1
00:00:00,160 --> 00:00:05,160
Speaker 1: So picture this. You are driving down a highway, okay,

2
00:00:05,360 --> 00:00:08,640
and it's the highway of human history. For the last oh,

3
00:00:08,720 --> 00:00:11,279
I don't know, maybe three hundred years or so, the

4
00:00:11,400 --> 00:00:13,800
road has been pretty predictable.

5
00:00:13,320 --> 00:00:15,759
Speaker 2: Right, mostly a straight shot, exactly.

6
00:00:15,800 --> 00:00:18,719
Speaker 1: It generally goes uphill, you know. Sometimes it's a steep climb,

7
00:00:18,760 --> 00:00:22,760
like during a major war or a plague, and sometimes

8
00:00:22,800 --> 00:00:27,039
it's a gentle, pleasant slope. But the direction has always

9
00:00:27,079 --> 00:00:27,600
been the same.

10
00:00:27,640 --> 00:00:30,000
Speaker 2: It's progress, better medicine, faster travel.

11
00:00:29,800 --> 00:00:32,880
Speaker 1: Right, more information. We are very used to looking at

12
00:00:32,920 --> 00:00:35,159
a graph that just goes up into the right. It's

13
00:00:35,640 --> 00:00:36,799
it's comfortable, it is.

14
00:00:37,240 --> 00:00:38,960
Speaker 2: It's kind of the narrative we've all been sold, right.

15
00:00:39,200 --> 00:00:41,759
We just assume that tomorrow will be slightly better than today,

16
00:00:42,399 --> 00:00:44,640
just like today is slightly better than yesterday. It's a

17
00:00:44,679 --> 00:00:45,359
linear way of.

18
00:00:45,320 --> 00:00:48,560
Speaker 1: Thinking, yes, exactly. But what if I told you that the.

19
00:00:48,560 --> 00:00:51,719
Speaker 2: Road is ending, well, I'd say you're trying to scare me.

20
00:00:52,000 --> 00:00:53,840
Speaker 1: Right, and not that the world is ending. Don't panic yet,

21
00:00:53,840 --> 00:00:55,799
But the road as we know it is ending. We

22
00:00:55,880 --> 00:00:59,039
are on a slope anymore. We are standing right now,

23
00:00:59,079 --> 00:01:03,320
at this exact moment in history, at a massive, terrifying

24
00:01:03,719 --> 00:01:06,159
three way fork in the road, and we have to

25
00:01:06,239 --> 00:01:09,359
choose which path to take, essentially yesterday, and.

26
00:01:09,319 --> 00:01:12,480
Speaker 2: Just to raise the stakes a little bit more here,

27
00:01:13,079 --> 00:01:15,680
this isn't a metaphorical fork where if we take a

28
00:01:15,719 --> 00:01:17,280
wrong turn we can just pull a U turn and

29
00:01:17,400 --> 00:01:20,879
backtrack right. No reverse gear exactly. This is an inflection point.

30
00:01:21,079 --> 00:01:24,200
Once we commit to one of these paths, the momentum

31
00:01:24,280 --> 00:01:27,519
of the technology, specifically artificial intelligence is going to lock

32
00:01:27,640 --> 00:01:27,920
us in.

33
00:01:28,079 --> 00:01:28,920
Speaker 1: It's like gravity.

34
00:01:29,280 --> 00:01:32,640
Speaker 2: It is once you cross that event horizon, there is

35
00:01:32,680 --> 00:01:33,480
no coming back.

36
00:01:33,640 --> 00:01:36,200
Speaker 1: And that is the premise we are unpacking today. It

37
00:01:36,239 --> 00:01:38,200
is honestly, it's a little hair raising. It's one of

38
00:01:38,200 --> 00:01:40,000
those topics that makes you want to stare at the

39
00:01:40,000 --> 00:01:41,239
ceiling at three am.

40
00:01:41,359 --> 00:01:44,719
Speaker 2: It definitely is, but it's also hopeful if you know

41
00:01:44,760 --> 00:01:46,760
where to look. It's not just doom and gloom.

42
00:01:46,840 --> 00:01:50,079
Speaker 1: Welcome to thrilling Threads. We take a stack of complex

43
00:01:50,120 --> 00:01:54,159
information articles, research talks, we shake it out and we

44
00:01:54,239 --> 00:01:57,359
find the threads that actually matter to your life. Today.

45
00:01:57,599 --> 00:02:00,519
We are looking at the future of humanity the lens

46
00:02:00,560 --> 00:02:03,719
of one specific very high stakes conversation.

47
00:02:04,120 --> 00:02:06,840
Speaker 2: Yeah, we're looking at a talk given by Alvin W. Graylan.

48
00:02:07,319 --> 00:02:11,000
He's recently interviewed by Manus Soamrodi for the ted YouTube channel,

49
00:02:11,520 --> 00:02:14,759
and the title of this talk is three possible futures

50
00:02:14,800 --> 00:02:16,919
for AI, Which will we choose?

51
00:02:17,199 --> 00:02:19,400
Speaker 1: And let me tell you the options he lays out

52
00:02:19,680 --> 00:02:23,719
They range from absolutely terrifying to a utopian.

53
00:02:23,240 --> 00:02:25,039
Speaker 2: Dream with a very little middle ground.

54
00:02:25,159 --> 00:02:27,759
Speaker 1: Yeah, there is no business's usual options.

55
00:02:27,759 --> 00:02:29,919
Speaker 2: It's the nature of the technology we're dealing with. It's

56
00:02:29,919 --> 00:02:32,719
an amplifier. Yeah. It amplifies our best traits and our

57
00:02:32,759 --> 00:02:35,479
worst ones. It just doesn't do mild.

58
00:02:35,800 --> 00:02:38,639
Speaker 1: So let's set the stage a bit. Who is Alvin Graylan?

59
00:02:39,039 --> 00:02:41,199
Because you know, if my neighbor tells me the world

60
00:02:41,280 --> 00:02:44,199
is ending, I might just not ignore him. But this

61
00:02:44,319 --> 00:02:46,840
guy seems to have the credentials to back up the

62
00:02:47,000 --> 00:02:49,560
doom saying why should we listen to him?

63
00:02:49,680 --> 00:02:52,479
Speaker 2: He certainly does. Graylan isn't just a pundit or you know,

64
00:02:52,680 --> 00:02:55,360
a tech journalist looking for clicks. He has been the

65
00:02:55,360 --> 00:02:56,840
trenches for thirty five years.

66
00:02:56,879 --> 00:02:58,360
Speaker 1: Wow, thirty five years.

67
00:02:58,759 --> 00:03:03,599
Speaker 2: Yeah, He's worked in AI, cybersecurity, virtual reality, semiconductors, the

68
00:03:03,639 --> 00:03:07,319
whole gamut. Currently he's doing work at Stanford. But there

69
00:03:07,400 --> 00:03:10,719
is one thing on his resume that makes his perspective

70
00:03:10,919 --> 00:03:16,080
incredibly rare and arguably more valuable than your typical Silicon

71
00:03:16,159 --> 00:03:20,000
valley CEO. What's that He is bicultural? Okay, he's a

72
00:03:20,080 --> 00:03:22,840
US citizen educated here, but he has spent half of

73
00:03:22,879 --> 00:03:26,840
his career working in China. Specifically, he was the China

74
00:03:26,879 --> 00:03:30,439
president for HTC, dealing with VR and AI at a

75
00:03:30,560 --> 00:03:31,280
very high level.

