat of

140
00:05:14,600 --> 00:05:19,760
terrorism on American soil.

141
00:05:17,759 --> 00:05:21,199
As he continued rising through the ranks

142
00:05:19,759 --> 00:05:23,240
of the FBI's counterterrorism

143
00:05:21,199 --> 00:05:25,479
department, he was cognizant of the

144
00:05:23,240 --> 00:05:28,600
importance of balancing the need to keep

145
00:05:25,480 --> 00:05:30,720
people safe with the infringement on

146
00:05:28,600 --> 00:05:32,560
their personal privacy. If you have a

147
00:05:30,720 --> 00:05:34,320
lot of order, there is very little

148
00:05:32,560 --> 00:05:36,519
liberty.

149
00:05:34,319 --> 00:05:39,120
And if you have a lot of liberty,

150
00:05:36,519 --> 00:05:41,039
there tends to be less order.

151
00:05:39,120 --> 00:05:43,680
And this great experiment that we call

152
00:05:41,040 --> 00:05:46,439
the United States of America has a

153
00:05:43,680 --> 00:05:49,400
perfect blend of ordered liberty.

154
00:05:46,439 --> 00:05:51,319
Ordered liberty, a balance between

155
00:05:49,399 --> 00:05:52,679
government surveillance and individual

156
00:05:51,319 --> 00:05:54,560
freedom.

157
00:05:52,680 --> 00:05:56,680
In the US, the government surveillance

158
00:05:54,560 --> 00:05:58,319
side of the seesaw is populated by lots

159
00:05:56,680 --> 00:06:00,519
of different intelligence agencies.

160
00:05:58,319 --> 00:06:02,519
There's local police and the FBI keeping

161
00:06:00,519 --> 00:06:04,879
an eye on people domestically,

162
00:06:02,519 --> 00:06:07,479
and the CIA keeping tabs on potential

163
00:06:04,879 --> 00:06:10,560
threats abroad. But in the 1990s, that

164
00:06:07,480 --> 00:06:12,560
system had a serious flaw.

165
00:06:10,560 --> 00:06:15,240
Each of those agencies gathered their

166
00:06:12,560 --> 00:06:17,959
own data, but sharing data between them

167
00:06:15,240 --> 00:06:19,840
was complicated. To understand how bad

168
00:06:17,959 --> 00:06:22,759
the data sharing situation was, think

169
00:06:19,839 --> 00:06:25,399
about it like this. Imagine the CIA,

170
00:06:22,759 --> 00:06:27,759
FBI, and local police are represented by

171
00:06:25,399 --> 00:06:31,039
individual jars, and each piece of

172
00:06:27,759 --> 00:06:33,680
information they gather is a marble.

173
00:06:31,040 --> 00:06:35,600
Each agency handles its own marbles. If

174
00:06:33,680 --> 00:06:37,120
the FBI wants to check out the CIA's

175
00:06:35,600 --> 00:06:40,000
marbles, they have to go through a

176
00:06:37,120 --> 00:06:42,040
complicated process of requesting

177
00:06:40,000 --> 00:06:43,319
specific ones. They have to know what

178
00:06:42,040 --> 00:06:44,960
they're looking for.

179
00:06:43,319 --> 00:06:47,879
And as you can see, the opening of this

180
00:06:44,959 --> 00:06:51,000
CIA bottle is very narrow. It's not easy

181
00:06:47,879 --> 00:06:53,519
to get an individual marble out.

182
00:06:51,000 --> 00:06:55,519
As John O'Neill probed the World Trade

183
00:06:53,519 --> 00:06:58,680
Center attack of 1993, he became

184
00:06:55,519 --> 00:07:01,079
interested in one particular marble.

185
00:06:58,680 --> 00:07:03,439
One that most counterterrorism experts

186
00:07:01,079 --> 00:07:05,079
at the time weren't paying attention to.

187
00:07:03,439 --> 00:07:07,639
I think if you asked most terrorism

188
00:07:05,079 --> 00:07:10,000
experts in the mid-1990s, well, what

189
00:07:07,639 --> 00:07:12,199
about this man, bin Laden?

190
00:07:10,000 --> 00:07:15,279
Most people in the mid-1990s would have

191
00:07:12,199 --> 00:07:18,360
said, "Ah, yes, the financier, the

192
00:07:15,279 --> 00:07:20,599
terrorist financier." In 1998, ABC

193
00:07:18,360 --> 00:07:22,680
America reporter Chris Isham, a personal

194
00:07:20,600 --> 00:07:25,000
friend of John O'Neill, trekked for 10

195
00:07:22,680 --> 00:07:27,040
days across Pakistan to interview the

196
00:07:25,000 --> 00:07:30,759
leader of a little-known extremist group

197
00:07:27,040 --> 00:07:34,200
called Al-Qaeda, Osama bin Laden.

