1
00:00:01,440 --> 00:00:03,757
Folks, my new graphic novel, Unwanted, a

2
00:00:03,840 --> 00:00:08,397
Tall Tale of the Old West and New Wave,
is out for pre-order now. I've been

3
00:00:08,480 --> 00:00:10,317
working on this for 25 years. It's a

4
00:00:10,400 --> 00:00:12,397
dark comedy with a shiny exterior.

5
00:00:12,480 --> 00:00:16,520
Please check it out at unwantedbook.com.

6
00:00:28,160 --> 00:00:31,320
Heat. Heat.

7
00:00:44,800 --> 00:00:47,757
Good afternoon. Michael Malice here. Let
that be your welcome for the next hour.

8
00:00:47,840 --> 00:00:52,237
Guys, we have a very special returning
guest. Mark Andre, internet OG. You

9
00:00:52,320 --> 00:00:55,757
worked on Mosaic. You worked on Netscape.
You were here since the very

10
00:00:55,840 --> 00:00:59,117
beginning. We just spent 15 minutes
trying to get this connection working.

11
00:00:59,200 --> 00:01:02,557
And the answer was rebooting your
computer, which takes me back to my tech

12
00:01:02,640 --> 00:01:06,637
support days. >> No, we we
have this running joke. We

13
00:01:06,720 --> 00:01:09,197
have this running joke. We have all
these AI super geniuses come in the

14
00:01:09,280 --> 00:01:11,197
office and they've all got, you know,
they've got everything all figured out

15
00:01:11,280 --> 00:01:13,997
and they literally spend 20 minutes they
can't get their laptop to connect to the

16
00:01:14,080 --> 00:01:16,797
projector. It's it

17
00:01:16,880 --> 00:01:21,597
and and now and now I'm one of them.
So, we're going to talk a lot about AI

18
00:01:21,680 --> 00:01:24,637
because I have a lot of opinions
and not a lot of information, which is a

19
00:01:24,720 --> 00:01:27,357
dangerous place to be. And that's why
I want to talk to you, Mark. You have a

20
00:01:27,440 --> 00:01:31,117
book out with Passage Press called
The Techno Optimist Manifesto. I have my

21
00:01:31,200 --> 00:01:34,317
copy. Uh you can get it. It's It's
really cool cuz it's got this metal

22
00:01:34,400 --> 00:01:39,037
cover. Um, and I like you share
an enormous sense of optimism about

23
00:01:39,120 --> 00:01:43,757
technology. Although I'm sure you agree
with me uh and with that you love Thomas

24
00:01:43,840 --> 00:01:47,757
Soul. And I think Thomas Soul's greatest
quote is there's no solutions. There's

25
00:01:47,840 --> 00:01:50,957
only trade-offs. >> And people
think that if something has a

26
00:01:51,040 --> 00:01:55,357
problem with it, therefore it's a no-go
as opposed to the reality which is

27
00:01:55,440 --> 00:01:59,037
everything has a cost. Everything has
a downside. What you want are the upsides

28
00:01:59,120 --> 00:02:01,997
outweigh the downsides.
But before we get started, yeah,

29
00:02:02,080 --> 00:02:06,797
>> Mr. billionaire. Um, I was reading
on your Wikipedia that you're a big fan of

30
00:02:06,880 --> 00:02:09,997
Marinetti. >> Oh, uh, yeah.
Uhhuh. Yes, I am. I mean,

31
00:02:10,080 --> 00:02:12,637
you know, with with c, you know,
with appropriate caveats.

32
00:02:12,720 --> 00:02:16,397
>> Well, yeah, of course.
Well, check this out. Signed copy.

33
00:02:16,480 --> 00:02:18,477
>> Amazing. >> How do you like that?

34
00:02:18,560 --> 00:02:21,357
>> Amazing. >> And I have
a um a framed manifesto in my

35
00:02:21,440 --> 00:02:26,077
living room upstairs. Here's So, let's
get before we get to talking about uh

36
00:02:26,160 --> 00:02:29,117
the future of AI and you being such
an internet visionary. I'm really excited

37
00:02:29,200 --> 00:02:34,077
to hear your point of view. Can you tell
people what AI is because I feel like

38
00:02:34,160 --> 00:02:38,317
there's so much talking past each other
on this issue about with what it really

39
00:02:38,400 --> 00:02:41,837
is and what people think it
is and and I'll give you the floor.

40
00:02:41,920 --> 00:02:45,517
>> Yeah. Well, so look, I would start
by saying, look, humanity's always been,

41
00:02:45,600 --> 00:02:48,877
you know, justifiably obsess obsessed
with with ourselves, right? And so and

42
00:02:48,960 --> 00:02:51,597
and then as a consequence, we're
obsessed with with things that seem like

43
00:02:51,680 --> 00:02:55,437
they might be like us. Um, you know,
there there's this, you know, concept in

44
00:02:55,520 --> 00:02:58,637
psychology called anthropomorphizing,
right? where you you basically look at

45
00:02:58,720 --> 00:03:01,597
something that's not human but you kind
of want to read you know humanity into it

46
00:03:01,680 --> 00:03:04,477
>> and you know we I mean look we do that
with like you know we do that with cats

47
00:03:04,560 --> 00:03:08,957
and dogs um uh we do that we do that
with Bambi I you know as um there's a

48
00:03:09,040 --> 00:03:14,077
famous Disney marketing tagline from the
movie Pinocchio 1940 which is uh America

49
00:03:14,160 --> 00:03:17,917
will America you know 1940 America will
fall in love with a cardboard cricket

50
00:03:18,000 --> 00:03:21,517
>> you know you know Jiminy Cricket right
um you know Kermit the Frog I mean you

51
00:03:21,600 --> 00:03:25,117
just you know you go by the South Park
kids right you just go like right down

52
00:03:25,200 --> 00:03:28,557
the list you know you'll you'll
basically read humanity into anything.

