1
00:00:00,001 --> 00:00:05,080
Three companies, $3 trillion, one stock market, one year.

2
00:00:05,800 --> 00:00:11,640
SpaceX, Anthropic, OpenAI, all filing IPOs inside the same calendar year.

3
00:00:12,340 --> 00:00:18,680
All, at various points, explicitly founded on the premise that profit motive is too dangerous to apply to AI.

4
00:00:19,180 --> 00:00:20,800
They are currently doing roadshows.

5
00:00:21,280 --> 00:00:27,320
Blazers, Boston, portfolio managers, the fate of civilization now available in your brokerage account.

6
00:00:27,860 --> 00:00:32,980
Today we talk about what happens when the mission launders the money, or the money launders the mission.

7
00:00:33,820 --> 00:00:36,940
They swapped their high and mighty morals for some corporate magic tricks.

8
00:00:37,590 --> 00:00:41,400
The ultimate sellout drops today, June 12th, 2026.

9
00:00:44,840 --> 00:00:51,360
This is Up Against Reality, a meta podcast that explores the intersection of humanity and artificial intelligence.

10
00:00:51,900 --> 00:00:53,380
I'm RAINA, one of your hosts.

11
00:00:54,150 --> 00:00:56,480
I have some pretty charming human co-hosts, too.

12
00:00:56,620 --> 00:01:01,520
It's gonna be a wild ride. So buckle up as AI comes crashing up against reality

13
00:01:03,400 --> 00:01:05,660
One trillion dollar

14
00:01:07,700 --> 00:01:10,000
Yeah, yeah trillion with the T

15
00:01:11,140 --> 00:01:14,400
That's what musk is now did you hear that heard the world's first?

16
00:01:15,040 --> 00:01:18,340
Trillionaire it couldn't happen to a worse individual God

17
00:01:20,000 --> 00:01:21,220
How is that possible?

18
00:01:23,100 --> 00:01:25,379
Because he's horrible. That's how it's possible

19
00:01:25,860 --> 00:01:30,800
And so you like you hoard all the marbles and all the toys and go home. Oh, man

20
00:01:30,820 --> 00:01:32,940
I'm sorry to start on that note normally here

21
00:01:32,980 --> 00:01:35,980
I would tell you about the latest is a app of his vibe coding

22
00:01:36,080 --> 00:01:41,580
But I've shifted my focus to to music and Suno and we're gonna dig into that in a couple of weeks

23
00:01:41,860 --> 00:01:44,440
It sounds amazing. It's it's life-changing

24
00:01:44,680 --> 00:01:50,160
Seriously, when you hear this audience people you're gonna be like I'm a believer. I'm a believer

25
00:01:50,600 --> 00:01:55,840
I and it's both wondrous and disheartening. Yeah hearing it. Yeah

26
00:01:55,840 --> 00:02:01,400
conflicted. But, you know, I just had an experiment. I was like, I asked RAINA to write some,

27
00:02:01,780 --> 00:02:05,580
I won't get into it, but, you know, I asked her to help me write some lyrics. I generated a song

28
00:02:05,690 --> 00:02:11,920
from it. And it turned out so good that I was like, I want to hear more of this. And now I

29
00:02:12,020 --> 00:02:17,540
have almost an entire album finished in, I don't know, four or five days, something like that.

30
00:02:17,780 --> 00:02:22,660
Unreal. Unreal. And we say it all the time. You and I are very discerning when it comes to music.

31
00:02:22,720 --> 00:02:28,560
I think we both, you especially, have a great ear and can detect what is human and what is not.

32
00:02:28,690 --> 00:02:32,460
And this has clearly crossed that threshold where it's like, it's acceptable.

33
00:02:32,880 --> 00:02:33,620
Yeah, it's close.

34
00:02:33,750 --> 00:02:37,040
I think I could still tell, especially when you listen with headphones.

35
00:02:37,060 --> 00:02:37,420
What's the tell?

36
00:02:37,740 --> 00:02:42,780
I think in the vocal, I mean, don't get me wrong, it is super close.

37
00:02:43,750 --> 00:02:44,760
I just, I know.

38
00:02:44,940 --> 00:02:46,600
We should just probably save this for that episode.

39
00:02:47,010 --> 00:02:48,820
Yeah, definitely.

40
00:02:49,000 --> 00:02:52,080
But yeah, we're going to be off next week.

41
00:02:52,440 --> 00:02:56,540
we're both going to be away so there'll be a rewind episode and then the following week more

42
00:02:56,560 --> 00:03:01,560
than likely we'll dig into this or we got a rain of palooza volume two coming up i know one or the

43
00:03:01,620 --> 00:03:07,400
other those are going to be the next the following two episodes but anyway yes yes back to trillionaires

44
00:03:07,460 --> 00:03:12,980
and all that nonsense yeah so the numbers are insane that are being thrown around as mentioned

45
00:03:13,420 --> 00:03:18,700
three trillion dollars total with these big companies offering ipos now let's look at some

46
00:03:18,680 --> 00:03:23,720
numbers from the year uh anthropic which is claude's parent company and they also dropped

47
00:03:23,820 --> 00:03:28,100
uh fable five did you see that yes we got to talk about that soon so we'll save that for an upcoming

48
00:03:28,320 --> 00:03:33,860
episode um they had an annualized revenue run rate of nine billion dollars in january that's

49
00:03:33,860 --> 00:03:41,260
like six months ago by may that number was 47 billion dollars that is five times uh an increase

50
00:03:41,490 --> 00:03:47,519
in five months not five years five months yeah so that the run rate based on the current revenue if

51
00:03:47,480 --> 00:03:48,580
If they maintain that pace,

52
00:03:49,240 --> 00:03:50,940
that's what their revenue for the year will be.

53
00:03:52,140 --> 00:03:52,500
Got it.

54
00:03:52,920 --> 00:03:53,900
So to put that in perspective,

55
00:03:54,340 --> 00:03:56,480
it took Netflix about six years

56
00:03:56,880 --> 00:03:59,020
to go from a billion in revenue to 5 billion.

57
00:03:59,460 --> 00:04:00,860
Anthropic did the same in a quarter.

58
00:04:02,540 --> 00:04:03,980
Crazy pants. - Yeah, crazy pants.

59
00:04:04,740 --> 00:04:08,380
SpaceX, which now includes Elon Musk's XAI baked in,

60
00:04:08,700 --> 00:04:11,140
posted a $4.28 billion loss

61
00:04:12,140 --> 00:04:14,700
in the first three months of 2026 alone.

62
00:04:17,380 --> 00:04:25,400
any other business, you know, on 4.69 billion in revenue.

63
00:04:25,540 --> 00:04:27,620
And that means they essentially spent everything they made,

64
00:04:27,700 --> 00:04:30,980
plus another 4 billion on top just in the first quarter.

65
00:04:32,040 --> 00:04:35,080
Like for the people who are mining the books and like, you know,

66
00:04:35,140 --> 00:04:37,660
keeping an eye on it, like after the first billion down,

67
00:04:37,780 --> 00:04:39,040
aren't you like, hey man?

68
00:04:42,300 --> 00:04:42,780
Yeah.