198
00:07:30,759 --> 00:07:36,399
He sat down, um, on this kind of

199
00:07:34,199 --> 00:07:38,839
bench covered in red fabric, and put a

200
00:07:36,399 --> 00:07:40,879
blanket kind of over his knee. It was

201
00:07:38,839 --> 00:07:43,039
like sitting at story time with an old

202
00:07:40,879 --> 00:07:46,365
uncle. But Chris Isham and his friend

203
00:07:43,040 --> 00:07:46,960
from the FBI

204
00:07:46,365 --> 00:07:48,960
>> [music]

205
00:07:46,959 --> 00:07:52,000
>> shared a growing interest in this old

206
00:07:48,959 --> 00:07:54,120
uncle in the Pakistani wilderness, who

207
00:07:52,000 --> 00:07:56,560
during the interview seemed pretty

208
00:07:54,120 --> 00:07:59,040
harmless. Bin Laden's handlers wouldn't

209
00:07:56,560 --> 00:08:02,040
allow anyone to translate the shake's

210
00:07:59,040 --> 00:08:04,160
answers. Miller didn't know what Bin

211
00:08:02,040 --> 00:08:07,400
Laden was saying, and the Al-Qaeda

212
00:08:04,160 --> 00:08:09,600
leader's monotonous, measured delivery

213
00:08:07,399 --> 00:08:10,879
was deceivingly calm. It was only when

214
00:08:09,600 --> 00:08:12,560
the interview was finished and the

215
00:08:10,879 --> 00:08:14,759
producers sent the tapes back to be

216
00:08:12,560 --> 00:08:17,480
translated that they learned what this

217
00:08:14,759 --> 00:08:19,480
seemingly harmless man was saying. We do

218
00:08:17,480 --> 00:08:22,040
not differentiate between those dressed

219
00:08:19,480 --> 00:08:24,600
in military uniforms and civilians.

220
00:08:22,040 --> 00:08:26,480
They are all targets. We predict a black

221
00:08:24,600 --> 00:08:29,040
day for America and the end of the

222
00:08:26,480 --> 00:08:31,560
United States. Hmm, maybe not as

223
00:08:29,040 --> 00:08:34,120
harmless as he seems.

224
00:08:31,560 --> 00:08:36,479
When he saw this interview, John O'Neill

225
00:08:34,120 --> 00:08:38,360
became fixated on Bin Laden.

226
00:08:36,479 --> 00:08:40,360
He was sure that the US was due for

227
00:08:38,360 --> 00:08:43,560
another domestic attack, and that Osama

228
00:08:40,360 --> 00:08:46,000
Bin Laden would be the person to do it.

229
00:08:43,559 --> 00:08:48,319
He tried hard to raise the alarm within

230
00:08:46,000 --> 00:08:49,639
the FBI, but it was outside of his

231
00:08:48,320 --> 00:08:51,040
remit.

232
00:08:49,639 --> 00:08:52,639
Chris Isham says that John started

233
00:08:51,039 --> 00:08:56,519
getting frustrated that he wasn't being

234
00:08:52,639 --> 00:08:59,600
taken seriously enough. He felt that the

235
00:08:56,519 --> 00:09:02,480
Saudis were definitely playing games and

236
00:08:59,600 --> 00:09:03,960
that that that the

237
00:09:02,480 --> 00:09:05,759
senior officials in the US government

238
00:09:03,960 --> 00:09:09,080
just didn't [music] get it. The problem

239
00:09:05,759 --> 00:09:11,240
was that John didn't have all the

240
00:09:09,080 --> 00:09:13,560
information. He couldn't see all the

241
00:09:11,240 --> 00:09:14,839
marbles. He could only look at the FBI's

242
00:09:13,559 --> 00:09:17,319
intelligence that they were gathering

243
00:09:14,839 --> 00:09:19,240
domestically, not the stuff the CIA was

244
00:09:17,320 --> 00:09:22,160
gathering in the Middle East.

245
00:09:19,240 --> 00:09:24,360
His frustration began to compound. John,

246
00:09:22,159 --> 00:09:26,399
because of his aggressive posture, his

247
00:09:24,360 --> 00:09:27,680
aggressive nature,

248
00:09:26,399 --> 00:09:30,159
his

249
00:09:27,679 --> 00:09:33,479
willingness to go forward

250
00:09:30,159 --> 00:09:35,879
when it may not be politically correct,

251
00:09:33,480 --> 00:09:37,399
I think a few people were just

252
00:09:35,879 --> 00:09:40,559
uncomfortable

253
00:09:37,399 --> 00:09:42,319
with John's aggressive style. In the FBI

254
00:09:40,559 --> 00:09:44,079
job which John had access to, there were

255
00:09:42,320 --> 00:09:46,960
a couple of concerning marbles. You can

256
00:09:44,080 --> 00:09:49,280
see those two little blue ones in here.

257
00:09:46,960 --> 00:09:51,360
One blue marble is that a man with known

258
00:09:49,279 --> 00:09:53,399
connections to extremist groups had

259
00:09:51,360 --> 00:09:55,360
recently been arrested by the FBI while

260
00:09:53,399 --> 00:09:57,039
attending a flight training school in

261
00:09:55,360 --> 00:09:59,399
Minnesota.

262
00:09:57,039 --> 00:10:01,759
They also had another blue marble

263
00:09:59,399 --> 00:10:04,199
that suggested a coordinated effort may

264
00:10:01,759 --> 00:10:06,639
be underway by Bin Laden to send people

265
00:10:04,200 --> 00:10:08,879
to the US [music] to obtain training

266
00:10:06,639 --> 00:10:10,679
after a number of suspicious people were

267
00:10:08,879 --> 00:10:12,399
seen attending flight schools in

268
00:10:10,679 --> 00:10:15,639
Arizona.

269
00:10:12,399 --> 00:10:19,600
Two very interesting blue marbles, but

270
00:10:15,639 --> 00:10:21,720
not enough to get the full picture.