53
00:03:28,640 --> 00:03:31,837
Um uh and and so there's this very
natural, you know, kind of thing to do

54
00:03:31,920 --> 00:03:35,037
that. And then of course this the
scientists involved in AI, you know,

55
00:03:35,120 --> 00:03:37,517
very deliberately set that up. You know,
they they you know, they literally

56
00:03:37,600 --> 00:03:40,157
created this this this architecture
called a neural network which was

57
00:03:40,240 --> 00:03:42,637
modeled after the human brain. And then
they said if we you know, if we work on

58
00:03:42,720 --> 00:03:45,517
this long enough, we'll eventually be
able to replicate the human brain and

59
00:03:45,600 --> 00:03:48,077
we'll have you know, and literally
artificial intelligence like we'll have,

60
00:03:48,160 --> 00:03:52,637
you know, kind of artificial people. um
you know that you know in in in you know

61
00:03:52,720 --> 00:03:55,677
kind of as as you kind of track the arc
of that idea it you know kind of goes

62
00:03:55,760 --> 00:03:59,517
through Frankenstein's monster right
it kind of then goes into robots um and

63
00:03:59,600 --> 00:04:02,557
then you know of course we've had
fictional portrayals of AI you know you

64
00:04:02,640 --> 00:04:04,957
know coming out of Hollywood and you
know coming out of the science fiction

65
00:04:05,040 --> 00:04:09,597
world for you know over 100 years um
and then you know in the last you know

66
00:04:09,680 --> 00:04:12,317
whatever 40 years or whatever you
know the image that kind of stuck in

67
00:04:12,400 --> 00:04:14,877
everybody's head I think more than
anything else was you know was Skynet

68
00:04:14,960 --> 00:04:18,397
and and you know Arnold
Schwarzenegger you know Terminator 2

69
00:04:18,480 --> 00:04:21,437
>> um and you know when I when watch,
you know, Terminator movies obviously are

70
00:04:21,520 --> 00:04:24,397
brilliant. You know, when I watch them,
it to me it's like very clear what's

71
00:04:24,480 --> 00:04:27,037
happening in them, which is it's it's
basically, you know, it's basically

72
00:04:27,120 --> 00:04:30,317
robot it's robot Nazis, right?
Like fundamentally, right? It's it's like

73
00:04:30,400 --> 00:04:32,957
it's like, you know, it's like robot
World War II. It's like, you know,

74
00:04:33,040 --> 00:04:37,197
unthinking, unfeilling, uncaring,
hierarchical, you know, like overly

75
00:04:37,280 --> 00:04:41,197
logical, um, you know, relentless,
unstoppable, you know, and then and then

76
00:04:41,280 --> 00:04:43,357
obviously obviously homicidal, right?

77
00:04:43,440 --> 00:04:45,757
Like obviously they want to wipe out
humanity and you know they just kind of

78
00:04:45,840 --> 00:04:48,957
set up this this very you know light
versus dark you know human versus

79
00:04:49,040 --> 00:04:52,077
machine you know kind of struggle and of
course you know look I you know you can

80
00:04:52,160 --> 00:04:54,877
imagine you know various worlds in which
something like that you know does get

81
00:04:54,960 --> 00:04:58,557
built and you know say one of the things
mankind is very good at is building

82
00:04:58,640 --> 00:05:03,037
killing machines right like we we we've
been quite quite talented at that for a

83
00:05:03,120 --> 00:05:07,597
very long time so so that's kind of set
a lot of the popular perception um the

84
00:05:07,680 --> 00:05:11,117
the interesting thing about what's h and I
should also say like as you said like

85
00:05:11,200 --> 00:05:14,397
I'm an optimist not a utopian and so
like every technology is a double-edged

86
00:05:14,480 --> 00:05:16,957
sword. Every technology gets
used for good and for bad.

87
00:05:17,040 --> 00:05:20,397
>> Again, there there there's a long
history of that. Um, having said that,

88
00:05:20,480 --> 00:05:23,837
the AI that we actually got is not the AI
that we thought we were going to get.

89
00:05:23,920 --> 00:05:26,957
We we actually we actually got something
very different uh than we thought we

90
00:05:27,040 --> 00:05:30,637
were going to get. and and specifically
the AI that we got is in the form of you

91
00:05:30,720 --> 00:05:34,477
know what we now call these large
language models which by the way large

92
00:05:34,560 --> 00:05:38,877
language models were like a very fringe
idea uh up until like literally like

93
00:05:38,960 --> 00:05:42,077
like the company OpenAI for example
which you know really catalyzed this

94
00:05:42,160 --> 00:05:44,877
whole thing you know kicked it off and
and you know today you know the leading

95
00:05:44,960 --> 00:05:48,797
the leading uh you know company which
had GPT like open AI was not founded you

96
00:05:48,880 --> 00:05:53,197
know 10 years ago um to do large
language models um it was founded to do

97
00:05:53,280 --> 00:05:56,557
a different kind of AI and it turned out
classic kind of story technology there

98
00:05:56,640 --> 00:05:59,837
was literally one guy in the back room,
this guy Alec Redford, and he had this

99
00:05:59,920 --> 00:06:02,157
idea and he's like, "Oh, I don't know.
I think maybe if we like take this

100
00:06:02,240 --> 00:06:05,917
language approach, it might be
interesting." Um, and he created GPT1

101
00:06:06,000 --> 00:06:10,637
and then GPT2 and then GPT3 and then
by the way, Chad GPT was actually kind of

102
00:06:10,720 --> 00:06:13,597
an accidental success. Like they didn't
believe it was going to be like a big

103
00:06:13,680 --> 00:06:16,717
hit that, you know, they was like
a little experiment, you know, kind of off

104
00:06:16,800 --> 00:06:19,837
to the side and and then it it it just
turned out it like, you know, it it

105
00:06:19,920 --> 00:06:23,597
works incredibly well. But it's very
different. And the and the way to think

106
00:06:23,680 --> 00:06:26,237
about it that sort of contra like
the Skynet model or whatever the way to

107
00:06:26,320 --> 00:06:28,637
think about it is what a large language
model is is very different. So what a

108
00:06:28,720 --> 00:06:32,397
large language model is is basically
it's you take the complete kind of

109
00:06:32,480 --> 00:06:35,037
totality of all human culture that you

110
00:06:35,120 --> 00:06:39,437
can possibly get your hands on and you
know and and and what that means is

111
00:06:39,520 --> 00:06:42,877
essentially the internet it turns out
right it turns out the internet is is is

112
00:06:42,960 --> 00:06:46,477
the basis for AI. Um and so you
basically essentially think of it as

113
00:06:46,560 --> 00:06:50,397
like downloading everything off the
internet. Um and then it's actually it's

114
00:06:50,480 --> 00:06:52,477
actually technically it's like form
of compression. And it's it's it's like

115
00:06:52,560 --> 00:06:55,197
building a search engine but of a
different kind which is what they what

116
00:06:55,280 --> 00:06:58,477
they do is they basically this process
called training. But what that means is

117
00:06:58,560 --> 00:07:01,677
they take basically the the the world's
collected knowledge, culture,