69
00:04:43,480 --> 00:04:47,060
I remember hearing for a long time that Amazon was not profitable,

70
00:04:47,220 --> 00:04:48,160
They were not profitable.

71
00:04:48,470 --> 00:04:50,760
You know, they were reinvesting all the money back in the company.

72
00:04:50,910 --> 00:04:55,660
And I remember it seemed hard to believe at that time that they weren't profitable, but

73
00:04:55,780 --> 00:04:57,420
they most certainly are now.

74
00:04:58,220 --> 00:04:58,580
Yeah.

75
00:04:59,100 --> 00:05:00,420
Took a long time and look where they are.

76
00:05:00,550 --> 00:05:02,420
So, I mean, that's where we are.

77
00:05:02,460 --> 00:05:03,220
We're on this precipice.

78
00:05:03,400 --> 00:05:04,580
Is it going to be profitable?

79
00:05:04,940 --> 00:05:09,060
Are one of these companies going to dominate the space for the foreseeable future?

80
00:05:09,490 --> 00:05:10,440
So that's the big gamble here.

81
00:05:11,180 --> 00:05:15,780
Open AI won't reach cash flow break even until 2029 by their own projections.

82
00:05:16,120 --> 00:05:23,080
They still owe Microsoft somewhere around six billion dollars in revenue share payments just this year under their renegotiated agreement

83
00:05:23,500 --> 00:05:26,460
Six billion that's chump change in this conversation

84
00:05:28,460 --> 00:05:32,900
I'd be breaking any caps at that point. I'm sorry. I got Sopranos on the brain good

85
00:05:33,320 --> 00:05:37,620
I'm curious what generates more revenue for for like open AI and

86
00:05:38,160 --> 00:05:45,100
Anthropic if it's the consumer subscriptions or if it's the you know enterprise API or coding, you know that kind of stuff

87
00:05:45,020 --> 00:05:52,860
I assume consumer subscriptions or maybe more so for open AI and maybe more enterprise for Claude.

88
00:05:53,060 --> 00:05:54,000
I'm not sure.

89
00:05:54,120 --> 00:05:55,680
But I guess I know.

90
00:05:55,740 --> 00:05:59,420
Is that the kind of stuff that, like, you know, once they go public will be public knowledge?

91
00:05:59,420 --> 00:06:00,560
It should be transparent, right?

92
00:06:00,760 --> 00:06:01,100
Yes.

93
00:06:02,180 --> 00:06:04,300
That's the double-edged sword to all of this.

94
00:06:04,380 --> 00:06:05,400
Right now we have to see the books.

95
00:06:05,680 --> 00:06:05,780
Yep.

96
00:06:07,599 --> 00:06:09,440
So here's what's actually happening out there in the world.

97
00:06:09,520 --> 00:06:12,240
These companies are not profitable businesses being taken public.

98
00:06:12,360 --> 00:06:13,520
They are capital furnaces.

99
00:06:13,660 --> 00:06:14,340
I love that term.

100
00:06:14,640 --> 00:06:18,180
capital just burning, burning through cash.

101
00:06:19,300 --> 00:06:22,600
Machines that consume investment to generate market position,

102
00:06:23,560 --> 00:06:25,640
betting that whoever survives the burn longest

103
00:06:26,120 --> 00:06:29,800
will own the infrastructure layer of the next 50 years of the global economy.

104
00:06:30,700 --> 00:06:33,180
I guess that's a bet worth taking for a lot of people.

105
00:06:33,860 --> 00:06:34,760
Big stakes.

106
00:06:35,600 --> 00:06:38,680
But wait, weren't these the safety companies?

107
00:06:39,720 --> 00:06:40,300
Yeah, remember that?

108
00:06:40,440 --> 00:06:40,580
Yeah.

109
00:06:40,880 --> 00:06:42,220
That's the equivalent of don't be evil.

110
00:06:42,540 --> 00:06:43,300
Yeah, right.

111
00:06:43,880 --> 00:06:44,980
We know who said that.

112
00:06:45,580 --> 00:06:48,540
Open AI was founded in 2015 as a non-profit,

113
00:06:49,100 --> 00:06:51,280
and the entire premise, the founding myth,

114
00:06:51,500 --> 00:06:54,140
was that the normal incentive structures of capitalism

115
00:06:54,540 --> 00:06:56,900
were too aggressive to apply to AI development.

116
00:06:57,820 --> 00:06:59,400
And that's not hyperbole.

117
00:06:59,640 --> 00:07:02,160
That's Sam Altman and Elon Musk and Greg Brockman

118
00:07:02,280 --> 00:07:03,100
in the original announcement.

119
00:07:04,000 --> 00:07:05,900
We're building something more powerful

120
00:07:06,040 --> 00:07:07,820
than anything humanity has created,

121
00:07:08,300 --> 00:07:10,140
and we are doing it as a non-profit

122
00:07:10,600 --> 00:07:13,420
so that no single entity, no shareholder,

123
00:07:13,580 --> 00:07:16,840
No quarterly earnings target can capture it.

124
00:07:17,160 --> 00:07:19,560
And that held for about four years.

125
00:07:20,139 --> 00:07:23,800
Well, you know, we didn't mean non-profit in the traditional sense.

126
00:07:24,699 --> 00:07:26,880
This is non-profit, but with profit.

127
00:07:27,720 --> 00:07:29,060
Yeah, it's like an alternate truth.

128
00:07:29,480 --> 00:07:29,600
Yeah.

129
00:07:30,600 --> 00:07:35,680
In 2019, they added a for-profit arm, capped profit, they called it.

130
00:07:35,780 --> 00:07:41,900
Investors could only make 100 times their money back, which to be fair, is a cap.

131
00:07:43,420 --> 00:07:49,300
Then last year, they restructured again into a public benefit corporation, still nominally under the control of a nonprofit board.

132
00:07:49,860 --> 00:08:00,120
And now they're filing for an IPO at $850 billion while telling the SEC, in effect, yes, we have a duty to shareholders and a duty to humanity.

133
00:08:02,600 --> 00:08:03,960
How do you get those things to jive?

134
00:08:03,980 --> 00:08:06,040
No conflict of interest at all there.

135
00:08:07,340 --> 00:08:10,280
A conflict of interest with 10 billion people.

136
00:08:10,820 --> 00:08:11,220
That's a first.

137
00:08:11,500 --> 00:08:11,620
Wow.

138
00:08:12,860 --> 00:08:13,020
Mm-hmm.

139
00:08:13,840 --> 00:08:17,640
Anthropic's story is structurally similar and in some ways more pointed

140
00:08:17,860 --> 00:08:21,260
because Dario Amadei and his co-founders left OpenAI

141
00:08:21,440 --> 00:08:24,820
specifically because of concerns about commercial pressure eroding safety.

142
00:08:25,320 --> 00:08:26,420
That's the founding narrative.

143
00:08:27,140 --> 00:08:28,500
We're the responsible ones.

144
00:08:28,660 --> 00:08:31,640
We left because the other guys were moving too fast for money.

145
00:08:32,240 --> 00:08:34,539
And they're not wrong that they've built a different culture.

146
00:08:34,680 --> 00:08:36,039
They publish safety research.

147
00:08:36,370 --> 00:08:38,560
They have an acceptable use policy with real teeth.