271
00:10:19,600 --> 00:10:24,360
Unbeknownst to John, over at the CIA,

272
00:10:21,720 --> 00:10:25,920
they had more blue marbles.

273
00:10:24,360 --> 00:10:27,840
The CIA knew that a number of

274
00:10:25,919 --> 00:10:30,639
high-profile Al-Qaeda operatives were

275
00:10:27,840 --> 00:10:32,440
attending planning meetings in Malaysia

276
00:10:30,639 --> 00:10:33,919
and intended to travel to the United

277
00:10:32,440 --> 00:10:36,160
States.

278
00:10:33,919 --> 00:10:39,079
John O'Neill never saw all the marbles.

279
00:10:36,159 --> 00:10:41,279
In mid-2001, he left the FBI partially

280
00:10:39,080 --> 00:10:43,240
out of frustration and took a high-paid

281
00:10:41,279 --> 00:10:45,836
job in the private sector, head of

282
00:10:43,240 --> 00:10:46,000
security for the World Trade Center.

283
00:10:45,836 --> 00:10:49,200
>> [music]

284
00:10:46,000 --> 00:10:53,839
>> He died in the center's South Tower on

285
00:10:49,200 --> 00:10:57,080
September 11th, 2001, just 3 weeks into

286
00:10:53,839 --> 00:10:58,240
the job. The coincidence there is just

287
00:10:57,080 --> 00:10:59,120
unbelievable. You couldn't write You

288
00:10:58,240 --> 00:11:01,399
couldn't

289
00:10:59,120 --> 00:11:03,480
You know, write about it. I mean, that's

290
00:11:01,399 --> 00:11:06,519
from the guy that identified Osama bin

291
00:11:03,480 --> 00:11:08,360
Laden as a villain that he was

292
00:11:06,519 --> 00:11:09,319
and then the fact that

293
00:11:08,360 --> 00:11:11,519
that

294
00:11:09,320 --> 00:11:15,520
Osama bin Laden was able to kill him.

295
00:11:11,519 --> 00:11:15,519
It's It's just amazing.

296
00:11:15,879 --> 00:11:19,519
Now, if you were around back then,

297
00:11:18,279 --> 00:11:21,759
you'll remember that there was

298
00:11:19,519 --> 00:11:23,679
understandable concern that something

299
00:11:21,759 --> 00:11:25,439
like this might happen again.

300
00:11:23,679 --> 00:11:27,439
A lot of changes were made to try and

301
00:11:25,440 --> 00:11:28,839
make people feel safe from terror

302
00:11:27,440 --> 00:11:30,080
attacks.

303
00:11:28,839 --> 00:11:31,240
And a lot of those changes were quite

304
00:11:30,080 --> 00:11:33,879
annoying.

305
00:11:31,240 --> 00:11:35,360
The sort of sense I had was that uh the

306
00:11:33,879 --> 00:11:38,360
way we were going with just, you know,

307
00:11:35,360 --> 00:11:41,680
ridiculous airport security checks and

308
00:11:38,360 --> 00:11:43,519
uh super intrusive um surveillance all

309
00:11:41,679 --> 00:11:45,719
the time, you know, wasn't really making

310
00:11:43,519 --> 00:11:48,439
us safer. This is tech billionaire Peter

311
00:11:45,720 --> 00:11:50,360
Thiel, who has some complicated opinions

312
00:11:48,440 --> 00:11:52,000
about how the world should be run, which

313
00:11:50,360 --> 00:11:54,080
we'll get into in the next episode of

314
00:11:52,000 --> 00:11:55,120
our series. But, let's just say at this

315
00:11:54,080 --> 00:11:57,040
stage

316
00:11:55,120 --> 00:11:59,279
he wasn't a fan of the government trying

317
00:11:57,039 --> 00:12:02,120
to intrude into people's lives in the

318
00:11:59,279 --> 00:12:03,399
name of trying to stop terrorism.

319
00:12:02,120 --> 00:12:06,080
He was concerned that if there were

320
00:12:03,399 --> 00:12:07,720
another attack along the lines of 9/11,

321
00:12:06,080 --> 00:12:10,333
the government would try to become even

322
00:12:07,720 --> 00:12:12,080
more intrusive. If the World Trade

323
00:12:10,332 --> 00:12:14,480
[music] Center would erode civil

324
00:12:12,080 --> 00:12:15,560
liberties as much as it did in 2001, I

325
00:12:14,480 --> 00:12:17,159
didn't even want to think what would

326
00:12:15,559 --> 00:12:19,559
happen if you had another terrorist

327
00:12:17,159 --> 00:12:22,039
attack. And so, you have to prevent it

328
00:12:19,559 --> 00:12:24,439
to to stop to stop more erosion. He was

329
00:12:22,039 --> 00:12:26,959
more in favor of the liberty side of

330
00:12:24,440 --> 00:12:28,018
John O'Neill's ordered liberty seesaw.

331
00:12:26,960 --> 00:12:28,160
Could one

332
00:12:28,018 --> 00:12:31,480
>> [music]

333
00:12:28,159 --> 00:12:33,279
>> do something from a libertarian or civil

334
00:12:31,480 --> 00:12:35,639
liberties point of view [music] that

335
00:12:33,279 --> 00:12:37,838
would still be, you know, tough on

336
00:12:35,639 --> 00:12:40,960
terrorism? He and a friend named Alex

337
00:12:37,839 --> 00:12:42,360
Karp started focusing on the marbles.