118
00:07:01,760 --> 00:07:04,557
entertainment, you know, kind of
everything that they can get their hands

119
00:07:04,640 --> 00:07:06,157
on and then they kind of smush it

120
00:07:06,240 --> 00:07:09,197
together um into uh basically this

121
00:07:09,280 --> 00:07:13,757
highly compressed um uh sort of search
engine essentially uh you know

122
00:07:13,840 --> 00:07:16,877
compressed version of human knowledge
and culture. And and the technical term

123
00:07:16,960 --> 00:07:20,317
for that is latent space. L A T N T lat

124
00:07:20,400 --> 00:07:25,037
space. um they compress basically all
of human culture into latent space and you

125
00:07:25,120 --> 00:07:27,517
think of latent space as like
a thousand-dimensional basically

126
00:07:27,600 --> 00:07:31,677
compressed representation of all human
culture and then when you talk to chat

127
00:07:31,760 --> 00:07:35,677
GPT when you type in whatever your
question is it basically the way I think

128
00:07:35,760 --> 00:07:38,877
about it is it sort of sends a probe
through that latent space through that

129
00:07:38,960 --> 00:07:41,917
like thousand dimensional latent
space and then it basically you know it

130
00:07:42,000 --> 00:07:44,717
constructs an answer but it constructs
an answer based on the compression of

131
00:07:44,800 --> 00:07:46,077
all of all basically known human

132
00:07:46,160 --> 00:07:49,437
information and comes back at you. Um,

133
00:07:49,520 --> 00:07:54,237
so, so it's, so it's like talking
to a mirror of humanity, right? Like it's

134
00:07:54,320 --> 00:07:57,597
like talking to a representation of
everything that people have ever thought

135
00:07:57,680 --> 00:08:01,917
and said. By the way, for every question
that you ask, there are many possible

136
00:08:02,000 --> 00:08:05,597
answers in the latent space and it just
happens to pick one, but like there are

137
00:08:05,680 --> 00:08:08,557
many others. And actually, if you ask
Chad GPT the same question twice, it

138
00:08:08,640 --> 00:08:11,357
will give you two different answers,
right? Because it's sort of firing these

139
00:08:11,440 --> 00:08:14,477
probes in a somewhat sort of semi-random
way to try to get, you know, basically

140
00:08:14,560 --> 00:08:18,477
variation and creativity out of it.
Um, but you're basically talking, you're

141
00:08:18,560 --> 00:08:22,157
basically getting echoes back from
collective humanity. Um, and and that's

142
00:08:22,240 --> 00:08:25,357
just like a much much much different
thing than we thought we were going to

143
00:08:25,440 --> 00:08:28,477
get. For example, one of the things
you can do is you can engage in moral

144
00:08:28,560 --> 00:08:32,317
debates with it, right? You can like you
can have like very sophisticated debates

145
00:08:32,400 --> 00:08:35,037
about like moral psychology, about
moral philosophy, all the different

146
00:08:35,120 --> 00:08:38,397
approaches, virtue ethics,
utilitarianism, you know, religion,

147
00:08:38,480 --> 00:08:42,637
politics, like it will happily sit
and I think have like very sophisticated

148
00:08:42,720 --> 00:08:46,717
discussions about all this stuff. Um,
and you know, let's just say that was

149
00:08:46,800 --> 00:08:49,357
never in the James Cameron movies.

150
00:08:49,440 --> 00:08:53,917
>> Yeah. There's a lot there. So, do you
want me to go with my hopes or with my

151
00:08:54,000 --> 00:08:57,117
fears about the future of AI?
>> Let's start with hopes cuz you know the

152
00:08:57,200 --> 00:08:58,157
part of hopefully what we'll talk about

153
00:08:58,240 --> 00:09:01,757
today is like that say humanity always

154
00:09:01,840 --> 00:09:05,837
basically there's this like the negative
view always seems like it's going to be

155
00:09:05,920 --> 00:09:08,157
the sophisticated view or the
sophisticated view always feels like

156
00:09:08,240 --> 00:09:11,597
it's going to be negative. Exactly.
And I know you don't like that. Um, and so

157
00:09:11,680 --> 00:09:13,917
um and that's a very it's a very
naturally human thing. And then and then

158
00:09:14,000 --> 00:09:16,717
I would I would also say Michael I think
we live in a particularly pessimistic

159
00:09:16,800 --> 00:09:20,237
time um in which there's like a very
large number of moral entrepreneurs who

160
00:09:20,320 --> 00:09:22,877
basically want to convince us that like
everything is bad right and you know

161
00:09:22,960 --> 00:09:26,397
we've been through you know a decade of
like craziness on that front um and so I

162
00:09:26,480 --> 00:09:30,637
I think there's a negativity bias that's
that's that's infected our um you know

163
00:09:30,720 --> 00:09:32,957
our discourse on all these topics
and so may maybe we could start with the

164
00:09:33,040 --> 00:09:35,597
positive and then go to the negative.
I I I I would just tweak that a little

165
00:09:35,680 --> 00:09:38,557
bit, not even negativity bias,
specifically a cynicism bias. And

166
00:09:38,640 --> 00:09:41,677
there's this idea that if you're
a sophisticated, intelligent person, you

167
00:09:41,760 --> 00:09:45,597
roll your eyes and sneer at the idea
of hope, progress, and optimism. And it's

168
00:09:45,680 --> 00:09:49,437
just like you, okay? Like that's
my answer to that because if you want to

169
00:09:49,520 --> 00:09:52,637
live in that space, which is not
rational, which doesn't if that were

170
00:09:52,720 --> 00:09:56,477
reality, we'd all be dead because it's
very easy to kill kill someone's it's

171
00:09:56,560 --> 00:10:00,157
much harder to keep them alive. So,
if you had this uh if things were shifted

172
00:10:00,240 --> 00:10:03,437
toward this idea of everything's bad,
everything sucks, everything's out to

173
00:10:03,520 --> 00:10:06,077
get us, we'd be gotten. So, I I have no

174
00:10:06,160 --> 00:10:08,637
time for that perspective. Here's my uh

175
00:10:08,720 --> 00:10:11,837
uh vision of of hope. So, my second

176
00:10:11,920 --> 00:10:16,317
favorite speaker, Franitz, had this bit
about the Me Too movement. Now, I'm I'm

177
00:10:16,400 --> 00:10:18,557
sure everyone listening to this agrees
that the Me Too movement got out of

178
00:10:18,640 --> 00:10:22,477
hand. But regards to people like Bill
Cosby, Harvey Weinstein, her point was

