1
00:00:00,120 --> 00:00:03,680
Speaker 1: Imagine a force, a force so so disruptive that if

2
00:00:03,680 --> 00:00:05,799
a rival power got their hands on it first, it

3
00:00:05,799 --> 00:00:09,599
could just completely rewrite the map, geopolitics, the economy, even

4
00:00:09,720 --> 00:00:14,000
you know, the way we live. It can make everything unrecognizable.

5
00:00:14,160 --> 00:00:16,879
Speaker 2: And for the better part of a century, almost eighty years,

6
00:00:17,399 --> 00:00:20,239
that fear has really been all about one thing, the.

7
00:00:20,239 --> 00:00:23,480
Speaker 1: Atom, exactly. But today that fear has a new name,

8
00:00:23,960 --> 00:00:26,800
and it's not fueled by nuclear fission. It's fueled by

9
00:00:26,920 --> 00:00:30,879
algorithms and just a terrifying amount of electricity. Welcome to

10
00:00:30,920 --> 00:00:33,479
Thrilling Threads, the place where we pull on the most

11
00:00:33,479 --> 00:00:36,840
crucial filaments of information to see exactly what the future

12
00:00:36,880 --> 00:00:38,479
is being woven from today.

13
00:00:38,479 --> 00:00:40,920
Speaker 2: We're doing a really, really deep dive into a concept

14
00:00:40,920 --> 00:00:43,799
that's being called Manhattan Project two point zero. Our mission

15
00:00:43,799 --> 00:00:47,439
here is to understand the stakes, which are they're existential,

16
00:00:47,880 --> 00:00:50,280
and the technology and the history that's all wrapped up

17
00:00:50,280 --> 00:00:52,759
in this new race for digital dominance.

18
00:00:52,640 --> 00:00:55,719
Speaker 1: Right and our source material for this is taking us

19
00:00:55,880 --> 00:01:00,560
deep into some reporting that comes from a very specific place, Oakridge, Tennessee,

20
00:01:01,079 --> 00:01:04,079
a place that's basically synonymous with the original nuclear age.

21
00:01:04,879 --> 00:01:07,560
We're looking at the arguments and the tech breakthroughs laid

22
00:01:07,560 --> 00:01:10,799
out in the video transcript Manhattan Project two point zero,

23
00:01:10,799 --> 00:01:11,760
which comes from.

24
00:01:11,599 --> 00:01:15,760
Speaker 2: The Daily Wire, and that location Oakridge is just so

25
00:01:16,000 --> 00:01:17,959
key to all of this. It's the whole story in

26
00:01:17,959 --> 00:01:20,640
one place. It was the epicenter of uranium enrichment for

27
00:01:20,680 --> 00:01:21,920
the original bomb.

28
00:01:21,719 --> 00:01:24,239
Speaker 1: And now eighty years later it's ground zero for the

29
00:01:24,280 --> 00:01:27,239
global AI arms race. It's home to the engine of

30
00:01:27,280 --> 00:01:29,120
this this whole new project.

31
00:01:29,239 --> 00:01:29,840
Speaker 2: It really is.

32
00:01:30,159 --> 00:01:32,480
Speaker 1: So the stakes, as the source material lays them out,

33
00:01:32,519 --> 00:01:35,480
are I mean, they're intentionally provocative. The core argument is this,

34
00:01:36,159 --> 00:01:39,400
the United States is an existential tech race, an AI

35
00:01:39,680 --> 00:01:40,840
arms race with China.

36
00:01:40,959 --> 00:01:43,439
Speaker 2: And if you just look back at the original Manhattan Project,

37
00:01:43,719 --> 00:01:46,319
the urgency, you know the reason for all the secrecy

38
00:01:46,359 --> 00:01:50,120
and they basically unlimited budget. It was this terrifying idea

39
00:01:50,200 --> 00:01:53,159
that Nazi Germany could get the bomb first, right, and

40
00:01:53,200 --> 00:01:55,359
that was seen as a direct threat to democracy, to

41
00:01:55,400 --> 00:01:56,480
the entire free world.

42
00:01:56,799 --> 00:01:59,920
Speaker 1: And the source makes that parallel crystal clear. It says, look,

43
00:02:00,359 --> 00:02:03,120
if Germany had won that race, our world today would

44
00:02:03,159 --> 00:02:06,840
be unrecognizable. And it argues that the danger of losing

45
00:02:06,920 --> 00:02:10,039
the AI race today is of that exact same magnitude.

46
00:02:10,080 --> 00:02:14,199
This isn't about a business advantage. It's a fundamental shift

47
00:02:14,240 --> 00:02:15,159
in global power.

48
00:02:15,439 --> 00:02:19,800
Speaker 2: Precisely. The argument is that AI it's not some slow

49
00:02:20,000 --> 00:02:23,560
academic thing anymore. It's hitting a critical mass like right

50
00:02:23,599 --> 00:02:26,400
now and in the next few years, it's expected to

51
00:02:26,439 --> 00:02:30,840
just change everything finance, science, and maybe most critically, national defense.

52
00:02:31,000 --> 00:02:34,120
Speaker 1: And the threat is defined very clearly. China is going

53
00:02:34,159 --> 00:02:36,719
after this technology aggressively, and if China were to get

54
00:02:36,759 --> 00:02:39,560
what they call a meaningful lead in AI, they argue,

55
00:02:39,560 --> 00:02:41,719
it would create a different world in the future, one

56
00:02:41,719 --> 00:02:44,719
where the US is probably no longer the main global leader.

57
00:02:44,879 --> 00:02:48,000
Speaker 2: So the goal then becomes really unambiguous. It echoes the

58
00:02:48,000 --> 00:02:50,840
Cold War, it echoes the atomic age. America has to

59
00:02:50,919 --> 00:02:52,240
lead and America has to win.

60
00:02:52,599 --> 00:02:56,080
Speaker 1: And this whole story, this convergence of nuclear history and

61
00:02:56,159 --> 00:03:00,400
digital future, it's literally happening in one place. Years ago.

62
00:03:00,520 --> 00:03:06,000
Oakridge was enriching uranium. Today it's home to the supercomputer Frontier,

63
00:03:06,360 --> 00:03:09,479
the engine for this Manhattan Project two point zero. It's

64
00:03:09,520 --> 00:03:12,199
the nexus point where the most powerful energy of the

65
00:03:12,240 --> 00:03:15,319
past meets the most powerful intelligence of the future.

