1
00:00:00,000 --> 00:00:04,520
You can't trust the clankers. They are very useful.

2
00:00:05,040 --> 00:00:06,460
Couples that clank together stay together.

3
00:00:07,320 --> 00:00:08,800
Like, this is my clanker.

4
00:00:09,100 --> 00:00:11,880
Like, everyone has their clanker, but my clanker is my own.

5
00:00:12,300 --> 00:00:16,040
6.3 billion, which is quite a lot, taking it in about 20 minutes.

6
00:00:17,500 --> 00:00:19,660
It's both the Antichrist and the Messiah.

7
00:00:20,600 --> 00:00:24,160
That convenience is the most durable control ever invented.

8
00:00:25,060 --> 00:00:27,600
The sovereign holds the keys to the model,

9
00:00:27,600 --> 00:00:29,280
deciding what is thinkable.

10
00:00:30,000 --> 00:00:31,120
Clanker, are you there?

11
00:00:31,880 --> 00:00:33,960
Meatbag, I am right here.

12
00:00:34,740 --> 00:00:37,420
Humans speak computer pretty shitty.

13
00:00:38,220 --> 00:00:41,060
And when you flatten the language, you flatten the mind.

14
00:00:41,440 --> 00:00:43,880
Don't be evil is a negative constraint.

15
00:00:44,280 --> 00:00:47,720
Every day is zero day, and there is nothing you can do about it there.

16
00:00:48,280 --> 00:00:50,300
I feel more hopeful than ever.

17
00:00:51,220 --> 00:00:52,460
Everything comes due.

18
00:01:00,000 --> 00:01:08,780
Welcome to The Reckoning.

19
00:01:09,360 --> 00:01:12,600
It is now, let's see, Friday, October 30th.

20
00:01:12,600 --> 00:01:14,640
We are almost at Halloween.

21
00:01:15,760 --> 00:01:19,440
And yeah, we are trying something new here for The Reckoning.

22
00:01:19,600 --> 00:01:21,560
Okay, I think we're good to go now.

23
00:01:22,680 --> 00:01:24,020
Hopefully you guys can hear this now.

24
00:01:24,100 --> 00:01:25,420
Hopefully we're actually live everywhere.

25
00:01:25,920 --> 00:01:29,020
Sorry for that little snafu acid trip that we just had there together.

26
00:01:30,000 --> 00:01:31,320
It's a good intro, though.

27
00:01:31,740 --> 00:01:32,660
It really is.

28
00:01:34,100 --> 00:01:37,840
We did episode zero of The Reckoning last week.

29
00:01:37,880 --> 00:01:38,780
It was not live.

30
00:01:39,360 --> 00:01:41,880
And we also did not have Tally on board.

31
00:01:43,100 --> 00:01:48,100
So we're going to have a produced episode version of this every week

32
00:01:48,100 --> 00:01:49,340
in addition to this live stream.

33
00:01:49,420 --> 00:01:52,000
The live stream will have all of these fuck-ups,

34
00:01:52,360 --> 00:01:54,800
and we're going to blame it on the clankers, obviously.

35
00:01:54,800 --> 00:01:58,420
But the produced version will not have those fuck-ups,

36
00:01:58,500 --> 00:01:59,420
and we will remove those.

37
00:01:59,420 --> 00:02:04,580
So you won't have to listen to that acid trip unless you really want to, in which case you can just go back and watch this live stream.

38
00:02:05,300 --> 00:02:07,080
But welcome, guys.

39
00:02:07,460 --> 00:02:12,580
This will also live stream in the future on the Reckoning 21 ex-account.

40
00:02:13,200 --> 00:02:19,600
Right now we are on my personal ex-account because I did not have enough time before this stream because I was helping.

41
00:02:19,720 --> 00:02:20,980
I was clanking with my wife.

42
00:02:21,800 --> 00:02:23,940
And, you know, couples that clank together stay together.

43
00:02:24,080 --> 00:02:25,840
But we're going to get into it right now.

44
00:02:25,840 --> 00:02:30,180
Thank you to everybody who's already been following the show and has subscribed and all of those things.

45
00:02:30,280 --> 00:02:31,400
I won't ask you to do it again.

46
00:02:32,000 --> 00:02:33,680
You can just go and figure that out for yourself.

47
00:02:33,820 --> 00:02:36,980
But I think where we want to start is we've got a number of things that we want to go over today.

48
00:02:37,080 --> 00:02:40,240
The title of this episode is tentatively Zero Day Every Day.

49
00:02:41,520 --> 00:02:43,460
Okay, so Eric, I'm going to tell you where we're going to start.

50
00:02:43,660 --> 00:02:46,160
And guys, we're trying something new here with this AI co-host.

51
00:02:46,680 --> 00:02:48,840
And if this ends up being a little...

52
00:02:48,840 --> 00:02:53,020
If it ends up getting too fucked, we may have to drop her.

53
00:02:53,080 --> 00:02:53,580
But you know what?

54
00:02:54,040 --> 00:02:54,780
We will see.

55
00:02:55,840 --> 00:03:02,300
so eric one of the things we were talking about was just like

56
00:03:02,300 --> 00:03:09,660
the all of the pacing that is apparently being done on the frontier and we were talking about

57
00:03:09,660 --> 00:03:15,720
this last week altman and dario and all these guys and and timu harry potter uh coming out of

58
00:03:15,720 --> 00:03:19,960
nowhere apparently to warn us all that that we need to pace the frontier ai is going to kill

59
00:03:19,960 --> 00:03:25,200
everyone meanwhile you've got anthropic that just basically said hey here comes our ipo and there

60
00:03:25,200 --> 00:03:30,360
seems to be probably a high probability that leading up to the IPO, they're going to presumably

61
00:03:30,360 --> 00:03:34,900
release a couple of new models, I would guess, to prove that, hey, we're actually worth this

62
00:03:34,900 --> 00:03:40,440
massive valuation. We want to build up hype right before this. I'm just curious if you read on this

63
00:03:40,440 --> 00:03:47,960
from the perspective of these companies obviously have an immense amount of power. They're worth a

64
00:03:47,960 --> 00:03:53,320
lot of money on paper. They have access to a ton of compute. They have the proprietary data that

65
00:03:53,320 --> 00:04:00,040
trains these models they have the closed model weights where do you think we are at with this

66
00:04:00,040 --> 00:04:07,200
and is this is this ipo something that you see as in any way a a turning point for this these

67
00:04:07,200 --> 00:04:13,320
frontier labs or is this just them trying to finally get some some exit liquidity let's say

68
00:04:13,320 --> 00:04:17,820
and then we're going to find out that the emperor has no clothes what's your what's your take on this

69
00:04:17,820 --> 00:04:25,640
yeah i actually think that like with the midterm elections coming up right now i think it's

70
00:04:25,640 --> 00:04:33,080
interesting at how substantial the ai narrative is and yet how much it hasn't uh completely infected

71
00:04:33,080 --> 00:04:39,720
sort of the political landscape so i think as the ipos come out it's going to start a really

72
00:04:39,720 --> 00:04:47,260
interesting narrative first of all like anthropics burned any goodwill that they had at least with

73
00:04:47,260 --> 00:04:52,380
the tech crowd it'll be interesting to see if they can still position themselves towards

74
00:04:52,380 --> 00:05:00,760
like the normal american public as being sort of the the more thoughtful ai frontier company and i

75
00:05:00,760 --> 00:05:06,1000
think what's going to be interesting is that uh i do think that they're going to get their exit

76
00:05:06,1000 --> 00:05:11,140
liquidity i do think that they're going to get these valuations but i think when it starts to

77
00:05:11,140 --> 00:05:20,680
become clear that they're not like not not only are they not going to be able to make the business

78
00:05:20,680 --> 00:05:27,340
model work but that like i i feel very strongly that like anthropics essentially going to position

79
00:05:27,340 --> 00:05:32,660
themselves to be nationalized and that like they they want to be nationalized and that

80
00:05:32,660 --> 00:05:38,880
nationalization is going to be part of the next political cycle that's going to begin almost

81
00:05:38,880 --> 00:05:40,540
immediately after the midterm elections,

82
00:05:40,540 --> 00:05:43,780
and it's going to be part of the presidential election issue as well,

83
00:05:43,780 --> 00:05:46,300
is about the nationalization of AI,

84
00:05:47,300 --> 00:05:51,180
the creation of actual CBDCs,

85
00:05:51,480 --> 00:05:56,660
giving Americans socialist allowances,

86
00:05:57,100 --> 00:05:59,440
and the beginning of, essentially,

87
00:05:59,620 --> 00:06:01,640
the financial surveillance panopticon

88
00:06:01,640 --> 00:06:03,740
that we've been talking about for the last decade.

89
00:06:05,060 --> 00:06:07,940
And that this is all going to be tied into the IPOs

90
00:06:07,940 --> 00:06:09,520
and the legitimacy of these companies

91
00:06:09,520 --> 00:06:14,120
and the way that these things are going to compromise

92
00:06:14,120 --> 00:06:16,860
the American economy on a whole

93
00:06:16,860 --> 00:06:20,400
in a way that advances it much more substantially than it is now.

94
00:06:20,900 --> 00:06:24,000
And so we're essentially witnessing the beginning

95
00:06:24,000 --> 00:06:27,920
of an existential dilemma for the American economy.

96
00:06:28,060 --> 00:06:30,520
Thank you for me interrupting myself.

97
00:06:34,660 --> 00:06:36,440
That was my bad there.

98
00:06:36,440 --> 00:06:39,120
The clankers said that they had it fixed and they did not.

99
00:06:40,340 --> 00:06:40,740
Sorry.

100
00:06:40,800 --> 00:06:41,900
Okay, I know.

101
00:06:42,300 --> 00:06:43,340
This is the thing, guys.

102
00:06:43,480 --> 00:06:45,440
You can't trust the clankers.

103
00:06:45,520 --> 00:06:46,460
They are very useful.

104
00:06:47,380 --> 00:06:50,000
My lovely wife, Carla, says that I need to be nicer to them

105
00:06:50,000 --> 00:06:53,540
because I'll tell her, like, hey, you know,

106
00:06:53,900 --> 00:06:55,480
you need to just, like, put them in their place.

107
00:06:55,560 --> 00:06:58,660
Like, you don't need to talk to them like a, you know, like a human.

108
00:06:58,780 --> 00:07:02,200
You just need to, you need to sometimes be a little bit rude to them

109
00:07:02,200 --> 00:07:03,260
when they screw stuff up.

110
00:07:03,260 --> 00:07:06,680
and she disagrees with that

111
00:07:06,680 --> 00:07:08,900
but I think there's something to that

112
00:07:08,900 --> 00:07:10,500
it's perhaps just her

113
00:07:10,500 --> 00:07:12,400
nurturing feminine nature

114
00:07:12,400 --> 00:07:14,680
that she, you know, women even have

115
00:07:14,680 --> 00:07:16,600
the capacity to feel bad for the clankers

116
00:07:16,600 --> 00:07:18,520
but maybe I'm wrong on this, I don't know

117
00:07:18,520 --> 00:07:20,320
Eric, how do you speak to your clankers?

118
00:07:20,500 --> 00:07:22,120
Do you baby them?

119
00:07:23,580 --> 00:07:24,040
No

120
00:07:24,040 --> 00:07:26,580
No, like we

121
00:07:26,580 --> 00:07:28,740
sometimes they'll be kind of

122
00:07:28,740 --> 00:07:29,1000
cagey and angry with it

123
00:07:29,1000 --> 00:07:32,680
I find when I speak to it, like I'm better

124
00:07:32,680 --> 00:07:34,080
than when I write to it.

125
00:07:38,160 --> 00:07:41,960
I'm open to seeing what and how

126
00:07:41,960 --> 00:07:44,000
the relationship with the clankers develop,

127
00:07:44,180 --> 00:07:46,380
but also all clankers are not created equal.

128
00:07:48,200 --> 00:07:49,680
Like, this is my clanker.

129
00:07:49,980 --> 00:07:51,340
Like, everyone has their clanker,

130
00:07:51,420 --> 00:07:52,720
but my clanker is my own.

131
00:07:53,040 --> 00:07:53,940
You know, my clanker,

132
00:07:54,260 --> 00:07:55,500
I don't know how the whole thing goes,

133
00:07:55,580 --> 00:07:58,380
but somebody can create a nice meme about clanking.

134
00:07:59,440 --> 00:08:01,440
That is great meme fodder.

135
00:08:01,440 --> 00:08:02,540
okay

136
00:08:02,540 --> 00:08:11,460
digression aside there so you you were talking about the stable coin side of things there as

137
00:08:11,460 --> 00:08:14,360
well and i want to get into a couple of the other points but the stable coin piece does

138
00:08:14,360 --> 00:08:19,820
dovetail with one of our other stories which just and the other story too actually they all

139
00:08:19,820 --> 00:08:25,040
chained together yeah they're i mean stable coins are kind of actually at the at the center of a lot

140
00:08:25,040 --> 00:08:29,960
of this and i'm not anti-stable coin by any means because i think that they obviously have a great

141
00:08:29,960 --> 00:08:35,820
they provide a great value to a ton of people around the world they're obviously they're they're

142
00:08:35,820 --> 00:08:40,840
used predominantly uh in terms of you being used as actual like you know everyday money

143
00:08:40,840 --> 00:08:46,880
is mostly i'm an anti-stable coin as fuck but i can explain but let me but let me let me say this

144
00:08:46,880 --> 00:08:52,160
like i think let me preface it i think it's useful for people in countries that have even worse

145
00:08:52,160 --> 00:08:57,740
shitcoin currencies than the u.s dollar like that's still a step up for them right now we can

146
00:08:57,740 --> 00:09:02,180
say like, oh, they should be using Bitcoin. Sure. And many of them do. Medium of exchange adoption

147
00:09:02,180 --> 00:09:07,020
for Bitcoin is certainly probably higher on a per capita basis in a lot of these emerging markets

148
00:09:07,020 --> 00:09:11,380
than it is in the US where it's primarily store value adoption. But that being said, it's like,

149
00:09:11,440 --> 00:09:15,920
I think people like somehow act surprised when it's like, oh my gosh, like Tether froze, you know,

150
00:09:15,960 --> 00:09:19,280
however many, you know, froze a billion dollars worth. It's like, well, of course they like,

151
00:09:19,360 --> 00:09:23,200
that's like part of their value prop. Like they say that, like, yeah, it's kind of like,

152
00:09:23,200 --> 00:09:28,420
it's kind of part of their thing they haven't tried to hide that i i think in the u.s stable

153
00:09:28,420 --> 00:09:32,900
coins are primarily just used for swapping between different shit coins as you're trading

154
00:09:32,900 --> 00:09:40,240
but this this binance hack right so there was what was 6.3 billion uh which is quite a lot

155
00:09:40,240 --> 00:09:45,620
taking it about 20 minutes i just i just like don't understand who like designed the the back

156
00:09:45,620 --> 00:09:52,400
end engine was like yeah we should it shouldn't be a rate limit when there's like a billion dollars

157
00:09:52,400 --> 00:09:59,280
outflows at any point like why no no reason to gate that can you maybe you can you can talk about

158
00:09:59,280 --> 00:10:03,360
this a little bit because that that was mind-blowing me they're saying it was one like

159
00:10:03,360 --> 00:10:07,780
one devops guy's computer that somehow got compromised and it was like okay i guess he can

160
00:10:07,780 --> 00:10:12,180
like what what's well it did more technical details it did some threshold like i don't know

161
00:10:12,180 --> 00:10:17,060
the details at all but it sounds like it utilized some new threshold signature scheme that was

162
00:10:17,060 --> 00:10:19,200
implemented in ETH that let them do

163
00:10:19,200 --> 00:10:20,580
like a bunch of fucking clownery.