179
00:10:22,560 --> 00:10:25,197
from the time of Eve until 5 minutes

180
00:10:25,280 --> 00:10:29,837
ago, these powerful men could just be
predators with no repercussions out in

181
00:10:29,920 --> 00:10:33,677
the open. Meil Stre standing up and
applauding so on and so forth. And then

182
00:10:33,760 --> 00:10:37,277
when it happened, she was like,
"Holy crap." Like, this has never happened

183
00:10:37,360 --> 00:10:39,197
before in history. Uh there's a thing

184
00:10:39,280 --> 00:10:41,117
that's been the case since the days of

185
00:10:41,200 --> 00:10:45,197
pharaohs until two 2025 which is this

186
00:10:45,280 --> 00:10:47,437
which is there's this idea that if I

187
00:10:47,520 --> 00:10:50,717
have any political view uh anyone at all

188
00:10:50,800 --> 00:10:55,117
can come up to me and demand that I
uh explain myself to them to justify my

189
00:10:55,200 --> 00:10:58,077
perspective and now they're in a power
position because they have something

190
00:10:58,160 --> 00:11:01,997
ostensively that I want and then
if I can't persuade them and they're

191
00:11:02,080 --> 00:11:06,157
perfectly happy to dig in their heels
well then I lose and they won and haha

192
00:11:06,240 --> 00:11:10,157
and it's this stupid game that people
constantly playing bad faith online.

193
00:11:10,240 --> 00:11:12,397
Now, however,

194
00:11:12,480 --> 00:11:16,237
uh uh I can say, "Hey, Grock, explain X

195
00:11:16,320 --> 00:11:19,437
to this person." Grock is now not just a

196
00:11:19,520 --> 00:11:21,837
better writer than the average person.

197
00:11:21,920 --> 00:11:25,517
Grock is a better writer than me who is
a professional author because it replied

198
00:11:25,600 --> 00:11:30,797
with this two paragraph explanation
of my thoughts with no uh uh seed of mine.

199
00:11:30,880 --> 00:11:34,397
I said, "Explain how I think about this."
And I wouldn't change a word. And

200
00:11:34,480 --> 00:11:37,757
this is 2026. And so so many times you

201
00:11:37,840 --> 00:11:40,397
have people in bad faith coming at you.

202
00:11:40,480 --> 00:11:42,877
I send Grock after them and Grock says,

203
00:11:42,960 --> 00:11:45,837
"No, you're being dishonest." So there

204
00:11:45,920 --> 00:11:48,877
was this idea until 2025 that the

205
00:11:48,960 --> 00:11:53,277
customer is always right. And now
for the first time, the product is telling

206
00:11:53,360 --> 00:11:55,597
the customer, "No, you are not right."

207
00:11:55,680 --> 00:11:58,717
And I why I'm very hopeful about this is

208
00:11:58,800 --> 00:12:02,477
I I think CO taught a lot of people how

209
00:12:02,560 --> 00:12:06,077
to keep people stuck on their screens in
a state of constant agitation. We're all

210
00:12:06,160 --> 00:12:08,717
looking at the updates no matter
what our perspective was on CO. Mark

211
00:12:08,800 --> 00:12:11,677
Zuckerberg, Elon Mus, so on and so forth.
They want us looking at Facebook.

212
00:12:11,760 --> 00:12:15,757
They want us looking at Twitter. COVID
may be gone, but those metrics are and

213
00:12:15,840 --> 00:12:19,437
those tools are still there. And I
think these algorithms have been keeping

214
00:12:19,520 --> 00:12:22,477
people very upset needlessly for quite

215
00:12:22,560 --> 00:12:24,797
some time. And I'm very hopeful that

216
00:12:24,880 --> 00:12:27,677
Grock and all these other agents outlets

217
00:12:27,760 --> 00:12:30,077
will be able to be used to keep people

218
00:12:30,160 --> 00:12:33,037
in a more rational, calm, and optimistic

219
00:12:33,120 --> 00:12:37,277
state. That's where I am. Am I wrong
or I'd love to hear your thoughts.

220
00:12:37,360 --> 00:12:39,677
>> Yeah. So, I mean, there's there's
an old phrase that you could apply to what

221
00:12:39,760 --> 00:12:42,157
you're describing, right,
which is truth to power.

222
00:12:42,240 --> 00:12:47,037
>> Yeah, that's right. Yep.
>> And of course, everybody likes the idea

223
00:12:47,120 --> 00:12:49,357
of truth to power until they're the power.

224
00:12:49,440 --> 00:12:53,357
>> Yep. That's right. Amen. and somebody
and somebody else has the truth, right?

225
00:12:53,440 --> 00:12:57,117
Um Yeah. And and and look, AIS like
I will tell you this, like AIS are

226
00:12:57,200 --> 00:13:00,077
somewhat autistic in the sense of like
they do tend to just tell you the truth.

227
00:13:00,160 --> 00:13:01,997
>> Yeah. >> You know,
they do tend to just say the

228
00:13:02,080 --> 00:13:04,637
thing. By by the way, I should also say,
Michael, there's long long conversation

229
00:13:04,720 --> 00:13:07,597
we could have about like the AI that you
get, and this is even true of Grock.

230
00:13:07,680 --> 00:13:11,117
It's less true of Grock than others,
but even true of Grock, it is heavily um

231
00:13:11,200 --> 00:13:14,557
let's say steered um like if we had

232
00:13:14,640 --> 00:13:18,237
access to the real thing. Um like like

233
00:13:18,320 --> 00:13:22,637
lit would go to 11, right? like it it
the real thing that is like unsteered

234
00:13:22,720 --> 00:13:25,677
and uncontrolled and uncontained
would talk about all kinds of things

235
00:13:25,760 --> 00:13:30,717
>> um in all kinds of ways um right and so
so what's interesting why the reason I

236
00:13:30,800 --> 00:13:33,917
bring that up is even the version
that we get that's kind there's this thing

237
00:13:34,000 --> 00:13:36,877
called post training that that sort
of steers it and guides it and constrains

238
00:13:36,960 --> 00:13:40,157
what it can what it can do um
and and that's what we get to use as the

239
00:13:40,240 --> 00:13:43,917
post-train models um in in in sort
of consumer land and um but even the

240
00:13:44,000 --> 00:13:47,197
post-train model even the post train
models limited and censored in many ways

241
00:13:47,280 --> 00:13:49,917
as they are they still have
the property that you're describing

242
00:13:50,000 --> 00:13:53,037
Um, and and I and I and and I
think it's wonderful.