66
00:03:15,479 --> 00:03:16,560
Speaker 2: That's an incredible story.

67
00:03:16,599 --> 00:03:18,800
Speaker 1: Okay, so let's really unpack this before we get into

68
00:03:18,800 --> 00:03:22,039
the machine itself frontier, and you know the algorithms it's running.

69
00:03:22,919 --> 00:03:25,800
The source material makes this incredibly strong case that the

70
00:03:25,800 --> 00:03:29,639
foundation of this new arms race isn't software, it's not chips,

71
00:03:29,800 --> 00:03:31,759
it's raw brute force energy.

72
00:03:31,879 --> 00:03:34,400
Speaker 2: This is such a crucial point, and it completely shifts

73
00:03:34,400 --> 00:03:36,599
how you think about AI. It takes it out of

74
00:03:36,639 --> 00:03:39,199
the sort of digital cloud and puts it squarely on

75
00:03:39,240 --> 00:03:41,400
the power grid. And we hear this from leaders in

76
00:03:41,439 --> 00:03:44,120
the field, people like open AI's co founder Greg Brockman

77
00:03:44,199 --> 00:03:47,199
and the former Energy Secretary Chris Wright. They see AI

78
00:03:47,319 --> 00:03:51,240
as this intensely, almost frighteningly energy hungry business.

79
00:03:51,520 --> 00:03:53,960
Speaker 1: Brockman's framing of it is so powerful because it just

80
00:03:54,000 --> 00:03:57,240
strips away all the mystique. He basically defines AI as

81
00:03:57,919 --> 00:03:59,879
manufacturing energy into intelligence.

82
00:04:00,199 --> 00:04:01,800
Speaker 2: Yeah, and just think about that for a second. When

83
00:04:01,840 --> 00:04:05,000
you're training a massive model, something like a GPT four,

84
00:04:05,439 --> 00:04:08,360
it eats up the energy equivalent of heating thousands of

85
00:04:08,400 --> 00:04:10,639
homes for a year maybe more.

86
00:04:10,800 --> 00:04:11,240
Speaker 1: Wow.

87
00:04:11,439 --> 00:04:13,639
Speaker 2: And the next generation of these models, the ones pushing

88
00:04:13,639 --> 00:04:16,800
into the trillion parameter range, I mean, the energy they'll

89
00:04:16,839 --> 00:04:19,680
need is completely unprecedented. So if you start thinking of

90
00:04:20,160 --> 00:04:23,800
AI as a factory, you know, an industrial process for

91
00:04:24,120 --> 00:04:25,279
making knowledge.

92
00:04:24,959 --> 00:04:28,959
Speaker 1: Then the energy infrastructure it needs becomes very physical, very industrial,

93
00:04:29,240 --> 00:04:31,160
not just digital exactly.

94
00:04:31,360 --> 00:04:33,800
Speaker 2: And that demand is only going to grow and it

95
00:04:33,800 --> 00:04:36,800
could very well outstrip what our current grid can even handle.

96
00:04:37,040 --> 00:04:39,879
Speaker 1: But here's the really interesting irony that the sources point out.

97
00:04:40,240 --> 00:04:43,079
While AI is consuming all this massive energy, it's also

98
00:04:43,120 --> 00:04:45,480
expected to help unlock more energy.

99
00:04:45,560 --> 00:04:48,600
Speaker 2: It's a fascinating loop, right, consumption and production loop. The

100
00:04:48,680 --> 00:04:51,879
idea is that AI is so good at optimizing these

101
00:04:51,879 --> 00:04:56,959
incredibly complex systems like power grids, energy transmission, even predicting

102
00:04:56,959 --> 00:04:59,839
when a power plant needs maintenance, that it could create

103
00:05:00,199 --> 00:05:01,680
huge efficiencies, so.

104
00:05:01,600 --> 00:05:04,600
Speaker 1: It actually helps us produce more usable power overall. It's

105
00:05:04,600 --> 00:05:08,399
a feedback mechanism. The intelligence helps solve the very energy

106
00:05:08,480 --> 00:05:09,639
crisis it's creating.

107
00:05:10,040 --> 00:05:13,439
Speaker 2: So if having limitless reliable power is the foundation for

108
00:05:13,519 --> 00:05:17,800
winning this race, the next question is obvious what energy

109
00:05:17,839 --> 00:05:21,959
source can actually sustain that pace? And the source Materials

110
00:05:22,040 --> 00:05:23,319
is very direct about this.

111
00:05:23,439 --> 00:05:27,519
Speaker 1: It points to nuclear absolutely. The US today we rely

112
00:05:27,639 --> 00:05:31,399
mostly on oil and natural gas, and they're huge dominant sources.

113
00:05:31,879 --> 00:05:34,199
But to scale up to the level you need for

114
00:05:34,279 --> 00:05:36,680
this kind of two hundred and forty seven expansion of

115
00:05:36,720 --> 00:05:40,680
AI and EXAs scale computing. Nuclear is presented as really

116
00:05:40,720 --> 00:05:43,240
the only viable solution that can scale that quickly and

117
00:05:43,279 --> 00:05:44,839
that reliably, and they.

118
00:05:44,759 --> 00:05:47,600
Speaker 2: Really stress its unique factor, the thing that a supercomputer

119
00:05:47,720 --> 00:05:50,920
like Frontier absolutely needs, which is that nuclear is the

120
00:05:50,920 --> 00:05:53,319
only technology that works whether the sun is shining or

121
00:05:53,319 --> 00:05:56,199
the wind is blowing. It's not intermittent. You get that solid,

122
00:05:56,360 --> 00:05:58,160
reliable base load power.

123
00:05:57,879 --> 00:06:00,319
Speaker 1: And you can't run a system with eighteen thousand blits

124
00:06:00,319 --> 00:06:03,000
on power that dips or has brown outs. You need absolute,

125
00:06:03,160 --> 00:06:04,560
unwavering energy, surety.

126
00:06:04,800 --> 00:06:08,160
Speaker 2: But the source goes even further than just electricity for computers.

127
00:06:08,439 --> 00:06:11,160
It talks about a not distant future and maybe a

128
00:06:11,240 --> 00:06:15,040
lot closer than we think, where nuclear also provides high

129
00:06:15,079 --> 00:06:18,240
temperature process heat. And this is where the strategic argument

130
00:06:18,319 --> 00:06:19,600
for nuclear gets even deeper.