338
00:12:40,960 --> 00:12:44,160
The government already had a lot of the

339
00:12:42,360 --> 00:12:46,360
information it needed to prevent terror

340
00:12:44,159 --> 00:12:47,668
attacks. It was just spread across all

341
00:12:46,360 --> 00:12:48,440
these different jars.

342
00:12:47,668 --> 00:12:50,080
>> [music]

343
00:12:48,440 --> 00:12:52,040
>> It wasn't using the information

344
00:12:50,080 --> 00:12:55,440
efficiently. Silicon Valley ought to be

345
00:12:52,039 --> 00:12:57,039
involved in in in fighting terrorism and

346
00:12:55,440 --> 00:12:58,839
protecting our civil liberties. Thiel

347
00:12:57,039 --> 00:13:01,279
and Karp founded a company called

348
00:12:58,839 --> 00:13:03,440
Palantir, named after Saruman's

349
00:13:01,279 --> 00:13:06,039
all-seeing glass ball from Lord of the

350
00:13:03,440 --> 00:13:07,600
Rings. A palantir is a dangerous tool,

351
00:13:06,039 --> 00:13:09,719
Saruman.

352
00:13:07,600 --> 00:13:12,279
Now, naming your nascent tech company

353
00:13:09,720 --> 00:13:13,839
after a tool used by the very bad guy

354
00:13:12,279 --> 00:13:15,679
trying to take over Middle-earth is an

355
00:13:13,839 --> 00:13:18,000
interesting decision, but we'll

356
00:13:15,679 --> 00:13:19,799
again, get to that a little bit later.

357
00:13:18,000 --> 00:13:21,919
As far as Alex Karp and Peter Thiel were

358
00:13:19,799 --> 00:13:24,599
concerned, the company was designed to

359
00:13:21,919 --> 00:13:27,199
break open all the jars of marbles,

360
00:13:24,600 --> 00:13:29,159
spill them out onto on surface, and make

361
00:13:27,200 --> 00:13:31,920
sense of whatever came out. It would

362
00:13:29,159 --> 00:13:33,399
allow humans to find needles in

363
00:13:31,919 --> 00:13:34,839
haystacks. So, make the data

364
00:13:33,399 --> 00:13:35,483
intelligible to you and me, which it's

365
00:13:34,840 --> 00:13:35,600
not.

366
00:13:35,484 --> 00:13:37,040
>> [music]

367
00:13:35,600 --> 00:13:39,320
>> And by doing that, it would allow them

368
00:13:37,039 --> 00:13:40,068
to find bad people trying to destroy our

369
00:13:39,320 --> 00:13:40,200
society

370
00:13:40,068 --> 00:13:42,399
>> [music]

371
00:13:40,200 --> 00:13:43,720
>> and could be used also to protect civil

372
00:13:42,399 --> 00:13:46,240
liberties by making the data sets

373
00:13:43,720 --> 00:13:48,560
transparent. Palantir's first investor

374
00:13:46,240 --> 00:13:50,519
was the CIA.

375
00:13:48,559 --> 00:13:52,919
It's been widely reported, though never

376
00:13:50,519 --> 00:13:55,399
officially confirmed, that in 2011

377
00:13:52,919 --> 00:13:57,519
Palantir played a key role in processing

378
00:13:55,399 --> 00:14:00,039
information which led to the killing of

379
00:13:57,519 --> 00:14:01,028
Osama bin Laden in his secret compound

380
00:14:00,039 --> 00:14:01,159
in Pakistan.

381
00:14:01,028 --> 00:14:03,799
>> [music]

382
00:14:01,159 --> 00:14:06,919
>> Palantir, uh, for example, effectively

383
00:14:03,799 --> 00:14:08,233
vectored Osama bin Laden's location. So,

384
00:14:06,919 --> 00:14:08,319
How exactly

385
00:14:08,234 --> 00:14:11,040
>> [music]

386
00:14:08,320 --> 00:14:12,600
>> does it work?

387
00:14:11,039 --> 00:14:15,719
For the first decade of Palantir's

388
00:14:12,600 --> 00:14:17,840
existence, basically nobody knew who

389
00:14:15,720 --> 00:14:19,840
they were or what they did. They didn't

390
00:14:17,840 --> 00:14:22,399
even start posting on social media until

391
00:14:19,840 --> 00:14:24,960
2015. A person can [music] see a pattern

392
00:14:22,399 --> 00:14:27,120
in a hundred things, but it's very hard

393
00:14:24,960 --> 00:14:29,280
for them to see a pattern in a million.

394
00:14:27,120 --> 00:14:31,240
The company embedded technical experts

395
00:14:29,279 --> 00:14:33,559
inside classified US intelligence

396
00:14:31,240 --> 00:14:35,600
operations and facilities, trying to

397
00:14:33,559 --> 00:14:38,199
access as many different information

398
00:14:35,600 --> 00:14:40,879
sources as possible and figure out how

399
00:14:38,200 --> 00:14:42,600
they relate to each other. Sticking with

400
00:14:40,879 --> 00:14:45,799
the [snorts] marble metaphor, it worked

401
00:14:42,600 --> 00:14:47,960
like this. Think of this plastic box as

402
00:14:45,799 --> 00:14:50,159
Palantir.