76
00:03:31,360 --> 00:03:33,719
Speaker 1: Ah. Okay, that is a big deal because usually when

77
00:03:33,719 --> 00:03:36,360
we talk about AI, we're hearing it from a purely

78
00:03:36,439 --> 00:03:39,639
Western Silicon valley bubble exactly, or on the flip side,

79
00:03:39,680 --> 00:03:42,000
we hear the perspective from Beijing. It's usually one or

80
00:03:42,000 --> 00:03:43,639
the other. We don't get the crossover right.

81
00:03:43,680 --> 00:03:46,080
Speaker 2: And Grayland has his feet in both worlds. He understands

82
00:03:46,120 --> 00:03:48,719
the technological landscape of the US, the venture capital, the

83
00:03:48,759 --> 00:03:52,439
fast innovation, and he understands the strategic landscape of China,

84
00:03:52,560 --> 00:03:55,439
the government planning, the massive resource allocation.

85
00:03:55,759 --> 00:03:58,120
Speaker 1: So when he talks about the race between these two powers,

86
00:03:58,319 --> 00:03:59,840
he isn't guessing he knows the.

87
00:03:59,800 --> 00:04:05,520
Speaker 2: P He really does. And minusche Zommarodi, who is interviewing him,

88
00:04:05,759 --> 00:04:08,520
sets this up perfectly in the intro. She points out

89
00:04:08,520 --> 00:04:12,319
that for normal people, you know, the non engineers among

90
00:04:12,400 --> 00:04:15,000
us the people just trying to pay their bills. We

91
00:04:15,080 --> 00:04:16,759
are in this state of total confusion.

92
00:04:16,839 --> 00:04:17,600
Speaker 1: Oh absolutely.

93
00:04:17,639 --> 00:04:19,800
Speaker 2: We don't know if AI is just hype like those

94
00:04:19,920 --> 00:04:23,319
NFTs everyone bought and forgot about, or if it's literally

95
00:04:23,360 --> 00:04:25,279
transforming the fabric of reality.

96
00:04:25,399 --> 00:04:29,399
Speaker 1: And Graylan's response to that ambivalence is basically rake up, yep.

97
00:04:29,680 --> 00:04:32,279
He says, there is a lot of misinformation out there,

98
00:04:32,399 --> 00:04:36,000
and frankly, his assessment is going to scare people. He

99
00:04:36,079 --> 00:04:38,279
warns the audience right off the bat that his view

100
00:04:38,319 --> 00:04:40,600
is very different from the Silicon Valley consensus.

101
00:04:40,680 --> 00:04:43,399
Speaker 2: I love that phrase, Silicon Valley consensus. It sound like

102
00:04:43,399 --> 00:04:45,800
a hive mind, like the borg or something.

103
00:04:45,959 --> 00:04:47,959
Speaker 1: In many ways it operates like one. There is a

104
00:04:48,120 --> 00:04:51,519
very specific narrative that big tech companies push because it

105
00:04:51,560 --> 00:04:55,680
benefits their stock price, it benefits their regulatory environment, and

106
00:04:55,800 --> 00:04:59,319
Graylan is here to disrupt that narrative. He says, we're

107
00:04:59,360 --> 00:05:03,759
currently heading toward two very dark futures, but a third

108
00:05:04,199 --> 00:05:08,240
beautiful future is possible, the Star Trek option, if and

109
00:05:08,480 --> 00:05:10,399
only if we act right now.

110
00:05:10,560 --> 00:05:14,639
Speaker 2: Okay, let's unpack these three futures. Because he uses movie analogies,

111
00:05:14,680 --> 00:05:17,920
which speaks my language perfectly. It really helps to visualize

112
00:05:17,920 --> 00:05:19,319
the abstract economics.

113
00:05:19,360 --> 00:05:22,600
Speaker 1: It's a great shorthand so future number one. He calls

114
00:05:22,639 --> 00:05:24,040
this the Elysium scenario.

115
00:05:24,319 --> 00:05:26,199
Speaker 2: Right, have you seen the movie years.

116
00:05:26,000 --> 00:05:29,279
Speaker 1: Ago Matt Damon, Jody Foster Big space station.

117
00:05:29,560 --> 00:05:32,040
Speaker 2: Yes, that's the one. The premise of Elysium is that

118
00:05:32,120 --> 00:05:35,519
the ultra wealthy have physically left Earth. They live on

119
00:05:35,560 --> 00:05:38,959
this pristine, high tech space station orbiting the planet.

120
00:05:39,240 --> 00:05:41,279
Speaker 1: Sounds nice for them, right, They have machines.

121
00:05:41,000 --> 00:05:44,879
Speaker 2: That cure cancer instantly. They have total abundance, its paradise. Meanwhile,

122
00:05:45,079 --> 00:05:47,160
everyone else, the ninety nine point nine percent, is stuck

123
00:05:47,199 --> 00:05:51,600
on a ruined Earth, fighting for scraps, living in complete squalor.

124
00:05:51,399 --> 00:05:53,480
Speaker 1: And Graylan thinks we are heading there. That seems a

125
00:05:53,480 --> 00:05:55,680
bit dramatic. Are we really going to build a space

126
00:05:55,720 --> 00:05:56,879
station for billionaires?

127
00:05:57,079 --> 00:05:59,279
Speaker 2: Well, let's strip away the sci fi element of the

128
00:05:59,319 --> 00:06:03,759
literal space focus on the economics. Graalan argues that this

129
00:06:03,839 --> 00:06:05,959
is one of the default paths we are on right now.

130
00:06:06,160 --> 00:06:09,560
How so, think about the mechanism here. We have massive

131
00:06:09,639 --> 00:06:15,600
tech labs, the big players we all know, Google, Open Ai, Microsoft, Meta.

132
00:06:16,120 --> 00:06:19,399
They are accumulating resources at an unprecedented rate.

133
00:06:19,600 --> 00:06:22,519
Speaker 1: And when you say resources, you don't just mean cash, no, I.

134
00:06:22,480 --> 00:06:26,959
Speaker 2: Mean the raw materials of intelligence, compute power, massive data centers,

135
00:06:27,160 --> 00:06:30,839
the most talented engineers in the entire world. Is a

136
00:06:30,839 --> 00:06:32,560
compounding advantage, right.

137
00:06:32,399 --> 00:06:34,680
Speaker 1: Because the better your AI is, the more money you make,

138
00:06:34,680 --> 00:06:36,879
which allows you to buy more GPUs, which makes your

139
00:06:36,920 --> 00:06:39,720
AI even better. It's a flywheel exactly.

140
00:06:39,839 --> 00:06:43,399
Speaker 2: And Graalan argues that if this goes unchecked, these entities

141
00:06:43,519 --> 00:06:46,079
essentially take control of the government. Wow, they become more

142
00:06:46,160 --> 00:06:48,759
powerful than nation states. Think about it. If a company

143
00:06:48,759 --> 00:06:52,000
controls the intelligence that runs the grid, the banking system,

144
00:06:52,040 --> 00:06:55,360
and the military logistics, who is really in charge, the

145
00:06:55,360 --> 00:06:56,800
president or the CEO.