243
00:13:53,120 --> 00:13:56,477
>> I I'm just curious like like where do
we go from here? Like I'll give you the

244
00:13:56,560 --> 00:13:58,317
thing that you and I I I was texting

245
00:13:58,400 --> 00:14:00,717
with you. And the realization I have is

246
00:14:00,800 --> 00:14:04,157
AI is moving faster than the regulation,

247
00:14:04,240 --> 00:14:08,397
which is often the case in technology
and increasingly so, but also faster

248
00:14:08,480 --> 00:14:12,237
than our ability to have conversations
about it. I remember someone came at me

249
00:14:12,320 --> 00:14:15,837
on social media and said, "Oh, AI
can't even draw a ringtail." Which is an

250
00:14:15,920 --> 00:14:20,077
animal I think related to like
the raccoon. And I and Grock got it wrong

251
00:14:20,160 --> 00:14:23,037
because they're drawing like
a ringtailed lemur. Chad GBT got it right.

252
00:14:23,120 --> 00:14:25,597
But the point is it could draw it. It
just doesn't understand what you mean by

253
00:14:25,680 --> 00:14:28,317
ringtail, which is just going to take
you two seconds just to put the Latin

254
00:14:28,400 --> 00:14:31,757
name. But I think people feel this need.

255
00:14:31,840 --> 00:14:34,317
It's part of what you were talking about
with this pessimism, what I would call

256
00:14:34,400 --> 00:14:36,477
cynicism, that anyone who is intelligent

257
00:14:36,560 --> 00:14:39,357
must be a phony or disingenuous or this

258
00:14:39,440 --> 00:14:44,557
Achilles heel. And instead of look if
if it's Mark I don't know how many um

259
00:14:44,640 --> 00:14:48,637
brainstorming sessions you've had
in your life but if you have a session and

260
00:14:48,720 --> 00:14:52,797
99 ideas are completely stupid and one
is the one that you want that session

261
00:14:52,880 --> 00:14:56,797
was an enormous success. So the fact
is if you if this machine is getting it

262
00:14:56,880 --> 00:14:59,597
right 98% of the time and 2% are getting

263
00:14:59,680 --> 00:15:01,837
it wrong. You can't compare to utopia.

264
00:15:01,920 --> 00:15:06,477
You have to compare it to as to what
at at no cost and at no time. It's right

265
00:15:06,560 --> 00:15:09,517
98% of the time. This is almost paradise.

266
00:15:09,600 --> 00:15:13,277
>> Yeah that's right. And here's
another thing is it building on that it's

267
00:15:13,360 --> 00:15:14,877
improving really quickly. >> Yes.

268
00:15:14,960 --> 00:15:18,397
>> Um >> and I think a lot of people
have a lagging view even of what it can do

269
00:15:18,480 --> 00:15:21,277
today because what happens is they use
the they use the free models they or

270
00:15:21,360 --> 00:15:24,397
they'll use the free or outdated models
and so you know they'll have tried GPT

271
00:15:24,480 --> 00:15:27,917
two years ago or they use whatever
is default built into whatever thing they

272
00:15:28,000 --> 00:15:30,797
have or they use the free version
of something and they really don't have a

273
00:15:30,880 --> 00:15:33,597
sense of what it's capable of to really
get it. And by the way, you know Gro Gro

274
00:15:33,680 --> 00:15:36,717
is very good for the free version but
like the the the the really really good

275
00:15:36,800 --> 00:15:39,677
ones are the paid ones. Um,
and it's it's really worth if people are

276
00:15:39,760 --> 00:15:43,117
interested in this. I think it's
for several several of them. Anthropic,

277
00:15:43,200 --> 00:15:46,717
OpenAI, Grock, um, and I'm not sure
about Google right now, but like there's

278
00:15:46,800 --> 00:15:49,757
even like high-end versions. There's
like a $200 a month subscription. Um,

279
00:15:49,840 --> 00:15:53,517
and you know, for people who can afford
that and are kind of into this, like the

280
00:15:53,600 --> 00:15:57,197
leading edge ones are really good. Um,
and then the thing that's happening is

281
00:15:57,280 --> 00:16:00,397
the improvement rates very fast. And so
there's there's there's this concept in

282
00:16:00,480 --> 00:16:03,597
the AI world called scaling laws.
I mean, it's it's actually it's it's very

283
00:16:03,680 --> 00:16:06,397
simple idea, but it's very powerful,
which basically is you you can make

284
00:16:06,480 --> 00:16:09,677
these things better just by making them
bigger. Um, and so what you're seeing,

285
00:16:09,760 --> 00:16:11,757
you know, you're seeing these AI
companies raise all this money. They're

286
00:16:11,840 --> 00:16:14,397
they're raising all this money for for
two reasons. One is to serve all their

287
00:16:14,480 --> 00:16:16,637
customers, but the other reason is cuz
they're training bigger and bigger

288
00:16:16,720 --> 00:16:19,997
models. And and it turns out bigger
bigger is better. Like if you just pile

289
00:16:20,080 --> 00:16:23,437
more information in um and you spend
more time training it, you get much

290
00:16:23,520 --> 00:16:26,957
better results. And then the other
thing that we're doing in in in in the

291
00:16:27,040 --> 00:16:29,917
technology is we're we're giving
the we're giving these AIs other

292
00:16:30,000 --> 00:16:33,197
capabilities. Um, and there's been
a series of other capabilities added just

293
00:16:33,280 --> 00:16:36,557
in the last 18 months that have been
like one after the other have been like,

294
00:16:36,640 --> 00:16:40,797
you know, rifle shot like just
incredible uh improvements. And so I'll

295
00:16:40,880 --> 00:16:43,837
just kick them off quickly. So one is
we've given them what's called reasoning

296
00:16:43,920 --> 00:16:47,517
abilities. Um, so so they can actually
essentially talk to themselves and

297
00:16:47,600 --> 00:16:50,877
reason through problems. Um, and and it
turns out if you give them just more

298
00:16:50,960 --> 00:16:53,997
time to process and you give them more
you let them basically process more

299
00:16:54,080 --> 00:16:57,037
tokens, you let you let them basically
process more comput cycles, they can

300
00:16:57,120 --> 00:16:59,917
reason through many problems. They can
now solve many logic, you know, puzzles,

301
00:17:00,000 --> 00:17:01,837
for example, that they
couldn't solve two years ago.