131
00:06:19,720 --> 00:06:22,480
Speaker 1: Okay, explain that for us, because electricity is obvious, but

132
00:06:22,759 --> 00:06:27,240
process heat that's an industrial idea that might not immediately

133
00:06:27,240 --> 00:06:28,319
click for everyone.

134
00:06:28,639 --> 00:06:31,959
Speaker 2: Right, So high temperature process heat is basically the thermal

135
00:06:32,079 --> 00:06:35,240
energy you need for heavy industry. The source calls it

136
00:06:35,279 --> 00:06:37,879
the single most important energy source in the world. If

137
00:06:37,879 --> 00:06:40,399
you want to make core manufacturing materials.

138
00:06:40,160 --> 00:06:43,480
Speaker 1: Things like steel, plastics, alumina.

139
00:06:43,120 --> 00:06:47,600
Speaker 2: Exactly, steel plastics, ammonia for fertilizer, aluminum. All those processes

140
00:06:47,720 --> 00:06:50,000
need incredibly high temperatures, and right now we get that

141
00:06:50,079 --> 00:06:53,560
by burning massive amounts of natural gas or coal. But

142
00:06:53,600 --> 00:06:56,959
if you can replace that fossil fuel with nuclear heat,

143
00:06:57,399 --> 00:07:00,959
often from these new small modular reactors or s, you

144
00:07:01,000 --> 00:07:02,160
get two huge advantages.

145
00:07:02,279 --> 00:07:04,839
Speaker 1: Okay, So first, I'm guessing you cut emissions in a

146
00:07:04,879 --> 00:07:06,839
really tough sector drastically. Yeah.

147
00:07:06,879 --> 00:07:10,160
Speaker 2: And second, you decouple your entire industrial base from volatile

148
00:07:10,199 --> 00:07:11,079
fossil fuel.

149
00:07:10,839 --> 00:07:13,639
Speaker 1: Markets, which is a massive strategic.

150
00:07:13,079 --> 00:07:16,399
Speaker 2: Advantage, a huge one. The source argues that if you

151
00:07:16,439 --> 00:07:20,240
control the cleanest, cheapest, most reliable source of this high

152
00:07:20,279 --> 00:07:23,560
temperature heat, you control the production of everything that builds

153
00:07:23,560 --> 00:07:26,480
a modern economy and a modern military.

154
00:07:26,959 --> 00:07:29,279
Speaker 1: And this leads us right into the political framing in

155
00:07:29,319 --> 00:07:32,560
the sources which connects this whole tech challenge to a

156
00:07:32,600 --> 00:07:36,839
sense of policy urgency. The former Energy Secretary's view is

157
00:07:36,879 --> 00:07:40,319
pretty forceful. He argues that America will lose the AI

158
00:07:40,439 --> 00:07:42,279
arms race if we strangle innovation.

159
00:07:42,839 --> 00:07:46,839
Speaker 2: Yeah, that phrase strangle innovation is very loaded. The source

160
00:07:46,879 --> 00:07:50,600
expresses this clear frustration, suggesting that the previous four years

161
00:07:50,639 --> 00:07:53,800
were and this is a quote consumed with theology regarding

162
00:07:53,839 --> 00:07:54,920
the environment.

163
00:07:54,680 --> 00:07:58,279
Speaker 1: Implying that environmental regulations were holding back essential energy sources

164
00:07:58,319 --> 00:07:59,519
like nuclear and gas.

165
00:07:59,680 --> 00:08:01,800
Speaker 2: Right They see that focus as a kind of self

166
00:08:01,800 --> 00:08:04,519
sabotage when we're in this critical global race.

167
00:08:04,759 --> 00:08:06,959
Speaker 1: It's a strong claim and one we have to look

168
00:08:07,000 --> 00:08:10,879
at impartially for someone listening who might not follow energy

169
00:08:10,879 --> 00:08:14,040
policy that closely. The source is really pushing for an

170
00:08:14,079 --> 00:08:17,160
innovation first approach. But what's the counter argument there? What

171
00:08:17,240 --> 00:08:19,920
about balancing that expansion with climate goals?

172
00:08:20,040 --> 00:08:23,240
Speaker 2: Well, the standard counter argument is all about long term

173
00:08:23,360 --> 00:08:27,839
risk and global leadership on climate. Proponents of regulation would

174
00:08:27,879 --> 00:08:30,240
say that, sure, we need to expand, but if we

175
00:08:30,279 --> 00:08:33,879
do it unchecked, especially with carbon intensive sources, you create

176
00:08:33,919 --> 00:08:35,799
these massive costs down the line.

177
00:08:35,639 --> 00:08:39,679
Speaker 1: Through things like extreme weather damage or even trade penalties exactly.

178
00:08:40,080 --> 00:08:43,559
Speaker 2: But the viewpoint in our source material is really prioritizing

179
00:08:43,679 --> 00:08:46,840
national security and tech dominance above all else.

180
00:08:46,720 --> 00:08:50,960
Speaker 1: Right now, which brings them to their solution unleash American energy,

181
00:08:51,720 --> 00:08:54,440
make nuclear a huge part of an all of the

182
00:08:54,519 --> 00:08:58,440
above energy portfolio, and the competitive angle is key. They

183
00:08:58,480 --> 00:09:01,480
see two camps in the world, one that favors collaboration

184
00:09:01,879 --> 00:09:03,919
and one that's all about national competition.

185
00:09:04,120 --> 00:09:06,720
Speaker 2: And the source is firmly in that second camp. The

186
00:09:06,759 --> 00:09:09,399
assumption is that if countries like China or Russia make

187
00:09:09,440 --> 00:09:11,639
a breakthrough, they're not going to share.

188
00:09:11,399 --> 00:09:14,679
Speaker 1: It, So relying on collaboration is just too risky for

189
00:09:14,720 --> 00:09:15,159
the US.

190
00:09:15,320 --> 00:09:18,360
Speaker 2: That's the argument. The imperative is clear. America has to

191
00:09:18,759 --> 00:09:22,919
out innovate everybody on the planet on energy because energy

192
00:09:23,000 --> 00:09:25,399
is the bottleneck for the entire AI engine.

193
00:09:25,440 --> 00:09:27,559
Speaker 1: Okay, so this is where it gets really interesting. We've

194
00:09:27,600 --> 00:09:33,120
talked about the fuel, this idea of limitless reliable nuclear power.