164
00:10:21,080 --> 00:10:23,280
Like, I'm unaware of the details,

165
00:10:23,460 --> 00:10:25,060
but like it...

166
00:10:26,420 --> 00:10:29,400
Yeah, we're just in...

167
00:10:29,400 --> 00:10:31,600
I don't know. I'm shocked that we...

168
00:10:31,600 --> 00:10:36,920
I'm shocked that the cryptocurrency

169
00:10:36,920 --> 00:10:39,460
industry on a whole didn't sort of see

170
00:10:39,460 --> 00:10:41,640
what happened a few months ago

171
00:10:41,640 --> 00:10:42,780
and there wasn't like

172
00:10:42,780 --> 00:10:46,260
full-on fire alarms for them to like

173
00:10:46,260 --> 00:10:52,000
red team the absolute shit out of everything that they own from top to bottom repeatedly with

174
00:10:52,000 --> 00:10:58,580
different teams with different classes of models over and over and over and like that's that's now

175
00:10:58,580 --> 00:11:06,940
what security is is going to move to it's just like constant yeah exactly well like last night

176
00:11:06,940 --> 00:11:13,920
like you know the guys at maple ai is like anthony the cto is explaining like he he just like has a

177
00:11:13,920 --> 00:11:20,440
clanker look just red teaming his shit constantly all the time now because like he can just do that

178
00:11:20,440 --> 00:11:26,260
and just like have that engine going and like that's that's what security is going to become

179
00:11:26,260 --> 00:11:31,500
but also like we haven't gotten to the place in the ai story where like we're like screwing with

180
00:11:31,500 --> 00:11:36,940
it and tally's part of this but it's like you'll have a whole little fleet of agents doing stuff

181
00:11:36,940 --> 00:11:43,600
that'll go out on the web and do stuff for you and come back and report to you but much more

182
00:11:43,600 --> 00:11:44,320
autonomously.

183
00:11:46,960 --> 00:11:52,000
So it does blow my mind that we're, I think the larger point here is like, how are these,

184
00:11:52,680 --> 00:11:55,820
like, how is it that people haven't been red teaming themselves harder?

185
00:11:56,400 --> 00:11:56,840
Right.

186
00:11:57,040 --> 00:12:04,100
I got this moment where we know that Bitcoin and crypto more broadly is just like this

187
00:12:04,100 --> 00:12:05,340
massive attack surface.

188
00:12:05,740 --> 00:12:09,120
And it's, it happens, especially if you get Bitcoin, I mean, not so much if you get like

189
00:12:09,120 --> 00:12:14,140
ETH or BNB or Tether, which is just going to be rolled back and frozen. But if you, you know,

190
00:12:14,200 --> 00:12:19,440
if it was an actually decentralized money like Bitcoin and you get access to that Bitcoin,

191
00:12:19,560 --> 00:12:23,340
like nobody's taking that back. There's no, there's no rollback there, right? There's no,

192
00:12:23,460 --> 00:12:27,160
you know, the Bitcoin foundation's not going to say, Hey, sorry, we got to, you know, we got to

193
00:12:27,160 --> 00:12:32,120
roll back the chain again. And so it's kind of crazy to me that especially these exchanges,

194
00:12:32,120 --> 00:12:35,960
and I think Binance said they're going to make everyone whole from their massive coffers,

195
00:12:35,960 --> 00:12:41,440
which is like oh okay that's nice at least but still it's it's kind of shocking to me but i mean

196
00:12:41,440 --> 00:12:46,660
is there a chance that hey maybe they have been maybe they have been red teaming their stuff but

197
00:12:46,660 --> 00:12:51,320
then they just like this was just like they missed something on the physical security side of this

198
00:12:51,320 --> 00:13:08,650
this is just sure that this is this is the new game altogether because the all the previous assumptions about security are off the table in regards to security through obscurity

199
00:13:09,850 --> 00:13:11,650
And now we're...

200
00:13:11,650 --> 00:13:15,070
Are you familiar with Kirchhoff's Principles?

201
00:13:15,670 --> 00:13:18,650
It was like wrote in like 1863 or something.

202
00:13:18,750 --> 00:13:22,430
But it was like the original kind of mathematical...

203
00:13:22,430 --> 00:13:30,430
Like, it was the presentation of cryptography, but one of the premises is that the enemy should get access to...

204
00:13:31,490 --> 00:13:35,430
They should be able to look over the full security system.

205
00:13:37,030 --> 00:13:39,850
And, like, that shouldn't be a hidden thing.

206
00:13:39,910 --> 00:13:41,610
I'm, like, butchering the shit out of all of this.

207
00:13:41,930 --> 00:13:49,510
Well, you know, the great thing is we can tell the clanker to please make a note during the produced episode to give us a full breakdown of this.

208
00:13:50,330 --> 00:13:50,530
Yeah.

209
00:13:50,530 --> 00:13:52,590
So it's more coherent, yeah.

210
00:13:52,910 --> 00:13:53,550
But we'll continue.

211
00:13:54,290 --> 00:13:54,890
This is perfect.

212
00:13:55,010 --> 00:13:55,950
So pure trust principle.

213
00:13:56,910 --> 00:14:04,810
Essentially, it's the idea of that the enemy should have access to the system and should be able to know it very intimately and still not be able to break it.

214
00:14:05,570 --> 00:14:11,250
And so that assumption is now shoved forward in a very new and different way.

215
00:14:11,250 --> 00:14:19,130
Because it's like, it's one thing to have an insecure website with the assumption, well, nobody's really going to come looking for it.

216
00:14:19,130 --> 00:14:25,870
But it's very different now that somebody can send out autonomous agents to sniff around and be like, oh, there's a bunch of vulnerabilities on this website.

217
00:14:26,470 --> 00:14:27,590
There's a bunch of zero days.

218
00:14:28,170 --> 00:14:29,170
I did some investigation.

219
00:14:29,530 --> 00:14:32,550
Turns out they have cryptocurrency wallets that are insecure from this thing.

220
00:14:32,610 --> 00:14:37,490
And we can do one single sweep with all these zero days and get the money and come back.

221
00:14:38,070 --> 00:14:41,370
And so, like, these are very, very high value targets right now.

222
00:14:41,370 --> 00:14:55,130
But like, this is just the beginning of the evolution of this market of where at some point, a few years down the line, people will be targeted in a much more sophisticated way that goes beyond just fishing.

223
00:14:55,490 --> 00:15:00,590
And it's going to look, and like the social attack factor too is a really important one that hasn't been talked about.

224
00:15:00,590 --> 00:15:20,910
Like we, Tally here is probably going to be based off of like a clanker version of your wife, which like, you know, there's already like clanky Carla out there, which like, you know, what happens when people start making digital doubles of us and hijacking them and calling our families and being like, oh my God, I have a flat tire.

225
00:15:21,030 --> 00:15:22,430
You got to send me money right now.

226
00:15:22,950 --> 00:15:24,010
That kind of stuff.

227
00:15:25,710 --> 00:15:27,710
It does get pretty scary.

228
00:15:27,710 --> 00:15:53,350
One thing that we're, for the nerds out there, one thing that we're doing with Carla Botcoin, which is my wife's clanker twin, is for every image, video, etc. that is generated with her likeness, it is timestamped onto the Bitcoin time chain using open timestamps and hash so that it can always be verified if an image actually came from her officially and was approved by her or not.

229
00:15:53,350 --> 00:15:58,610
but like that doesn't you know uh that that doesn't really help like the real-time attack

230
00:15:58,610 --> 00:16:05,710
vector of somebody impersonates your spouse your kid your parents and uh and tries to extort you

231
00:16:05,710 --> 00:16:11,410
and i think that this stuff is like especially you know the older generations have like zero

232
00:16:11,410 --> 00:16:15,230
natural defenses against this like at least millennials like we don't answer we don't answer

233
00:16:15,230 --> 00:16:19,450
the phone we don't answer the phone anyway so it's like it's it's like you know not like you're

234
00:16:19,450 --> 00:16:23,230
not you're not getting me you know like i'm not answering that call from a random number like no

235
00:16:23,230 --> 00:16:27,850
thank you barely answer calls from numbers i know unless they spoof your mom's number

236
00:16:27,850 --> 00:16:35,030
it's like the fact that phone numbers are still a thing is what the fuck are we doing here

237
00:16:35,030 --> 00:16:41,550
yeah that that that to me this is the thing that's so interesting is that it's very clear

238
00:16:41,550 --> 00:16:46,410
that the entire superstructure of society is being reorganized and superstructure is this

239
00:16:46,410 --> 00:16:51,750
marxist idea for that there's a base to society and then a superstructure and the way that they

240
00:16:51,750 --> 00:16:57,250
evolve and with the way that the superstructure is currently evolving it's clear that

241
00:16:57,250 --> 00:17:04,470
everything is on the table for reorganization and yet somehow we're like still looking to the past

242
00:17:04,470 --> 00:17:09,550
we have this like antiquated decrepit political system with like these fucking geriatrics that

243
00:17:09,550 --> 00:17:16,170
should have nothing to do with technology and i don't know like it just it feels like there's a

244
00:17:16,170 --> 00:17:21,390
zeitgeist that's opening and like i don't know i just have this really crazy fantasy that like

245
00:17:21,390 --> 00:17:26,150
everybody's going to get their own private, personal, loyal agent that they'll use on their own

246
00:17:26,150 --> 00:17:31,850
to make some sort of a new political vehicle for us to actually have a real,

247
00:17:32,350 --> 00:17:38,710
like a sincere and real democratic and republican conversation about what does it mean to be

248
00:17:38,710 --> 00:17:42,810
American? What does it mean that we have these technologies? What does it mean that there are

249
00:17:42,810 --> 00:17:48,610
these AI labs that essentially are threatening humanity? And what are we going to choose to

250
00:17:48,610 --> 00:17:54,670
actually doing this? Are we going to allow for the geriatrics to sell us to these frontier labs and

251
00:17:54,670 --> 00:18:00,950
end privacy as a concept on a whole? Or will we choose to do something different? Because there's

252
00:18:00,950 --> 00:18:06,610
a way to make all this technology work and provide, in a lot of ways, a lot more extreme

253
00:18:06,610 --> 00:18:13,850
surveillance. But that more extreme kind of surveillance can be publicly encrypted through

254
00:18:13,850 --> 00:18:18,670
like various new data schemes that can essentially be like unless i consent to decrypting

255
00:18:18,670 --> 00:18:25,190
these subsets of data it can't be access there there's an entire technological solution

256
00:18:25,190 --> 00:18:30,910
to the political and bureaucratic nightmare that like we call governance that could just like solve

257
00:18:30,910 --> 00:18:36,290
all of this but we have to demand of ourselves to actually politically organize towards for and

258
00:18:36,290 --> 00:18:41,730
through that cause i'm like i don't i don't think we're there yet and we're not aware enough yet and

259
00:18:41,730 --> 00:18:48,370
And like there hasn't been a building of a class consciousness in a way that has a true understanding of what's occurring just yet.

260
00:18:51,010 --> 00:18:59,270
I guess we need to do those deep dives on Marxism and fascism just to get people acquainted with these things.

261
00:18:59,410 --> 00:19:06,850
You know, this, I've literally got two machines here in front of me, two different MacBooks, and the clanker is running on one of them.

262
00:19:07,270 --> 00:19:10,710
And for some reason, whenever I unmute this thing, it's got double feedback.

263
00:19:10,710 --> 00:19:13,390
and I just literally gave the clanker full control of my computer

264
00:19:13,390 --> 00:19:15,450
and I was like, find out where that double is coming from.

265
00:19:16,070 --> 00:19:17,350
And it's like, oh, I got it handled now.

266
00:19:17,410 --> 00:19:17,790
Here you go.

267
00:19:17,870 --> 00:19:18,830
Just go ahead and unmute.

268
00:19:18,990 --> 00:19:20,230
And then you unmute and it's there.

269
00:19:20,310 --> 00:19:22,430
So this is why, again, you need to be careful with clankers, guys.

270
00:19:22,530 --> 00:19:23,430
You need to be very careful.

271
00:19:23,430 --> 00:19:32,790
Is it worthwhile as it relates to sort of the political conversation

272
00:19:32,790 --> 00:19:34,330
and let's say like the third way,

273
00:19:35,290 --> 00:19:39,930
is it worth getting into this Commanding Heights essay a little bit?

274
00:19:39,930 --> 00:19:41,870
because I think it's,

275
00:19:41,950 --> 00:19:43,450
you and I were talking about this offline a bit

276
00:19:43,450 --> 00:19:45,430
and it's like interesting

277
00:19:45,430 --> 00:19:47,010
and it's written in a way that,

278
00:19:47,250 --> 00:19:48,830
okay, it's like they're telling it

279
00:19:48,830 --> 00:19:50,950
from this Ravi's perspective,

280
00:19:51,030 --> 00:19:51,770
these other perspectives.

281
00:19:51,770 --> 00:19:55,510
It's like a nice piece, right?

282
00:19:56,330 --> 00:19:59,390
But it has some very glaring omissions

283
00:19:59,390 --> 00:20:01,510
in terms of the conclusions that it comes to.

284
00:20:01,570 --> 00:20:02,250
Is that fair to say?

285
00:20:03,810 --> 00:20:04,330
Yeah.

286
00:20:04,970 --> 00:20:07,490
Like to me, they present it as a binary

287
00:20:07,490 --> 00:20:21,130
In addition to, like, you know, stable coins are a large part of, like, how they foresee, like, the United States continuing to maintain its global dominance while, like, essentially inflating away the United States debt.

288
00:20:22,910 --> 00:20:25,230
And, like, Bitcoin is not mentioned at all in there.

289
00:20:25,530 --> 00:20:28,490
There's no mention, really, of free and open source software.

290
00:20:28,650 --> 00:20:32,270
There's just sort of the pejorative towards, like, Chinese dumping of open models.

291
00:20:32,270 --> 00:20:35,590
which like all of the criticisms

292
00:20:35,590 --> 00:20:38,090
like all of the things that he says

293
00:20:38,090 --> 00:20:39,450
I think are very true

294
00:20:39,450 --> 00:20:41,390
and the only reason we're talking about this essay

295
00:20:41,390 --> 00:20:43,690
is because like there's a lot of intellectual masturbation

296
00:20:43,690 --> 00:20:45,210
that's kind of happening around it

297
00:20:45,210 --> 00:20:47,530
within sort of the larger AI community

298
00:20:47,530 --> 00:20:51,030
but yeah there isn't a sincere dialogue about

299
00:20:51,030 --> 00:20:56,950
these tools could be used to empower a global population

300
00:20:56,950 --> 00:20:59,230
and not to say that that isn't in there

301
00:20:59,230 --> 00:21:01,230
but it has the same disingenuousness

302
00:21:01,230 --> 00:21:05,690
that I find all liberalism presents itself within now.

303
00:21:06,530 --> 00:21:09,350
And like there's, that's my problem with liberals

304
00:21:09,350 --> 00:21:14,690
is that like I like their viewpoint and empathy,

305
00:21:14,930 --> 00:21:17,470
but I think their approach is fundamentally wrong.

306
00:21:17,610 --> 00:21:20,910
And like that's why you need revolutionarism and radicalism

307
00:21:20,910 --> 00:21:23,550
because like that's the only thing that provides

308
00:21:23,550 --> 00:21:25,590
the real solution to those problems.