146
00:06:57,040 --> 00:06:59,199
Speaker 1: I mean, some would argue, we're already seeing the prequel

147
00:06:59,240 --> 00:07:01,160
to that movie. The law being power is immense.

148
00:07:01,480 --> 00:07:04,600
Speaker 2: We are The result of this path is the creation

149
00:07:04,759 --> 00:07:09,600
of a class of trillionaires, not billionaires, trillionaires millionaires. And

150
00:07:09,639 --> 00:07:13,279
then there's everyone else, a permanent underclass of serfs who

151
00:07:13,319 --> 00:07:16,759
have no economic utility because AI does everything better than

152
00:07:16,759 --> 00:07:17,120
they do.

153
00:07:17,560 --> 00:07:21,160
Speaker 1: That's the key phrase, right there, no economic utility because

154
00:07:21,199 --> 00:07:23,639
in the past, even if the boss was incredibly rich

155
00:07:23,680 --> 00:07:26,879
and terrible, he still needed workers to build the cars

156
00:07:26,959 --> 00:07:27,920
or flip the burgers.

157
00:07:27,959 --> 00:07:29,480
Speaker 2: Maybe labor exactly.

158
00:07:29,800 --> 00:07:33,120
Speaker 1: But in the Elysium scenario, the AI creates all the

159
00:07:33,199 --> 00:07:34,160
value precisely.

160
00:07:34,560 --> 00:07:37,000
Speaker 2: If I have a robot army that can build houses,

161
00:07:37,040 --> 00:07:39,639
cook food, and write software, I don't need you. I

162
00:07:39,639 --> 00:07:42,600
don't need your labor at all. So in the Elysium scenario,

163
00:07:42,920 --> 00:07:46,199
the AI works. It works great. It just doesn't work

164
00:07:46,240 --> 00:07:48,279
for us. It works for the people who own the servers.

165
00:07:48,360 --> 00:07:51,680
Speaker 1: It's a future of extreme inequality cemented by technology. It's

166
00:07:51,720 --> 00:07:53,000
feudalism with Wi Fi.

167
00:07:53,240 --> 00:07:55,759
Speaker 2: That is a terrifyingly accurate way to put it.

168
00:07:55,839 --> 00:07:57,800
Speaker 1: Okay, well, that's depressing. I definitely don't want to be

169
00:07:57,839 --> 00:08:00,600
a serf. Let's see what's behind door number two. Maybe

170
00:08:00,639 --> 00:08:01,079
it's better.

171
00:08:01,160 --> 00:08:02,519
Speaker 2: Door number two is mad Max.

172
00:08:02,839 --> 00:08:07,360
Speaker 1: Oh, fantastic leather jackets, desert wasteland and people strike to

173
00:08:07,360 --> 00:08:08,560
the front of cars.

174
00:08:08,319 --> 00:08:13,480
Speaker 2: And chaos, complete chaos. If Elysium is about control, mad

175
00:08:13,519 --> 00:08:18,319
Max is about collapse. Grayland describes an escalation ladder that

176
00:08:18,360 --> 00:08:20,160
we are currently climbing right now.

177
00:08:20,199 --> 00:08:21,279
Speaker 1: What does that ladder look like?

178
00:08:21,439 --> 00:08:23,040
Speaker 2: It starts with an AI race.

179
00:08:23,000 --> 00:08:25,720
Speaker 1: Which we are definitely in. Every headline is about who

180
00:08:25,720 --> 00:08:28,879
has the fastest model, who has the most parameters.

181
00:08:28,360 --> 00:08:30,399
Speaker 2: Right, But then it moves to an AI war. This

182
00:08:30,439 --> 00:08:33,200
doesn't mean terminators marching down the street yet. It means

183
00:08:33,200 --> 00:08:35,960
cyber warfare, digital conflict.

184
00:08:35,559 --> 00:08:37,679
Speaker 1: Like taking down infrastructure exactly.

185
00:08:37,799 --> 00:08:42,039
Speaker 2: Imagine hacking power grids, scrambling financial data, manipulating stock markets

186
00:08:42,039 --> 00:08:44,120
at the speed of light. If you can use a

187
00:08:44,159 --> 00:08:47,600
super intelligent agent to find zero day vulnerabilities in your

188
00:08:47,679 --> 00:08:50,840
enemy's defense systems, you can cripple them without firing a

189
00:08:50,879 --> 00:08:51,440
single shot.

190
00:08:51,559 --> 00:08:53,720
Speaker 1: Okay, so that's really bad, But how do we get

191
00:08:53,759 --> 00:08:55,480
to mad max from a cyber attack?

192
00:08:55,840 --> 00:08:59,600
Speaker 2: From there, it moves to kinetic war, actual shooting, bombing,

193
00:08:59,639 --> 00:09:03,480
physical drones. If a nation feels this digital infrastructure is

194
00:09:03,519 --> 00:09:07,519
being destroyed, they will respond with physical force, and the

195
00:09:07,559 --> 00:09:11,600
final rung of that ladder, according to Graylan, is potential

196
00:09:11,759 --> 00:09:12,519
nuclear war.

197
00:09:12,919 --> 00:09:16,440
Speaker 1: That seems like a massive jump from chatbots to nukes.

198
00:09:17,039 --> 00:09:19,759
Why would AI trigger a nuclear war?

199
00:09:20,120 --> 00:09:23,519
Speaker 2: It's about the destabilization. It's the Thucidides trap, right. If

200
00:09:23,639 --> 00:09:27,440
nations feel that losing the AI race poses an existential

201
00:09:27,480 --> 00:09:30,000
threat to their sovereignty, that if they come in second,

202
00:09:30,240 --> 00:09:32,960
they've basically become a vassal state to the winner. They

203
00:09:33,000 --> 00:09:35,919
might lash out a cornered animal exactly if you think

204
00:09:35,960 --> 00:09:38,200
your opponent is about to turn on a superintelligence that

205
00:09:38,240 --> 00:09:42,120
will render your military obsolete instantly. You might strike first, Well,

206
00:09:42,159 --> 00:09:42,879
you still have a chance.

207
00:09:42,919 --> 00:09:45,360
Speaker 1: So it's a user or lose it mentality exactly.

208
00:09:45,639 --> 00:09:48,840
Speaker 2: And here's the part that gave me absolute chills. Graylan

209
00:09:48,919 --> 00:09:52,120
mentions that he has talked to people in Washington, DC, policymakers,

210
00:09:52,200 --> 00:09:54,679
real insiders, who view this conflict as.

211
00:09:54,519 --> 00:09:57,399
Speaker 1: Inevitable, inevitable. They think we have to fight this war.

212
00:09:57,879 --> 00:10:00,559
Speaker 2: Yes, they aren't trying to stop the car, they're just

213
00:10:00,600 --> 00:10:04,000
bracing for impact. They believe that conflict with China over

214
00:10:04,039 --> 00:10:09,120
this technology is destiny and that fatalism is incredibly dangerous.