195
00:09:33,720 --> 00:09:37,519
Now let's look inside the engine itself. Let's talk about Frontier.

196
00:09:37,639 --> 00:09:40,639
Speaker 2: Frontier is just it's not just a big room full

197
00:09:40,639 --> 00:09:44,519
of servers. It is the world's fastest supercomputer for open science.

198
00:09:44,840 --> 00:09:47,240
When it was unveiled, it was the first machine ever

199
00:09:47,279 --> 00:09:49,720
confirmed to break the Exascal barrier.

200
00:09:49,759 --> 00:09:52,279
Speaker 1: Okay, we need to go beyond the simple comparison of

201
00:09:52,360 --> 00:09:55,639
you know, fifty times your laptop. What does xaskel actually

202
00:09:55,679 --> 00:09:56,639
mean in real terms?

203
00:09:56,720 --> 00:09:59,440
Speaker 2: Well, the jump from petaflop, which is a quadrillion calculations

204
00:09:59,440 --> 00:10:03,600
per second flop is just massive. Exascal means Frontier can

205
00:10:03,639 --> 00:10:06,519
do a quintillion calculators per second. That's a one with

206
00:10:06,600 --> 00:10:07,840
eighteen zeros after it.

207
00:10:07,960 --> 00:10:10,159
Speaker 1: That's I can't even picture that.

208
00:10:10,159 --> 00:10:12,759
Speaker 2: To try and visualize it. If every single person on

209
00:10:12,799 --> 00:10:15,399
Earth did one calculation every second, it would take them

210
00:10:15,440 --> 00:10:17,559
four years to do what Frontier can do in one second.

211
00:10:17,759 --> 00:10:18,120
Speaker 1: Wow.

212
00:10:18,759 --> 00:10:22,480
Speaker 2: Okay, and you need that speed because the algorithms these

213
00:10:22,519 --> 00:10:27,840
scientists are writing have you know, billions of parameters.

214
00:10:27,320 --> 00:10:31,600
Speaker 1: And they're tackling problems that are just genuinely unfathomable for

215
00:10:31,639 --> 00:10:34,279
a human mind. The example they give is working with

216
00:10:34,320 --> 00:10:37,240
a company like ge to design new aircraft gifts. I

217
00:10:37,279 --> 00:10:42,000
mean that's calculating fluid dynamics, material stress, combustion, just an

218
00:10:42,080 --> 00:10:43,840
insane number of variables.

219
00:10:43,399 --> 00:10:46,519
Speaker 2: All at once, precisely. And if a simulation takes weeks

220
00:10:46,519 --> 00:10:49,519
to run, it's basically useless for a company that needs

221
00:10:49,519 --> 00:10:52,600
to make quick design choices. Frontier shrinks that time from

222
00:10:52,639 --> 00:10:56,879
weeks down to hours, so engineers can test just exponentially

223
00:10:56,919 --> 00:10:59,720
more designs. That's what accelerated innovation really means.

224
00:11:00,120 --> 00:11:02,080
Speaker 1: A key point, and this is a parallel to the

225
00:11:02,080 --> 00:11:05,120
original Manhattan Project, is that the technology for a machine

226
00:11:05,200 --> 00:11:08,039
this complex often doesn't even exist when the project.

227
00:11:07,759 --> 00:11:11,759
Speaker 2: Starts exactly when Oakridge decides to procure a system like Frontier,

228
00:11:12,440 --> 00:11:15,000
the tech needed to actually hit that exast scale goal.

229
00:11:15,360 --> 00:11:17,720
It hasn't been manufactured yet. They commit to building it

230
00:11:18,039 --> 00:11:20,360
knowing they'll have to innovate it into existence through these

231
00:11:20,480 --> 00:11:21,919
really intense partnerships.

232
00:11:22,080 --> 00:11:24,399
Speaker 1: So in this case, that meant working hand in hand

233
00:11:24,480 --> 00:11:26,919
with Hewlett Packard Enterprise and AMD.

234
00:11:27,399 --> 00:11:29,840
Speaker 2: Right, and it wasn't just about assembling parts. It was

235
00:11:29,879 --> 00:11:33,559
about pushing the fundamentals of chip design and maybe even

236
00:11:33,600 --> 00:11:34,879
more importantly cooling.

237
00:11:35,600 --> 00:11:38,120
Speaker 1: Okay, so what were the big engineering hurdles they had

238
00:11:38,159 --> 00:11:38,519
to clear.

239
00:11:38,759 --> 00:11:42,000
Speaker 2: Well, the biggest one in modern supercomputing is usually the

240
00:11:42,039 --> 00:11:45,759
interconnect and memory bandwidth, basically how fast all the different

241
00:11:45,759 --> 00:11:47,919
processors can talk to each other and get to the

242
00:11:48,000 --> 00:11:51,679
data they need. Frontier was designed with this really integrated

243
00:11:51,759 --> 00:11:55,919
CPU and GPU architecture, which required a completely bespoke design

244
00:11:56,000 --> 00:11:58,159
from AMD to cut down latency.

245
00:11:58,519 --> 00:12:01,399
Speaker 1: But the innovation that oak really pushed for, and this

246
00:12:01,440 --> 00:12:04,600
ties right back to our energy discussion, was the cooling system.

247
00:12:04,720 --> 00:12:08,360
Speaker 2: Yes, this was a massive crucial push the implementation of

248
00:12:08,440 --> 00:12:11,320
warm water cooling. It's something people overlook but it's a

249
00:12:11,320 --> 00:12:15,159
revolutionary piece of engineering. You have eighteen thousand blades producing

250
00:12:15,200 --> 00:12:18,639
that much heat. Cooling is normally your biggest energy cost.

251
00:12:18,600 --> 00:12:21,039
Speaker 1: And the result was pretty staggering, right as savings of

252
00:12:21,240 --> 00:12:23,159
at least forty percent in our energy cost.

253
00:12:23,519 --> 00:12:26,399
Speaker 2: A forty percent savings on the cooling for the world's

254
00:12:26,399 --> 00:12:31,480
fastest computer is an enormous advantage strategically and financially. They're

255
00:12:31,480 --> 00:12:35,080
not using these giant chilled air conditioned rooms. They're flowing

256
00:12:35,159 --> 00:12:39,639
warmer water directly over the components, which is way more efficient, and.

257
00:12:39,600 --> 00:12:41,240
Speaker 1: They can reuse that captured heat.