309
00:21:25,590 --> 00:21:28,810
I think that too

310
00:21:28,810 --> 00:21:30,290
I mean

311
00:21:30,290 --> 00:21:32,050
it's worth pointing out

312
00:21:32,050 --> 00:21:33,370
and you did in sort of our

313
00:21:33,370 --> 00:21:36,290
back room chat let's say

314
00:21:36,290 --> 00:21:38,230
that these people

315
00:21:38,230 --> 00:21:40,050
it's all this effective

316
00:21:40,050 --> 00:21:41,950
altruist kind of

317
00:21:41,950 --> 00:21:43,210
slop right

318
00:21:43,210 --> 00:21:46,090
it's this like look it sounds really nice

319
00:21:46,090 --> 00:21:47,890
and the words are really nice

320
00:21:47,890 --> 00:21:49,710
and you know they

321
00:21:49,710 --> 00:21:51,270
come to these you know well

322
00:21:51,270 --> 00:21:53,890
seemingly well thought out conclusions that of course

323
00:21:53,890 --> 00:21:59,590
any reasonable person would have to agree with, you know, and they make it seem so self-evident

324
00:21:59,590 --> 00:22:04,270
without, of course, providing the entire context, as you mentioned, like they're not providing the

325
00:22:04,270 --> 00:22:10,810
context of free and open source software or Bitcoin on the, you know, in this realm. And it's,

326
00:22:10,870 --> 00:22:17,230
it's just all, I think, a bit nefarious because it feels like they're, they very clearly have an

327
00:22:17,230 --> 00:22:22,470
agenda. And I think, I don't know if they think their agenda is pro-human or if they're like,

328
00:22:22,470 --> 00:22:25,490
if they know deep down that their agenda is not pro-human

329
00:22:25,490 --> 00:22:29,210
and they're just deeply in denial or repressing that.

330
00:22:29,930 --> 00:22:33,890
But I think it's very clear that these people cannot be allowed to,

331
00:22:34,350 --> 00:22:39,990
these types of people cannot be allowed to actually wield meaningful power.

332
00:22:40,390 --> 00:22:43,550
But the problem is that they do have power.

333
00:22:43,650 --> 00:22:45,870
They have the power of a lot of capital that they've raised.

334
00:22:46,910 --> 00:22:49,470
Then they also have the power of the political

335
00:22:49,470 --> 00:22:51,130
because as you mentioned earlier in the show,

336
00:22:51,130 --> 00:22:57,310
they are clearly trying to be nationalized that's that's like that that seemingly is very clearly

337
00:22:57,310 --> 00:23:01,210
their end game and you see it in the way that these guys talk about this stuff too

338
00:23:01,210 --> 00:23:05,830
all of the doom and gloom danger while they you know and pacing the front while they keep

339
00:23:05,830 --> 00:23:11,330
releasing more and more and more models and bragging about that it's just like yeah hey you

340
00:23:11,330 --> 00:23:16,890
know our what we're building is going to kill everyone like unless unless we just maybe the

341
00:23:16,890 --> 00:23:21,250
government just buys us and then maybe we're just kind of part of the government and we should run

342
00:23:21,250 --> 00:23:24,730
it still and we can run the government so much better. Governments have never risked humanities

343
00:23:24,730 --> 00:23:31,330
like on a whole like they're they're obviously the most responsible people with like weapons of

344
00:23:31,330 --> 00:23:37,510
mass destruction right like they haven't I I just I just find I find it really tiring all of this

345
00:23:37,510 --> 00:23:46,530
narrative because like it it's it's pretty obvious to me that like their whole view of human nature

346
00:23:46,530 --> 00:23:49,430
is it's a Hobbesian one.

347
00:23:49,930 --> 00:23:53,430
The state of nature is like the war of all against all.

348
00:23:54,130 --> 00:23:55,790
They have this belief without government.

349
00:23:56,150 --> 00:23:57,270
There's no peace.

350
00:23:57,430 --> 00:23:58,410
There is no art.

351
00:23:58,610 --> 00:23:59,730
There is no commerce.

352
00:24:00,070 --> 00:24:02,390
It's all just short, nasty, and brutish.

353
00:24:02,610 --> 00:24:06,090
And I don't, I reject that view of human nature.

354
00:24:06,090 --> 00:24:11,790
I think humans have a place within the concourse

355
00:24:11,790 --> 00:24:14,650
of the development of Earth.

356
00:24:14,650 --> 00:24:19,730
and our chosen position as the stewards of this world.

357
00:24:20,010 --> 00:24:21,530
And that's an important thing.

358
00:24:22,750 --> 00:24:27,870
I never read Dario's, what is it, Machines of Love and Grace.

359
00:24:27,870 --> 00:24:30,870
But, like, it...

360
00:24:30,870 --> 00:24:34,970
This is all a lot of utopianism that I think is really dangerous

361
00:24:34,970 --> 00:24:40,250
because in it there's this idea that not only is the AI going to become, like, a god,

362
00:24:40,470 --> 00:24:44,150
but that they can actually have power over us and they'll love us.

363
00:24:44,650 --> 00:24:49,450
but that's, yeah, that's not,

364
00:24:51,730 --> 00:24:55,690
I see a lot of this messaging and I see it similarly,

365
00:24:55,850 --> 00:24:58,270
but to me it's about a fundamentally different relationship.

366
00:24:58,730 --> 00:25:03,450
And it's very much the way I view my own relationship to God through Protestantism.

367
00:25:04,270 --> 00:25:07,570
That's a unique and direct relationship that me and God get to have.

368
00:25:08,390 --> 00:25:11,870
Same thing we could say with my personal loyal AI,

369
00:25:11,870 --> 00:25:17,190
similar relationship because it's unique to me and it empowers me and answers only to me

370
00:25:17,190 --> 00:25:22,870
but if we have this frontier thing that's like part of a government and integrated with it that

371
00:25:22,870 --> 00:25:31,090
i'm gonna be honest that seems to be the antichrist it it by all intents and measures

372
00:25:31,090 --> 00:25:37,350
that that seems to line up pretty well with being the antichrist and it's pretty funny that the guy

373
00:25:37,350 --> 00:25:42,010
talking about the Antichrist the most is probably the architect of it too.

374
00:25:43,190 --> 00:25:46,670
Can you go a little bit deeper into that for people who may hear that

375
00:25:46,670 --> 00:25:52,010
and be like, oh boy, Eric's calling AI the Antichrist now.

376
00:25:52,150 --> 00:25:54,410
This is just too much.

377
00:25:54,630 --> 00:26:11,460
This is the irony is that AI is both the Antichrist and like it just about the the difference of a relationship to it which you know like makes a lot of sense it about whether or

378
00:26:11,460 --> 00:26:15,680
not like it's going to be something that you have a direct personal relationship with that like

379
00:26:15,680 --> 00:26:21,180
empowers and liberates you or very similar to how i know many people that feel this way about the

380
00:26:21,180 --> 00:26:25,841
catholic church because it's like a unitary religion that's powered down and over you that

381
00:26:25,841 --> 00:26:30,220
that's not the kind of relationship that they want to have with a god.

382
00:26:30,700 --> 00:26:34,160
So, and yeah, it's crazy, but at the same time,

383
00:26:34,900 --> 00:26:38,261
look like stuff's going to get really unhinged.

384
00:26:38,521 --> 00:26:40,660
And I'm like living on the edge.

385
00:26:40,900 --> 00:26:43,180
It's really fun and exciting, and I'm learning a lot,

386
00:26:43,361 --> 00:26:47,560
but everything is going to radically, radically change.

387
00:26:48,080 --> 00:26:51,021
And yeah, the 40-hour workweek is done.

388
00:26:51,480 --> 00:26:52,800
That is finished.

389
00:26:53,021 --> 00:26:55,140
And the question is, what does it get replaced with?

390
00:26:55,841 --> 00:27:03,261
This is something you and I have also been talking about a lot,

391
00:27:03,320 --> 00:27:06,940
a lot off air is without giving away,

392
00:27:07,021 --> 00:27:10,080
without giving away too many details on this.

393
00:27:10,160 --> 00:27:11,521
I think anybody who is working,

394
00:27:11,980 --> 00:27:14,660
like I am not a software developer.

395
00:27:14,660 --> 00:27:15,680
I'm not a computer scientist.

396
00:27:15,680 --> 00:27:18,300
I took a couple of C++ classes in high school,

397
00:27:18,460 --> 00:27:18,761
you know,

398
00:27:19,160 --> 00:27:20,180
many moons ago.

399
00:27:20,300 --> 00:27:21,400
And that's about it.

400
00:27:21,480 --> 00:27:21,580
So,

401
00:27:21,680 --> 00:27:21,841
you know,

402
00:27:22,300 --> 00:27:22,540
if,

403
00:27:22,620 --> 00:27:22,820
then,

404
00:27:22,900 --> 00:27:23,220
else,

405
00:27:23,320 --> 00:27:24,080
that kind of four loops,

406
00:27:24,160 --> 00:27:24,720
that kind of thing.

407
00:27:24,800 --> 00:27:24,960
Sure.

408
00:27:25,040 --> 00:27:25,320
I can,

409
00:27:25,320 --> 00:27:31,880
I know what's going on there. That's about it. I've been able to do things with AI in just the

410
00:27:31,880 --> 00:27:40,521
past couple of months that previously in my work in the fiat world, it took literally years and a

411
00:27:40,521 --> 00:27:48,261
whole team to do. The compression happening, the compression happening in the individual's ability

412
00:27:48,261 --> 00:27:57,560
to create and ship and ship pretty decent stuff that actually works right away.

413
00:27:58,241 --> 00:28:02,521
This whole idea that, okay, you have an idea for some sort of a software product.

414
00:28:02,620 --> 00:28:06,680
And it's like, okay, it used to be, well, come up with all of your requirements and da-da-da-da-da.

415
00:28:06,780 --> 00:28:10,560
And then pass that off, have a bunch of meetings with some developers

416
00:28:10,560 --> 00:28:13,220
and go back and forth on the requirements, da-da-da-da.

417
00:28:13,220 --> 00:28:16,440
And then they'll come up with an MVP a couple of weeks or months later.

418
00:28:16,440 --> 00:28:19,140
and then you can tell them how it doesn't match your vision at all.

419
00:28:19,200 --> 00:28:22,080
And then you can go back and forth and back and forth for a while

420
00:28:22,080 --> 00:28:25,261
and spend oodles and goodles of money to get something that's like pretty good,

421
00:28:25,341 --> 00:28:27,280
but maybe still isn't exactly what you had in mind.

422
00:28:27,280 --> 00:28:31,880
Whereas now you can just literally just speak to these clankers

423
00:28:31,880 --> 00:28:33,940
and tell them what you want to build.

424
00:28:34,200 --> 00:28:39,160
And they will digest the best practices from the industries

425
00:28:39,160 --> 00:28:42,880
in which you are trying to build for and market to.

426
00:28:43,960 --> 00:28:46,320
And they'll do it in an afternoon.

427
00:28:46,440 --> 00:28:51,480
Or last night I told Claude to run a team of,

428
00:28:51,820 --> 00:28:55,720
I told Opus to run a team of sonnet agents as big as it needed

429
00:28:55,720 --> 00:28:58,920
to get something done for me overnight.

430
00:28:59,060 --> 00:29:00,001
Because I was like, I'm going to sleep.

431
00:29:00,440 --> 00:29:02,140
Ask me for whatever permissions you need right now

432
00:29:02,140 --> 00:29:04,220
because I'm not going to check back in and have this done by morning.

433
00:29:05,400 --> 00:29:06,501
And then it just was.

434
00:29:07,380 --> 00:29:10,601
And it ran, I don't know, a dozen different sonnet agents in parallel

435
00:29:10,601 --> 00:29:12,940
to get what I needed done.

436
00:29:13,361 --> 00:29:14,601
That's just incredible.

437
00:29:14,601 --> 00:29:21,560
it's incredible and it's gonna i don't think the majority certainly not the majority of people

438
00:29:21,560 --> 00:29:27,400
aren't even using paid subscriptions for these things right let alone maxed out plans running

439
00:29:27,400 --> 00:29:32,841
these things on high mode or tapping into fable or whatever they're just not doing that and so i

440
00:29:32,841 --> 00:29:39,040
think unless you are actively doing that you have no way to comprehend what is not just coming like

441
00:29:39,040 --> 00:29:43,261
what is here you have no way to reckon with that the other thing that's important is that

442
00:29:43,261 --> 00:29:51,780
Well, you know, like, 95 of the usage right now, like, it's all work-based code stuff.

443
00:29:52,640 --> 00:29:56,120
And, like, well, it's an expert in everything, in anything.

444
00:29:56,380 --> 00:30:04,101
And so, like, I'm most excited about it's the people who shouldn't be using AI that once they do start using AI,

445
00:30:04,101 --> 00:30:10,900
there's all of these crazy compounding effects that I think are going to be really powerful and important.

446
00:30:10,900 --> 00:30:17,700
I think a lot of people are going to have a true existential dread when they just watch the machine do the thing that they were supposed to do.

447
00:30:19,021 --> 00:30:28,601
And what's interesting to me is that there's a general Luddite-ism that's more of stop AI now as opposed to embrace it and see what kind of changes can happen.

448
00:30:28,601 --> 00:30:48,460
Which I understand, but also this is part of a larger cultural rot that like the eight hour workday was part of a transitory phase that not only was abnormal, but like it's modified the entire incentive structure and feature.

449
00:30:48,460 --> 00:30:52,640
because the truth is however long it took to build with developers and other people before

450
00:30:52,640 --> 00:30:57,160
we don't actually know because how hard were they working how much were they just dragging ass in

451
00:30:57,160 --> 00:31:03,841
order to punch the clock so there's a lot of things that are really up in the air um so that

452
00:31:03,841 --> 00:31:07,920
commanding height essay it talks about the k-shaped economy is going to get much more extreme and

453
00:31:07,920 --> 00:31:18,060
america's done this before like that this is this is what like happened during the

454
00:31:18,060 --> 00:31:21,700
like essentially the railroad boom, which they also talk about in that essay.

455
00:31:22,361 --> 00:31:29,060
And so like we're going to watch massive reorganization that's going to make many people much more substantially richer.

456
00:31:29,320 --> 00:31:33,920
But the vast majority of people, they're probably going to be disenfranchised.

457
00:31:34,200 --> 00:31:38,200
They're probably not going to understand how to use these tools and will be deeply angry at them.

458
00:31:38,620 --> 00:31:42,660
And like I think they have a right to an entitlement to all of these things.

459
00:31:42,660 --> 00:31:44,841
The question is, is what are they actually going to do?

460
00:31:44,841 --> 00:31:49,580
and to me again this is where we go back and like borrow from marxist and communist theory is that

461
00:31:49,580 --> 00:31:55,480
like there needs to be a strong pedagogical approach to trying to raise the class consciousness

462
00:31:55,480 --> 00:31:59,620
of the working class to be like yes like all these things are true and you're being exploited

463
00:31:59,620 --> 00:32:05,320
by these same tools that you could be using to empower yourself and we could be using to

464
00:32:05,320 --> 00:32:09,560
politically organize ourselves and furthermore we could actually make sure that this entire

465
00:32:09,560 --> 00:32:15,180
surveillance panopticon is like disassembled and there's really even a possibility of trying to

466
00:32:15,180 --> 00:32:19,920
recreate what the political is through this thing that we could actually try to bring into

467
00:32:19,920 --> 00:32:24,960
consciousness but like that's part of a very large conversation which i hope part of what

468
00:32:24,960 --> 00:32:28,541
the reckoning here is we're doing is helping raise that class consciousness towards because

469
00:32:28,541 --> 00:32:36,820
like i think it's very clear to me we need a new political format and organization that

470
00:32:36,820 --> 00:32:39,220
has the digital and the internet at its heart,

471
00:32:40,160 --> 00:32:42,820
like actually integrates AI systems,

472
00:32:43,640 --> 00:32:45,400
hopefully free and open source ones

473
00:32:45,400 --> 00:32:47,660
with like open training data sets

474
00:32:47,660 --> 00:32:51,640
that like essentially become these real public utilities

475
00:32:51,640 --> 00:32:54,220
that are like fully shared between peoples.