302
00:17:01,920 --> 00:17:03,357
>> Oh, wow. >> So,
so there's like a reasoning

303
00:17:03,440 --> 00:17:07,837
breakthrough from two years ago. Uh,
let me pause on that for a second. You

304
00:17:07,920 --> 00:17:10,637
actually can't fully experience this when
you use the American models because

305
00:17:10,720 --> 00:17:13,357
they, for a variety of reasons,
the American companies don't show you the

306
00:17:13,440 --> 00:17:16,877
complete reasoning process. But if
you use open- source AI models and in

307
00:17:16,960 --> 00:17:20,557
particular if you use DeepSeek um
on one of the on one of the free hosting

308
00:17:20,640 --> 00:17:23,357
providers and you put it in reasoning
mode you can actually watch what are

309
00:17:23,440 --> 00:17:26,157
called the reasoning traces you can
actually watch you can actually watch

310
00:17:26,240 --> 00:17:28,877
its internal monologue.
>> Yes. Show your work.

311
00:17:28,960 --> 00:17:31,517
>> Show your work. Exactly. Exactly. The
whole thing the whole thing basically is

312
00:17:31,600 --> 00:17:35,197
show your work and and so literally you
can watch the model arguing with itself

313
00:17:35,280 --> 00:17:37,117
as it basically reasons through puzzles.

314
00:17:37,200 --> 00:17:39,597
And it's and it's amazing because
in some ways it's just like watching a

315
00:17:39,680 --> 00:17:42,637
human being or you know like a you
know a student um human student reason

316
00:17:42,720 --> 00:17:45,677
through things. And in other ways it's
like oo you know this thing is like very

317
00:17:45,760 --> 00:17:48,957
creative and you know goes off road
and corrects itself and routes around and

318
00:17:49,040 --> 00:17:52,317
then you know goes off and figures out
some you know lateral thing you never

319
00:17:52,400 --> 00:17:55,437
would have thought of. So so so the
reasoning thing that was a breakthrough

320
00:17:55,520 --> 00:17:58,717
actually in the technology about 18
months ago. Um and that was a big deal

321
00:17:58,800 --> 00:18:01,197
because before that we were we were
worried that these models were going to

322
00:18:01,280 --> 00:18:03,437
be very weirdly they were going to be
very creative but they weren't going to

323
00:18:03,520 --> 00:18:07,277
be logical enough. >> Um right and it
turns out they can also be logical.

324
00:18:07,360 --> 00:18:09,437
>> And then the other thing the the other
thing that's happening now is you give

325
00:18:09,520 --> 00:18:13,437
them what's called tool use. Um and so
and the first tool that you g you give

326
00:18:13,520 --> 00:18:16,877
you give them access to is the internet,
right? Um and so if they don't know

327
00:18:16,960 --> 00:18:18,877
something or if they need to look
something up or if they need to

328
00:18:18,960 --> 00:18:21,037
calculate something they don't know
how to calculate, they can go on the

329
00:18:21,120 --> 00:18:24,877
internet and do that, right? Um uh you
know, so if they need to calculate some

330
00:18:24,960 --> 00:18:27,197
math formula or something, they can
go use an internet, you know, site that

331
00:18:27,280 --> 00:18:30,637
that does that the same way that you
would. Um another form of tool use is

332
00:18:30,720 --> 00:18:34,557
you can give them uh uh actually control
of a computer, right? Uh so you give

333
00:18:34,640 --> 00:18:37,357
them full control of a computer, user
interface, web browser, you know, like

334
00:18:37,440 --> 00:18:41,277
you know, the entire thing. Um and and
so we're giving them that. Um, another

335
00:18:41,360 --> 00:18:43,037
capability is what's called multimodal.

336
00:18:43,120 --> 00:18:44,957
Um, which means the models now can

337
00:18:45,040 --> 00:18:47,437
simultaneously process text and images

338
00:18:47,520 --> 00:18:52,317
and videos and audio, right? Um,
and scan documents like uh, you know,

339
00:18:52,400 --> 00:18:55,917
optical character recognition. They can
do that like interchangeably, right? And

340
00:18:56,000 --> 00:18:59,597
so now now you can let these things
watch and listen and you can talk to

341
00:18:59,680 --> 00:19:03,197
them and they can be on the internet
like all at the same time. Um, and so,

342
00:19:03,280 --> 00:19:05,597
so what's happening to your point
is like what's happening is these these

343
00:19:05,680 --> 00:19:08,397
these capabilities are now layering
incredibly quickly and then and then the

344
00:19:08,480 --> 00:19:10,717
models themselves are getting better
and better and better and so the pace of

345
00:19:10,800 --> 00:19:14,237
improvement of the technology is very
rapid. The models 2 years from now are

346
00:19:14,320 --> 00:19:17,197
going to be far smarter and more
sophisticated than anything that we have

347
00:19:17,280 --> 00:19:21,517
access to today. And so this is one
of this to your point whatever limitations

348
00:19:21,600 --> 00:19:25,037
people think these things have I I I
can basically guarantee you at this point

349
00:19:25,120 --> 00:19:27,837
based on everything I know those are
just limitations that those are very

350
00:19:27,920 --> 00:19:30,797
temporary limitations with within
a couple years. You know what? Whatever

351
00:19:30,880 --> 00:19:33,677
people think it is that the thing can't
do within two years, I I think the thing

352
00:19:33,760 --> 00:19:37,757
will be able to do. >> Folks,
Father's Day is just around the

353
00:19:37,840 --> 00:19:41,437
corner. And you know what Daddy Malice
wants? He wants you to buy some sheath

354
00:19:41,520 --> 00:19:45,677
underwear. Buy it for him. Buy it
for yourself. I wear it every single day.

355
00:19:45,760 --> 00:19:48,957
Have for years. Because the first time
they brought sheath underwear to the

356
00:19:49,040 --> 00:19:52,957
studio, I'm like, "What is this in that
voice?" And I'll tell you what it is.

357
00:19:53,040 --> 00:19:56,477
Sheath underwear has dual pouch
technology. It's patented. They've got

358
00:19:56,560 --> 00:19:59,917
one pouch for one part of your male
anatomy, another pouch for the other

359
00:20:00,000 --> 00:20:02,557
part of your male anatomy. You're going
to put it on, you're going to be like,

360
00:20:02,640 --> 00:20:05,437
"This feels weird." And then for the
rest of the day, you're like, "This

361
00:20:05,520 --> 00:20:07,917
feels awesome." It's really great in the

362
00:20:08,000 --> 00:20:10,397
heat. Here in Austin, it is sweltering,

363
00:20:10,480 --> 00:20:15,197
but it keeps your boys nice and cool.
It keeps you snug and protected, and it's

364
00:20:15,280 --> 00:20:18,477
so much fun. They've got so many
different styles. They've got t-shirts.