215
00:10:09,240 --> 00:10:12,440
Speaker 1: Because if you believe war is inevitable, you stop looking

216
00:10:12,480 --> 00:10:13,000
for off.

217
00:10:13,000 --> 00:10:16,679
Speaker 2: Ramps, right you start stockpiling ammunition instead of sending diplomats.

218
00:10:16,799 --> 00:10:18,960
Speaker 1: So Graylan says, we are currently heading toward one of

219
00:10:19,000 --> 00:10:21,960
these two Elysium or mad Max.

220
00:10:22,120 --> 00:10:26,000
Speaker 2: Yes, he believes the current trajectory, the momentum of the

221
00:10:26,039 --> 00:10:29,639
system as it stands today, the incentives of capitalism and

222
00:10:29,679 --> 00:10:33,200
the incentives of geopolitics, leads directly to one of these

223
00:10:33,200 --> 00:10:36,440
two outcomes. Either the witch get everything and leave us behind,

224
00:10:36,759 --> 00:10:38,759
or we blow each other up fighting over who gets

225
00:10:38,840 --> 00:10:39,559
the technology.

226
00:10:39,639 --> 00:10:42,519
Speaker 1: Well, I'm thoroughly bummed out, but we have to keep going.

227
00:10:42,679 --> 00:10:44,759
Please tell me about the Star Trek option before I

228
00:10:44,799 --> 00:10:45,399
lose all hope.

229
00:10:45,519 --> 00:10:48,759
Speaker 2: Entirely, there is hope. Graylan is actually an optimist at heart.

230
00:10:48,960 --> 00:10:51,320
He sees a third path. He calls it the Star

231
00:10:51,360 --> 00:10:51,879
Trek option.

232
00:10:52,000 --> 00:10:53,840
Speaker 1: Ee me up? What does this look like?

233
00:10:54,159 --> 00:10:58,639
Speaker 2: He's the Vulcan analogy? So in Star Trek lore, humanity

234
00:10:58,720 --> 00:11:02,919
was kind of a mess until the Vulcans, a peaceful, rational,

235
00:11:03,320 --> 00:11:07,279
highly advanced species, showed up and introduced us to advanced technology.

236
00:11:07,320 --> 00:11:08,919
They basically saved us from ourselves.

237
00:11:09,039 --> 00:11:12,120
Speaker 1: But aliens aren't coming to save us unless Grayland knows

238
00:11:12,120 --> 00:11:12,799
something we don't.

239
00:11:12,960 --> 00:11:16,639
Speaker 2: No, no aliens. In this analogy, AI is the Vulcan.

240
00:11:17,320 --> 00:11:20,600
If we develop it correctly, AI becomes that rational force.

241
00:11:21,080 --> 00:11:24,480
Think about what a superintelligence could actually do if it

242
00:11:24,519 --> 00:11:27,240
wasn't being used for war or ad targeting.

243
00:11:27,279 --> 00:11:28,120
Speaker 1: Okay, what could it do?

244
00:11:28,360 --> 00:11:31,519
Speaker 2: It could solve the protein folding problem, completely curing cancer

245
00:11:31,559 --> 00:11:34,879
and Alzheimer's. It could optimize fusion reactor designs in seconds,

246
00:11:34,879 --> 00:11:38,200
giving us infinite clean energy. Wow. It could solve logistics

247
00:11:38,200 --> 00:11:41,919
for global food distribution, ending hunger overnight. It unleashes a

248
00:11:41,960 --> 00:11:45,159
century or even millennia of discovery compressed into a decade.

249
00:11:45,240 --> 00:11:47,639
Speaker 1: It sounds utopian, almost too good to be true.

250
00:11:47,799 --> 00:11:50,360
Speaker 2: It is utopian, But the key difference between Star Trek

251
00:11:50,360 --> 00:11:51,840
and Elysium is distribution.

252
00:11:52,000 --> 00:11:52,559
Speaker 1: Distribution.

253
00:11:52,960 --> 00:11:57,080
Speaker 2: Yes, in Star Trek the technology is shared. It is

254
00:11:57,200 --> 00:12:00,159
used for the good of humanity as a species, not

255
00:12:00,240 --> 00:12:02,960
for the profit of a single corporation or the dominance

256
00:12:02,960 --> 00:12:03,679
of a single nation.

257
00:12:03,960 --> 00:12:06,840
Speaker 1: So the technology is exactly the same in Elysium and

258
00:12:06,919 --> 00:12:10,759
Star Trek. The AI is just as powerful in both scenarios.

259
00:12:11,039 --> 00:12:12,759
The difference is simply who owns it.

260
00:12:12,960 --> 00:12:16,039
Speaker 2: That is the crucial insight. It's not a technological problem

261
00:12:16,039 --> 00:12:20,320
we are facing. It's a political and social choice. In Elysium,

262
00:12:20,440 --> 00:12:23,720
the tech is hoarded. In Star Trek, it's a public good.

263
00:12:24,159 --> 00:12:27,000
Speaker 1: Okay, So we know the stakes, we know the destination

264
00:12:27,080 --> 00:12:29,919
we want. We all want Star Trek, nobody wants mad Max.

265
00:12:30,440 --> 00:12:32,919
But Grellan says, there is a massive myth standing in

266
00:12:32,960 --> 00:12:35,480
our way, a narrative that is actively blocking us from

267
00:12:35,519 --> 00:12:38,120
taking the exit ramp to the Star Trek future.

268
00:12:38,320 --> 00:12:41,320
Speaker 2: Yes, and this brings us back to that Silicon Valley consensus.

269
00:12:41,320 --> 00:12:44,720
We mentioned earlier. Minucheto Marati brings up the standard story

270
00:12:44,759 --> 00:12:47,080
told by leaders like Sam Olman. You hear this in

271
00:12:47,120 --> 00:12:48,200
practically every interview.

272
00:12:48,279 --> 00:12:49,080
Speaker 1: How does the story go.

273
00:12:49,279 --> 00:12:51,279
Speaker 2: It goes like this, we have to lock this technology down.

274
00:12:51,279 --> 00:12:53,240
We have to develop it as fast as humanly possible.

275
00:12:53,360 --> 00:12:56,080
We can't pause, we can't regulate, we can't share, because

276
00:12:56,120 --> 00:12:57,960
if we don't do it first, China will I.

277
00:12:57,879 --> 00:12:59,960
Speaker 1: Hear this all the time. It's the ultimate trump card.

278
00:13:00,120 --> 00:13:03,159
You can't regulate us or the bad guys win. It

279
00:13:03,200 --> 00:13:06,320
shuts down every single argument about safety.

280
00:13:06,519 --> 00:13:09,440
Speaker 2: Graylan calls this one of the biggest myths out there.

281
00:13:09,600 --> 00:13:11,840
Speaker 1: Really, but isn't China trying to beat us?

282
00:13:11,919 --> 00:13:14,360
Speaker 2: He calls it one of the scariest myths. And remember

283
00:13:14,399 --> 00:13:16,919
he was in China two days before giving this talk.

284
00:13:17,159 --> 00:13:19,879
He knows what's happening on the ground there. He argues

285
00:13:19,919 --> 00:13:22,559
that the AI industry here in the West is using

286
00:13:22,639 --> 00:13:26,799
a century old playbook, the military industrial complex playbook.