403
00:14:47,960 --> 00:14:52,280
Rather than everything being siloed in

404
00:14:50,159 --> 00:14:54,838
individual jars, Palantir had the

405
00:14:52,279 --> 00:14:57,480
capacity to empty all the marbles out

406
00:14:54,839 --> 00:14:57,480
like this.

407
00:14:57,919 --> 00:15:03,719
And then sort the marbles more logically

408
00:15:00,879 --> 00:15:06,639
and identify patterns

409
00:15:03,720 --> 00:15:09,800
and connections. So, it meant that the

410
00:15:06,639 --> 00:15:12,559
CIA marbles

411
00:15:09,799 --> 00:15:13,519
could be connected with the FBI marbles

412
00:15:12,559 --> 00:15:16,959
here.

413
00:15:13,519 --> 00:15:20,199
And it meant that the NSA marbles

414
00:15:16,960 --> 00:15:20,200
could be mixed in as well.

415
00:15:20,639 --> 00:15:25,439
Sorted, organized. Gee, they have a lot

416
00:15:23,039 --> 00:15:27,879
of marbles. Like really what it is is

417
00:15:25,440 --> 00:15:29,880
creating a unified view

418
00:15:27,879 --> 00:15:31,039
over lots of disparate data sources

419
00:15:29,879 --> 00:15:32,879
which don't otherwise make sense

420
00:15:31,039 --> 00:15:35,319
together. The reason they had so little

421
00:15:32,879 --> 00:15:37,960
public presence is that they really only

422
00:15:35,320 --> 00:15:39,560
had one customer, the US government.

423
00:15:37,960 --> 00:15:41,280
>> Fighting ISIS, stopping human

424
00:15:39,559 --> 00:15:43,159
trafficking, supporting money laundering

425
00:15:41,279 --> 00:15:45,919
investigations. Working with the special

426
00:15:43,159 --> 00:15:47,879
forces, law enforcement, the DOJ.

427
00:15:45,919 --> 00:15:50,279
According to their marketing, Palantir

428
00:15:47,879 --> 00:15:53,080
was all about organizing information in

429
00:15:50,279 --> 00:15:55,439
a way that humans would find more

430
00:15:53,080 --> 00:15:57,480
intuitive. Humans are incredible at

431
00:15:55,440 --> 00:15:59,000
insight. It's just surfacing the right

432
00:15:57,480 --> 00:16:01,039
information for them to see. They were

433
00:15:59,000 --> 00:16:03,279
effectively creating an enormous

434
00:16:01,039 --> 00:16:05,519
self-updating PowerPoint slide which

435
00:16:03,279 --> 00:16:06,919
could be used by humans

436
00:16:05,519 --> 00:16:09,000
to

437
00:16:06,919 --> 00:16:10,919
find patterns,

438
00:16:09,000 --> 00:16:13,480
find a pattern there,

439
00:16:10,919 --> 00:16:14,919
circle that, connect that,

440
00:16:13,480 --> 00:16:16,080
and kind of figure out what was really

441
00:16:14,919 --> 00:16:18,120
going on.

442
00:16:16,080 --> 00:16:19,879
Using Palantir software was like looking

443
00:16:18,120 --> 00:16:22,360
down on the world from above with

444
00:16:19,879 --> 00:16:25,120
information displayed on a large flat

445
00:16:22,360 --> 00:16:26,600
map that you could interact with. An ICE

446
00:16:25,120 --> 00:16:28,440
official said it's basically like a

447
00:16:26,600 --> 00:16:30,240
Google's map interface where you can

448
00:16:28,440 --> 00:16:32,400
look around the United States, you can

449
00:16:30,240 --> 00:16:34,519
zoom in on targets. You then click on an

450
00:16:32,399 --> 00:16:36,120
individual person and it brings up their

451
00:16:34,519 --> 00:16:37,519
name, a photo.

452
00:16:36,120 --> 00:16:39,159
>> We'll get into how it's being used for

453
00:16:37,519 --> 00:16:40,279
immigration enforcement in our next

454
00:16:39,159 --> 00:16:43,679
episode, but looking at it

455
00:16:40,279 --> 00:16:45,399
dispassionately, it's clearly impressive

456
00:16:43,679 --> 00:16:47,079
software.

457
00:16:45,399 --> 00:16:48,439
Way better than the PowerPoint

458
00:16:47,080 --> 00:16:49,400
nightmares they used to have to deal

459
00:16:48,440 --> 00:16:51,240
with.

460
00:16:49,399 --> 00:16:54,519
But programming these maps which

461
00:16:51,240 --> 00:16:57,639
Palantir calls the ontology was slow,

462
00:16:54,519 --> 00:17:00,879
painstaking work. It involved writing

463
00:16:57,639 --> 00:17:03,399
enormous amounts of code in incredibly

464
00:17:00,879 --> 00:17:05,599
complicated software.

465
00:17:03,399 --> 00:17:07,279
And talking to a lot of different people

466
00:17:05,599 --> 00:17:09,838
so that they didn't overemphasize the

467
00:17:07,279 --> 00:17:12,279
importance of one particular data source

468
00:17:09,838 --> 00:17:14,399
or ignore another one.