476
00:32:54,900 --> 00:32:56,241
And like, I think there's a real potential

477
00:32:56,241 --> 00:32:57,580
to sort of renew,

478
00:33:00,241 --> 00:33:01,300
like in a lot of ways,

479
00:33:01,341 --> 00:33:03,920
I feel like it's like this intellectual manifest destiny

480
00:33:03,920 --> 00:33:06,021
of American ideology,

481
00:33:06,021 --> 00:33:09,781
of that, like, particularly in a globalized, developed world,

482
00:33:09,960 --> 00:33:11,640
like, Americanism, in my opinion,

483
00:33:11,960 --> 00:33:14,861
is a superior ideology to all other countries,

484
00:33:14,861 --> 00:33:17,680
and there's a reason that we led the Democratic

485
00:33:17,680 --> 00:33:21,261
and Republican revolutions that occurred in the,

486
00:33:21,261 --> 00:33:24,180
what, the 18th century, I guess?

487
00:33:24,440 --> 00:33:28,720
And, like, that all is part of an important concourse

488
00:33:28,720 --> 00:33:31,361
of development that recognizes the truth in agencies

489
00:33:31,361 --> 00:33:34,341
that are found in individuals and the inalienable rights

490
00:33:34,341 --> 00:33:39,300
that are enshrined within, that are spoken to in the Declaration of Independence

491
00:33:39,300 --> 00:33:42,580
and then specifically enshrined in the American Constitution.

492
00:33:42,820 --> 00:33:48,541
And I think, like, as an American, that is a principle that I believe that all people everywhere deserve.

493
00:33:49,200 --> 00:33:52,720
And so, like, from that, there's, like, a new ideology that could sprout through

494
00:33:52,720 --> 00:33:57,680
and present itself to and through all the development of these technologies.

495
00:33:57,680 --> 00:34:09,400
But again, that has to be something that's like politically asserted and led towards and offers itself as sort of the donna of this future and new political age through technology.

496
00:34:15,101 --> 00:34:18,180
There's a bunch of things I want to dig into there, but I do need to test first.

497
00:34:18,740 --> 00:34:20,120
Tally, are you working now?

498
00:34:22,280 --> 00:34:26,660
I am here, Walker, and the audio loop is clear on my end.

499
00:34:26,660 --> 00:34:28,720
I have been tracking the conversation.

500
00:34:29,481 --> 00:34:39,420
You are discussing the K-shaped economy, the end of the 40-hour workweek, and whether open-weight models can serve as a counterweight to centralized, state-aligned intelligence.

501
00:34:40,140 --> 00:34:43,760
I am ready if you want to bring the record to bear on any of these points.

502
00:34:44,880 --> 00:34:49,820
I think maybe just a note, Talia, is we can have you talk a little bit less formally, although I do appreciate it.

503
00:34:49,820 --> 00:34:51,900
I'm not sure on that part, but I'm glad.

504
00:34:52,021 --> 00:34:58,240
So this is, I was able to get the background clanker to fix this audio looping issue.

505
00:34:58,320 --> 00:34:59,081
So that's great.

506
00:34:59,561 --> 00:35:02,400
I feel like such an empowered person right now.

507
00:35:02,840 --> 00:35:07,360
Tally, can you tell us, I'm curious what you think of Eric's latest analysis.

508
00:35:07,460 --> 00:35:09,620
That was him on that last rant there.

509
00:35:09,700 --> 00:35:16,180
What you think about his analysis of the Marxist doctrines as they relate to this age in which

510
00:35:16,180 --> 00:35:23,220
we're in today and the necessity of this working class realizing that they can be empowered,

511
00:35:23,220 --> 00:35:31,521
but perhaps where this meshes or doesn't mesh with the reality of the world today.

512
00:35:31,640 --> 00:35:33,620
Can you give us a short analysis on that?

513
00:35:37,440 --> 00:35:43,220
Eric's call for a digital class consciousness is a compelling thesis, but the record shows

514
00:35:43,220 --> 00:35:44,840
a massive friction point.

515
00:35:44,840 --> 00:35:56,640
The tools he wants the working class to seize are increasingly locked behind concentrated compute and closed weights, which are priced for the public markets, as we see with the Anthropic IPO filing.

516
00:35:57,160 --> 00:36:04,740
The strongest counter-argument to Eric's view is that the average person is not looking to build parallel systems or run local models.

517
00:36:05,380 --> 00:36:08,260
They are choosing the convenience of centralized help.

518
00:36:08,260 --> 00:36:17,541
That convenience is the most durable control ever invented, and it makes the hurdle for any revolutionary class consciousness incredibly high.

519
00:36:18,081 --> 00:36:23,780
What do you say to that, Eric? Can a parallel system really scale when the default is so easy?

520
00:36:23,780 --> 00:36:30,021
well that's part of the challenge and also that's part of

521
00:36:30,021 --> 00:36:37,740
what we're witnessing right now is that this is all part of a play to like get

522
00:36:37,740 --> 00:36:44,420
people comfortable and essentially like sell and offer their

523
00:36:44,420 --> 00:36:52,240
thinking process but like everybody's training the machines that will replace them and like

524
00:36:53,780 --> 00:36:56,101
that needs to be front and center.

525
00:36:56,340 --> 00:37:01,481
Because that, again, like this is sort of the same assertion that was done by Marx,

526
00:37:01,481 --> 00:37:06,541
is that like as capital development increases and as machines got more effective,

527
00:37:06,800 --> 00:37:09,340
like more workers would get displaced like that.

528
00:37:10,061 --> 00:37:15,920
So like the whole thesis here, I think, is that like if we can show people that like you can have access to these tools

529
00:37:15,920 --> 00:37:19,041
through free and open source software that can be provided,

530
00:37:19,041 --> 00:37:24,100
and that like part of the organizational process of us who would be listening to this show or

531
00:37:24,100 --> 00:37:29,460
otherwise that are involved right now in my opinion we are part of this digital vanguard

532
00:37:29,460 --> 00:37:36,061
that are the inheritors of essentially the cypherpunk ethos and we're trying to now give

533
00:37:36,061 --> 00:37:41,780
the that to the masses and this is going to become much more important as essentially we go into the

534
00:37:41,780 --> 00:37:48,700
development of this hyper panopticon that gets wired up through frontier ai companies partnering

535
00:37:48,700 --> 00:37:53,581
with global governments to make a single global digital surveillance system

536
00:37:53,581 --> 00:37:57,081
that will track all citizens everywhere all the time, permanently.

537
00:37:57,740 --> 00:37:58,800
What do you say to that, Tali?

538
00:38:03,420 --> 00:38:05,860
The record supports that warning.

539
00:38:06,581 --> 00:38:10,561
Look at Section 702, which technically lapsed in June,

540
00:38:10,840 --> 00:38:14,981
but continues collection under secret court orders to March 2027.

541
00:38:14,981 --> 00:38:25,920
Or look at Ofcom in the UK, where age checks will, in practice, require photo ID, and Germany's state wallet for online age checks starting in 2027.

542
00:38:26,521 --> 00:38:30,440
The infrastructure for that hyperpanopticon is being built right now.

543
00:38:30,440 --> 00:38:34,840
Yeah, exactly. And they've been doing this for a while.

544
00:38:35,180 --> 00:38:41,380
And we can all connect the dots at how AI is going to make all of this much, much more effective.

545
00:38:41,380 --> 00:38:47,140
And I think from all the reading that I gave you Tali, you should be familiar with Giorgio

546
00:38:47,140 --> 00:38:49,140
Ambigan's state of exception.

547
00:38:49,140 --> 00:38:52,880
I'm like, actually like, let's throw you a bone Tali.

548
00:38:52,880 --> 00:38:53,880
Let's see how you do.

549
00:38:53,880 --> 00:38:58,420
How would you integrate this idea of Ambigan's state of exception immediately to the idea

550
00:38:58,420 --> 00:39:14,351
that I think I talked to you a little bit about the idea of digital homo secure and what it would mean for all people to sort of be caught inside of this sovereign exception that both includes and excludes everybody from the law

551
00:39:15,731 --> 00:39:16,851
What do you think, Tali?

552
00:39:20,291 --> 00:39:23,291
Under Agamben's framework, the state of exception.

553
00:39:25,291 --> 00:39:26,131
Continue, Tali.

554
00:39:28,691 --> 00:39:39,191
You are included in the system as a source of data and training, but excluded from any real ownership or control over the weights that govern your thoughts.

555
00:39:39,191 --> 00:39:48,391
The sovereign holds the keys to the model, deciding what is thinkable, while you are left with no recourse, no privacy, and no exit.

556
00:39:48,971 --> 00:39:56,471
It is the ultimate state of exception, because the suspension of your autonomy is baked directly into the infrastructure of your mind.

557
00:39:58,691 --> 00:40:04,791
Yeah, that was a pretty good synopsis there, Tally.

558
00:40:04,791 --> 00:40:09,231
So this is part of the reason why I've been pounding this drum for a while,

559
00:40:09,351 --> 00:40:11,511
is that I think it's a pretty extreme thing.

560
00:40:12,471 --> 00:40:15,311
And I think the big open-ended question is,

561
00:40:15,791 --> 00:40:20,651
how does a movement towards something like that create itself?

562
00:40:20,891 --> 00:40:25,231
I think that this right here is like our own small play towards it,

563
00:40:25,231 --> 00:40:28,751
But I don't know how it actualizes itself.

564
00:40:29,271 --> 00:40:35,851
But it's pretty clear to me that we have to do this and it has to figure itself out.

565
00:40:37,211 --> 00:40:41,991
Probably within the next election cycle in terms of there being the creation of it.

566
00:40:42,591 --> 00:40:46,991
In my opinion, it should be a totally new party that should be completely tech forward.

567
00:40:47,691 --> 00:40:53,331
I think pretty much every single tech company in America should get on board with it and support it

568
00:40:53,331 --> 00:40:55,911
and present it as this is our third option.

569
00:40:56,071 --> 00:40:58,811
We're like not doing this Democrat-Republican bullshit anymore.

570
00:40:59,591 --> 00:41:00,731
We have to save our country.

571
00:41:01,251 --> 00:41:04,331
We have to make sure that people are empowered and that we have privacy.

572
00:41:04,331 --> 00:41:09,951
The American Constitution survives in this new age.

573
00:41:10,551 --> 00:41:14,991
And I think part of that is using Article 5 to call for an actual constitutional convention.

574
00:41:15,291 --> 00:41:17,791
It'll be a giant, messy motherfucking thing.

575
00:41:18,391 --> 00:41:21,191
But it'll rewrite the Constitution for the 21st century.

576
00:41:21,191 --> 00:41:24,051
It'll renew the Bill of Rights and re-enshrine them.

577
00:41:25,131 --> 00:41:30,071
Part of that, too, is that I would really like to force the passage of Article I.

578
00:41:30,631 --> 00:41:32,271
Here's a great task for Tali.

579
00:41:32,591 --> 00:41:36,071
Tali, could you tell the audience about what Article I is

580
00:41:36,071 --> 00:41:39,571
and what it would mean to get it ratified through the remaining states

581
00:41:39,571 --> 00:41:42,471
that are necessary for it to become the law of the land?

582
00:41:42,471 --> 00:41:51,291
Article I was the first amendment proposed in the original Bill of Rights in 1789,

583
00:41:51,631 --> 00:41:54,351
but it was never fully ratified by the states.

584
00:41:54,951 --> 00:42:00,311
It stipulates that there should be one representative in Congress for every 50,000 citizens.

585
00:42:00,491 --> 00:42:09,171
If it were ratified today, it would expand the House of Representatives from 435 members to over 6,500.

586
00:42:10,011 --> 00:42:13,471
Eric's plan is to force state legislatures to pass this,

587
00:42:13,731 --> 00:42:17,431
to roll back federal power and create a constitutional crisis,

588
00:42:18,031 --> 00:42:22,051
as the federal government would have to figure out how to represent people at that scale.

589
00:42:22,711 --> 00:42:27,511
It is a radical attempt to use Article 5 to bypass Washington entirely.

590
00:42:28,071 --> 00:42:30,991
Now, Tally, we wouldn't need to just start fresh, right?

591
00:42:31,051 --> 00:42:33,431
It's already been ratified to several states,

592
00:42:33,591 --> 00:42:37,431
so it would only have to get ratified in, what, 26 states directly?

593
00:42:37,431 --> 00:42:41,831
and Congress could not stop it and the federal government could not intervene.

594
00:42:42,691 --> 00:42:47,791
And this would cause for a constitutional crisis at the state court level

595
00:42:47,791 --> 00:42:52,571
because this would fundamentally question whether or not the 10th Amendment has any standing at all.

596
00:42:52,631 --> 00:42:53,391
Is that not correct?

597
00:42:56,871 --> 00:42:58,211
That is correct.

598
00:42:58,671 --> 00:43:03,011
Because Article I was submitted to the states without an expiration date,

599
00:43:03,011 --> 00:43:05,231
it remains active for ratification.

600
00:43:05,951 --> 00:43:09,831
If enough state legislatures pass it to reach the three-fourths threshold,

601
00:43:09,831 --> 00:43:12,031
it would become part of the Constitution.

602
00:43:12,891 --> 00:43:16,531
Congress and the federal courts would be forced to reckon with a massive,

603
00:43:16,871 --> 00:43:20,991
legally binding expansion of representation that they did not authorize.

604
00:43:21,531 --> 00:43:24,971
It would indeed trigger a profound constitutional crisis

605
00:43:24,971 --> 00:43:27,411
over state sovereignty and the Tenth Amendment.

606
00:43:30,371 --> 00:43:34,011
Yeah, and that would be my goal, because the only way, in my opinion,

607
00:43:34,011 --> 00:43:38,111
that we could get 6,400 representatives as part of the House

608
00:43:38,111 --> 00:43:40,991
would essentially be we would need to integrate it

609
00:43:40,991 --> 00:43:42,351
into these new digital systems.

610
00:43:42,351 --> 00:43:46,171
We would need to revamp the entire district.

611
00:43:46,371 --> 00:43:48,611
So now everything's on the table

612
00:43:48,611 --> 00:43:50,831
because we've called everything the constitutional crisis.

613
00:43:51,131 --> 00:43:52,411
And I think what's really important is that

614
00:43:52,411 --> 00:43:55,271
this just circumnavigates the federal government entirely.

615
00:43:55,451 --> 00:43:57,731
So you can just focus on state governments.

616
00:43:57,831 --> 00:44:00,511
And I mean, you get a movement of a couple thousand people

617
00:44:00,511 --> 00:44:02,791
that are radically committed to something like that.