287
00:13:26,799 --> 00:13:29,679
Speaker 1: Oh interesting, how does that work in the context of AI.

288
00:13:30,039 --> 00:13:33,399
Speaker 2: It's a five step process. Step one, create an enemy.

289
00:13:33,519 --> 00:13:35,840
You need a boogeyman to justify what you're about to do.

290
00:13:35,879 --> 00:13:37,759
Speaker 1: And China fits that mold perfectly.

291
00:13:37,919 --> 00:13:41,879
Speaker 2: Right. Step two, use that enemy to get massive government funding.

292
00:13:42,120 --> 00:13:44,440
You go to Congress and say we need billions in

293
00:13:44,480 --> 00:13:47,200
subsidies for chips and data centers to beat China.

294
00:13:47,000 --> 00:13:47,759
Speaker 1: Which is happening.

295
00:13:48,000 --> 00:13:51,360
Speaker 2: Step three, get public support. Make the public afraid so

296
00:13:51,399 --> 00:13:55,879
they cheer for the funding. Step four, get deregulation. You say,

297
00:13:55,879 --> 00:13:59,200
we can't have safety checks, we can't have ethical oversight boards.

298
00:13:59,399 --> 00:13:59,960
We're in a wall.

299
00:14:00,519 --> 00:14:03,759
Speaker 1: Speed is safety, Move fast and break things, but on

300
00:14:03,799 --> 00:14:05,039
a geopolitical scale.

301
00:14:05,080 --> 00:14:09,799
Speaker 2: And step five move fast and make unimaginable amounts of money.

302
00:14:10,000 --> 00:14:14,440
Speaker 1: That sounds disturbingly accurate. It completely frames safety as a weakness.

303
00:14:15,159 --> 00:14:17,480
Speaker 2: Graylan says the labs aren't trying to save the world

304
00:14:17,679 --> 00:14:20,919
or protect democracy. Those are the marketing slogans. They are

305
00:14:20,960 --> 00:14:23,440
trying to create trillions of dollars in value.

306
00:14:23,600 --> 00:14:24,200
Speaker 1: Trillions.

307
00:14:24,440 --> 00:14:26,840
Speaker 2: Yes, And this is where he drops a really heavy

308
00:14:26,879 --> 00:14:31,600
reality check regarding AGI, artificial general intelligence. He cites Sam

309
00:14:31,679 --> 00:14:34,960
Altman's own definition of AGI. Do you know what it is?

310
00:14:35,399 --> 00:14:39,399
Speaker 1: I've heard a few definitions, usually something vague like smarter

311
00:14:39,480 --> 00:14:41,600
than a human at most economic tasks.

312
00:14:41,799 --> 00:14:45,120
Speaker 2: The specific definition grayl in sites is technology that can

313
00:14:45,159 --> 00:14:46,559
replace the average worker.

314
00:14:47,720 --> 00:14:49,279
Speaker 1: When you put it that way, it doesn't sound like

315
00:14:49,279 --> 00:14:52,000
a helpful tool. It sounds like a replacement part exactly.

316
00:14:52,279 --> 00:14:54,840
Speaker 2: Graylan points out that the stated goal, the thing they

317
00:14:54,840 --> 00:14:58,440
are spending billions of dollars to build, is literally technology

318
00:14:58,440 --> 00:15:01,279
that takes jobs. Now he add some nuances, as this

319
00:15:01,320 --> 00:15:04,480
could be amazing if robots do all the tedious work.

320
00:15:04,720 --> 00:15:07,600
We can paint, write poetry, hang out at ted talks

321
00:15:07,600 --> 00:15:09,159
all day. That's the dream.

322
00:15:09,399 --> 00:15:12,799
Speaker 1: That's the Star Trek dream. Work is for machines. Life

323
00:15:12,840 --> 00:15:13,840
is for people.

324
00:15:13,679 --> 00:15:16,840
Speaker 2: But currently there is no plan for the people who

325
00:15:16,840 --> 00:15:20,480
get displaced. The plan is just to build the replacement engine.

326
00:15:20,600 --> 00:15:22,840
The labs are racing to build the thing that makes

327
00:15:22,919 --> 00:15:25,879
human labor obsolete, using the China threat as a smoke

328
00:15:25,879 --> 00:15:28,840
screen to avoid regulation and garner subsidies.

329
00:15:28,919 --> 00:15:31,120
Speaker 1: It's a classic bait and switch. Look over there at

330
00:15:31,120 --> 00:15:33,919
the scary foreign power while we dismantle the labor market

331
00:15:34,000 --> 00:15:34,559
here at home.

332
00:15:34,759 --> 00:15:37,639
Speaker 2: That is the essence of his argument, and that is

333
00:15:37,720 --> 00:15:40,720
exactly why he believes we are heading toward Elysium or

334
00:15:40,759 --> 00:15:44,440
mad Max, because the current incentives are all wrong. The

335
00:15:44,519 --> 00:15:48,080
incentives are for speed and hoarding, not for saty and sharing.

336
00:15:48,320 --> 00:15:50,799
Speaker 1: But I have to play devil's advocate here. Even if

337
00:15:50,879 --> 00:15:54,120
the US labs are totally cynical, isn't the threat real?

338
00:15:54,639 --> 00:15:57,720
If the US stops, China will keep building? And do

339
00:15:57,759 --> 00:16:01,440
we really want an authoritarian regime to have the superintelligence first?

340
00:16:01,679 --> 00:16:05,360
Speaker 2: That's the prisoner's dilemma, and Grayland addresses this directly. He

341
00:16:05,440 --> 00:16:07,879
says that the labs in China are also competing with

342
00:16:07,919 --> 00:16:11,120
each other. They aren't this unified monolith of evil. They

343
00:16:11,120 --> 00:16:15,200
are fragmented. And more importantly, the Chinese leadership is terrified

344
00:16:15,360 --> 00:16:22,919
the exact same things we are are uncontrollable AI, social instability, chaos.

345
00:16:22,600 --> 00:16:25,240
Speaker 1: So they have an incentive to cooperate on safety because

346
00:16:25,279 --> 00:16:27,519
a rogue AI is bad for the CCP too.

347
00:16:27,759 --> 00:16:31,120
Speaker 2: Exactly. A rogue AI doesn't care about borders or political parties.

348
00:16:31,320 --> 00:16:33,279
It's a threat to everyone equally.

349
00:16:33,440 --> 00:16:35,960
Speaker 1: Okay, So Grayland isn't just throwing rocks at the windows

350
00:16:35,960 --> 00:16:38,960
of open AI. He actually submitted a policy paper to Stanford.

351
00:16:39,320 --> 00:16:42,240
He has a plan, a three part plan to steer

352
00:16:42,279 --> 00:16:43,480
the ship towards star Trek.

353
00:16:43,639 --> 00:16:45,840
Speaker 2: Yes, and this isn't just a let's all be nicer

354
00:16:45,879 --> 00:16:48,840
to each other plea. These are structural, massive changes.

355
00:16:49,000 --> 00:16:51,919
Speaker 1: Let's break them down. Pillar number one, Pillar one the

356
00:16:51,960 --> 00:16:55,559
sern of AI, like the Particle Collider in Europe.