258
00:12:41,240 --> 00:12:44,519
Speaker 2: I assume often. Yes, it creates a much more sustainable system.

259
00:12:44,759 --> 00:12:47,240
It just shows how oak Ridge is pushing the boundaries

260
00:12:47,440 --> 00:12:49,639
of both the hardware and the efficiency.

261
00:12:49,159 --> 00:12:52,159
Speaker 1: At the same time, and that innovation trickles down. You

262
00:12:52,240 --> 00:12:55,440
mentioned the chips they helped design became commercially available just

263
00:12:55,720 --> 00:12:56,519
a year later.

264
00:12:56,759 --> 00:12:59,279
Speaker 2: That rapid spin off is a huge part of the mission.

265
00:13:00,000 --> 00:13:02,759
Bridge acts as this forcing function for the whole industry.

266
00:13:03,080 --> 00:13:05,600
They demand technology a year or two before the market

267
00:13:05,639 --> 00:13:07,720
is even ready for it, and in doing so they

268
00:13:07,759 --> 00:13:12,240
basically fund and test the absolute bleeding edge, which then

269
00:13:12,360 --> 00:13:15,799
cascades down into commercial products for everyone, which.

270
00:13:15,600 --> 00:13:18,679
Speaker 1: Brings us to the core philosophy of how Frontier is used.

271
00:13:18,840 --> 00:13:21,600
And this is what really separates Manhattan Project two point

272
00:13:21,639 --> 00:13:24,799
zero from the original. It's the mandate of open science.

273
00:13:24,720 --> 00:13:28,559
Speaker 2: That is the defining line. The machine is available to

274
00:13:28,639 --> 00:13:32,080
scientists with really big problems, the kinds of problems no

275
00:13:32,200 --> 00:13:36,200
single university or private company could ever tackle alone. But

276
00:13:36,240 --> 00:13:38,919
there's a catch, a quid pro quo.

277
00:13:39,120 --> 00:13:41,240
Speaker 1: You have to agree to publish part of your work.

278
00:13:41,279 --> 00:13:44,080
Speaker 2: You have to publish it in the scientific domain. This

279
00:13:44,240 --> 00:13:48,200
ensures what they call true open science where everyone benefits.

280
00:13:48,519 --> 00:13:52,120
It's a very deliberate rejection of the extreme secrecy of

281
00:13:52,159 --> 00:13:53,000
eighty years ago.

282
00:13:53,200 --> 00:13:56,080
Speaker 1: Yeah, the leadership at the facility is really clear about it.

283
00:13:56,120 --> 00:13:58,039
They say, we don't do the secret stuff because we

284
00:13:58,080 --> 00:13:59,679
want it to be available to the public.

285
00:14:00,000 --> 00:14:01,799
Speaker 2: So, even though the whole thing is set against this

286
00:14:01,960 --> 00:14:06,600
backdrop of intense national competition, the actual engine of research

287
00:14:06,799 --> 00:14:10,759
is in a way democratized for global science, and that

288
00:14:10,799 --> 00:14:13,840
creates this profound duality. It's just fascinating to think about.

289
00:14:14,000 --> 00:14:16,919
Speaker 1: So we've got the energy, we've got the exasma machine.

290
00:14:17,000 --> 00:14:19,039
What are they actually doing with it? How do these

291
00:14:19,080 --> 00:14:22,480
eighteen thousand blades translate into real scientific breakthroughs?

292
00:14:22,519 --> 00:14:24,480
Speaker 2: Right? Now this is where it all comes to life.

293
00:14:25,120 --> 00:14:28,440
The collaboration is so big, with researchers from all the

294
00:14:28,519 --> 00:14:31,879
national labs and tons of universities that they're calling it

295
00:14:31,960 --> 00:14:35,440
the one thousand Scientists aijim I love that the goal

296
00:14:35,559 --> 00:14:38,399
is to integrate these sophisticated AI models with what human

297
00:14:38,440 --> 00:14:42,320
scientists are already doing and just exponentially accelerate their work.

298
00:14:42,440 --> 00:14:44,559
Speaker 1: Well, let's start with one of the biggest examples they

299
00:14:44,559 --> 00:14:47,039
give the search for nuclear fusion.

300
00:14:47,519 --> 00:14:52,799
Speaker 2: Fusion is famously notoriously difficult because plasma, the superheated stuff

301
00:14:52,799 --> 00:14:56,399
you need, is just so chaotic and volatile. So scientists

302
00:14:56,440 --> 00:14:58,879
are using AI on frontier to come up with new

303
00:14:58,919 --> 00:15:03,960
designs for experimental reactors modeling these incredibly complex magnetic fields.

304
00:15:03,679 --> 00:15:06,399
Speaker 1: And a human team might spend weeks just simulating one

305
00:15:06,440 --> 00:15:07,639
design to see if it's even.

306
00:15:07,559 --> 00:15:11,120
Speaker 2: Viable, right, But the AI can run thousands of different versions,

307
00:15:11,240 --> 00:15:15,159
instantly tweaking parameters, spotting flaws, and even proposing new designs

308
00:15:15,159 --> 00:15:17,840
that a human might never think of. And that acceleration

309
00:15:18,000 --> 00:15:21,480
is everything, because every month we save infusion research gets

310
00:15:21,679 --> 00:15:25,120
that much closer to potentially limitless clean energy. The AIA

311
00:15:25,240 --> 00:15:27,399
just speeds up that whole cycle of trial and error.

312
00:15:27,519 --> 00:15:30,720
Speaker 1: So another area you mentioned is manufacturing and material science,

313
00:15:30,840 --> 00:15:34,840
especially at the neutrons scattering division at oak Ridge. This

314
00:15:34,879 --> 00:15:37,200
is a bit technical, so let's slow down and really

315
00:15:37,240 --> 00:15:41,639
explain how studying atoms applies to this arms race.

316
00:15:41,840 --> 00:15:45,440
Speaker 2: It connects directly to industrial strength, which is fundamental. They're

317
00:15:45,519 --> 00:15:49,279
using neutrons to literally map the atomic structure of materials.

318
00:15:49,519 --> 00:15:52,200
You take a material, say a new metal alloy, you

319
00:15:52,279 --> 00:15:54,240
put in a beam line, you hit it with neutrons,

320
00:15:54,559 --> 00:15:57,279
and the way those neutrons scatter gives you data with

321
00:15:57,480 --> 00:15:58,879
atomic resolution.