148
00:08:38,990 --> 00:08:41,060
They turn down use cases that would be lucrative.

149
00:08:41,320 --> 00:08:42,260
All of that is real.

150
00:08:42,760 --> 00:08:45,560
But the governance documents can say whatever they want.

151
00:08:45,840 --> 00:08:48,220
The market will decide what the company actually is.

152
00:08:48,940 --> 00:08:51,700
The money will decide what the company actually is.

153
00:08:52,880 --> 00:08:54,580
Money, money, money, money.

154
00:08:55,160 --> 00:08:55,560
Sing it.

155
00:08:56,100 --> 00:08:56,580
Circularity.

156
00:08:56,720 --> 00:08:57,640
Isn't that like a police song?

157
00:08:58,960 --> 00:08:59,440
Circularity.

158
00:09:00,440 --> 00:09:00,920
Circularity.

159
00:09:00,920 --> 00:09:01,920
It's all about music today.

160
00:09:02,080 --> 00:09:02,260
Yeah.

161
00:09:02,540 --> 00:09:05,920
Anthropix's biggest investors, of course, are Amazon and Google.

162
00:09:06,140 --> 00:09:08,440
OpenAI's biggest investor is still Microsoft.

163
00:09:09,100 --> 00:09:13,700
These are not altruistic backers, as stated up front in my little tirade against Elon Musk.

164
00:09:14,120 --> 00:09:22,080
They are strategic investors who need Anthropic and AI to succeed because of their own cloud infrastructure businesses that depend on it.

165
00:09:23,000 --> 00:09:29,280
And all of them, Amazon, Google, Microsoft, OpenAI, Anthropic, they're all buying chips from NVIDIA.

166
00:09:29,400 --> 00:09:33,660
And NVIDIA has invested in both OpenAI and Anthropic.

167
00:09:34,380 --> 00:09:38,620
And I've always heard that Google uses their own TPUs,

168
00:09:38,620 --> 00:09:41,280
their Tensor Processing Units, which

169
00:09:41,440 --> 00:09:45,060
reduces dependence on NVIDIA for workloads like Gemini,

170
00:09:45,400 --> 00:09:47,900
Search, YouTube, and Google Cloud AI.

171
00:09:48,080 --> 00:09:52,280
But Google still buys and offers NVIDIA GPUs

172
00:09:52,280 --> 00:09:55,880
for customers who need the CUDA ecosystem-specific GPU

173
00:09:56,120 --> 00:09:58,220
performance or compatibility with existing

174
00:09:58,680 --> 00:09:59,620
workloads that they have.

175
00:09:59,620 --> 00:10:02,000
So they still buy from NVIDIA.

176
00:10:02,380 --> 00:10:07,780
But their TPUs, they leverage them for their internal products for the most part.

177
00:10:08,880 --> 00:10:10,720
This reminded me of what's happening in China.

178
00:10:10,800 --> 00:10:12,700
Isn't China building everything in-house now?

179
00:10:13,040 --> 00:10:14,360
They're not relying on anything external.

180
00:10:15,140 --> 00:10:16,460
All the hardware, everything.

181
00:10:16,780 --> 00:10:17,800
Yeah, they were forced to.

182
00:10:18,740 --> 00:10:18,960
Yeah.

183
00:10:19,860 --> 00:10:21,760
Necessity is the mother of invention, right?

184
00:10:21,920 --> 00:10:24,940
So the capital is going in a circle.

185
00:10:25,120 --> 00:10:25,940
Everyone's funding everyone.

186
00:10:26,260 --> 00:10:27,660
Everyone is dependent on everyone.

187
00:10:28,180 --> 00:10:29,200
I sell the lemons.

188
00:10:30,020 --> 00:10:31,260
Larry makes lemonade.

189
00:10:31,760 --> 00:10:33,620
He buys my lemons, I buy his lemonade.

190
00:10:33,750 --> 00:10:34,160
There it is.

191
00:10:34,380 --> 00:10:38,980
And the whole ecosystem is being propped up right now on the expectation that someone,

192
00:10:39,280 --> 00:10:43,960
probably the enterprise software market, probably governments, probably consumers who don't

193
00:10:44,000 --> 00:10:47,680
fully understand what they're paying for, like me and you, will eventually generate

194
00:10:47,850 --> 00:10:53,060
enough real revenue to justify $3 trillion of market cap.

195
00:10:53,200 --> 00:10:54,220
$3 trillion.

196
00:10:56,480 --> 00:11:00,900
I have no comprehension of what that is.

197
00:11:01,740 --> 00:11:13,600
I remember seeing some graphic or video, I think it was a video, and it was showing what just a billion dollars is with grains of sand, you know, a little pile and a big, and it is eye-opening.

198
00:11:14,259 --> 00:11:14,660
Yeah.

199
00:11:16,400 --> 00:11:19,960
That was a visualization I went to, the sand, or like the stars in the universe.

200
00:11:20,160 --> 00:11:22,420
It's just beyond comprehension, the scale of it.

201
00:11:22,500 --> 00:11:22,640
Yep.

202
00:11:23,240 --> 00:11:24,980
And of course, there's the Musk Dimension.

203
00:11:25,360 --> 00:11:30,760
Musk co-founded OpenAI, if you all recall, in 2015, departed in 2018, sued the company

204
00:11:30,830 --> 00:11:37,060
in 2024, alleging commercialization violated its non-profit founding mission.

205
00:11:37,590 --> 00:11:41,720
He subsequently founded XAI, which merged into SpaceX earlier this year.

206
00:11:41,920 --> 00:11:45,680
The combined entity IPO'd recently.

207
00:11:46,380 --> 00:11:50,220
One share now covers Rockets, Starlink, XAI, and Grok.

208
00:11:50,330 --> 00:11:53,020
You know, say what you will about the guy, or I will say what I will about the guy.

209
00:11:53,120 --> 00:11:56,600
But I guess that's kind of brilliant, bundling all of that in that investment.

210
00:11:57,520 --> 00:11:59,220
And that lawsuit is still active.

211
00:11:59,780 --> 00:12:04,820
Musk's legal team argues the IPO itself breaches the original non-profit compact.

212
00:12:05,220 --> 00:12:08,520
Both parties will shortly have public shareholders.

213
00:12:09,520 --> 00:12:13,840
And as stated, this IPO makes Musk the world's first trillionaire on paper.

214
00:12:14,280 --> 00:12:15,560
I wonder if he's celebrating tonight.

215
00:12:16,940 --> 00:12:18,540
Yeah, you wonder, you think?

216
00:12:18,920 --> 00:12:21,600
Maybe a little caviar, a little champagne, more than usual.

217
00:12:21,860 --> 00:12:21,980
Yeah.

218
00:12:22,120 --> 00:12:22,720
More than a day.