469
00:17:12,279 --> 00:17:17,199
It also didn't entirely solve the marble

470
00:17:14,400 --> 00:17:19,600
jar problem because each industry needed

471
00:17:17,199 --> 00:17:21,240
information laid out in different ways,

472
00:17:19,599 --> 00:17:23,679
the systems didn't talk to each other

473
00:17:21,240 --> 00:17:25,519
perfectly. For instance, Gotham,

474
00:17:23,680 --> 00:17:27,439
Palantir's military and law enforcement

475
00:17:25,519 --> 00:17:28,960
software, was different from its

476
00:17:27,439 --> 00:17:30,320
logistics software, which was called

477
00:17:28,960 --> 00:17:32,319
Foundry.

478
00:17:30,319 --> 00:17:34,839
But then something came along that put

479
00:17:32,319 --> 00:17:36,960
Palantir's marble sorting tech into

480
00:17:34,839 --> 00:17:40,079
hyperdrive. Let's start with this

481
00:17:36,960 --> 00:17:41,920
language AI chat GPT.

482
00:17:40,079 --> 00:17:44,319
What is it?

483
00:17:41,920 --> 00:17:45,640
Great question. In 2022, people were

484
00:17:44,319 --> 00:17:48,599
starting to get excited about the

485
00:17:45,640 --> 00:17:51,040
potential uses of large language models

486
00:17:48,599 --> 00:17:53,759
like chat GPT. It can create custom

487
00:17:51,039 --> 00:17:55,119
code. It can create entire books that

488
00:17:53,759 --> 00:17:57,559
are written from scratch that are not

489
00:17:55,119 --> 00:18:00,000
plagiarized. It can write essays. Wow,

490
00:17:57,559 --> 00:18:01,119
that really is amazing. But at Palantir,

491
00:18:00,000 --> 00:18:02,519
they were quickly coming to a

492
00:18:01,119 --> 00:18:05,119
realization.

493
00:18:02,519 --> 00:18:07,359
They'd been building software that was

494
00:18:05,119 --> 00:18:09,000
to be used by humans.

495
00:18:07,359 --> 00:18:11,599
But the way they'd built it was also

496
00:18:09,000 --> 00:18:13,240
perfect for large language models. We

497
00:18:11,599 --> 00:18:15,439
were pleasantly surprised to see how

498
00:18:13,240 --> 00:18:17,839
much the world we had been building for

499
00:18:15,440 --> 00:18:20,000
met its moment with LLMs. Palantir's

500
00:18:17,839 --> 00:18:23,159
chief technology officer, Shyam Sankar,

501
00:18:20,000 --> 00:18:24,799
said that LLMs, large language models,

502
00:18:23,160 --> 00:18:25,880
weren't just great at reading their data

503
00:18:24,799 --> 00:18:28,039
maps,

504
00:18:25,880 --> 00:18:29,720
but needed their data maps in order to

505
00:18:28,039 --> 00:18:31,960
be reliable. It's like, "Wow, you

506
00:18:29,720 --> 00:18:34,240
actually cannot unleash the value of an

507
00:18:31,960 --> 00:18:37,319
LLM without these things." They had,

508
00:18:34,240 --> 00:18:39,400
purely by luck, spent 20 years building

509
00:18:37,319 --> 00:18:41,960
a system that organized data in a way

510
00:18:39,400 --> 00:18:44,440
that could be intuitively read by the

511
00:18:41,960 --> 00:18:47,120
new wave of AI systems.

512
00:18:44,440 --> 00:18:49,720
And it was laid out in an ideal way for

513
00:18:47,119 --> 00:18:52,000
the AI system to produce reliable

514
00:18:49,720 --> 00:18:54,319
information. By January, it was quite

515
00:18:52,000 --> 00:18:56,640
obvious that we needed to tear up all

516
00:18:54,319 --> 00:18:58,480
our road maps and get excited about how

517
00:18:56,640 --> 00:19:00,360
we could incorporate LLMs into our

518
00:18:58,480 --> 00:19:02,160
software to provide a whole new series

519
00:19:00,359 --> 00:19:05,079
of experiences. They developed something

520
00:19:02,160 --> 00:19:07,279
called the Palantir AIP, which stands

521
00:19:05,079 --> 00:19:08,599
for artificial intelligence platform.

522
00:19:07,279 --> 00:19:10,799
AIP

523
00:19:08,599 --> 00:19:13,359
is your

524
00:19:10,799 --> 00:19:15,519
AI operating system. Now, individual

525
00:19:13,359 --> 00:19:18,359
users didn't need to look at the marble

526
00:19:15,519 --> 00:19:20,839
map at all. You can just use AIP, which

527
00:19:18,359 --> 00:19:23,639
sits on top of the marble map and looks

528
00:19:20,839 --> 00:19:24,720
at it for you, far faster than any human

529
00:19:23,640 --> 00:19:26,680
could.

530
00:19:24,720 --> 00:19:28,519
We start with a military operator

531
00:19:26,680 --> 00:19:29,720
responsible for monitoring activity

532
00:19:28,519 --> 00:19:31,920
within [music] Eastern Europe.

533
00:19:29,720 --> 00:19:32,640
>> To use it, all you have to do is ask it

534
00:19:31,920 --> 00:19:34,519
a question.

535
00:19:32,640 --> 00:19:36,840
>> What enemy units are in the region?

536
00:19:34,519 --> 00:19:39,240
>> The human operator can ask AIP in plain

537
00:19:36,839 --> 00:19:41,799
English to deploy surveillance drones.