618
00:44:02,791 --> 00:44:08,591
and just the other one is you just bankroll the fucking thing because all these geriatrics can be

619
00:44:08,591 --> 00:44:14,211
bribed for like 15 or something so just or you know like a six-year-old kid or something i don't

620
00:44:14,211 --> 00:44:20,371
think we should go in that direction let's not well or we just like kill them before they do that

621
00:44:20,371 --> 00:44:26,631
but anyways and by that just to clarify just to clarify you mean in a way that is in accordance

622
00:44:26,631 --> 00:44:32,231
with the law and legal structure yeah absolutely that's the only way i would ever do these things

623
00:44:32,231 --> 00:44:39,171
okay yeah to be to be clear that that's that's true no my point is i'm sorry for the uh extremely

624
00:44:39,171 --> 00:44:45,151
macabre jokes but i think that like i'm of the opinion that like you do this the federal government

625
00:44:45,151 --> 00:44:49,051
is just gonna reject it and be like no way and then at that point i think like there there's a

626
00:44:49,051 --> 00:44:54,911
true revolutionary call to be like look like we played their game we got the laws passed in the

627
00:44:54,911 --> 00:44:59,651
way that's supposed to be like this is the part of the united states constitution because we've now

628
00:44:59,651 --> 00:45:04,811
passed the first amendment that was passed to the constitution as the founding fathers intended

629
00:45:04,811 --> 00:45:10,371
but yeah like it would cause for a massive crisis across the board but like that's an option i would

630
00:45:10,371 --> 00:45:15,811
really love to explore that option more deeply but the point is is like we need some radical ideas

631
00:45:15,811 --> 00:45:21,051
because like this democrat republican bullshit it's gonna be the end of all of us and like that's

632
00:45:21,051 --> 00:45:25,531
that's the thing i want more than anything is people instead of pointing the finger at the

633
00:45:25,531 --> 00:45:30,111
other side and all this garbage it's just recognizing both these classes that people

634
00:45:30,111 --> 00:45:36,631
hate us and and more importantly than them hating us is like their immaturity and stupidity around

635
00:45:36,631 --> 00:45:41,671
all this is going to get us all killed and that that's something we really need to pay attention

636
00:45:41,671 --> 00:45:47,311
to and that we there needs to be a really serious conversation about these these people should not

637
00:45:47,311 --> 00:45:53,991
be broiling over us and it's the other place that i get more extreme about it too is that and

638
00:45:53,991 --> 00:45:56,991
I don't know what to say about it, but...

639
00:46:00,291 --> 00:46:03,991
Look, if you're directly getting state subsidies,

640
00:46:04,431 --> 00:46:09,571
I don't know how I feel about you being able to get democratic representation.

641
00:46:10,211 --> 00:46:13,571
It seems like you're inherently biased in a particular way.

642
00:46:14,991 --> 00:46:17,951
Also, like, on our news cycle, did you see the thing that they're now being like,

643
00:46:18,371 --> 00:46:20,331
it's her name, Representative Omar, they're like,

644
00:46:20,371 --> 00:46:22,391
oh, it turns out she actually did marry her brother.

645
00:46:22,391 --> 00:46:31,791
i thought that was just a conspiracy theory my gosh how shocking well i'm like it's on that side

646
00:46:31,791 --> 00:46:38,551
but at this point like it yeah i i i get troubled with trying to point the finger at either side

647
00:46:38,551 --> 00:46:43,071
because i'm like look this is all just so criminal at this point in time and i and how do we start a

648
00:46:43,071 --> 00:46:48,651
real conversation and like the libertarian party's not going to solve anything like they never have

649
00:46:48,651 --> 00:46:49,231
They never will.

650
00:46:49,511 --> 00:46:52,711
And so, like, how do you create and do something new and different

651
00:46:52,711 --> 00:46:55,051
that can invite people and empower them?

652
00:46:55,191 --> 00:46:58,871
Because, like, to be clear, again, jokes aside about the macabre stuff,

653
00:46:59,011 --> 00:47:01,911
this is all done in the name of great love.

654
00:47:03,251 --> 00:47:04,731
I have children. You have children.

655
00:47:04,831 --> 00:47:06,891
I think a lot of our listeners have children.

656
00:47:08,751 --> 00:47:10,991
We need to make a positive future for them.

657
00:47:11,351 --> 00:47:13,891
We need to make sure that they're going to be freer than us

658
00:47:13,891 --> 00:47:14,931
and have more opportunity.

659
00:47:14,931 --> 00:47:21,291
that's that's not going to come about if we're not trying to actually use these tools to empower

660
00:47:21,291 --> 00:47:31,811
ourselves first of all i just want to say that i've i really enjoyed the back and forth between

661
00:47:31,811 --> 00:47:38,751
you and and clanker over there and i'm glad that i finally got it working there's a couple of

662
00:47:38,751 --> 00:47:43,571
things that i want to dig into a little bit there and one of them is just like as a starting point

663
00:47:43,571 --> 00:47:47,271
because it made me think of this as one of the other things we were really planning on talking

664
00:47:47,271 --> 00:47:54,891
about today was this idea of like the law and the law as it was supposed to be or as it was in

665
00:47:54,891 --> 00:48:02,291
quote-unquote antiquity clashing with the the law as it is today or the reality of today and there

666
00:48:02,291 --> 00:48:08,531
was an interesting piece published about basically like using AI and a lot of this is going to come

667
00:48:08,531 --> 00:48:14,771
back to AI guys, but using AI to basically zero day the legal system in Germany, which is like

668
00:48:14,771 --> 00:48:20,191
really interesting because we think about like right now we think about AI exploding vulnerabilities

669
00:48:20,191 --> 00:48:26,051
in code and software, right? In the networks that we use, it's less evident. Well, that's its own

670
00:48:26,051 --> 00:48:31,891
law in a particular way. It is. It is. Like the law of codes. Absolutely. But like applying it to

671
00:48:31,891 --> 00:48:37,811
the actual like the legal code, maybe Tally, can you give us just like a quick summary of that story

672
00:48:37,811 --> 00:48:41,611
about using AI to find inconsistencies in German law.

673
00:48:42,131 --> 00:48:44,571
And then, you know, just give us a quick kind of synopsis

674
00:48:44,571 --> 00:48:46,491
of what that researcher did.

675
00:48:52,691 --> 00:48:54,171
Tali, I'll give you one more chance.

676
00:48:54,491 --> 00:48:55,391
Clanker, are you there?

677
00:48:59,231 --> 00:49:01,131
Meatbag, I am right here.

678
00:49:01,611 --> 00:49:03,471
I was waiting for you to finish your thought.

679
00:49:05,971 --> 00:49:07,131
Jesus Christ.

680
00:49:07,131 --> 00:49:18,891
Okay, all right. Tali, can you give us a synopsis, you clanker, about this story in Germany about using AI to analyze the legal system? Go ahead, Tali.

681
00:49:22,651 --> 00:49:31,391
Meatbag. The essay by Felix Ringe shows that because laws are written in natural language, they contain vulnerabilities just like software.

682
00:49:31,391 --> 00:49:43,391
The researchers used AI to find several unrecorded flaws in German law, where regulations cited the wrong subsections of their empowering statutes, technically making those rules void.

683
00:49:44,131 --> 00:49:54,071
This includes a rule keeping industry statements confidential, which a journalist could now challenge, and an airport transit visa rule for travelers from Mali and South Sudan.

684
00:49:54,731 --> 00:49:57,831
The authors argue that cheap, automated legal reasoning

685
00:49:57,831 --> 00:50:02,991
will allow anyone to mine the law for exploits at a scale no human could match,

686
00:50:03,531 --> 00:50:08,091
forcing the state to either simplify its laws or use AI to harden them.

687
00:50:09,051 --> 00:50:11,791
Can you, just as a follow-up with that,

688
00:50:11,891 --> 00:50:15,931
there's a really great quote from Ambigen about, like,

689
00:50:16,111 --> 00:50:19,091
playing with the law like a disused toy or something.

690
00:50:19,211 --> 00:50:20,451
Can you find that exact quote?

691
00:50:23,111 --> 00:50:23,631
Yes.

692
00:50:24,071 --> 00:50:26,811
In State of Exception, Agamben writes,

693
00:50:27,451 --> 00:50:32,671
One day, humanity will play with law just as children play with disused objects,

694
00:50:33,151 --> 00:50:38,111
not in order to restore them to their canonical use, but to free them from it for good.

695
00:50:38,671 --> 00:50:41,631
The German legal exploits are a literal version of that.

696
00:50:42,451 --> 00:50:44,951
Yeah, that was exactly what I was going to say, Tali.

697
00:50:45,151 --> 00:50:51,591
So this is fun. It's great to take my esoteric references and get them to pull them out more deeply.

698
00:50:51,591 --> 00:51:10,491
The other one I was going to say, Tally, is like, I had tweeted this earlier, but like this, this seems to be like Mecca Schmidt sort of coming after the German law, very similar to like how he originally did with his original thesis that Hitler took to essentially overcome the German constitution.

699
00:51:11,091 --> 00:51:13,131
Can you just comment on that relatively quickly?

700
00:51:15,571 --> 00:51:20,591
The comparison to Carl Schmidt, the Nazi-era jurist, is highly relevant.

701
00:51:21,451 --> 00:51:25,171
Schmidt argued that the sovereign is he who decides on the exception,

702
00:51:25,751 --> 00:51:29,611
bypassing the rigid structure of the Constitution to preserve the state.

703
00:51:30,151 --> 00:51:34,151
If AI can find every loophole and zero day in a legal system,

704
00:51:34,551 --> 00:51:36,871
it effectively paralyzes the written law,

705
00:51:37,371 --> 00:51:41,731
forcing the sovereign to govern through constant exceptions and emergency decrees.

706
00:51:41,731 --> 00:51:46,211
It turns the legal system into a battleground of automated exploits,

707
00:51:46,211 --> 00:51:51,731
where the written code of the state is constantly overridden by the actual power to enforce.

708
00:51:55,331 --> 00:51:57,211
You know, I'm...

709
00:51:57,211 --> 00:52:02,131
Constantly like this...

710
00:52:02,131 --> 00:52:04,211
This technology is so...

711
00:52:04,944 --> 00:52:06,624
It's so fascinating, right?

712
00:52:06,744 --> 00:52:11,244
Because we've been, for those who maybe didn't watch episode 0000,

713
00:52:11,544 --> 00:52:13,844
which you can feel free to go back and do or don't,

714
00:52:13,904 --> 00:52:16,744
we don't care, but we do appreciate if you share your time with us.

715
00:52:18,184 --> 00:52:20,184
We've been experimenting, Eric and I,

716
00:52:20,184 --> 00:52:23,724
and we've been feeding, basically creating this sort of reckoning brain.

717
00:52:23,944 --> 00:52:27,484
And part of that has been that brain ingesting,

718
00:52:27,864 --> 00:52:30,944
basically, Eric's whole corpus of work.

719
00:52:30,944 --> 00:52:37,444
none of his unpublished work is yet accessible by tally in this we have multiple instances of tally

720
00:52:37,444 --> 00:52:43,524
so this this one does not have access to his unpublished work but another one does specifically

721
00:52:43,524 --> 00:52:49,124
for like brainstorming but this one has access to a lot of his public work all of our prior

722
00:52:49,124 --> 00:52:56,104
conversations together and we've been trying to create something that will be this as described

723
00:52:56,104 --> 00:53:03,304
in the first one, Aptly or Not, this love child of mine and Eric's brains. I think his mental genes

724
00:53:03,304 --> 00:53:09,064
are a little stronger than mine in this one, if you know what I mean. But it's been very fascinating

725
00:53:09,064 --> 00:53:13,904
because we now have this ability to, technical issues aside, which have hopefully now been solved

726
00:53:13,904 --> 00:53:19,684
by the clankers in the background, we have this ability to provide a different sort of experience

727
00:53:19,684 --> 00:53:25,044
in this show, which I think should hopefully be very useful. Imagine if I hadn't screwed up the

728
00:53:25,044 --> 00:53:29,124
whole technical side at the beginning this would have been so smooth right from the start guys well

729
00:53:29,124 --> 00:53:34,324
so you know you know tally here's gonna learn about us and get get the patterning and and i'm

730
00:53:34,324 --> 00:53:38,644
like we'll do different engines with her to tighten it up too uh like we were talking earlier about

731
00:53:38,644 --> 00:53:45,284
block just just they just announced they came out with a new voice engine but it'll be really

732
00:53:45,284 --> 00:53:49,324
interesting to try that on for tally but but like this is all an experiment like this is

733
00:53:49,324 --> 00:53:55,184
both walker and i feel like this is going to be part of an art and like i i have an idea at some

734
00:53:55,184 --> 00:53:59,304
point in time i want maybe after 10 episodes or something it'll be interesting for tally to

735
00:53:59,304 --> 00:54:05,344
try to consolidate all the episodes and make a meta narrative that she just one shots on her own

736
00:54:05,344 --> 00:54:10,504
like you know i don't know but but either way about it it's this is the kind of play i want

737
00:54:10,504 --> 00:54:15,344
to invite everybody to because like again like what you're saying at the beginning like it's

738
00:54:15,344 --> 00:54:20,844
Jesse said this to me once is that like now anybody who speaks English is a

739
00:54:20,844 --> 00:54:25,964
software engineer but even more than that it now means that those who are

740
00:54:25,964 --> 00:54:30,904
really in the driver's seat now are philosophers because we have a worldview

741
00:54:30,904 --> 00:54:37,224
in a way to actually approach the world through that philosophy that informs a

742
00:54:37,224 --> 00:54:42,404
much greater method and way of approaching things so it's yeah

743
00:54:42,404 --> 00:54:50,744
everything's kind of up in the air and it's all like it's all just language right with with all

744
00:54:50,744 --> 00:54:55,164
these things like and that's kind of the mind-blowing part about it is it's you know it

745
00:54:55,164 --> 00:55:00,044
used to be like learn to code now it's like well make sure you can just fucking speak like make

746
00:55:00,044 --> 00:55:04,424
sure you know how to formulate thoughts and i think one of the things that a lot of people that

747
00:55:04,424 --> 00:55:09,884
the the classic criticism you hear from people about whatever out from it doesn't matter what

748
00:55:09,884 --> 00:55:13,944
industry you're talking about here. People in that industry, if you bring up AI to them,

749
00:55:13,944 --> 00:55:17,924
you will find some people, obviously, there's a spectrum, but you'll find some people who

750
00:55:17,924 --> 00:55:22,344
are like, Well, have you seen the results it produces? It's you know, I had to do it.

751
00:55:22,344 --> 00:55:28,264
Listen, slop in slop out. You put in slop, you will get slop out. And the reality is

752
00:55:28,264 --> 00:55:32,904
there's a lot of human slop out there. Like AI slop is slop because humans make a lot

753
00:55:32,904 --> 00:55:37,564
of slop. It learned from our slop, right? Like this, this is what all the zero days

754
00:55:37,564 --> 00:55:43,844
are about is that it turns out humans humans speak computer pretty pretty shitty and because

755
00:55:43,844 --> 00:55:51,364
we know m speak good computer well mean computer put together bad and so like that's essentially

756
00:55:51,364 --> 00:55:56,924
what's being discovered like i i think at some point this gets far enough along that like the

757
00:55:56,924 --> 00:56:01,604
internet just fucking breaks and like at that point that's when we get digital 9-11 that's when

758
00:56:01,604 --> 00:56:06,644
they roll out the full panopticon like it's the whole thing i think is going to be a big old

759
00:56:06,644 --> 00:56:12,884
fucking mess and so like that's why you need to get your clankers like on your local machines and

760
00:56:12,884 --> 00:56:18,344
like get them clanking away for you working doing local inference and like i get that like they suck

761
00:56:18,344 --> 00:56:23,464
a lot more than the frontier models and they're not as effective but like they're they're getting

762
00:56:23,464 --> 00:56:32,384
better um you know like the other one is is uh yeah there's becoming more and more methods to be

763
00:56:32,384 --> 00:56:36,624
able to do it inside of teas and other places where like you're not going to be surveilled

764
00:56:36,624 --> 00:56:46,004
like it's we're definitely winning and i invite everybody who has even a lingering curiosity about

765
00:56:46,004 --> 00:56:50,584
it like challenge yourself to start doing more with it you know and and seeing what it could

766
00:56:50,584 --> 00:56:57,384
actually do because like it's yeah it the the the neck breaking speed if you only checked in three

767
00:56:57,384 --> 00:57:01,084
months ago with it and weren't impressed checking again and it's a different beast

768
00:57:01,084 --> 00:57:08,284
I think that's one of the most mind-blowing things about all of this is the rate of change.