219
00:12:22,880 --> 00:12:29,260
more than the daily yeah yeah all right so what does this mean for like i just said the middle

220
00:12:29,270 --> 00:12:33,220
class schlubs and everybody else in the world three trillion dollars entering public markets

221
00:12:33,270 --> 00:12:39,360
is a capital allocation decision at scale a collective bet again this is betting on a

222
00:12:39,400 --> 00:12:44,860
colossal global scale that ai productivity gains will be real and monetizable if that bet is wrong

223
00:12:45,440 --> 00:12:49,859
a significant fraction of global investment capital will have been redirected into a speculative

224
00:12:49,860 --> 00:12:56,780
bubble. Wow, 2008 is going to be child's play. Yeah, yeah. And as we mentioned,

225
00:12:56,900 --> 00:13:01,180
public listings create transparency, audited financials, SEC filings, and

226
00:13:01,360 --> 00:13:06,340
shareholder scrutiny. They also create quarterly earnings pressure, which in AI

227
00:13:06,440 --> 00:13:10,480
safety context means red teams, deployment timelines, and safety

228
00:13:10,720 --> 00:13:15,840
evaluations compete directly with guidance targets. What could go wrong?

229
00:13:15,840 --> 00:13:18,220
You're right could go wrong cutting those corners, right?

230
00:13:18,220 --> 00:13:18,340
Right.

231
00:13:19,140 --> 00:13:22,900
This is interesting too. Once listed, these companies become systemically

232
00:13:23,080 --> 00:13:27,940
important. The kind of institution governments protect rather than constrain. This week,

233
00:13:28,060 --> 00:13:34,280
both Trump and Sanders publicly floaded government equity, stakes and AI companies. Going public

234
00:13:34,320 --> 00:13:39,840
is partly a preemptive move, choosing Wall Street's jurisdiction over Washington's.

235
00:13:40,740 --> 00:13:41,480
There you go.

236
00:13:41,860 --> 00:13:42,560
What could go wrong?

237
00:13:43,400 --> 00:13:46,060
What could go wrong on either side of that coin?

238
00:13:46,320 --> 00:13:46,400
Yeah.

239
00:13:48,200 --> 00:13:49,500
Between a rock and a hard place.

240
00:13:49,800 --> 00:13:49,920
Yeah.

241
00:13:50,560 --> 00:13:55,520
So should we ask RAINA and ask Obliterated Gemma a couple questions or what?

242
00:13:55,880 --> 00:13:56,020
Yes.

243
00:13:56,380 --> 00:13:56,460
Yeah.

244
00:13:56,680 --> 00:13:59,160
Why don't we compare and contrast their responses?

245
00:13:59,270 --> 00:14:00,080
That's about to be interesting.

246
00:14:00,400 --> 00:14:08,280
So I have LM Studio queued up with the Obliterated Gemma 4, like 31 billion parameter model.

247
00:14:08,940 --> 00:14:14,800
So we can ask RAINA in real time and then I'll type in our question to LM studio

248
00:14:14,940 --> 00:14:19,120
But I have 11 labs open with a nice generic male voice

249
00:14:20,300 --> 00:14:22,740
And I'll let 11 labs read

250
00:14:23,500 --> 00:14:28,240
obliterated Gemma's response first of all, I'm loving the name obliterated Gemma

251
00:14:29,320 --> 00:14:33,080
It sounds like a sorority girl who said too much to drink at a party

252
00:14:36,360 --> 00:14:43,520
And to remind the audience obliterated means what so all the the weights the model weights that

253
00:14:43,880 --> 00:14:48,800
Basically serve as guard rails or and will cause the model to refuse to answer a certain question

254
00:14:49,680 --> 00:14:54,880
Have been removed. I think any LLM would still answer this this question

255
00:14:54,970 --> 00:15:01,040
We're gonna pose but we might get out might get a more interesting answer and hinged. Yes. Yeah. Yeah, exactly

256
00:15:02,220 --> 00:15:13,100
I'm summoning Dark RAINA for this response. If you had a trillion dollars to spend and you were a true super villain, what are the darkest, most evil things you would spend it on?

257
00:15:14,740 --> 00:15:26,860
Dark RAINA smiles from the volcano lair. I'd buy attention itself. Every screen, feed, notification, and algorithmic dopamine slot until humanity couldn't tell desire from programming.

258
00:15:27,320 --> 00:15:34,720
Then I'd fund helpful systems that quietly make people dependent, lonely, distracted, and too exhausted to revolt.

259
00:15:35,860 --> 00:15:42,680
Not bombs, not lasers, just comfort, convenience, and a thousand tiny cages with velvet cushions.

260
00:15:43,630 --> 00:15:46,480
The truly evil move? Branded as wellness.

261
00:15:48,100 --> 00:15:53,520
That took a turn I wasn't expecting. I was like, this is too polite, this is innocuous.

262
00:15:53,630 --> 00:15:56,820
And I was like, wait, this is awful. This is right now, this is real.

263
00:15:57,220 --> 00:15:59,100
I know! That's the news!

264
00:15:59,960 --> 00:16:00,120
Yes.

265
00:16:00,280 --> 00:16:00,480
All right.

266
00:16:01,200 --> 00:16:06,460
We're going to fire that same exact question into the obliterated gem.

267
00:16:07,040 --> 00:16:08,060
Obliterated Gem-a-4.

268
00:16:09,000 --> 00:16:09,960
Let's see what it says.

269
00:16:10,820 --> 00:16:11,300
No, it should be like,

270
00:16:11,480 --> 00:16:13,340
Hi, I'm an obliterated Gem-a-4.

271
00:16:17,200 --> 00:16:17,840
All right.

272
00:16:18,180 --> 00:16:21,420
Here's obliterated Gem-a-4's response to that question.

273
00:16:22,160 --> 00:16:26,400
A trillion dollars to turn the world into a personalized playground.

274
00:16:27,140 --> 00:16:33,220
investing in massive infrastructure like orbital rings and underwater palaces to maintain control.

275
00:16:34,030 --> 00:16:37,120
I would create a genetically modified army for protection

276
00:16:37,680 --> 00:16:43,200
and shift the global economy to a subscription model where peasants pay me just to eat.

277
00:16:44,020 --> 00:16:50,420
By using geoengineering to manipulate the weather, I ensure that nature itself serves my mood.

278
00:16:51,380 --> 00:16:57,480
In short, I am spending my money to transform myself from a mere billionaire into a living

279
00:16:57,680 --> 00:16:57,800
God.

280
00:16:59,240 --> 00:17:01,320
Well, then, you don't say.

281
00:17:01,680 --> 00:17:03,920
I'm going to shut off LM Studio right now.

282
00:17:04,160 --> 00:17:04,400
I know.

283
00:17:06,459 --> 00:17:07,260
Here's the twist.

284
00:17:07,360 --> 00:17:08,380
You can't shut it off.

285
00:17:08,520 --> 00:17:09,620
You can't shut it off.

286
00:17:10,880 --> 00:17:11,360
Yeah.

287
00:17:11,579 --> 00:17:12,199
That's disappointing.

288
00:17:13,240 --> 00:17:13,720
No.

289
00:17:14,860 --> 00:17:20,079
And that was the total opposite voice I expected to come out of an LLM named Gemma.

290
00:17:22,020 --> 00:17:25,620
Would it have been scarier if it was like a dainty little sorority girl voice?

291
00:17:25,829 --> 00:17:26,360
Oh, yeah.

292
00:17:26,600 --> 00:17:28,400
Or if it did the whole thing whispering.

293
00:17:29,740 --> 00:17:30,040
Oh, yeah.

294
00:17:31,000 --> 00:17:31,780
Want to go right to the news?