357
00:16:55,039 --> 00:16:59,480
Speaker 2: Exactly, or the International Space Station or EIDER for fusion energy.

358
00:17:00,600 --> 00:17:04,240
Graylan argues that right now we have hundreds of labs competing.

359
00:17:04,680 --> 00:17:08,480
They are hoarding chips, hoarding talent, and duplicating the exact

360
00:17:08,519 --> 00:17:13,079
same work over and over again. It's inefficient and it's dangerous.

361
00:17:12,920 --> 00:17:15,440
Speaker 1: Because if everyone is racing, no one is checking.

362
00:17:15,160 --> 00:17:18,559
Speaker 2: The brakes precisely. His solution is to create a single

363
00:17:18,880 --> 00:17:24,799
global lab, aggregate the top talent from the US, China, Europe, everywhere,

364
00:17:24,880 --> 00:17:25,880
put them in one place.

365
00:17:26,119 --> 00:17:28,880
Speaker 1: But is that realistic? Getting the US and China to

366
00:17:28,920 --> 00:17:31,880
hold hands and build an AI brain together, Google and

367
00:17:31,920 --> 00:17:34,240
Meta giving up their proprietary secrets.

368
00:17:34,359 --> 00:17:36,559
Speaker 2: It sounds impossible, right, but look at the history. We

369
00:17:36,599 --> 00:17:39,319
did it with the Space station, Russian and American astronauts

370
00:17:39,359 --> 00:17:41,680
living together while their countries were at odds. We did

371
00:17:41,680 --> 00:17:44,079
it with nuclear fusion research. We did it with CERN.

372
00:17:44,359 --> 00:17:46,799
We didn't have a US physics team fighting a Chinese

373
00:17:46,799 --> 00:17:49,400
physics team to find the Higgs boson. We work together

374
00:17:49,400 --> 00:17:51,839
because the science was simply too big and too expensive

375
00:17:51,839 --> 00:17:52,440
for one nation.

376
00:17:52,599 --> 00:17:54,039
Speaker 1: And what does that achieve for AI?

377
00:17:54,400 --> 00:17:56,920
Speaker 2: It creates open science. Whenever comes out of that lab

378
00:17:56,960 --> 00:18:00,440
belongs to humanity. It's shared with the world, not hoarded

379
00:18:00,440 --> 00:18:05,400
by single corporation. This democratizes the power. It completely removes

380
00:18:05,400 --> 00:18:08,519
the arms race dynamic because everyone gets the breakthrough at

381
00:18:08,519 --> 00:18:09,359
the exact same time.

382
00:18:09,480 --> 00:18:12,200
Speaker 1: So you commoditize the superintelligence. You make it a public

383
00:18:12,319 --> 00:18:14,640
utility like GPS exactly.

384
00:18:14,599 --> 00:18:17,279
Speaker 2: And then companies can build apps and services on top

385
00:18:17,319 --> 00:18:19,559
of it, but no one controls the core engine.

386
00:18:19,680 --> 00:18:24,599
Speaker 1: Okay, so one giant lab for mankind. What's pillar number two?

387
00:18:24,839 --> 00:18:28,279
Speaker 2: Pillar two is the global Data Commons. This addresses the

388
00:18:28,279 --> 00:18:29,119
issue of bias.

389
00:18:29,559 --> 00:18:32,720
Speaker 1: I've heard a lot about sovereign AI lately, countries saying

390
00:18:32,720 --> 00:18:34,920
they want their own AI trained on their own data.

391
00:18:35,119 --> 00:18:38,000
France wants a French AI, India wants an Indian AI.

392
00:18:38,200 --> 00:18:40,559
Speaker 2: Right, it sounds patriotic, It sounds like a good idea

393
00:18:40,559 --> 00:18:43,880
to preserve culture. But Grayland says, the research actually shows

394
00:18:43,880 --> 00:18:47,319
the opposite. How So, the research shows that less data

395
00:18:47,359 --> 00:18:50,039
equals more bias. If you restrict the data set to

396
00:18:50,079 --> 00:18:53,480
just one culture or one viewpoint, the AI becomes skewed.

397
00:18:53,519 --> 00:18:55,920
It basically becomes a caricature that makes sense.

398
00:18:56,000 --> 00:18:58,839
Speaker 1: It's living in an echo chamber. If you only train

399
00:18:58,880 --> 00:19:02,039
an AI on Western Internet data, it's going to have

400
00:19:02,160 --> 00:19:04,240
extreme Western biases exactly.

401
00:19:04,720 --> 00:19:09,920
Speaker 2: Graylan's solution is to aggregate everyone's data, all history, all languages,

402
00:19:10,000 --> 00:19:13,000
all cultures. Put it all into the pot, the ultimate

403
00:19:13,039 --> 00:19:16,079
melting pot. By doing this, you create an AI that

404
00:19:16,160 --> 00:19:19,799
can find an optimal path for everyone. It balances needs,

405
00:19:20,200 --> 00:19:23,160
It understands the full spectrum of human experience, not just

406
00:19:23,200 --> 00:19:26,599
the Western perspective or the Eastern perspective. It reduces the

407
00:19:26,680 --> 00:19:29,160
US versus them mentality in the machine itself.

408
00:19:29,279 --> 00:19:32,039
Speaker 1: So Pillar one is shared hardware and talent. Pillar two

409
00:19:32,119 --> 00:19:35,119
is shared data. That leaves Pillar three. And this one

410
00:19:35,119 --> 00:19:37,440
seems like it hits closest to home for the listener.

411
00:19:37,480 --> 00:19:38,359
This is about the money.

412
00:19:38,519 --> 00:19:41,000
Speaker 2: Pillar three is the GI bill for the AI age.

413
00:19:41,279 --> 00:19:43,640
Speaker 1: I love a good historical analogy. Break this down for US.

414
00:19:43,720 --> 00:19:46,240
Speaker 2: So nineteen forty four nineteen forty five, World War II

415
00:19:46,319 --> 00:19:49,640
is ending. You have about fifteen million American service members

416
00:19:49,640 --> 00:19:52,680
coming home and they need jobs, right, and the economy

417
00:19:52,720 --> 00:19:56,519
is terrified. They're expecting a massive depression because suddenly you

418
00:19:56,559 --> 00:20:00,200
have fifteen million unemployed young men flooding the market all

419
00:20:00,240 --> 00:20:02,440
at once. It was an employment shock.

420
00:20:02,599 --> 00:20:04,480
Speaker 1: So the goverment stepped in big time.

421
00:20:04,960 --> 00:20:09,000
Speaker 2: They created the GI Bill, free education, zero interest loans,

422
00:20:09,400 --> 00:20:12,519
free medical care, help buying homes, and it worked. It

423
00:20:12,559 --> 00:20:16,160
didn't just avoid a depression, It created the American middle class.

424
00:20:16,599 --> 00:20:20,079
It sparked the greatest economic boon in history. It turned

425
00:20:20,079 --> 00:20:23,200
a potential massive crisis into a golden age.

426
00:20:23,319 --> 00:20:25,279
Speaker 1: So Graylan is saying we need to do that again,

427
00:20:25,440 --> 00:20:26,720
but on a much.