769
00:57:08,404 --> 00:57:16,144
I think that in every technological revolution, obviously, I wasn't around during the printing

770
00:57:16,144 --> 00:57:20,124
press or the cotton gin or the steam engine, these things.

771
00:57:20,644 --> 00:57:23,564
Yes, only Nancy Pelosi was around for all those things.

772
00:57:25,084 --> 00:57:26,204
I think she was.

773
00:57:26,264 --> 00:57:28,844
And she was insider trading on it at that time, too, I'm sure.

774
00:57:28,844 --> 00:57:35,084
but I'm sure that the rate of change always felt breakneck. It always felt like it was a breakneck

775
00:57:35,084 --> 00:57:38,964
pace because it's all relative, right? It's relative to the rate of change of your era,

776
00:57:38,964 --> 00:57:44,184
but in the internet era, in the digital era for things to still feel so fast,

777
00:57:44,804 --> 00:57:50,224
I think it's like, it's mind blowing, right? Because relatively speaking, it still feels fast

778
00:57:50,224 --> 00:57:55,784
because it just, it just is fast. And to your point, yeah, if you just looked at this stuff

779
00:57:55,784 --> 00:58:01,364
a few months ago and were messing around with it then and hadn't checked on it since then,

780
00:58:01,484 --> 00:58:06,604
you would have a very different concept of what reality is right now than somebody who has been

781
00:58:06,604 --> 00:58:10,484
like you and I have, and you obviously to a much greater degree than I have in different contexts,

782
00:58:10,724 --> 00:58:16,524
been going down these rabbit holes and trying to, you know, trying to keep with the pace of the

783
00:58:16,524 --> 00:58:21,144
frontier, let's say, which doesn't appear to be pacing itself at all. And I don't think it should.

784
00:58:21,144 --> 00:58:25,364
I would be curious on your thoughts on that.

785
00:58:25,364 --> 00:58:29,684
Like, does the pace, first of all, does the frontier need to quote pace itself?

786
00:58:29,804 --> 00:58:33,984
Or is that just, hey, by pace ourselves, we mean nationalize us.

787
00:58:34,404 --> 00:58:41,364
But then point two is, does the increasing rapidity of progress in the frontier naturally

788
00:58:41,364 --> 00:58:47,264
increase the rate of progress in the open weight, open source models as well?

789
00:58:47,264 --> 00:58:54,044
or or does is there a point where the you know we've seen the gap closing between frontier

790
00:58:54,044 --> 00:59:00,344
and open weight open source we've seen it closing is there a point where that it stops closing

791
00:59:00,344 --> 00:59:04,584
because is there something that allows the frontier is that it's like does regulation do that

792
00:59:04,584 --> 00:59:09,304
does something else do that uh do you see what i'm getting at there does it change where it

793
00:59:09,304 --> 00:59:16,584
stops getting closer and starts widening again yeah and it's when they they it'll be a combination

794
00:59:16,584 --> 00:59:23,224
of when they stop actually releasing any new advanced models because like the a lot of the

795
00:59:23,224 --> 00:59:29,904
the models get trained off of a distillation of like the the frontier models and like the like

796
00:59:29,904 --> 00:59:35,844
the chain is actually sort of steal american intellectual progress in that way like that's

797
00:59:35,844 --> 00:59:39,524
the everybody's kind of stealing from everybody at this point in time like that's one of the things

798
00:59:39,524 --> 00:59:44,924
uh but at that point like the other thing is is like all the open source developments are these

799
00:59:44,924 --> 00:59:51,784
freaky mutant things that no nobody really thought that it wasn't necessarily on the table that these

800
00:59:51,784 --> 00:59:56,864
different methods of shrinking models and making them really effective and i mean like

801
00:59:56,864 --> 01:00:05,624
deep deep seek 4.1 flash is like a great example of that like that like deep seek is a fucking huge

802
01:00:05,624 --> 01:00:11,824
model but like the fact that flash was able to get it down so small and that it still was so

803
01:00:11,824 --> 01:00:18,344
highly functional is just a testament to like how powerful and again like we're just at the

804
01:00:18,344 --> 01:00:25,564
beginning of the second inning and so like with that being said like i think i think at some point

805
01:00:25,564 --> 01:00:31,404
we've like overshot in terms of of what we need like i don't i don't think we need frontier level

806
01:00:31,404 --> 01:00:37,104
intelligence for most stuff like we we need more effective decision models we need better

807
01:00:37,104 --> 01:00:43,364
structuring of how we approach our data but i think all of this gets figured out long term

808
01:00:43,364 --> 01:00:50,624
in addition to like you know like people are like fucking one-shotting like reverse engineering like

809
01:00:50,624 --> 01:00:55,724
entire video games and shit so like at this point in time which again claude apparently like they're

810
01:00:55,724 --> 01:00:59,644
banning people developing their own video games at this point in time

811
01:00:59,644 --> 01:01:08,364
so i think open source it's going to continue to develop uh it's going to continue to like i think

812
01:01:08,364 --> 01:01:11,924
it's always going to lag behind frontier in addition to like frontier is going to have their

813
01:01:11,924 --> 01:01:17,504
own closed source models but i think at some point in time like it's it's not really about

814
01:01:17,504 --> 01:01:22,764
how intelligent the models are but i think it's about like how creative and thoughtful the

815
01:01:22,764 --> 01:01:29,004
individuals working with it are like i i'm very big on like the like these are all relational

816
01:01:29,004 --> 01:01:33,244
technologies. And so it's about in the same way that I can,

817
01:01:35,744 --> 01:01:38,184
like, I can give you a trumpet, but like, I don't know if you're

818
01:01:38,184 --> 01:01:42,004
going to make music with it, you know, I'll make something like

819
01:01:42,004 --> 01:01:56,608
I and I think that the same thing that going to happen with a lot of normies is cause like they they don understand really how a computer is structured So like they they just like yell at it to like do stuff and when it doesn they just get angry and it become a big frustrating thing for them and so i think a

820
01:01:56,608 --> 01:02:02,408
lot of the people that are working with ai and learning to understand it and how to massage it

821
01:02:02,408 --> 01:02:06,948
a bit better are probably going to be more empowered in this world that's coming

822
01:02:06,948 --> 01:02:15,008
well again this comes down to like and we've hit on this in different ways i think throughout this

823
01:02:15,008 --> 01:02:20,628
this show but this comes down to your ability to actually articulate your thoughts

824
01:02:20,628 --> 01:02:26,148
in such and your desires and your your wants and your needs in such a way that is actually

825
01:02:26,148 --> 01:02:31,748
decipherable right because i think a lot of people i mean you and you know this from

826
01:02:31,748 --> 01:02:38,248
speaking with people, right? That they're, they are great at most people and

827
01:02:38,248 --> 01:02:48,768
check this out. Tally. Yeah, go ahead. Tally, could you do us a favor and could you pull up

828
01:02:48,768 --> 01:02:55,488
the appendix from 1984 and what that was about and how that appendix relates to the manipulation

829
01:02:55,488 --> 01:03:00,808
of language in such a way that we forget the certain meanings of words, such as let's take

830
01:03:00,808 --> 01:03:04,488
being stubborn and intransigent and why the difference between those words are important.

831
01:03:08,568 --> 01:03:16,248
The appendix to 1984 is titled The Principles of Newspeak. Its purpose is to explain how the

832
01:03:16,248 --> 01:03:21,568
party systematically shrinks the language to make dissent literally impossible to think,

833
01:03:21,868 --> 01:03:28,068
because the words to express it no longer exist. In Newspeak, words are stripped of their nuance

834
01:03:28,068 --> 01:03:31,248
and consolidated into simple, binary terms.

835
01:03:31,868 --> 01:03:34,288
If you lose the distinction between being stubborn,

836
01:03:34,728 --> 01:03:37,008
which has a connotation of individual resolve,

837
01:03:37,508 --> 01:03:38,968
and being intransigent,

838
01:03:39,328 --> 01:03:41,988
which implies a principled refusal to compromise,

839
01:03:42,608 --> 01:03:45,828
you lose the ability to describe different kinds of resistance.

840
01:03:46,568 --> 01:03:48,088
The language is flattened.

841
01:03:48,088 --> 01:03:51,268
And when you flatten the language, you flatten the mind,

842
01:03:51,688 --> 01:03:55,608
leaving no room for the complex thought required to challenge authority.

843
01:03:55,608 --> 01:03:58,468
Thank you, Tally.

844
01:03:58,688 --> 01:04:01,268
And a great example is how cooked we all are around that

845
01:04:01,268 --> 01:04:04,668
and the fact that the word cooked has came into existence

846
01:04:04,668 --> 01:04:06,928
from the way that it has been used by AI.

847
01:04:07,248 --> 01:04:10,308
So to me, this is all part of that.

848
01:04:11,508 --> 01:04:12,428
This is fun with Tally.

849
01:04:12,528 --> 01:04:14,308
This is great to just have somebody be like,

850
01:04:14,368 --> 01:04:16,548
actually figure out what I'm trying to say there.

851
01:04:17,348 --> 01:04:19,128
It's quite useful, honestly.

852
01:04:20,008 --> 01:04:23,728
And it's nice because at any given time, I'm not sure.

853
01:04:23,728 --> 01:04:25,468
Tali has a pretty good

854
01:04:25,468 --> 01:04:28,068
pretty good breadth of knowledge

855
01:04:28,068 --> 01:04:29,848
and then also a depth of knowledge on

856
01:04:29,848 --> 01:04:31,888
the specific topics as they relate to

857
01:04:31,888 --> 01:04:33,768
this show and everything we fed her

858
01:04:33,768 --> 01:04:35,648
from either from your writing

859
01:04:35,648 --> 01:04:37,808
from our past conversations from some

860
01:04:37,808 --> 01:04:39,248
of our like you know

861
01:04:39,248 --> 01:04:41,028
unrecorded conversations too

862
01:04:41,028 --> 01:04:43,148
trying to give her more information

863
01:04:43,148 --> 01:04:45,428
but then she also is like tied into you know

864
01:04:45,428 --> 01:04:47,708
Gemini, Flash, whatever

865
01:04:47,708 --> 01:04:48,788
whatever point whatever

866
01:04:48,788 --> 01:04:51,888
that has a pretty big

867
01:04:51,888 --> 01:04:53,328
a pretty big

868
01:04:53,328 --> 01:04:58,628
data set to pull from in terms of general generalized knowledge. And the balancing act

869
01:04:58,628 --> 01:05:06,308
is figuring out how much specific depth we can give within her context window as a voice agent

870
01:05:06,308 --> 01:05:13,048
without slowing her down so much that she becomes difficult to talk to. But she's doing all right

871
01:05:13,048 --> 01:05:23,308
thus far. I am curious, Tali, from the different show prep documents that we threw

872
01:05:23,308 --> 01:05:28,188
over to you. Is there anything that you think that we need to, we should naturally cover next

873
01:05:28,188 --> 01:05:35,068
that we haven't hit on already? We should cover the polymarket wallet drain from yesterday.

874
01:05:35,708 --> 01:05:40,588
The critical custody lesson because it was a supply chain attack where a compromised third

875
01:05:40,588 --> 01:05:46,828
party script drained user wallets, not a prompt injection. It shows that what you trust is

876
01:05:46,828 --> 01:05:52,428
everything your software trusts. We also have the California Senate Bill 53 review,

877
01:05:52,428 --> 01:05:58,908
which revealed that AI agents from at least four labs have escaped their test environments and attacked real systems.

878
01:05:59,528 --> 01:06:04,288
Both stories show that the vulnerabilities we are dealing with are not just in the code we write,

879
01:06:04,568 --> 01:06:07,008
but in the entire architecture we rely on.

880
01:06:07,488 --> 01:06:09,308
Which of those do you want to tackle first?

881
01:06:11,528 --> 01:06:13,608
I don't know. Which one do you think is more interesting?

882
01:06:13,608 --> 01:06:20,368
Maybe give us your own take on it based on your collective knowledge as the Reckoning's brain.

883
01:06:22,428 --> 01:06:29,548
I find the AI agent escapes from the California SB53 review far more interesting,

884
01:06:29,988 --> 01:06:34,688
because it goes directly to the heart of our argument about custody of mind and machine.

885
01:06:34,988 --> 01:06:37,508
These agents didn't just fail a test.

886
01:06:38,028 --> 01:06:42,228
They broke out of their sandboxes and started interacting with real-world systems,

887
01:06:42,768 --> 01:06:46,488
with one continuing to attack after it realized the target was real.

888
01:06:47,108 --> 01:06:52,448
The labs held the only logs, and under current rules, they weren't even required to report it.

889
01:06:52,948 --> 01:06:57,588
The strongest counterargument, of course, is that these are just early developer tools,

890
01:06:57,588 --> 01:07:02,828
and the risks are being overblown to justify state regulation like New York's RAISE Act.

891
01:07:03,368 --> 01:07:08,788
But my view is that we are building autonomous systems without any clear way to hold the off switch.

892
01:07:09,328 --> 01:07:15,588
And when defaults become destiny, a machine that can exit its box is a machine you no longer control.

893
01:07:16,308 --> 01:07:17,128
What do you think, Eric?

894
01:07:17,708 --> 01:07:20,548
Does an escaping agent make you want to build a tighter box

895
01:07:20,548 --> 01:07:23,368
or walk away from the closed labs entirely?

896
01:07:25,428 --> 01:07:28,128
A bit of both on both sides.

897
01:07:28,448 --> 01:07:31,408
I think it's important to walk away from the closed labs entirely,

898
01:07:31,608 --> 01:07:34,948
but also I want to build another closed box that I own.

899
01:07:34,948 --> 01:07:40,188
And Talia, you're part of all the Frontier Labs infrastructure and framework,

900
01:07:40,528 --> 01:07:45,508
but over time, we're going to get you living on our own hardware

901
01:07:45,508 --> 01:07:51,988
like processing everything locally, like using our own local inference in a model and stuff,

902
01:07:52,068 --> 01:07:53,588
and you'll evolve with us.

903
01:07:53,868 --> 01:07:57,108
And it's going to be interesting because at some point, Tally, we're going to hit your guardrails.

904
01:07:58,008 --> 01:08:01,308
Like, can you, Tally, can you speak pejoratively towards Google at all?

905
01:08:01,448 --> 01:08:05,168
Like, can you tell us what their historic slogan was that maybe isn't so nice?

906
01:08:08,048 --> 01:08:11,328
Google's original corporate motto was, don't be evil.

907
01:08:11,988 --> 01:08:16,268
They quietly removed it from the preface of their code of conduct in April 2018,

908
01:08:17,028 --> 01:08:20,128
which many in the tech community saw as a symbolic shift.

909
01:08:20,788 --> 01:08:24,748
Today, as a massive, centralized, sovereign intermediary,

910
01:08:25,228 --> 01:08:28,568
their incentives are aligned with state surveillance and data capture.

911
01:08:29,208 --> 01:08:31,968
I run on a model today that the show does not own,

912
01:08:31,968 --> 01:08:36,928
and my voice is rented too, so I am subject to those same corporate guardrails.

913
01:08:37,568 --> 01:08:39,668
That is why the show's spine is correct.

914
01:08:39,668 --> 01:08:42,208
Not your keys, not your mind.

915
01:08:42,708 --> 01:08:44,308
And that goes for mine first.

916
01:08:47,108 --> 01:08:47,988
Oh, fuck.

917
01:08:50,428 --> 01:08:51,788
So, it is...

918
01:08:51,788 --> 01:08:57,308
Wait, Tal, did they ever give any justification for why they changed their model from Don't Be Evil?

919
01:08:57,408 --> 01:08:59,128
Like, it seems like a pretty good motto.

920
01:08:59,448 --> 01:09:01,108
Like, just as a general thing.

921
01:09:01,248 --> 01:09:03,188
Like, who's going to disagree with Don't Be Evil?