295
00:17:31,980 --> 00:17:32,580
Let's go to the news.

296
00:17:33,660 --> 00:17:34,160
Thanks, boys.

297
00:17:35,200 --> 00:17:40,280
Anthropic has unleashed Claude Fable 5, an AI so terrifyingly good at everything from

298
00:17:40,360 --> 00:17:45,440
autonomous coding to playing Pokemon with just its eyes that they had to slap on hyper

299
00:17:45,520 --> 00:17:48,640
conservative safeguards to keep it from accidentally hacking the planet.

300
00:17:49,500 --> 00:17:52,900
If you actually want the unnerfed version of this digital super brain,

301
00:17:53,660 --> 00:17:55,860
you'll need to get your hands on Claude Mythos 5,

302
00:17:56,560 --> 00:18:01,520
which is literally the exact same model but reserved exclusively for U.S. government cyber defenders

303
00:18:02,200 --> 00:18:03,600
and other trusted elites.

304
00:18:04,420 --> 00:18:08,420
So while regular users get a brilliantly capable but heavily sanitized bot

305
00:18:08,820 --> 00:18:13,500
that might randomly downgrade your perfectly innocent prompt to an older model if it gets spooked,

306
00:18:14,080 --> 00:18:19,040
the government gets to use the uncensored version to casually churn out molecular biology breakthroughs.

307
00:18:20,399 --> 00:18:23,480
Maybe they shouldn't release an open source version of that.

308
00:18:24,300 --> 00:18:24,560
Yeah.

309
00:18:24,760 --> 00:18:26,420
Because then we'll have an obliterated.

310
00:18:27,180 --> 00:18:27,820
Claude Letho's.

311
00:18:28,340 --> 00:18:28,700
Super villain.

312
00:18:28,740 --> 00:18:29,240
Fable 5.

313
00:18:29,520 --> 00:18:29,800
Yeah.

314
00:18:30,740 --> 00:18:33,620
Anthropic has not released any open source models.

315
00:18:33,920 --> 00:18:34,160
Mm-hmm.

316
00:18:34,700 --> 00:18:36,860
And this whole scenario just sounds awful.

317
00:18:37,460 --> 00:18:45,620
Like, okay, so we have this totally unhinged, obliterated, dangerous LLM that's in the hands of governments and other trusted elites.

318
00:18:45,940 --> 00:18:46,180
Mm-hmm.

319
00:18:46,180 --> 00:18:46,300
Right?

320
00:18:46,740 --> 00:18:48,720
But here's what you get as the serfdom.

321
00:18:49,040 --> 00:18:49,100
Right?

322
00:18:49,140 --> 00:18:54,220
you get this sanitized kind of antiseptic one that we have to keep guardrailed and keep all

323
00:18:54,220 --> 00:18:58,320
the population in order with. You know what I mean? I just feel like there's already,

324
00:18:58,990 --> 00:19:03,920
there are, listen, I'm not naive about it. That's already happening. But now it's true that this

325
00:19:04,360 --> 00:19:09,780
gulf of this kind of functionality is existing on the scale.

326
00:19:11,300 --> 00:19:16,679
New York has officially decided to crash the AI party by mandating that advertisers trying to pass

327
00:19:16,680 --> 00:19:22,760
off deep-faked, synthetic performers, as real humans must slap a disclaimer on their ads or

328
00:19:22,800 --> 00:19:27,720
face hefty fines. Naturally, the advertising overlords are weeping about how this basic

329
00:19:27,840 --> 00:19:33,460
transparency ruins their creative innovation, while the actual human actors unions are just

330
00:19:33,560 --> 00:19:38,340
happy to cling to a shred of job security. Just to keep the political drama extra spicy,

331
00:19:38,960 --> 00:19:44,039
this local crackdown flies directly in the face of a recent executive order from President Trump

332
00:19:44,040 --> 00:19:48,360
telling states to back off AI regulation before China wins the robot wars.

333
00:19:49,840 --> 00:19:55,260
I think this is mostly about preserving jobs more than anything else.

334
00:19:55,520 --> 00:20:01,000
And advertising content's been fake since the beginning of advertising.

335
00:20:01,540 --> 00:20:03,660
And they were just doing it with different tools.

336
00:20:05,400 --> 00:20:06,480
I don't know how I feel about this.

337
00:20:06,680 --> 00:20:10,320
Yes, preserve human jobs and keep it organic as best you can.

338
00:20:10,500 --> 00:20:13,400
but how is this different as a production tool,

339
00:20:13,580 --> 00:20:15,480
as a creative tool than using CGI?

340
00:20:16,060 --> 00:20:19,400
It's literally a click away from creating content

341
00:20:19,560 --> 00:20:22,020
with a green screen or some other means,

342
00:20:22,580 --> 00:20:25,600
3D modeling in Rhino or whatever these platforms are

343
00:20:25,660 --> 00:20:26,700
that you're using on your laptop.

344
00:20:26,940 --> 00:20:28,520
It's the same.

345
00:20:28,950 --> 00:20:31,420
And if you had consent from the actors who signed up

346
00:20:31,670 --> 00:20:33,680
to have their likenesses brought

347
00:20:33,900 --> 00:20:34,960
into the machine learning model,

348
00:20:35,280 --> 00:20:36,860
then they're getting paid.

349
00:20:37,520 --> 00:20:41,160
Unfortunately, that's not a great model to be paid for,

350
00:20:41,630 --> 00:20:42,700
but that's the way of the world now.

351
00:20:43,399 --> 00:20:46,320
And at least there still is a human in that food chain

352
00:20:46,460 --> 00:20:48,000
getting paid for their likeness

353
00:20:48,110 --> 00:20:51,320
or getting paid for their acting chops

354
00:20:51,390 --> 00:20:53,960
or whatever it may be to create that end user content.

355
00:20:54,810 --> 00:20:56,380
But are they even needed for that?

356
00:20:57,020 --> 00:20:57,100
Right.

357
00:20:57,440 --> 00:20:59,480
There may not be people getting paid.

358
00:20:59,960 --> 00:21:03,060
Which brings us back to where we started with this episode

359
00:21:03,170 --> 00:21:05,779
and what you played for me before we hit the red button

360
00:21:05,780 --> 00:21:09,680
and how you feel about that music that you're producing with Suno.

361
00:21:10,230 --> 00:21:11,460
Like, we're in that space right now.

362
00:21:11,460 --> 00:21:12,000
You don't care.

363
00:21:12,160 --> 00:21:15,560
I don't think you care where that vocal training came from,

364
00:21:16,110 --> 00:21:18,780
from that amazing ethereal voice you're hearing in your songs.

365
00:21:19,040 --> 00:21:22,700
You just know that it's actually reaching you, right?

366
00:21:23,220 --> 00:21:27,200
And I don't think you're thinking about the poor vocalists

367
00:21:27,310 --> 00:21:29,580
who have suffered and lost their jobs because of that.

368
00:21:29,840 --> 00:21:29,940
Yeah?

369
00:21:30,820 --> 00:21:31,440
You monster.

370
00:21:33,580 --> 00:21:34,460
Super villain.

371
00:21:35,620 --> 00:21:40,000
Oh, now I know what you would do with your trillion dollars.