538
00:19:39,240 --> 00:19:43,799
>> Task the MQ-9 to capture video of this

539
00:19:41,799 --> 00:19:46,399
location. The drone footage shows an

540
00:19:43,799 --> 00:19:48,039
enemy T-80 main battle tank. Then the

541
00:19:46,400 --> 00:19:49,960
operator can ask it to look at what

542
00:19:48,039 --> 00:19:52,039
options are available in the area to

543
00:19:49,960 --> 00:19:54,160
destroy that tank. Generate three

544
00:19:52,039 --> 00:19:56,079
courses of action to target this enemy

545
00:19:54,160 --> 00:19:58,120
equipment. Send these three options to

546
00:19:56,079 --> 00:20:00,439
my commander for review. Our commander

547
00:19:58,119 --> 00:20:03,000
selects a course of action. Approve

548
00:20:00,440 --> 00:20:05,279
course of action three.

549
00:20:03,000 --> 00:20:07,160
Until pretty recently, executing this

550
00:20:05,279 --> 00:20:09,559
kind of operation would have required

551
00:20:07,160 --> 00:20:11,400
big maps being rolled out on tables and

552
00:20:09,559 --> 00:20:13,879
little wooden figures pushed around on

553
00:20:11,400 --> 00:20:16,320
top, phone calls right up and down the

554
00:20:13,880 --> 00:20:18,680
chain of command. No man makes a perfect

555
00:20:16,319 --> 00:20:21,599
plan, Keith. AIP allows it all to be

556
00:20:18,680 --> 00:20:23,680
done as easily as asking ChatGPT to plan

557
00:20:21,599 --> 00:20:26,000
your holiday to Queensland. Of course,

558
00:20:23,680 --> 00:20:27,920
Matt. I can help you plan your holiday

559
00:20:26,000 --> 00:20:30,640
to Queensland. You're going to have such

560
00:20:27,920 --> 00:20:32,724
a great time. Not only can AI be used to

561
00:20:30,640 --> 00:20:32,840
operate the Palantir software,

562
00:20:32,723 --> 00:20:35,079
>> [music]

563
00:20:32,839 --> 00:20:37,039
>> it can also be used to design it. What

564
00:20:35,079 --> 00:20:39,000
once took software engineers years of

565
00:20:37,039 --> 00:20:41,240
observing operations and writing code

566
00:20:39,000 --> 00:20:44,039
now takes weeks.

567
00:20:41,240 --> 00:20:46,960
This rapid increase in efficiency meant

568
00:20:44,039 --> 00:20:49,559
that suddenly Palantir had a lot of

569
00:20:46,960 --> 00:20:52,559
extra time and resources on its hands.

570
00:20:49,559 --> 00:20:54,399
So, it started expanding. [music]

571
00:20:52,559 --> 00:20:55,879
Previously, pretty much only the US

572
00:20:54,400 --> 00:20:57,120
government could afford Palantir's

573
00:20:55,880 --> 00:21:00,040
services. [music]

574
00:20:57,119 --> 00:21:02,079
Now, it's getting into everything.

575
00:21:00,039 --> 00:21:03,839
ExxonMobil is adding powerful new

576
00:21:02,079 --> 00:21:04,720
features by partnering with Palantir

577
00:21:03,839 --> 00:21:06,128
Technologies.

578
00:21:04,720 --> 00:21:08,759
>> Today, we're in the world of customer

579
00:21:06,128 --> 00:21:09,480
[music] service focusing on AIP customer

580
00:21:08,759 --> 00:21:11,200
service engine.

581
00:21:09,480 --> 00:21:13,720
>> So, what do you think your local Coles

582
00:21:11,200 --> 00:21:16,000
has in common with the CIA?

583
00:21:13,720 --> 00:21:17,680
Well, they both use Palantir. With

584
00:21:16,000 --> 00:21:19,920
Palantir as the platform for making

585
00:21:17,680 --> 00:21:21,560
cricket the number one team bat and ball

586
00:21:19,920 --> 00:21:23,160
sport in [music] Europe. I have to say

587
00:21:21,559 --> 00:21:25,240
the European cricket network has got to

588
00:21:23,160 --> 00:21:27,080
be the most out of pocket of Palantir's

589
00:21:25,240 --> 00:21:28,400
clients. With our entrepreneurial

590
00:21:27,079 --> 00:21:30,480
intuition, we can now use [music]

591
00:21:28,400 --> 00:21:32,680
Palantir to ask the right questions of

592
00:21:30,480 --> 00:21:35,519
the data all in one place. Well, if they

593
00:21:32,680 --> 00:21:37,039
can sort out the FBI and CIA's marbles,

594
00:21:35,519 --> 00:21:38,759
they can probably handle the European

595
00:21:37,039 --> 00:21:41,920
cricket networks.

596
00:21:38,759 --> 00:21:43,879
This rapid expansion in clients made

597
00:21:41,920 --> 00:21:46,120
markets extremely excited about

598
00:21:43,880 --> 00:21:48,280
Palantir's potential future profits.

599
00:21:46,119 --> 00:21:50,359
>> stocks have moved pretty ferociously

600
00:21:48,279 --> 00:21:53,944
here um in the last couple of months. In

601
00:21:50,359 --> 00:21:58,399
2023, Palantir stock jumped 167%

602
00:21:53,944 --> 00:22:00,559
[music] in 2024, it jumped by 340%.