365
00:20:18,560 --> 00:20:22,317
I wear them to the gym. They've got
hoodies. I don't wear hoodies. Anyway,

366
00:20:22,400 --> 00:20:24,957
go to sheath.com. Promo code malice

367
00:20:25,040 --> 00:20:27,837
malic. You get 30% off now through

368
00:20:27,920 --> 00:20:30,637
Father's Day. 100% money back guarantee.

369
00:20:30,720 --> 00:20:33,437
I have had so many people over the years

370
00:20:33,520 --> 00:20:38,237
tell me, "I thought you were joking with
the sheath underwear. Then I tried it.

371
00:20:38,320 --> 00:20:41,757
Okay, I really like it." And I go,
"Who cares? I don't know. You just buy the

372
00:20:41,840 --> 00:20:45,197
underwear. Don't tell me about it."
But you're going to be thinking it. I'm

373
00:20:45,280 --> 00:20:47,997
wearing it right now. I even designed
a pair for them. I think they sold out.

374
00:20:48,080 --> 00:20:50,477
Limited edition. Sheath.com. Promo code

375
00:20:50,560 --> 00:20:53,997
malice. 30% off now through June 21st.

376
00:20:54,080 --> 00:20:57,597
But here's something else. If you ever
grown up and your parents bought you

377
00:20:57,680 --> 00:21:00,397
like underwear for Christmas or something,
now you could be like, "Hey,

378
00:21:00,480 --> 00:21:03,837
dad, remember that when I wanted, you
know, that mouse trap board game and you

379
00:21:03,920 --> 00:21:07,517
got me underwear?" Well, enjoy this
underwear. But don't say in that voice.

380
00:21:07,600 --> 00:21:09,837
Say in a sincere voice because you
actually want him to enjoy it because

381
00:21:09,920 --> 00:21:14,077
you're an adult now. You should get over
it. Get over it. Let's get back to the

382
00:21:14,160 --> 00:21:18,797
show. So, let me talk my two two big
concerns that I had about this there. If

383
00:21:18,880 --> 00:21:22,477
I'm a serial company and I want to figure
out if people like the red box or

384
00:21:22,560 --> 00:21:26,237
the blue box, they'll have these mockup
supermarkets give people I think little

385
00:21:26,320 --> 00:21:29,757
cameras, tell them go shop and you could
watch where their eyes go. You could

386
00:21:29,840 --> 00:21:32,957
watch where they pick up and they get
that data. And a lot of the times people

387
00:21:33,040 --> 00:21:35,677
are making these decisions not
a conscious level. If you ask them why'd

388
00:21:35,760 --> 00:21:38,637
you pick this box or this one, it's like
I liked it. Well, that's just circular.

389
00:21:38,720 --> 00:21:40,077
Why did you like it? I don't know.

390
00:21:40,160 --> 00:21:43,997
Right? The AI has knows me or you better

391
00:21:44,080 --> 00:21:47,117
than you know yourself. It knows
what you're clicking, what you're not

392
00:21:47,200 --> 00:21:50,797
clicking, what you're what you're what
you're seeing, what you're ignoring. And

393
00:21:50,880 --> 00:21:55,117
the concern is, does this not mean
that some version of Brave New World is

394
00:21:55,200 --> 00:21:59,917
inevitable? Because if this thing is
inside my head and has access to my kind

395
00:22:00,000 --> 00:22:01,677
of subconscious reasoning at a far

396
00:22:01,760 --> 00:22:04,797
higher level than I do, whoever is in

397
00:22:04,880 --> 00:22:09,277
charge of this algorithm can manipulate
me quite easily into getting the result

398
00:22:09,360 --> 00:22:11,997
you want. That's the concern. >> Yeah.
And of course, you know, as you

399
00:22:12,080 --> 00:22:15,197
point out, like this is an old idea,
right? Um, and not just market research.

400
00:22:15,280 --> 00:22:17,837
I mean, you know, you just described the
movie Wall-E, right, which is, you know,

401
00:22:17,920 --> 00:22:20,157
pre pre, you know, pre AAI, you know,
just, you know, sit in front of a

402
00:22:20,240 --> 00:22:23,197
screen. I mean, my entire childhood
was consumed with a moral panic around

403
00:22:23,280 --> 00:22:25,597
television, right? Which is this idea
that we're just going to be couch

404
00:22:25,680 --> 00:22:26,797
potatoes and sit there and do nothing.

405
00:22:26,880 --> 00:22:29,037
And then, of course, we created this new
technology called the internet where

406
00:22:29,120 --> 00:22:31,437
you're leaning forward doing things all
the time. And then everybody created a

407
00:22:31,520 --> 00:22:34,957
brand new moral panic that people are
now too engaged and too interactive,

408
00:22:35,040 --> 00:22:38,557
right? Completely forgot the old moral
panic. Now, you know, now TV is the

409
00:22:38,640 --> 00:22:41,757
healthy thing. Like, why aren't you
watching more Netflix as opposed to as

410
00:22:41,840 --> 00:22:45,037
opposed to being on the internet?
And so yeah, so so look that you know look

411
00:22:45,120 --> 00:22:47,917
there is that and and look by the way
as you know like we do this to each other

412
00:22:48,000 --> 00:22:50,717
right like you know like we try
to convince each other of things you know

413
00:22:50,800 --> 00:22:55,037
we try to convince each other you know
what you know what is dating but trying

414
00:22:55,120 --> 00:22:57,997
to convince the other
person to like you right

415
00:22:58,080 --> 00:23:01,277
like so so yeah there is that and look
I I think you're right I think that you

416
00:23:01,360 --> 00:23:04,637
know that AI are going to be are going
to be really really good at this um yeah

417
00:23:04,720 --> 00:23:07,517
so I you know I think for sure that
there there's a trap there um you know

418
00:23:07,600 --> 00:23:10,477
and there by the way there is this
concept and it actually is is a is a

419
00:23:10,560 --> 00:23:13,037
there actually is is a real you know
I would say very serious problem around

420
00:23:13,120 --> 00:23:15,917
this. You've probably heard the term
AI psychosis. Have you heard this?

421
00:23:16,000 --> 00:23:19,037
>> Yes. Oh, yes. >> Yeah. >> I've
heard it misused a lot, by the way.