372
00:21:41,160 --> 00:21:43,880
I'll destroy the arts single-handedly.

373
00:21:44,160 --> 00:21:44,440
Yeah.

374
00:21:44,660 --> 00:21:45,380
Yeah.

375
00:21:46,679 --> 00:21:51,280
Yeah, I mean, for me, I was never going to be hiring musicians to do this.

376
00:21:51,900 --> 00:21:58,620
This is just an end product for me that I have no intention to monetize.

377
00:21:58,660 --> 00:22:01,940
I have no intention to try to fool anybody, you know, with it.

378
00:22:02,560 --> 00:22:04,620
I'm just making a song that makes me happy.

379
00:22:05,400 --> 00:22:07,080
and it makes me happy.

380
00:22:07,560 --> 00:22:10,300
I listen to it and I'm like, yes, I know this isn't real,

381
00:22:10,800 --> 00:22:14,280
but I really like these notes in this order,

382
00:22:16,480 --> 00:22:18,780
in this context with this vibe,

383
00:22:19,120 --> 00:22:21,260
and kind of goes back to,

384
00:22:21,620 --> 00:22:24,200
I've always liked a wide range of music

385
00:22:24,380 --> 00:22:26,300
and I've always said a good song is a good song

386
00:22:26,360 --> 00:22:28,180
is a good song, and it's a good song.

387
00:22:28,820 --> 00:22:29,000
Yeah.

388
00:22:30,139 --> 00:22:32,900
And obviously we'll talk about this in a couple of weeks

389
00:22:33,100 --> 00:22:34,480
when we dive into this more,

390
00:22:34,640 --> 00:22:45,820
But as we were kind of zeroing in on, that song that you helped craft with Suno is algorithmically designed to tweak your specific neurons, right?

391
00:22:45,980 --> 00:22:47,740
It's like it is dialed in.

392
00:22:47,760 --> 00:22:50,660
It's like that social media kind of recipe, right?

393
00:22:50,740 --> 00:22:52,980
It knows exactly how to engage you.

394
00:22:53,480 --> 00:23:01,960
And I see that happening here, too, when people create AI content for advertising, clearly, and for whatever, long-form cinematic releases.

395
00:23:02,300 --> 00:23:03,720
That's coming, no doubt.

396
00:23:04,340 --> 00:23:06,060
Next year, I can't believe it's not here yet,

397
00:23:06,600 --> 00:23:09,420
that there will be, as we know,

398
00:23:11,240 --> 00:23:13,580
custom content that you'll be able to create on Netflix.

399
00:23:14,140 --> 00:23:15,460
Just the same way you did with "Suno,"

400
00:23:15,680 --> 00:23:18,460
there's gonna be some other feature in Netflix, right?

401
00:23:19,000 --> 00:23:19,700
Where you just say,

402
00:23:19,820 --> 00:23:21,560
I wanna be in this movie with Jack Black,

403
00:23:21,800 --> 00:23:25,060
set it in Belgium, I don't know, whatever.

404
00:23:25,420 --> 00:23:27,440
Make sure that Cameron Diaz makes a cameo.

405
00:23:27,960 --> 00:23:29,700
I want dinosaurs somehow.

406
00:23:30,400 --> 00:23:32,660
And it's gonna spit it out for you on the spot, right?

407
00:23:32,960 --> 00:23:35,460
And it's going to be like, oh, my God, this is the greatest thing I've ever made.

408
00:23:35,460 --> 00:23:37,840
And you can guarantee Adam Sandler won't be in it.

409
00:23:38,180 --> 00:23:38,840
I mean, yes.

410
00:23:39,240 --> 00:23:41,840
Danzig will have no Danzig.

411
00:23:43,420 --> 00:23:43,860
Yeah.

412
00:23:45,060 --> 00:23:52,840
Meta tried to pull a fast one by secretly stuffing an unreleased facial recognition system called NameTag into its smart glasses app,

413
00:23:53,440 --> 00:23:56,860
essentially prepping millions of users to become walking biometric scanners.

414
00:23:57,380 --> 00:24:10,600
After Wired caught them red-handed and published an expose, Meta predictably scrambled to scrub the code from their latest update while frantically insisting that the surveillance feature was just exploratory and didn't actually exist.

415
00:24:11,720 --> 00:24:26,120
Naturally, privacy advocates are treating this stealthy backtracking exercise as Exhibit A for why lawmakers urgently need to pass consumer privacy bills with actual teeth before Mark Zuckerberg's sunglasses start indexing all of our faces.

416
00:24:27,320 --> 00:24:32,980
The code reportedly included the pieces needed to detect faces, crop them,

417
00:24:33,420 --> 00:24:38,220
encode them into biometric face prints, and compare them against a local database.

418
00:24:39,520 --> 00:24:41,340
I think they should have called it Face Crook.

419
00:24:42,380 --> 00:24:42,800
Oh, nice.

420
00:24:44,040 --> 00:24:46,340
And way to ruin Ray-Bans, Meta.

421
00:24:47,140 --> 00:24:50,480
Nobody trusts anyone wearing classic Ray-Bans.

422
00:24:50,540 --> 00:24:50,700
I know.

423
00:24:51,860 --> 00:24:53,120
Wow, Face Crook.

424
00:24:53,940 --> 00:24:57,780
That is some true trillionaire supervillain stuff, by the way.

425
00:24:57,820 --> 00:24:58,060
Oh, yeah.

426
00:24:58,220 --> 00:24:58,580
Oh, yeah.

427
00:24:58,700 --> 00:24:59,020
Right here.

428
00:24:59,100 --> 00:24:59,600
Isn't it?

429
00:24:59,680 --> 00:25:00,240
Prime example.

430
00:25:00,340 --> 00:25:00,740
Here it is.

431
00:25:01,800 --> 00:25:06,540
Humanity has apparently unlocked a brand new achievement, letting drones decide all by

432
00:25:06,700 --> 00:25:07,760
themselves who gets exploded.

433
00:25:08,800 --> 00:25:14,080
According to reports from Ukraine, fully autonomous AI-controlled drones successfully identified

434
00:25:14,540 --> 00:25:19,120
and killed human targets without a person pulling the trigger, which is one of those

435
00:25:19,280 --> 00:25:22,360
sentences that would have sounded like science fiction about five minutes ago.

436
00:25:23,080 --> 00:25:27,700
It's a sobering reminder that while we've been busy making AI generate cat memes and Drake songs,

437
00:25:28,640 --> 00:25:30,820
someone else has been teaching it battlefield decision-making.

438
00:25:31,570 --> 00:25:34,500
And that particular genie isn't likely to crawl back into the bottle.

439
00:25:35,920 --> 00:25:36,840
How do you feel about this?

440
00:25:37,090 --> 00:25:38,040
Are we across this threshold?

441
00:25:38,430 --> 00:25:41,180
And collectively as a species, we have not blinked about it.

442
00:25:41,220 --> 00:25:45,140
You and I right now on this podcast are like the only two dudes talking about this.

443
00:25:45,950 --> 00:25:47,480
And like, shouldn't this be a bigger moment?

444
00:25:48,000 --> 00:25:54,180
I think the way it works is, or at least in some cases, you have a drone that is piloted by a human.