603
00:21:58,400 --> 00:22:02,460
I think there's a golden path right now

604
00:22:00,559 --> 00:22:02,519
for them to monetize

605
00:22:02,460 --> 00:22:03,640
>> [music]

606
00:22:02,519 --> 00:22:06,079
>> what we view as potentially a

607
00:22:03,640 --> 00:22:07,880
trillion-dollar market opportunity. It's

608
00:22:06,079 --> 00:22:10,240
become one of the fastest-growing

609
00:22:07,880 --> 00:22:11,800
companies on Earth with clients all over

610
00:22:10,240 --> 00:22:13,400
the world, including government

611
00:22:11,799 --> 00:22:14,680
contracts in lots of different

612
00:22:13,400 --> 00:22:16,519
countries.

613
00:22:14,680 --> 00:22:18,640
And it's basically just a glorified

614
00:22:16,519 --> 00:22:20,839
version of that PowerPoint slide General

615
00:22:18,640 --> 00:22:22,680
McChrystal was considering during the

616
00:22:20,839 --> 00:22:24,639
war in Afghanistan.

617
00:22:22,680 --> 00:22:26,960
It was designed to help keep the balance

618
00:22:24,640 --> 00:22:29,600
that FBI agent John O'Neill thought was

619
00:22:26,960 --> 00:22:30,840
so important. And this great experiment

620
00:22:29,599 --> 00:22:32,000
that we call the United States of

621
00:22:30,839 --> 00:22:35,159
America

622
00:22:32,000 --> 00:22:36,519
has a perfect blend of ordered liberty.

623
00:22:35,160 --> 00:22:38,920
He wanted to make sure that the seesaw

624
00:22:36,519 --> 00:22:41,039
didn't go too far one way or the other.

625
00:22:38,920 --> 00:22:43,279
If you have a lot of order, there is

626
00:22:41,039 --> 00:22:45,480
very little liberty.

627
00:22:43,279 --> 00:22:47,759
And if you have a lot of liberty,

628
00:22:45,480 --> 00:22:48,920
there tends to be less order. But a lot

629
00:22:47,759 --> 00:22:51,640
of people are starting to think that

630
00:22:48,920 --> 00:22:55,120
Palantir is tipping the balance towards

631
00:22:51,640 --> 00:22:58,080
order and away from liberty. Ladies and

632
00:22:55,119 --> 00:22:59,759
gentlemen, this is scary. This is us

633
00:22:58,079 --> 00:23:01,599
moving towards what China is doing to

634
00:22:59,759 --> 00:23:04,000
their citizens. People are really quite

635
00:23:01,599 --> 00:23:05,000
scared of this company. The Palantir

636
00:23:04,000 --> 00:23:07,640
surveillance state.

637
00:23:05,000 --> 00:23:09,559
>> Palantir is a definite CIA front

638
00:23:07,640 --> 00:23:11,520
company. There are more conspiracy

639
00:23:09,559 --> 00:23:13,679
theories swirling around about Palantir

640
00:23:11,519 --> 00:23:15,720
than any other Silicon Valley company.

641
00:23:13,680 --> 00:23:17,960
>> Peter Thiel and his company Palantir are

642
00:23:15,720 --> 00:23:20,240
some of the most evil forces on our

643
00:23:17,960 --> 00:23:20,759
planet conspiring against the American

644
00:23:20,240 --> 00:23:23,640
people.

645
00:23:20,759 --> 00:23:25,720
>> Palantir is like a surveillance and

646
00:23:23,640 --> 00:23:29,080
predictive policing firm.

647
00:23:25,720 --> 00:23:32,039
>> Palantir, an organization that is run

648
00:23:29,079 --> 00:23:33,919
also by a Zionist psychopath. And that's

649
00:23:32,039 --> 00:23:36,799
because even though it really is just a

650
00:23:33,920 --> 00:23:38,800
very fancy PowerPoint slide which makes

651
00:23:36,799 --> 00:23:40,639
intelligence, military, logistics,

652
00:23:38,799 --> 00:23:43,319
customer service, and cricketing

653
00:23:40,640 --> 00:23:46,520
operations more efficient, its leaders

654
00:23:43,319 --> 00:23:48,720
have decided to create a kind of weird

655
00:23:46,519 --> 00:23:50,480
public image. We have dedicated our

656
00:23:48,720 --> 00:23:51,720
company to the service of the West and

657
00:23:50,480 --> 00:23:54,000
the United States of [music] America.

658
00:23:51,720 --> 00:23:56,519
Palantir is here

659
00:23:54,000 --> 00:23:59,680
to disrupt and when it's necessary to

660
00:23:56,519 --> 00:24:02,359
scare enemies and on occasion kill them.

661
00:23:59,680 --> 00:24:03,640
And we hope you're in favor of that. So,

662
00:24:02,359 --> 00:24:06,559
I think it's important to spend some

663
00:24:03,640 --> 00:24:09,960
time talking about this guy, Alex Karp,

664
00:24:06,559 --> 00:24:12,079
and what he wants to do with his company

665
00:24:09,960 --> 00:24:13,319
that now controls so much of the world's

666
00:24:12,079 --> 00:24:14,960
data.

667
00:24:13,319 --> 00:24:16,480
But we're not going to do it today.

668
00:24:14,960 --> 00:24:18,960
We're going to spend our whole next

669
00:24:16,480 --> 00:24:22,480
episode on it. And we hope you're in

670
00:24:18,960 --> 00:24:22,480
favor of that. We'll see you then.

671
00:24:35,730 --> 00:24:37,751
>> [music]