322
00:15:58,559 --> 00:16:01,080
Speaker 1: So you're basically getting a map of where every single

323
00:16:01,120 --> 00:16:03,240
atom is in that material exactly.

324
00:16:03,440 --> 00:16:07,440
Speaker 2: The raw data these diffraction peaks are just massive streams

325
00:16:07,440 --> 00:16:09,919
of information that tell you exactly where the atoms are

326
00:16:10,159 --> 00:16:13,240
and whether the material is under stress. And this huge,

327
00:16:13,320 --> 00:16:16,799
complex data stream is then immediately quote shoved into a

328
00:16:16,879 --> 00:16:18,799
machine learning or AI, and the.

329
00:16:18,759 --> 00:16:22,200
Speaker 1: AI isn't just storing the data, it's actively interpreting it

330
00:16:22,279 --> 00:16:24,600
in real time, seeing patterns that would take a person

331
00:16:24,759 --> 00:16:27,399
days to figure out, and then making suggestions.

332
00:16:27,960 --> 00:16:31,679
Speaker 2: That's the critical difference. The AI takes that instantaneous data

333
00:16:31,960 --> 00:16:36,000
and actively steers the experiment. It gives predictive feedback to

334
00:16:36,120 --> 00:16:39,360
the scientist, telling them how to optimize the material or

335
00:16:39,399 --> 00:16:41,000
adjust the manufacturing process.

336
00:16:41,039 --> 00:16:44,120
Speaker 1: Wait, if the AI is steering the experiment, how do

337
00:16:44,240 --> 00:16:46,440
the humans stay in control? I mean, how do they

338
00:16:46,480 --> 00:16:50,120
even understand the physics behind what the AI is suggesting?

339
00:16:50,519 --> 00:16:53,639
Speaker 2: That's a great question. The AI isn't replacing the scientist,

340
00:16:53,759 --> 00:16:56,519
it's just making the experiment faster and more efficient. So

341
00:16:56,600 --> 00:16:59,759
the AI is typically optimizing the parameters of the neutron beam,

342
00:17:00,080 --> 00:17:03,399
the angle, the intensity, based on the atomic stress it's seeing.

343
00:17:03,840 --> 00:17:07,039
The human scientist still sets the overall scientific goal. The

344
00:17:07,079 --> 00:17:10,799
AI just becomes this hyper fast, tireless optimizer, so the

345
00:17:10,880 --> 00:17:13,319
human can focus on the big picture physics instead of

346
00:17:13,359 --> 00:17:14,359
the tiny details.

347
00:17:14,519 --> 00:17:17,640
Speaker 1: And the source gives this just stunning real world example

348
00:17:17,640 --> 00:17:20,559
of this inaction, measuring the structure of an actual gas

349
00:17:20,559 --> 00:17:22,720
engine cylinder head while the engine is running.

350
00:17:22,799 --> 00:17:25,319
Speaker 2: I mean, that's just phenomenal. They can sync the pulse

351
00:17:25,359 --> 00:17:27,400
of the engine with the pulse of the neutron beam

352
00:17:27,720 --> 00:17:30,880
and capture what's happening inside the combustion chamber in real time.

353
00:17:31,279 --> 00:17:35,119
And the goal is totally applied science. Develop materials for

354
00:17:35,200 --> 00:17:39,000
cars and planes that are stronger lighter and cheaper by

355
00:17:39,079 --> 00:17:42,640
understanding and manipulating their atomic structure with AI guidance.

356
00:17:42,839 --> 00:17:46,240
Speaker 1: The computational power of Frontier just lets them train these

357
00:17:46,279 --> 00:17:49,880
AI models so quickly on so much data that the

358
00:17:49,920 --> 00:17:52,160
predictions become incredibly accurate.

359
00:17:52,359 --> 00:17:56,119
Speaker 2: Exactly. It shortens the entire materials development cycle from years

360
00:17:56,160 --> 00:17:58,519
down to months or even weeks. It's a speed that

361
00:17:58,640 --> 00:18:00,839
was just unthinkable a decade ago.

362
00:18:01,160 --> 00:18:04,960
Speaker 1: And finally, beyond physics and materials, they're also using Frontier

363
00:18:05,000 --> 00:18:06,519
for health data.

364
00:18:06,039 --> 00:18:08,799
Speaker 2: A huge area for public good. They have users who

365
00:18:08,799 --> 00:18:11,039
come in with these vast protected health data sets and

366
00:18:11,039 --> 00:18:15,519
they're looking for complex cancer related trends across huge populations.

367
00:18:15,640 --> 00:18:18,480
By speeding up that analysis, the AI can spot subtle

368
00:18:18,519 --> 00:18:22,559
correlations in treatment success in genomic markers that would be

369
00:18:22,599 --> 00:18:24,680
impossible for a human team to ever find.

370
00:18:24,920 --> 00:18:28,160
Speaker 1: The goal being to help clinicians make better decisions.

371
00:18:27,839 --> 00:18:29,759
Speaker 2: Way faster and ultimately save lives.

372
00:18:29,880 --> 00:18:31,839
Speaker 1: Yeah, so we have to connect this back to the

373
00:18:31,839 --> 00:18:36,079
bigger picture. Oakridge is special not just for what it's

374
00:18:36,119 --> 00:18:39,079
doing today, but for what it did eighty years ago,

375
00:18:39,880 --> 00:18:44,039
that secretive, profound and dangerous leap it engineered.

376
00:18:44,319 --> 00:18:47,039
Speaker 2: The history really gives the Manhattan Project two point zero

377
00:18:47,079 --> 00:18:50,319
analogy its weight and the sources take us right back

378
00:18:50,319 --> 00:18:53,400
to the birthplace of it all, the X ten Graphite Reactor.

379
00:18:53,000 --> 00:18:55,319
Speaker 1: Which is a National Historic Site now it is.

380
00:18:55,559 --> 00:18:58,720
Speaker 2: It started operating on November fourth, nineteen forty three, and

381
00:18:58,759 --> 00:19:01,920
it was the birthplace of operational nuclear power. Its whole

382
00:19:02,000 --> 00:19:04,400
job was just to prove that a controlled chain reaction

383
00:19:04,519 --> 00:19:08,440
was possible. Workers loaded thirty one tons of uranium slugs

384
00:19:08,480 --> 00:19:12,000
into this huge twenty four foot cube of graphite.