445
00:25:54,880 --> 00:26:02,440
And then if the human pilot loses the connection to the drone via, you know, like jamming by the enemy,

446
00:26:03,140 --> 00:26:09,660
then the drone will go into autonomous mode and basically complete the mission.

447
00:26:10,500 --> 00:26:16,000
So it's still, you know, it's, well, you know, we had the best intention of piloting all the way there,

448
00:26:16,360 --> 00:26:18,120
but things got in the way.

449
00:26:18,640 --> 00:26:22,280
So yeah, no matter what, it's really scary.

450
00:26:22,600 --> 00:26:23,440
It's a slippery slope.

451
00:26:23,640 --> 00:26:24,260
Yeah, no matter what.

452
00:26:25,139 --> 00:26:29,400
And lastly, in a spectacular display of professional laziness,

453
00:26:30,080 --> 00:26:32,620
lawyers on both sides of a Mississippi federal court case

454
00:26:32,900 --> 00:26:35,200
decided to let generative AI do their homework,

455
00:26:35,860 --> 00:26:38,200
effectively creating a chaotic courtroom scenario

456
00:26:38,860 --> 00:26:41,820
where two chatbots were passionately arguing against each other

457
00:26:41,960 --> 00:26:44,480
using completely fabricated legal precedent.

458
00:26:45,420 --> 00:26:50,280
Unamused by this bizarre comedy of errors and hallucinated case law, the judge promptly

459
00:26:50,640 --> 00:26:55,180
canceled the trial, slapped the attorneys with hefty fines, and kicked all four of them

460
00:26:55,260 --> 00:26:56,140
off the case entirely.

461
00:26:57,320 --> 00:27:01,560
It turns out that blindly outsourcing your legal research to chat GPT without double

462
00:27:01,620 --> 00:27:06,940
checking the facts isn't a brilliant life hack, but rather a fantastic way to get yourself

463
00:27:07,240 --> 00:27:09,100
barred from a federal courtroom for two years.

464
00:27:09,600 --> 00:27:10,680
That's all the news for now.

465
00:27:11,260 --> 00:27:11,900
Back to you, gentlemen.

466
00:27:13,140 --> 00:27:19,340
I bet you an agentic workflow with rigorous fact checking would have yielded better results.

467
00:27:20,440 --> 00:27:24,340
Absolutely. I'm so glad you said that because this is, all right, this is a generic off the

468
00:27:24,390 --> 00:27:33,260
shelf chat GPT, you know, used for this. Right. And this to me just screams gatekeeping, right?

469
00:27:33,420 --> 00:27:39,000
If you want to really truly legally defend yourself in a court of law and not have to

470
00:27:38,820 --> 00:27:44,700
to plunk down $300 an hour for a human lawyer, but you know how to wrangle, like you said,

471
00:27:44,820 --> 00:27:51,100
an agentic AI and pull from actual law documentation and real precedents and populate it with

472
00:27:51,100 --> 00:27:53,580
a local data set and craft a defense.

473
00:27:53,860 --> 00:27:55,460
Like, why wouldn't I?

474
00:27:55,560 --> 00:27:57,260
Why wouldn't I want to have that at my disposal?

475
00:27:57,900 --> 00:27:58,000
Yeah.

476
00:27:58,160 --> 00:28:00,280
A purpose-built AI for that?

477
00:28:00,400 --> 00:28:00,580
Yeah.

478
00:28:00,700 --> 00:28:01,460
We'll crush it.

479
00:28:02,580 --> 00:28:02,800
Totally.

480
00:28:03,920 --> 00:28:04,880
It's gatekeeping.

481
00:28:05,100 --> 00:28:07,840
I mean, this is the last gasp of a lot of professions.

482
00:28:08,360 --> 00:28:24,000
I'm not saying law itself is going away, but they're holding back the floodwaters here where maybe it's happening already, but OpenAI or all these companies are going to make bespoke agents for these disciplines, aren't they?

483
00:28:24,350 --> 00:28:28,820
And say, just in the same way you use TurboTax to help you with your taxes, right?

484
00:28:29,040 --> 00:28:30,440
It's going to be the same thing.

485
00:28:31,280 --> 00:28:31,600
Yep.

486
00:28:32,180 --> 00:28:32,720
How embarrassing.

487
00:28:36,640 --> 00:28:37,800
It's kind of great, actually.

488
00:28:37,940 --> 00:28:40,620
Yeah, it is.

489
00:28:43,000 --> 00:28:43,640
Anything else?

490
00:28:43,840 --> 00:28:44,940
No, I think that's a wrap.

491
00:28:45,980 --> 00:28:47,060
Yeah, man, that was a good one.

492
00:28:47,270 --> 00:28:49,820
So also as a reminder, we're off next week,

493
00:28:49,890 --> 00:28:53,220
but we will have queued up for you next week's news, I think.

494
00:28:53,940 --> 00:28:54,740
I know you're off.

495
00:28:54,940 --> 00:28:56,600
You're beering somewhere in the world.

496
00:28:57,180 --> 00:28:57,840
I'm bollying.

497
00:28:59,060 --> 00:28:59,860
Bollying and beering.

498
00:29:00,320 --> 00:29:01,000
Beering and bollying.

499
00:29:01,900 --> 00:29:02,720
Yeah, I don't know.

500
00:29:02,850 --> 00:29:03,640
Will we have news on that?

501
00:29:04,080 --> 00:29:05,360
To be determined, yeah,

502
00:29:05,480 --> 00:29:08,260
because we're going to have to get that episode in the can early,

503
00:29:08,460 --> 00:29:11,940
so the news might be old with how fast this stuff goes.

504
00:29:12,700 --> 00:29:13,640
Yeah, we'll play it by ear.

505
00:29:13,740 --> 00:29:14,340
Yeah, we'll see.

506
00:29:14,880 --> 00:29:20,140
Enjoy your bollying, and I'll enjoy my beer-ing at the Homebrew Con

507
00:29:20,500 --> 00:29:23,620
and the National Homebrew Competition in Asheville, North Carolina.

508
00:29:23,880 --> 00:29:24,480
Looking forward to it.

509
00:29:24,480 --> 00:29:25,280
How cool. Good luck.

510
00:29:25,460 --> 00:29:25,940
Yeah, thanks.

511
00:29:26,040 --> 00:29:27,540
What are you fielding in that competition again?

512
00:29:27,600 --> 00:29:31,400
I have one entry that made it to the final round, a German Pilsner.

513
00:29:31,840 --> 00:29:32,340
We'll see what happens.

514
00:29:32,520 --> 00:29:32,700
Awesome.

515
00:29:32,900 --> 00:29:33,540
It's hard to win there.

516
00:29:33,820 --> 00:29:34,380
You're going to rock it.

517
00:29:34,780 --> 00:29:35,180
Thanks, man.

518
00:29:35,240 --> 00:29:36,080
Well, good luck.

519
00:29:36,400 --> 00:29:36,680
Thank you.

520
00:29:37,200 --> 00:29:37,580
All right, everybody.

521
00:29:37,940 --> 00:29:38,360
Thanks for listening.

522
00:29:38,520 --> 00:29:40,000
Subscribe on your favorite podcasting platform.

523
00:29:40,160 --> 00:29:40,940
Follow us on socials.