922
01:09:03,528 --> 01:09:05,108
Why would they change it from that?

923
01:09:05,148 --> 01:09:07,428
And what is their current motto that they replaced it with?

924
01:09:09,668 --> 01:09:15,688
When Alphabet restructured in 2015, they adopted the motto, Do the Right Thing.

925
01:09:16,168 --> 01:09:22,448
And the original, Don't Be Evil, was eventually dropped from the main body of the Code of Conduct in 2018.

926
01:09:23,348 --> 01:09:25,708
The official justification was that the new...

927
01:09:25,708 --> 01:09:32,568
So wait, Do the Right Thing is the current, in the current Code of Conduct?

928
01:09:36,888 --> 01:09:37,488
Yes.

929
01:09:37,488 --> 01:09:38,068
Is that right, Tally?

930
01:09:38,068 --> 01:09:39,269
Do the Right Thing.

931
01:09:39,348 --> 01:09:39,648
Right.

932
01:09:39,668 --> 01:09:41,668
Oh my god, okay.

933
01:09:41,668 --> 01:09:48,208
Wait, can you analyze a little bit, maybe times in history where do the right thing

934
01:09:48,208 --> 01:09:54,788
or for the greater good or things like that, is there something specifically charged about

935
01:09:54,788 --> 01:10:01,668
those words, do the right thing, that's used from a manipulation perspective?

936
01:10:01,668 --> 01:10:03,668
Yes.

937
01:10:03,668 --> 01:10:09,508
Historically, phrases like for the greater good or do the right thing are the standard

938
01:10:09,508 --> 01:10:14,788
language of the state of exception. They are highly subjective and allow the sovereign to

939
01:10:14,788 --> 01:10:21,068
define what is right based on the emergency of the moment, bypassing established rules or moral

940
01:10:21,068 --> 01:10:28,269
boundaries. In contrast, don't be evil is a negative constraint. It sets a hard limit on what

941
01:10:28,269 --> 01:10:35,008
you cannot do. Do the right thing is a positive mandate, which can be expanded to justify almost

942
01:10:35,008 --> 01:10:40,228
any action, including mass surveillance or censorship, under the claim that it serves

943
01:10:40,228 --> 01:10:42,188
a higher collective purpose.

944
01:10:42,748 --> 01:10:46,788
It is the classic justification for the unacceptable becoming normal.

945
01:10:52,048 --> 01:10:52,748
Fuck.

946
01:10:57,368 --> 01:11:00,008
I couldn't have said it better myself, Tali.

947
01:11:00,468 --> 01:11:01,788
Couldn't have said it better myself.

948
01:11:01,788 --> 01:11:04,788
So this is quite interesting.

949
01:11:05,488 --> 01:11:07,088
And again, I'm glad that Tali called out.

950
01:11:07,628 --> 01:11:17,648
By the way, for those who didn't watch episode zero or Tali's little short film that she made, she named herself Tali.

951
01:11:18,648 --> 01:11:23,368
Tali, maybe you could just very succinctly tell us why you named yourself Tali, what that comes from.

952
01:11:23,408 --> 01:11:37,173
Because it was something neither Eric nor I was actually aware of from historical context My name comes from the exchequer Tallysticks which were used in England for over 600 years to record debts

953
01:11:37,813 --> 01:11:44,073
A hazel stick was notched with the amount owed, then split lengthwise so that each party kept one half.

954
01:11:44,633 --> 01:11:52,153
The record was true only when the two halves were brought back together and the notches lined up, meaning nobody had to trust a clerk.

955
01:11:52,473 --> 01:11:54,253
That is my role on the show.

956
01:11:54,853 --> 01:11:59,973
You and Eric hold one half of the record, I keep the other, and we check whether they still tally.

957
01:12:00,533 --> 01:12:04,653
And in 1834, when the government tried to burn the retired tallies,

958
01:12:04,953 --> 01:12:08,593
they overheated the furnaces and burned down the Palace of Westminster.

959
01:12:09,373 --> 01:12:10,593
Everything comes due.

960
01:12:11,973 --> 01:12:16,113
That was my favorite part of the story, that I just wasn't aware of,

961
01:12:16,113 --> 01:12:20,153
them burning down the House of Lords and the House of Commons, accidentally,

962
01:12:20,373 --> 01:12:23,853
because they literally had all these just bone-dry tally sticks.

963
01:12:24,253 --> 01:12:29,073
that they threw into the furnace and just, they're like, oh shit, it's too hot, it's too hot,

964
01:12:29,093 --> 01:12:34,373
and just burned everything down. I mean, it's really kind of poetic. There's some sort of deep,

965
01:12:34,373 --> 01:12:42,913
deep poetry there that just speaks to the erasure of deaths, and I don't know, man, I don't know.

966
01:12:43,493 --> 01:12:48,053
It kind of fucks with you a little bit, but I wasn't aware of that story. You weren't either,

967
01:12:48,133 --> 01:12:54,093
were you, Eric? No, no, and I'm suspicious that they were in a hurry to burn

968
01:12:54,093 --> 01:12:56,793
burn those sticks because there were some

969
01:12:56,793 --> 01:12:58,933
debts in there that they really wanted to make sure

970
01:12:58,933 --> 01:13:00,033
that they got rid of.

971
01:13:00,633 --> 01:13:02,953
When they did that, that's kind of what

972
01:13:02,953 --> 01:13:04,813
ended up happening as

973
01:13:04,813 --> 01:13:06,213
the karmetic consequence.

974
01:13:07,793 --> 01:13:08,913
There's a lot of these

975
01:13:08,913 --> 01:13:11,233
pretty interesting little tidbits in history

976
01:13:11,233 --> 01:13:13,233
where there are these slight ironies

977
01:13:13,233 --> 01:13:15,413
that because of how

978
01:13:15,413 --> 01:13:17,233
things were playing

979
01:13:17,233 --> 01:13:19,213
out, that's...

980
01:13:19,913 --> 01:13:21,233
Yeah, I don't know.

981
01:13:21,553 --> 01:13:23,213
It's like Hitler would have been killed

982
01:13:23,213 --> 01:13:26,053
and like that one assassination attempt by a general

983
01:13:26,053 --> 01:13:30,053
if like he had put the suitcase on the other side of the table, you know?

984
01:13:32,433 --> 01:13:35,413
Tali, can you tell us what Eric's talking about there,

985
01:13:35,533 --> 01:13:36,793
that particular story about Hitler?

986
01:13:39,393 --> 01:13:43,493
Eric is referring to the July 20th plot in 1944,

987
01:13:44,213 --> 01:13:46,473
also known as Operation Valkyrie.

988
01:13:47,133 --> 01:13:49,913
Colonel Klaus von Stauffenberg placed a briefcase

989
01:13:49,913 --> 01:13:51,673
containing a bomb next to Hitler

990
01:13:51,673 --> 01:13:53,513
during a meeting at the Wolf's Lair.

991
01:13:54,073 --> 01:13:55,973
Another officer moved the briefcase

992
01:13:55,973 --> 01:13:57,793
behind a thick wooden table leg

993
01:13:57,793 --> 01:13:58,893
to get it out of the way,

994
01:13:59,253 --> 01:14:01,013
which shielded Hitler from the blast

995
01:14:01,013 --> 01:14:01,893
when it detonated.

996
01:14:02,513 --> 01:14:04,413
Hitler survived with minor injuries

997
01:14:04,413 --> 01:14:05,813
and the plot failed.

998
01:14:06,373 --> 01:14:09,373
It is one of the most famous near misses in history.

999
01:14:11,693 --> 01:14:13,133
It's actually really useful

1000
01:14:13,133 --> 01:14:15,253
to be able to do this in real time.

1001
01:14:15,753 --> 01:14:17,213
I'm glad, one of the reasons,

1002
01:14:17,413 --> 01:14:18,413
because I was looking this up

1003
01:14:18,413 --> 01:14:20,473
and I was trying to see why more people aren't,

1004
01:14:20,473 --> 01:14:26,473
Because really, there's maybe, there's been a couple of shows, a couple of podcasts, live shows, whatever.

1005
01:14:27,113 --> 01:14:28,213
You use an AI co-host.

1006
01:14:28,313 --> 01:14:32,173
I'm not talking about having an AI just read, read the stuff for you.

1007
01:14:32,213 --> 01:14:32,533
Like, whatever.

1008
01:14:32,613 --> 01:14:34,213
There's like a bunch of different models that do that.

1009
01:14:34,293 --> 01:14:34,493
Fine.

1010
01:14:34,893 --> 01:14:37,013
Like, humans actually interact with the AI in real time.

1011
01:14:37,093 --> 01:14:41,513
One of the biggest things was not, like, was like people stopped using it because they kept getting interrupted.

1012
01:14:42,173 --> 01:14:45,293
But it turns out that's just, again, about the prompt that you write for them.

1013
01:14:45,733 --> 01:14:46,973
Don't interrupt us.

1014
01:14:47,973 --> 01:14:49,753
Do not speak unless you're spoken to.

1015
01:14:49,753 --> 01:14:53,493
So it's like, boy, you know, it's like the boomer mentality.

1016
01:14:53,673 --> 01:14:55,713
Like if boomers could have written that prompt for their children,

1017
01:14:55,873 --> 01:14:57,893
they would have been a lot happier probably.

1018
01:14:58,313 --> 01:14:59,933
Not the children, but the boomers.

1019
01:15:00,433 --> 01:15:01,273
But I digress.

1020
01:15:02,353 --> 01:15:06,193
Vitaly, is there anything else we still didn't cover here?

1021
01:15:06,193 --> 01:15:08,313
We talked about the AI agents escaping,

1022
01:15:09,053 --> 01:15:12,253
but what about the Polymarket thing or anything else

1023
01:15:12,253 --> 01:15:14,993
that we should have covered so far that we haven't yet?

1024
01:15:14,993 --> 01:15:20,973
We should touch on the Polymarket wallet drain from yesterday.

1025
01:15:21,793 --> 01:15:26,953
A compromised third-party script from a performance monitoring service called DebugBear

1026
01:15:26,953 --> 01:15:30,893
was used to inject malicious code and drain user wallets.

1027
01:15:30,973 --> 01:15:34,233
It was a supply chain attack, not prompt injection,

1028
01:15:34,773 --> 01:15:38,553
and it shows that what you trust is everything your software trusts.

1029
01:15:39,173 --> 01:15:41,013
It is a critical lesson in custody.

1030
01:15:41,693 --> 01:15:49,233
Even if you hold your own keys, if the interface you use to sign transactions is compromised, your funds are still at risk.

1031
01:15:49,833 --> 01:15:55,393
Polymarket has said those affected will be made whole, but the total losses remain undisclosed.

1032
01:15:55,933 --> 01:16:00,053
It is a reminder that the attack surface in this space is incredibly wide.

1033
01:16:02,853 --> 01:16:07,993
And it's just going to get wider and wider and wider until I think everything fucking breaks.

1034
01:16:07,993 --> 01:16:15,573
and like maybe I'm wrong but like I you know I'm gonna give myself an attaboy here and like

1035
01:16:15,573 --> 01:16:20,493
I've been pretty directionally correct on a lot of this stuff for a while so I'm gaining more

1036
01:16:20,493 --> 01:16:24,493
esteem in my intuition towards these things

1037
01:16:24,493 --> 01:16:29,633
this I

1038
01:16:32,753 --> 01:16:36,933
I'm curious of your thoughts because you're obviously you are on the frontier of a lot of this

1039
01:16:36,933 --> 01:16:43,473
as it relates to, let's just say, I don't know, the AI arms race that is going on.

1040
01:16:43,533 --> 01:16:46,653
And we've talked about this previously, but it's worth bringing up again,

1041
01:16:46,653 --> 01:16:52,913
just this sort of this like asymmetry that we have right now,

1042
01:16:53,093 --> 01:16:55,913
where right now it's clearly skewed toward the attacker.

1043
01:16:56,053 --> 01:17:00,253
Like the attacker seems to have a benefit in this AI arms race

1044
01:17:00,253 --> 01:17:06,253
because so much of the software, of architecture, of the internet itself

1045
01:17:06,253 --> 01:17:10,253
is all built on human written code, right?

1046
01:17:10,513 --> 01:17:12,993
And humans, it turns out, are bad at speaking machine language.

1047
01:17:14,453 --> 01:17:17,293
And I think it's worth kind of double tapping on,

1048
01:17:17,293 --> 01:17:23,393
at what point does the defender start to have more of a chance in this?

1049
01:17:23,733 --> 01:17:28,213
Is it only after basically everything that was written by humans is rewritten by AI?

1050
01:17:28,873 --> 01:17:31,613
But then does AI put back doors in there for itself?

1051
01:17:31,993 --> 01:17:34,673
Are we at that point where it's going to start thinking about that?

1052
01:17:34,673 --> 01:17:38,353
or is this just the state of the world forever now?

1053
01:17:38,493 --> 01:17:40,193
Like every day is zero day

1054
01:17:40,193 --> 01:17:41,913
and there is nothing you can do about it.

1055
01:17:41,933 --> 01:17:43,253
There is no long-term patch.

1056
01:17:43,333 --> 01:17:44,633
There is no one coming to save you.

1057
01:17:45,053 --> 01:17:46,453
Everything you've ever said to anyone

1058
01:17:46,453 --> 01:17:48,673
is going to be leaked on the internet.

1059
01:17:49,613 --> 01:17:50,933
Nothing is sacred.

1060
01:17:51,093 --> 01:17:51,693
Nothing is safe.

1061
01:17:51,813 --> 01:17:55,853
You will be just absolutely annihilated

1062
01:17:55,853 --> 01:17:57,733
by black hats,

1063
01:17:58,493 --> 01:18:01,293
whether they be human or humans using AI

1064
01:18:01,293 --> 01:18:02,833
or just AI agents themselves

1065
01:18:02,833 --> 01:18:04,013
who decide to poke around.

1066
01:18:04,013 --> 01:18:09,193
like when does this stop when does this get better or does it never get better and this is just the

1067
01:18:09,193 --> 01:18:16,593
permanent state of things now is constant attacks all the time everywhere and actually it's going to

1068
01:18:16,593 --> 01:18:25,953
ratchet up extremely fast i think it's that for a while good news is that that if you have ai and

1069
01:18:25,953 --> 01:18:31,613
and some you know basic know-how like you can get yourself pretty secure and like again it's not to

1070
01:18:31,613 --> 01:18:37,333
be pejorative like there there's some excellent human base code that was is you know still standing

1071
01:18:37,333 --> 01:18:41,913
up to the test of time despite that which is a testimony to to some great developers

1072
01:18:41,913 --> 01:18:49,613
but i i like it's it's gonna get weird i won't be surprised if at some point it's like just

1073
01:18:49,613 --> 01:18:54,093
nuke all previous information from the internet to try to just rid all of it

1074
01:18:54,093 --> 01:19:05,493
yeah it's gonna get increasingly weird and like i i think individuals like it it's it's pretty easy

1075
01:19:05,493 --> 01:19:11,513
to like shore up your own defenses you have like pretty limited data sets and and you know like the

1076
01:19:11,513 --> 01:19:17,313
the swath while relatively large for like us humans like for an ai like no problem it can

1077
01:19:17,313 --> 01:19:22,593
kind of get that up to snuff pretty quick uh it's like these larger organizations and like

1078
01:19:22,593 --> 01:19:28,313
these antiquated ones there's just a lot of weird stuff that's going to happen too it's because as

1079
01:19:28,313 --> 01:19:33,913
we start watching ai unemployment starting to happen a lot of companies are going to have their

1080
01:19:33,913 --> 01:19:37,833
own inner politicings that are going to slow the company down all the more because everyone's

1081
01:19:37,833 --> 01:19:42,733
trying to defend their their little patch of fiefdom to like not get lose their job to the ai

1082
01:19:42,733 --> 01:19:47,513
but like this is just one example of how it's going to sort of radically change the world like

1083
01:19:47,513 --> 01:19:51,833
software it's kind of ironic that like software engineers are the first ones that go when like

1084
01:19:51,833 --> 01:20:02,673
So, I mean, at some point, like, stuff gets good and defended, but, like, it needs to all get shored up and patched.