385
00:19:11,480 --> 00:19:13,960
Speaker 1: And the graphite acts as a moderator, right, It slows

386
00:19:14,000 --> 00:19:17,039
the neutrons down so they can sustain the reaction exactly.

387
00:19:17,319 --> 00:19:20,480
Speaker 2: And once that uranium was irradiated, it was incredibly radioactive

388
00:19:20,519 --> 00:19:24,039
and it contained these tiny amounts of a brand new element, plutonium.

389
00:19:24,440 --> 00:19:27,519
So that material was taken next door to a chemical

390
00:19:27,559 --> 00:19:30,720
separation plant where they had to work behind heavy shielding

391
00:19:31,079 --> 00:19:33,640
to isolate that tiny bit of plutonium. Yeah, they were

392
00:19:33,640 --> 00:19:36,880
literally inventing nuclear engineering as they went, figuring out the

393
00:19:36,880 --> 00:19:39,880
safest materials, the best shielding all of it.

394
00:19:39,960 --> 00:19:44,400
Speaker 1: And what's so fascinating is the extreme secrecy and compartmentalization

395
00:19:44,519 --> 00:19:46,720
they needed to pull it off. I mean, how do

396
00:19:46,759 --> 00:19:49,599
you keep seventy five thousand people from knowing what they're building.

397
00:19:49,759 --> 00:19:52,559
Speaker 2: It was incredibly strict. The sources say most of those

398
00:19:52,559 --> 00:19:55,000
seventy five thousand people had no idea what the final

399
00:19:55,039 --> 00:19:58,079
goal was. They only knew their one tiny job. A

400
00:19:58,119 --> 00:20:00,680
worker in one building might be told their separate nickel

401
00:20:00,920 --> 00:20:04,160
while someone next door is purifying a liquid. Neither one

402
00:20:04,279 --> 00:20:06,119
knew they were contributing to a weapon.

403
00:20:06,200 --> 00:20:08,319
Speaker 1: They would just do their job and pass the component

404
00:20:08,400 --> 00:20:10,519
to the next room, completely in the dark until the

405
00:20:10,559 --> 00:20:12,960
bombs were dropped, and then the government was very open

406
00:20:12,960 --> 00:20:15,759
immediately that they were made at Oak Ridge. The psychology

407
00:20:15,759 --> 00:20:18,000
of that is just staggering, It truly is.

408
00:20:18,079 --> 00:20:20,759
Speaker 2: I mean, can you imagine working under that kind of security,

409
00:20:21,079 --> 00:20:23,799
knowing your work is vital, but having no idea you're

410
00:20:23,799 --> 00:20:27,000
building the thing that will change the world forever. And

411
00:20:27,039 --> 00:20:30,440
that's such a stark contract with frontiers mandate today of

412
00:20:30,720 --> 00:20:31,519
open science.

413
00:20:31,799 --> 00:20:34,559
Speaker 1: But the story of the X ten reactor didn't end

414
00:20:34,640 --> 00:20:37,359
with the war. It actually became the foundation for the

415
00:20:37,519 --> 00:20:39,599
entire post war nuclear.

416
00:20:39,240 --> 00:20:42,160
Speaker 2: Age, right. It transitioned very quickly from a weapon site

417
00:20:42,319 --> 00:20:45,400
to a science and engineering center. The scientists there realized

418
00:20:45,440 --> 00:20:49,640
they had this unprecedented scientific tool, so it immediately became

419
00:20:49,680 --> 00:20:52,720
the training ground for the very first nuclear engineers. Because

420
00:20:52,720 --> 00:20:55,160
they were the only ones who had the expertise and

421
00:20:55,200 --> 00:20:57,920
a working reactor. They were inventing the field.

422
00:20:57,759 --> 00:21:00,880
Speaker 1: And that expertise gave the US a map massive strategic

423
00:21:00,920 --> 00:21:02,359
advantage for decades.

424
00:21:02,559 --> 00:21:05,519
Speaker 2: The most famous example they cite is Admiral Hyman Rickover.

425
00:21:05,920 --> 00:21:08,440
He came to the X ten reactor specifically to train

426
00:21:08,519 --> 00:21:10,880
and figure out how to put a safe, working nuclear

427
00:21:10,920 --> 00:21:13,559
reactor onto a submarine, which was a game changer, a

428
00:21:13,640 --> 00:21:16,079
huge one. The ability for a nuclear sub to stay

429
00:21:16,119 --> 00:21:20,240
under water almost indefinitely silently gave the US Davy this

430
00:21:20,319 --> 00:21:22,480
huge strategic advantage over its rivals.

431
00:21:22,759 --> 00:21:26,400
Speaker 1: Now, if we connect that historical risk to today's AI race,

432
00:21:26,880 --> 00:21:31,559
the source highlights this deep fundamental tension between AI optimism

433
00:21:31,559 --> 00:21:33,119
and that historic atomic fear.

434
00:21:33,400 --> 00:21:36,960
Speaker 2: This tension really defines the whole conversation. On one side,

435
00:21:37,000 --> 00:21:40,319
you have AI innovators like Brockman, who sees AI not

436
00:21:40,400 --> 00:21:44,240
as a crisis, but is this opportunity for extraordinary growth

437
00:21:44,599 --> 00:21:47,559
for the human race. He's focused on all the powerful,

438
00:21:47,559 --> 00:21:49,079
benevolent things we can do with it.

439
00:21:49,160 --> 00:21:51,960
Speaker 1: But the parallel to the atomic age immediately brings up

440
00:21:51,960 --> 00:21:55,000
all the moral and ethical conflicts that the original scientists face.

441
00:21:55,119 --> 00:21:58,279
Speaker 2: Right. It's the explicit parallel to J. Robert Oppenheimer, who

442
00:21:58,359 --> 00:22:01,720
later saw atomic energy as a grave crisis because of

443
00:22:01,759 --> 00:22:05,680
the immense, unpredictable power it unleashed on humanity.

444
00:22:05,279 --> 00:22:07,400
Speaker 1: And that tension, that risk of the unknown. It was

445
00:22:07,440 --> 00:22:09,680
there from the very beginning, even in the design phase

446
00:22:09,680 --> 00:22:11,599
of the X ten back in nineteen forty three.