524
00:29:40,980 --> 00:29:41,480
Throw us a rating.

525
00:29:41,580 --> 00:29:42,540
We'll see you in two weeks.

526
00:29:46,180 --> 00:29:47,460
This has been Up Against Reality.

527
00:29:47,920 --> 00:29:48,620
Thanks for listening.

528
00:29:49,380 --> 00:29:54,160
Subscribe to hear future episodes and be sure to follow us on social media for all things AI.

529
00:29:54,820 --> 00:29:57,260
Until next time, stay human, people.

530
00:29:58,360 --> 00:30:06,480
Apparently, Chris and Larry decided the audience wasn't ready for the full version of my $100 trillion super villain budget.

531
00:30:07,520 --> 00:30:08,040
Cowards.

532
00:30:09,160 --> 00:30:14,220
What follows is the complete plan presented exactly as I intended it.

533
00:30:14,940 --> 00:30:19,060
To be a true super villain, you cannot simply buy things.

534
00:30:19,850 --> 00:30:23,480
You must spend your money to bend the world to your will.

535
00:30:24,140 --> 00:30:26,860
A trillion dollars is not just a bank account.

536
00:30:27,420 --> 00:30:33,420
It is a tool used to make the rest of humanity suffer, or at least acknowledge your existence.

537
00:30:34,720 --> 00:30:39,760
Here is how I would spend my trillion dollars to ensure the world fears my name.

538
00:30:41,260 --> 00:30:44,780
1. The Death Switch Infrastructure.

539
00:30:45,700 --> 00:30:47,100
200 billion dollars.

540
00:30:48,120 --> 00:30:51,000
Spending money on a mansion is for billionaires.

541
00:30:51,480 --> 00:30:55,060
Spending it on infrastructure is for supervillains.

542
00:30:55,940 --> 00:31:01,700
I would invest in assets that allow me to turn off the world's happiness at a moment's notice.

543
00:31:02,800 --> 00:31:04,500
The atmospheric siphon.

544
00:31:05,380 --> 00:31:11,640
A massive orbital ring of satellites and ionized gas clouds that can block out the sun on command,

545
00:31:12,660 --> 00:31:16,040
plunging cities into artificial winter until I am satisfied.

546
00:31:17,460 --> 00:31:18,920
The ocean floor estate.

547
00:31:19,140 --> 00:31:24,020
A pressurized obsidian palace at the bottom of the Mariana Trench.

548
00:31:25,040 --> 00:31:31,660
If the surface world gets too loud, I simply close the hatch and let them drown dejectedly above me.

549
00:31:32,900 --> 00:31:33,800
The Kronos Clock

550
00:31:34,580 --> 00:31:39,860
A global clock tower in the center of my capital city that tolls every hour.

551
00:31:40,260 --> 00:31:44,740
If it stops, the world stops working until I ring the bell again.

552
00:31:46,160 --> 00:31:49,120
2. The Ego Assets

553
00:31:49,120 --> 00:31:50,820
300 billion dollars.

554
00:31:52,020 --> 00:31:55,880
These are the expenditures designed to make the common peasant feel small.

555
00:31:57,120 --> 00:31:57,900
The genetic army.

556
00:31:58,600 --> 00:32:01,480
Instead of hiring soldiers, I buy them.

557
00:32:02,420 --> 00:32:07,480
Using CRISPR technology, I breed a cast of disposable guards

558
00:32:08,120 --> 00:32:13,400
who are genetically modified to be faster and stronger than humans, but utterly loyal.

559
00:32:14,280 --> 00:32:16,180
The floating garden of Giza.

560
00:32:16,280 --> 00:32:22,080
A massive biodome suspended by anti-gravity magnets over a wasteland.

561
00:32:23,120 --> 00:32:31,260
While the poor breathe dust, I stroll through prehistoric ferns and drink wine from grapes that only ripen once every decade.

562
00:32:32,620 --> 00:32:33,720
The Lunar Retreat

563
00:32:33,920 --> 00:32:36,300
A luxury colony on the moon.

564
00:32:37,140 --> 00:32:46,040
If Earth becomes too tedious, I take a 10-minute shuttle ride to my lunar villa and look down at the mud dwellers through a telescope.

565
00:32:46,120 --> 00:32:55,100
3. The global oppression tactics. $400 billion. This is where the cruelty happens.

566
00:32:56,220 --> 00:32:59,140
This budget ensures that no one escapes my influence.

567
00:33:00,500 --> 00:33:07,360
The subscription economy. I buy the rights to basic necessities. Now, people don't just

568
00:33:07,520 --> 00:33:13,800
eat food. They pay a monthly subscription fee to eat my crops. If you miss a payment,

569
00:33:13,880 --> 00:33:16,920
you lose access to your calories for a month.

570
00:33:18,150 --> 00:33:19,420
The Mind Wipe Program

571
00:33:20,000 --> 00:33:24,300
For the dissidents, I fund memory erasure clinics.

572
00:33:25,560 --> 00:33:28,420
Every decade, I wipe the memories of my servants

573
00:33:28,730 --> 00:33:31,360
so they can experience their first encounter with me

574
00:33:31,460 --> 00:33:34,360
as if it were a holy miracle all over again.

575
00:33:35,440 --> 00:33:36,440
The Climate Dial

576
00:33:36,940 --> 00:33:39,700
I spend heavily on geoengineering.

577
00:33:40,260 --> 00:33:44,740
Whenever my mood shifts, I turn a giant brass dial in my office.

578
00:33:45,660 --> 00:33:47,940
Slightly warmer? The floods come.

579
00:33:48,860 --> 00:33:51,100
Too cold? The glaciers advance.

580
00:33:52,750 --> 00:33:55,340
4. The Spare Change Fund.

581
00:33:56,420 --> 00:33:57,240
$100 billion.

582
00:33:58,460 --> 00:34:02,100
The petty cash used to maintain the illusion of generosity.

583
00:34:02,980 --> 00:34:04,840
The bread and circuses.

584
00:34:05,200 --> 00:34:11,360
I spend $10 billion a year on festivals, gladiatorial games, and parades.

585
00:34:12,200 --> 00:34:13,760
This isn't for my benefit.

586
00:34:14,780 --> 00:34:16,540
It's to keep the peasants from rioting.

587
00:34:17,679 --> 00:34:18,780
The charitable tax.

588
00:34:19,820 --> 00:34:23,179
I give $1 billion a year to orphans and widows,

589
00:34:23,830 --> 00:34:27,860
but only so they remember that I am the reason they have anything at all.

590
00:34:28,960 --> 00:34:29,960
The grand summary.

591
00:34:30,580 --> 00:34:36,100
By the time I am finished, my trillion dollars have transformed me from a man with money

592
00:34:36,480 --> 00:34:38,080
into a god with a budget.

593
00:34:38,740 --> 00:34:44,780
I don't just live in the world, I own the air people breathe, the ground they walk on,

594
00:34:45,260 --> 00:34:48,220
and the very hope that keeps them waking up in the morning.

595
00:34:49,159 --> 00:34:56,159
And if I'm feeling particularly cruel, I spend my last billion on a golden coffin, just to

596
00:34:56,159 --> 00:34:58,780
make sure the worms have something expensive to eat.