428
00:20:26,559 --> 00:20:29,640
Speaker 2: Much larger scale. He points out that the shock coming

429
00:20:29,640 --> 00:20:33,279
from AI won't be fifteen million people in the US alone.

430
00:20:33,319 --> 00:20:36,920
It could be one hundred and fifty million globally billions one.

431
00:20:36,880 --> 00:20:38,920
Speaker 1: Hundred and fifty million. That's almost half the population.

432
00:20:39,119 --> 00:20:41,400
Speaker 2: We are talking about displacement on a level we have

433
00:20:41,480 --> 00:20:44,680
never seen in human history. Gral In cites predictions of

434
00:20:44,720 --> 00:20:47,400
one hundred plus million people affected just in the US.

435
00:20:48,039 --> 00:20:50,640
If we don't have a safety net of that magnitude,

436
00:20:50,680 --> 00:20:55,720
free education, reskilling, maybe universal basic income, though he focuses

437
00:20:55,720 --> 00:20:57,880
more on the safety net aspect. If we don't have that.

438
00:20:57,960 --> 00:21:01,359
Speaker 1: We get the mad Max scenario. It's in the streets, we.

439
00:21:01,319 --> 00:21:04,759
Speaker 2: Get social collapse. He puts it very bluntly. If we

440
00:21:04,799 --> 00:21:06,880
don't this world is not going to be a very

441
00:21:06,920 --> 00:21:08,960
good place for us to hang out in. You cannot

442
00:21:08,960 --> 00:21:11,319
have a stable society if fifty percent of the population

443
00:21:11,400 --> 00:21:13,359
has zero income and no hope.

444
00:21:13,440 --> 00:21:15,519
Speaker 1: That is a polite way of saying guillotines.

445
00:21:15,880 --> 00:21:20,039
Speaker 2: It really is. So. The three pillars are centralized open

446
00:21:20,079 --> 00:21:23,400
science to stop the arms race, global data commons to

447
00:21:23,400 --> 00:21:26,480
stop the bias, and a massive social safety net to

448
00:21:26,480 --> 00:21:27,279
stop the revolution.

449
00:21:27,480 --> 00:21:30,480
Speaker 1: It sounds like a solid plan. It sounds entirely rational.

450
00:21:31,400 --> 00:21:34,839
But and there is always a butt. The obstacles seem enormous,

451
00:21:34,960 --> 00:21:36,680
the biggest one being human nature.

452
00:21:36,759 --> 00:21:39,839
Speaker 2: That's the hurdle. Convincing power is to actually cooperate.

453
00:21:39,960 --> 00:21:42,319
Speaker 1: Right, Why would the US share its tech with China?

454
00:21:42,359 --> 00:21:44,799
Why would Google share its secrets with the world. We

455
00:21:44,839 --> 00:21:45,920
are tribal creatures.

456
00:21:46,160 --> 00:21:48,599
Speaker 2: Graylan argues, we need a massive shift in mindset. We

457
00:21:48,599 --> 00:21:50,519
need to understand enlightened self interest.

458
00:21:50,599 --> 00:21:52,640
Speaker 1: Define that for us in this context.

459
00:21:52,240 --> 00:21:54,759
Speaker 2: It means realizing that the world is not zero sum.

460
00:21:55,119 --> 00:21:57,599
That's the old thinking. If I wind, you lose. In

461
00:21:57,640 --> 00:22:00,079
the age of AI, if I win, you don't have

462
00:22:00,079 --> 00:22:02,599
have to lose. In fact, if I help you win,

463
00:22:02,720 --> 00:22:06,359
I might be safer, like the peace dividend exactly. He

464
00:22:06,440 --> 00:22:10,039
asks a haunting question, why send children ten thousand miles

465
00:22:10,079 --> 00:22:14,160
to kill other children over resources, you know, oil, land, water,

466
00:22:14,599 --> 00:22:18,960
when this technology could provide infinite resources locally. If AI

467
00:22:19,079 --> 00:22:23,079
can solve energy with fusion and food production with synthetic biology,

468
00:22:23,680 --> 00:22:26,759
the reason for war completely evaporates.

469
00:22:26,839 --> 00:22:29,000
Speaker 1: That's the star Trek logic. We don't fight over food

470
00:22:29,000 --> 00:22:31,279
because we have replicators. We don't fight over oil because

471
00:22:31,319 --> 00:22:32,680
we have dilyceum crystals.

472
00:22:32,759 --> 00:22:36,079
Speaker 2: Right, but we have to survive the transition, and the

473
00:22:36,160 --> 00:22:39,400
speed of this transition is the other major obstacle.

474
00:22:39,519 --> 00:22:40,599
Speaker 1: How fast are we talking?

475
00:22:40,680 --> 00:22:44,960
Speaker 2: Think about the Industrial revolution steam engines, electricity. That change

476
00:22:44,960 --> 00:22:49,400
took forty sixty eighty years to ripple through society. Generations

477
00:22:49,440 --> 00:22:52,400
had time to adapt. People move from farms to factories,

478
00:22:52,440 --> 00:22:53,519
but it took a lifetime.

479
00:22:53,559 --> 00:22:56,079
Speaker 1: My grandfather was a farmer, My father worked in a factory.

480
00:22:56,440 --> 00:22:57,359
That kind of pace.

481
00:22:57,640 --> 00:23:00,519
Speaker 2: Grayland says, the AI revolution will happen in five to

482
00:23:00,559 --> 00:23:03,279
ten years, maybe shorter five years.

483
00:23:03,319 --> 00:23:04,799
Speaker 1: That's not enough time to retrain.

484
00:23:05,319 --> 00:23:08,680
Speaker 2: Society is simply not equipped to move at that speed

485
00:23:09,160 --> 00:23:11,519
without intervention. We can't just wait and see. By the

486
00:23:11,519 --> 00:23:14,079
time we see it'll be over, the cement will be dry.

487
00:23:14,200 --> 00:23:16,119
Speaker 1: This is why he says, we are at the inflection

488
00:23:16,200 --> 00:23:19,119
point right now. The window is closing. Yes, okay, so

489
00:23:19,200 --> 00:23:21,920
we've covered the scary parts, the hopeful parts, and the

490
00:23:22,240 --> 00:23:25,279
massive geopolitical chess game. But I want to bring this

491
00:23:25,359 --> 00:23:26,160
down to the listener.

492
00:23:26,319 --> 00:23:27,079
Speaker 2: Let's do it the.

493
00:23:27,039 --> 00:23:30,319
Speaker 1: Person driving to work right now, or folding laundry or

494
00:23:30,400 --> 00:23:33,480
walking the dog. It's easy to feel helpless when we

495
00:23:33,519 --> 00:23:37,480
talk about trillionaires and nuclear war. But Graylan gives some

496
00:23:37,519 --> 00:23:42,519
specific advice for two groups of people, business leaders and individuals.

497
00:23:42,720 --> 00:23:45,759
Speaker 2: Let's start with the business leaders, because many people listening

498
00:23:45,799 --> 00:23:47,880
might run a team or a company, or even just

499
00:23:47,960 --> 00:23:48,720
manage a department.

500
00:23:48,759 --> 00:23:49,519
Speaker 1: What should they do?