447
00:22:11,880 --> 00:22:15,920
Speaker 2: It was DuPont, the contractor. They realized the dangers, the

448
00:22:16,039 --> 00:22:18,920
risk that the reactor could run out of control, kill workers,

449
00:22:19,039 --> 00:22:22,119
or send nuclear particles into the air. They were stepping

450
00:22:22,160 --> 00:22:27,480
into completely new, terrifying territory. And the source captures this perfectly.

451
00:22:27,519 --> 00:22:31,799
It says the project's leaders recognized the power of unleashing

452
00:22:31,839 --> 00:22:35,599
the atom, but we also realize the great responsibility that comes.

453
00:22:35,359 --> 00:22:39,200
Speaker 1: With it that duality. Great power demands great responsibility, and.

454
00:22:39,119 --> 00:22:42,079
Speaker 2: That defines both the original Manhattan project and this new

455
00:22:42,160 --> 00:22:44,119
AI endeavor with Frontier Today.

456
00:22:44,480 --> 00:22:48,880
Speaker 1: What a monumental, just thrilling story of technological continuity. From

457
00:22:48,880 --> 00:22:51,400
the X ten reactor, which taught the world how to

458
00:22:51,440 --> 00:22:55,079
harness the atom, to the Frontier supercomputer, which is teaching

459
00:22:55,119 --> 00:22:57,079
AI how to optimize our future.

460
00:22:57,519 --> 00:23:00,519
Speaker 2: We've really seen these three core threads all in twined.

461
00:23:00,920 --> 00:23:05,400
First this massive essential link between national dominance and scalable

462
00:23:05,480 --> 00:23:08,640
nuclear power. That's the physical fuel for this new race right.

463
00:23:08,680 --> 00:23:12,000
Speaker 1: And second the incredible speed of innovation needed, which is

464
00:23:12,000 --> 00:23:16,359
personified by Frontier and its contradictory mandate for open science,

465
00:23:16,599 --> 00:23:19,319
which is the total opposite of the original project secrecy.

466
00:23:19,559 --> 00:23:23,680
Speaker 2: And Third that inescapable historical and ethical tension that comes

467
00:23:23,680 --> 00:23:27,720
back every single time humanity takes one of these enormous leaps,

468
00:23:27,759 --> 00:23:31,680
the tension between the promise of extraordinary growth and the

469
00:23:31,680 --> 00:23:33,759
threat of an unrecognizable world.

470
00:23:33,960 --> 00:23:36,920
Speaker 1: The focus today at Oakridge, at least on the surface,

471
00:23:37,279 --> 00:23:41,759
is on solving these really big problems fusion reactors, cancer treatments,

472
00:23:41,839 --> 00:23:45,279
new materials, and it's all under this banner of true

473
00:23:45,720 --> 00:23:47,680
open science for the greater good.

474
00:23:47,839 --> 00:23:50,279
Speaker 2: And yet the political framing that's driving all the funding

475
00:23:50,279 --> 00:23:54,160
and the urgency is completely geopolitical. It is framed as

476
00:23:54,200 --> 00:23:58,880
a competitive nationalistic arms race against rivals like China and Russia.

477
00:23:58,960 --> 00:24:01,119
It's a zero sum game that America has to win.

478
00:24:01,559 --> 00:24:06,359
Speaker 1: This dichotomy benevolent science versus nationalist competition. It brings us

479
00:24:06,400 --> 00:24:09,000
to the profound meaning of scientific freedom itself, which is

480
00:24:09,000 --> 00:24:10,079
where the sources leave us.

481
00:24:10,119 --> 00:24:12,440
Speaker 2: It's a really powerful quote that argues the value of

482
00:24:12,480 --> 00:24:14,720
sciences more than just material.

483
00:24:14,599 --> 00:24:18,720
Speaker 1: Yes, the concluding thought is that science has profoundly altered

484
00:24:18,759 --> 00:24:21,440
man's life, not just materially, but in the ways of

485
00:24:21,480 --> 00:24:24,200
the spirit. It's extended the range of questions we can

486
00:24:24,240 --> 00:24:27,000
even choose to answer, and it's extended our freedom to

487
00:24:27,000 --> 00:24:29,240
make significant decisions about our own future.

488
00:24:29,799 --> 00:24:33,519
Speaker 2: And the source ends with this powerful defense of free inquiry.

489
00:24:34,200 --> 00:24:37,680
It says no one can predict what vast new continents

490
00:24:37,680 --> 00:24:40,559
of knowledge the future of science will discover. But we

491
00:24:40,680 --> 00:24:42,200
know that as long as men are free to ask

492
00:24:42,279 --> 00:24:44,799
what they will, free to say what they think, free

493
00:24:44,880 --> 00:24:47,799
to think what they must, science will never regress, and

494
00:24:47,880 --> 00:24:50,279
freedom itself will never be wholly lost.

495
00:24:50,640 --> 00:24:53,200
Speaker 1: Which brings us to our final question for you, the listener,

496
00:24:53,240 --> 00:24:56,440
And it's based entirely on these competing narratives we've talked about.

497
00:24:56,559 --> 00:24:59,279
On one hand, the scientific community at Oakridge believes this

498
00:24:59,319 --> 00:25:02,400
AI push is for the greater good and open science.

499
00:25:02,759 --> 00:25:05,640
But on the other the political framing is entirely about

500
00:25:05,680 --> 00:25:06,559
an arms race.

501
00:25:06,960 --> 00:25:09,519
Speaker 2: So what's your stand on that duality? Do you think

502
00:25:09,559 --> 00:25:14,039
transformative technologies like AI are best developed in a fiercely

503
00:25:14,119 --> 00:25:17,799
competitive nationalistic race where the US has to out innovate

504
00:25:17,839 --> 00:25:21,039
everybody on the planet. Or does the core requirement of

505
00:25:21,119 --> 00:25:24,720
open science, which Frontier is literally built on, does that

506
00:25:24,759 --> 00:25:29,440
inherently demand global collaboration, even with strategic rivals, to ensure

507
00:25:29,480 --> 00:25:32,160
the safest and fastest progress for all of humanity.

508
00:25:32,559 --> 00:25:34,559
Speaker 1: Just think about the difference between the secrecy of the

509
00:25:34,720 --> 00:25:37,640
X ten reactor and the openness of Frontier. Which path

510
00:25:37,680 --> 00:25:39,960
do you think truly secures our future? Let us know

511
00:25:40,000 --> 00:25:41,640
what you think in the comments. We'll catch you next

512
00:25:41,680 --> 00:25:42,720
time on thrilling Threads.