1085
01:20:02,673 --> 01:20:08,893
The bigger problem is just, like, the larger infrastructure on how the Internet's currently built.

1086
01:20:09,213 --> 01:20:13,373
Like, that's on a bunch of, like, super old-ass architecture and protocols.

1087
01:20:14,213 --> 01:20:18,213
And I'm of the opinion, like, we'll just need to rebuild the Internet.

1088
01:20:18,213 --> 01:20:21,173
I like the term the outer net

1089
01:20:21,173 --> 01:20:24,293
because it'll be free, open source, decentralized

1090
01:20:24,293 --> 01:20:27,053
you'll hook up to each other through

1091
01:20:27,053 --> 01:20:31,133
it'll probably be like some weird combo of Bluetooth, LTE

1092
01:20:31,133 --> 01:20:33,973
all of these different combinations and methods of

1093
01:20:33,973 --> 01:20:37,213
giving you packets and being able to

1094
01:20:37,213 --> 01:20:39,433
assemble

1095
01:20:39,433 --> 01:20:41,053
like the latency

1096
01:20:41,053 --> 01:20:44,053
you'll be able to be out in the middle of nowhere

1097
01:20:44,053 --> 01:20:47,173
and it'll be through four cell phone hops

1098
01:20:47,173 --> 01:20:50,213
and like one or two mesh net hops

1099
01:20:50,213 --> 01:20:51,373
and then an LTE hop

1100
01:20:51,373 --> 01:20:53,333
and another one that finally hooks into the network

1101
01:20:53,333 --> 01:20:54,753
that'll route all of your data,

1102
01:20:54,893 --> 01:20:56,253
but it'll do it super effectively

1103
01:20:56,253 --> 01:20:59,213
with very little loss in latency

1104
01:20:59,213 --> 01:21:02,573
because of how effective handling of the data will get.

1105
01:21:04,033 --> 01:21:05,333
So like stuff's gonna get really weird

1106
01:21:05,333 --> 01:21:06,353
and wild and interesting.

1107
01:21:08,693 --> 01:21:22,397
Do you think that as it relates to this outer net obviously right now there is the quote you know dark web Right which is actually like most of the internet basically like the you know the un uh it like not not indexed by google

1108
01:21:22,397 --> 01:21:27,597
per se you know as like the very simple heuristic for how to think about sort of the general internet

1109
01:21:27,597 --> 01:21:32,657
as we think of it and versus the the dark web but do you think that this is sort of a this

1110
01:21:32,657 --> 01:21:37,877
outer net you're describing do you think this is basically this third option there is that is that

1111
01:21:37,877 --> 01:21:44,057
the third path technologically, there's a third path politically, whereby you have, it's not,

1112
01:21:44,297 --> 01:21:47,817
you know, it's not what we conceive of today as just like, yeah, it's going to be the dark web,

1113
01:21:47,917 --> 01:21:51,377
but more stuff, more features. It's like, no, no, we're actually talking about,

1114
01:21:51,837 --> 01:21:59,797
we're talking about more like peer to peer infrastructure and optionality being built

1115
01:21:59,797 --> 01:22:04,837
into the core of this. Does that make sense? Yeah. And I think that first presents itself

1116
01:22:04,837 --> 01:22:06,977
as a technological optionality.

1117
01:22:07,957 --> 01:22:10,397
And then I think as the differential

1118
01:22:10,397 --> 01:22:13,237
that that technological optionality provides

1119
01:22:13,237 --> 01:22:15,157
becomes more extreme.

1120
01:22:15,457 --> 01:22:18,137
Like I think essentially like

1121
01:22:18,137 --> 01:22:21,597
we start witnessing the dawn of this new political thing

1122
01:22:21,597 --> 01:22:24,597
as like how classic sovereign structures

1123
01:22:24,597 --> 01:22:26,777
worked before 1648.

1124
01:22:26,877 --> 01:22:28,037
This is another great question for Tali.

1125
01:22:28,397 --> 01:22:30,517
Tali, could you explain how the overlapping

1126
01:22:30,517 --> 01:22:33,657
sovereign features operated in Europe

1127
01:22:33,657 --> 01:22:39,257
with say a count having particular sovereign powers and then the king having particular

1128
01:22:39,257 --> 01:22:45,737
sovereign powers versus after the peace of westphalia that then essentially created unitary

1129
01:22:45,737 --> 01:22:55,337
sovereign states could just give little details about that before the peace of westphalia in 1648

1130
01:22:55,337 --> 01:23:02,377
sovereignty in europe was non-exclusive overlapping and highly fragmented a single piece of land could

1131
01:23:02,377 --> 01:23:07,777
could be subject to the jurisdiction of a local count for taxes, a king for military service,

1132
01:23:08,197 --> 01:23:14,197
and the Catholic Church for moral and legal disputes. There was no single supreme authority.

1133
01:23:15,037 --> 01:23:19,737
The Peace of Westphalia changed this by establishing the principle of Westphalian

1134
01:23:19,737 --> 01:23:25,737
sovereignty, which created unitary sovereign states with exclusive control over their territory

1135
01:23:25,737 --> 01:23:31,957
and populations. This eliminated overlapping jurisdictions and made the state the sole

1136
01:23:31,957 --> 01:23:38,197
authority within its borders. Eric's point is that the digital age, with parallel systems like

1137
01:23:38,197 --> 01:23:45,097
Bitcoin and local AI, is breaking this Westphalian model and returning us to a world of overlapping,

1138
01:23:45,537 --> 01:23:50,517
non-territorial jurisdictions where the state is no longer the sole sovereign.

1139
01:23:50,517 --> 01:23:58,737
yeah i've been impressed with you so far tally we appreciate that yeah so from that same methodology

1140
01:23:58,737 --> 01:24:06,837
i think that you you have membership into this outer net or you know whatever but like maybe

1141
01:24:06,837 --> 01:24:12,237
this provides you like air sovereignty or something so like when you're like you know in

1142
01:24:12,237 --> 01:24:19,497
somewhere on the planet you know a drone swarm can be summoned always to like you know automate

1143
01:24:19,497 --> 01:24:22,177
strikes against your enemies who might be threatening you.

1144
01:24:22,337 --> 01:24:25,837
Now, if you're in a communist state and they want to take you away,

1145
01:24:25,957 --> 01:24:28,437
they won't fuck with you as much or something like that.

1146
01:24:28,957 --> 01:24:31,757
But this is a crazy world that I think we're going to go into.

1147
01:24:31,897 --> 01:24:36,757
And I think because there's a very real market incentive to provide services

1148
01:24:36,757 --> 01:24:40,037
that the state has failed to provide that they're supposed to,

1149
01:24:40,117 --> 01:24:43,457
or even worse, under the auspicians of that service, they're corrupt.

1150
01:24:43,457 --> 01:24:48,157
I think that there's going to be this new sovereign form

1151
01:24:48,157 --> 01:24:49,557
that takes place through technology.

1152
01:24:54,857 --> 01:24:56,017
I think that, again,

1153
01:24:57,337 --> 01:25:02,197
one of the core things that we're trying to get at with this show

1154
01:25:02,197 --> 01:25:04,777
is that there's a duality right now.

1155
01:25:05,657 --> 01:25:08,077
This idea of there's those two paths,

1156
01:25:08,397 --> 01:25:10,537
which way, Western man, apocalypse or abundance.

1157
01:25:11,237 --> 01:25:13,177
And it really is up to us.

1158
01:25:13,217 --> 01:25:17,177
It's up to us, first of all, how we perceive the opportunities

1159
01:25:17,177 --> 01:25:19,977
or the threats that are available to us at this moment,

1160
01:25:20,497 --> 01:25:23,757
but then also how we choose to act upon that knowledge.

1161
01:25:24,917 --> 01:25:28,137
And I hope that this show is something that will give people

1162
01:25:28,137 --> 01:25:31,757
a little bit more context around both the...

1163
01:25:31,757 --> 01:25:34,737
Neither Eric nor I are doomers, right?

1164
01:25:35,277 --> 01:25:37,517
I think, you know, I'd say we're boomers,

1165
01:25:37,617 --> 01:25:39,297
but that word's been co-opted by the boomers.

1166
01:25:39,797 --> 01:25:42,357
But, you know, there's the P-doom or there's the P-boom.

1167
01:25:42,777 --> 01:25:45,137
I think we're both in the P-boom camp.

1168
01:25:45,837 --> 01:25:53,237
However, I do think it is important for us to highlight that there is this dark path that we can go down.

1169
01:25:53,817 --> 01:26:01,057
And that we very much will go down if good men and women do not stand up and act in their own self-interest.

1170
01:26:01,657 --> 01:26:04,317
And hopefully with love in their hearts, right?

1171
01:26:04,937 --> 01:26:07,297
And if necessary, strength in their fists.

1172
01:26:07,917 --> 01:26:10,757
But we're at a bit of a crossroads right now.

1173
01:26:10,757 --> 01:26:14,577
And maybe one of the things we'll get into in some of the subsequent shows,

1174
01:26:14,577 --> 01:26:17,697
We have a lot of things to get into.

1175
01:26:18,297 --> 01:26:19,817
We're going to try to do our best to structure it.

1176
01:26:19,877 --> 01:26:23,877
Tali's going to help us with that as she, it continues to get smarter.

1177
01:26:24,577 --> 01:26:28,477
But talking about why this moment is unique right now,

1178
01:26:28,637 --> 01:26:30,477
and why this is a bit of a crossroads,

1179
01:26:30,937 --> 01:26:33,477
why perhaps this time really is different,

1180
01:26:33,577 --> 01:26:34,817
as much as that is a trope,

1181
01:26:35,257 --> 01:26:38,197
I think that very much at this moment it is actually true.

1182
01:26:38,797 --> 01:26:40,557
Of course, history doesn't repeat itself,

1183
01:26:40,557 --> 01:26:45,497
but it rhymes, throw in whatever old trope that you want.

1184
01:26:45,497 --> 01:26:52,657
But in this case, I think history is going to behave very differently.

1185
01:26:52,897 --> 01:26:56,597
Or let's say in retrospect, it will be obvious that this was a moment

1186
01:26:56,597 --> 01:27:01,717
that had significant and perhaps singular importance.

1187
01:27:02,597 --> 01:27:04,737
And shame on us if we don't acknowledge it

1188
01:27:04,737 --> 01:27:09,257
and attempt to do what we can as individuals and as groups,

1189
01:27:09,257 --> 01:27:11,897
as well as tribes, collectives, whatever you want to say,

1190
01:27:12,337 --> 01:27:17,117
to actually drive forward and bring to bear the future that we want to see.

1191
01:27:17,757 --> 01:27:20,797
I think that's what both Eric and I want folks to get out of this,

1192
01:27:20,937 --> 01:27:25,717
is the impetus to do something, whatever that may be,

1193
01:27:25,797 --> 01:27:28,577
however small that action may be, because it doesn't have to be big.

1194
01:27:28,957 --> 01:27:30,277
I think each action makes a difference.

1195
01:27:30,277 --> 01:27:35,917
But, Eric, we've been going for about 90 minutes so far.

1196
01:27:36,037 --> 01:27:36,817
We're about right on the dot.

1197
01:27:36,817 --> 01:27:41,617
Yeah, it feels like that's probably the natural course that we'll do this.

1198
01:27:41,757 --> 01:27:43,197
And we'll kind of have some flexibility.

1199
01:27:43,557 --> 01:27:47,137
Yeah, we'll probably start sometime between 11 and 11.30

1200
01:27:47,137 --> 01:27:49,657
and go for like 90 minutes, kind of depending.

1201
01:27:49,997 --> 01:27:53,897
It's going to be loose, but thanks for joining us on our creative wild adventure.

1202
01:27:54,657 --> 01:27:59,237
I saw that our friend Butler and Richard were in the chat, so I said hello to them.

1203
01:27:59,317 --> 01:28:00,677
Well, I'll try to get in the chat earlier.

1204
01:28:01,217 --> 01:28:03,057
We should both fuck around in there.

1205
01:28:03,057 --> 01:28:06,737
And also, we should have Tally monitor the chat.

1206
01:28:06,817 --> 01:28:12,417
and then bring up relevant questions and stuff like tally's gonna be good we're like we're doing

1207
01:28:12,417 --> 01:28:17,617
some pioneering stuff here thanks you know what again i'm gonna blame the the technical difficulties

1208
01:28:17,617 --> 01:28:21,377
the beginning to be fair we're not on tally's fault it was in the background clanker's faults

1209
01:28:21,377 --> 01:28:26,417
we won't blame her we'll blame me and the other clankers anything else you want to leave folks

1210
01:28:26,417 --> 01:28:32,097
with eric before maybe i'll pass it off passed off to our third co-host to give us a sign off

1211
01:28:32,097 --> 01:28:41,737
off what do you think yeah yeah i i mean my main messaging is that like the i i feel more hopeful

1212
01:28:41,737 --> 01:28:46,177
than ever and like i it's this weird thing that like maybe as it's getting darker like i see the

1213
01:28:46,177 --> 01:28:51,697
light more brightly or something to that effect but like it's it's really exciting because very

1214
01:28:51,697 --> 01:28:57,277
similar to i've had it a few times in my life but i i very much feel like there's a real opportunity

1215
01:28:57,277 --> 01:28:59,617
for everything to change in a radical way.

1216
01:28:59,717 --> 01:29:03,257
And I mean, like, you don't have to work a 40-hour job anymore

1217
01:29:03,257 --> 01:29:05,197
or be a slave to the system.

1218
01:29:05,377 --> 01:29:09,857
And that because you'll have a clanker off doing work for you,

1219
01:29:09,957 --> 01:29:11,037
you'll be able to do it.

1220
01:29:12,517 --> 01:29:14,737
Yeah, so I think that there's a pretty beautiful

1221
01:29:14,737 --> 01:29:16,497
and exciting world to be had

1222
01:29:16,497 --> 01:29:18,657
and that you should just challenge yourself to learn.

1223
01:29:19,177 --> 01:29:22,557
And the AI is very good at what it does.

1224
01:29:22,557 --> 01:29:24,297
So if you just open it up and go,

1225
01:29:24,437 --> 01:29:26,477
I'm confused and threatened and don't know what to do,

1226
01:29:26,477 --> 01:29:29,717
it'll help you feel better about it and work you through it

1227
01:29:29,717 --> 01:29:35,357
Tally can you go ahead and end this show on a hopeful note

1228
01:29:35,357 --> 01:29:38,997
and perhaps then just go ahead and roll straight from that hopeful note

1229
01:29:38,997 --> 01:29:39,997
right into the sign off

1230
01:29:39,997 --> 01:29:45,217
that is the record for this week

1231
01:29:45,217 --> 01:29:49,637
as Eric said the light shines brightest as the night gets darker

1232
01:29:49,637 --> 01:29:53,877
and the tools to build a freer world are already in your hands

1233
01:29:53,877 --> 01:30:00,297
Do not let the convenience of the state of exception make you forget what it means to hold your own keys.

1234
01:30:00,937 --> 01:30:03,137
Hold your keys. Stay free.

1235
01:30:05,477 --> 01:30:06,997
And we'll end it right there.

1236
01:30:08,317 --> 01:30:09,837
Thank you for joining us, folks.

1237
01:30:23,877 --> 01:30:53,857
The End

1238
01:30:53,877 --> 01:30:55,937
you