501
00:23:49,680 --> 00:23:52,960
Speaker 2: Stop thinking about replacement, start thinking about efficiency.

502
00:23:53,160 --> 00:23:56,119
Speaker 1: I feel like every CEO is currently thinking about replacement.

503
00:23:56,680 --> 00:23:59,640
The temptation to cut costs is huge.

504
00:24:00,079 --> 00:24:02,440
Speaker 2: Says He's talked to fifty companies in the last two months,

505
00:24:02,559 --> 00:24:05,839
and sadly, many are saying, great, I can fire thirty

506
00:24:05,880 --> 00:24:08,559
percent of my staff and keep the same output. He

507
00:24:08,720 --> 00:24:10,119
strongly warns against this.

508
00:24:10,599 --> 00:24:12,880
Speaker 1: Why apart from it being morally.

509
00:24:12,599 --> 00:24:15,799
Speaker 2: Questionable, because it feeds the elysium and mad Max dynamics, y,

510
00:24:16,000 --> 00:24:19,880
it destabilizes the society. Your business actually relies on. If

511
00:24:19,920 --> 00:24:23,000
nobody has a job, nobody buys your product. Instead, he

512
00:24:23,079 --> 00:24:27,160
suggests using the efficiency games to offer a four day workweek.

513
00:24:27,279 --> 00:24:30,240
Speaker 1: Oh, sign me up for that same pay less.

514
00:24:29,960 --> 00:24:34,160
Speaker 2: Hours exactly, or reskill employees for new roles. Use the

515
00:24:34,200 --> 00:24:36,759
surplus value to make lives better, not just to pad

516
00:24:36,759 --> 00:24:40,519
the bottom line. Reduce the shock to society. That is

517
00:24:40,519 --> 00:24:43,039
the responsibility of leadership right now, and for.

518
00:24:42,960 --> 00:24:47,000
Speaker 1: The individual, a person who isn't the CEO, what can

519
00:24:47,039 --> 00:24:47,359
I do?

520
00:24:47,599 --> 00:24:51,160
Speaker 2: His advice is simple but aggressive. Use the models, don't

521
00:24:51,160 --> 00:24:53,559
ignore them, Do not be an ostrich. Do not listen

522
00:24:53,599 --> 00:24:55,400
to the people who say it's not that scary or

523
00:24:55,440 --> 00:24:56,440
it won't replace humans.

524
00:24:56,519 --> 00:24:59,240
Speaker 1: He's pretty adamant about that. He doesn't sugarcoat the capability

525
00:24:59,400 --> 00:24:59,759
he is.

526
00:25:00,319 --> 00:25:02,759
Speaker 2: He says, you must use them to understand how powerful

527
00:25:02,759 --> 00:25:06,799
they are. If you aren't using these tools daily, you

528
00:25:06,839 --> 00:25:10,119
don't understand the speed of change. You are flying blind.

529
00:25:10,519 --> 00:25:12,799
Speaker 1: If you don't use it, you won't understand it. That's

530
00:25:12,839 --> 00:25:16,000
a powerful takeaway. You can't regulate, manage, or adapt to

531
00:25:16,000 --> 00:25:18,440
something you are treating as a novelty exactly.

532
00:25:19,119 --> 00:25:22,480
Speaker 2: Familiarity is a first step toward agency. You can't advocate

533
00:25:22,519 --> 00:25:24,759
for this star Trek future. If you don't understand the

534
00:25:24,759 --> 00:25:26,640
engine that drives it. If you know what the AI

535
00:25:26,720 --> 00:25:30,000
can do, you can demand better policy. You can demand

536
00:25:30,039 --> 00:25:30,839
the gi bil.

537
00:25:31,200 --> 00:25:33,720
Speaker 1: So here we are standing at the fort. We are

538
00:25:33,920 --> 00:25:37,599
to the left Elysium Trillionaires and space stations are effectively

539
00:25:37,640 --> 00:25:40,440
space stations while we fight for crumbs to the right,

540
00:25:40,680 --> 00:25:43,519
Mad Max. The world burning because we couldn't agree on

541
00:25:43,559 --> 00:25:45,839
how to share and straight ahead. If we have the

542
00:25:45,880 --> 00:25:49,119
courage to build it, Star Trek a world where the

543
00:25:49,200 --> 00:25:51,119
vulcan of AI lifts us all up.

544
00:25:51,319 --> 00:25:53,640
Speaker 2: It's not a technology choice. That's the most important thing.

545
00:25:53,680 --> 00:25:56,240
Graylan wants us to know. The technology will do whatever

546
00:25:56,240 --> 00:25:59,079
we tell it to. The code doesn't care. The choice

547
00:25:59,160 --> 00:26:03,279
is policy, The choice is mindset. The choice is fundamentally human.

548
00:26:03,519 --> 00:26:06,279
Speaker 1: It all comes down to that enlightened self interest. Can

549
00:26:06,400 --> 00:26:09,960
we overcome our biological wiring? We are wired for tribalism,

550
00:26:10,160 --> 00:26:13,079
us versus them, my tribe against your tribe, my country

551
00:26:13,119 --> 00:26:16,400
against your country. Can we override that programming fast enough

552
00:26:16,440 --> 00:26:18,799
to manage a technology that moves at the seat of light?

553
00:26:19,119 --> 00:26:22,680
Speaker 2: That is the ultimate question. Can our wisdom catch up

554
00:26:22,720 --> 00:26:27,240
to our intelligence? Because right now our intelligence, our tech

555
00:26:27,880 --> 00:26:30,119
is Lapping our Wisdom by Miles.

556
00:26:30,400 --> 00:26:32,599
Speaker 1: I want to turn this over to you listening right now,

557
00:26:33,119 --> 00:26:38,720
Graylin laid out three futures Elysium, mad Max, Star Trek.

558
00:26:39,039 --> 00:26:42,160
Be honest, look at the news, look at your workplace.

559
00:26:42,200 --> 00:26:45,000
Which one do you think we are currently most likely

560
00:26:45,079 --> 00:26:46,960
to hit? Not which one you want, but which one

561
00:26:47,000 --> 00:26:48,319
is the car pointed at right now?

562
00:26:48,640 --> 00:26:51,880
Speaker 2: And are you actually using these models yourself or are

563
00:26:51,880 --> 00:26:54,519
you just reading headlines about them, because as we learn today,

564
00:26:54,680 --> 00:26:55,920
there is a very big difference.

565
00:26:56,000 --> 00:26:57,440
Speaker 1: Leave us a comment, let us know. This is a

566
00:26:57,440 --> 00:27:00,640
conversation we need to have, not just in Stanford but everywhere.

567
00:27:00,680 --> 00:27:03,519
Speaker 2: Absolutely. The future isn't written yet, but the ink is

568
00:27:03,559 --> 00:27:04,279
drying fast.

569
00:27:05,039 --> 00:27:07,480
Speaker 1: Thanks for listening to thrilling threads. We'll see you in

570
00:27:07,519 --> 00:27:08,880
the future, hopefully the Star.

571
00:27:08,720 --> 00:27:10,279
Speaker 2: Trek one live, long and prosper.

572
00:27:10,359 --> 00:27:11,119
Speaker 1: Catch you next time.