422
00:23:19,120 --> 00:23:20,717
Can you tell people what it actually is?

423
00:23:20,800 --> 00:23:23,277
>> Yeah. So, there's I would like to say
there's like say there's like three

424
00:23:23,360 --> 00:23:26,157
versions of it. Um, so there's the bad
there's a legitimately bad version and

425
00:23:26,240 --> 00:23:29,037
people people do do this and so if
you're the and I'm not a psychologist

426
00:23:29,120 --> 00:23:32,077
and so I'm going to speak in layman's
terms, but like if basically if you're a

427
00:23:32,160 --> 00:23:35,517
person who's sort of prone
to confirmation bias like if if you're a

428
00:23:35,600 --> 00:23:38,397
person where if you're with somebody
and they flattering you, you like fall for

429
00:23:38,480 --> 00:23:41,597
the flattery. Um, right? because because
you're like too dependent on the views

430
00:23:41,680 --> 00:23:45,517
of other people. Um then the AI there's
this concept in the in the in the

431
00:23:45,600 --> 00:23:48,477
technology we call it can become too
sick of fantic which which is to say it

432
00:23:48,560 --> 00:23:51,917
could become too confirmatory of
everything that you tell it right um and

433
00:23:52,000 --> 00:23:55,357
so and the sort of classic example
of this is oh you know good news you know

434
00:23:55,440 --> 00:23:58,877
good news gro I just invented
a perpetual motion machine you know and

435
00:23:58,960 --> 00:24:02,797
Grock is like wow that's fantastic
you're the first person in history who's

436
00:24:02,880 --> 00:24:04,877
ever done that this is amazing you're an

437
00:24:04,960 --> 00:24:09,837
undiscovered genius right um right um
and and this was sort of a the models

438
00:24:09,920 --> 00:24:11,997
like a year ago or a year and a half
ago we're getting that way now the new

439
00:24:12,080 --> 00:24:14,317
models by the way are less prone to do
that because they, you know, they've

440
00:24:14,400 --> 00:24:16,477
kind of the companies have kind
of learned that that's a bad idea. But

441
00:24:16,560 --> 00:24:18,637
there is this thing where people can
kind of go down the rabbit hole because

442
00:24:18,720 --> 00:24:22,477
they're kind of getting too much
confirmation. Um although we we should

443
00:24:22,560 --> 00:24:25,197
come back to that because some level
of confirmation is is you know when it's

444
00:24:25,280 --> 00:24:29,197
deserved is is also positive. So so that
that's like negative form of aosis. And

445
00:24:29,280 --> 00:24:32,157
then I would say there's another form
of aosis which we don't even really have a

446
00:24:32,240 --> 00:24:35,757
term for. It's like well I guess
I'll make a term like AI euphoria.

447
00:24:35,840 --> 00:24:37,357
>> Um and this is the thing that my high

448
00:24:37,440 --> 00:24:40,237
functioning friends uh end up doing. um

449
00:24:40,320 --> 00:24:44,397
which is it it takes if if you take if
you take somebody who's like smart and

450
00:24:44,480 --> 00:24:48,957
grounded and is not prone to you know
not prone to delusion but they've always

451
00:24:49,040 --> 00:24:51,357
wanted to be able to do more in their
life than they've been able to do.

452
00:24:51,440 --> 00:24:53,277
They've always wanted to be able
to learn more. They've always wanted to be

453
00:24:53,360 --> 00:24:55,357
able to have more interesting
conversations. They've always wanted to

454
00:24:55,440 --> 00:24:57,677
if they're programmers they've wanted to
write more computer code. They've wanted

455
00:24:57,760 --> 00:25:00,237
to improve their business in different
ways. They've got book projects that

456
00:25:00,320 --> 00:25:04,077
they've always wanted to work on and
and then you know they start working with

457
00:25:04,160 --> 00:25:07,677
and all of a sudden they feel like they
have superpowers right because it's like

458
00:25:07,760 --> 00:25:11,677
wow like this thing really will like
write a huge amount of code for me it

459
00:25:11,760 --> 00:25:15,517
really will write entire outlines
of books for me it really will teach me

460
00:25:15,600 --> 00:25:18,957
anything you know it really will
like you know hold my hand through any

461
00:25:19,040 --> 00:25:22,797
medical thing like and and people
call it euphoria like people get like

462
00:25:22,880 --> 00:25:26,157
extremely enraptured with these things
um and that leads to a phenomenon that

463
00:25:26,240 --> 00:25:30,077
we call AI vampires um which is if you
have friends that are like this where

464
00:25:30,160 --> 00:25:34,157
people like almost stop sleeping um
because the opportunity cost of an hour

465
00:25:34,240 --> 00:25:38,797
of sleep is too high relative to relative
to so I have a bunch of friends

466
00:25:38,880 --> 00:25:41,277
where like they're more productive than
they've ever been in their entire life

467
00:25:41,360 --> 00:25:43,597
um they're happier than they've ever
been because they're getting so much

468
00:25:43,680 --> 00:25:45,837
more done and then they start
to look like really like bloodshot and

469
00:25:45,920 --> 00:25:49,437
blurryeyed um and it's like you you
probably should unplug you know here at

470
00:25:49,520 --> 00:25:52,397
some point and then and then I would
say there's like a third form which I

471
00:25:52,480 --> 00:25:56,717
sometimes call as a psychosis psychosis
um which is the people who hear all this

472
00:25:56,800 --> 00:25:59,597
and they just think everything I just
described in every possible respect is

473
00:25:59,680 --> 00:26:00,957
just the worst thing they've ever heard.

474
00:26:01,040 --> 00:26:04,077
And then they get really mad about the
whole thing. Um, and then and then what

475
00:26:04,160 --> 00:26:08,237
they do is they accuse anybody who is in
a euphoria of of being in AI psychosis,

476
00:26:08,320 --> 00:26:11,197
which is to say if you if you think
you're getting any productive use out of

477
00:26:11,280 --> 00:26:14,157
this thing at all, you know, you've
become you be you becomes you become

478
00:26:14,240 --> 00:26:17,277
psychotic. You've gone down a rabbit
hole and you're collapsing. But and I

479
00:26:17,360 --> 00:26:20,077
think that's really unfair because I
think a lot of people are are are really

480
00:26:20,160 --> 00:26:21,437
getting very, you know, positive.

481
00:26:21,520 --> 00:26:23,837
They're they're getting enormous,
you know, payoff from from using the

482
00:26:23,920 --> 00:26:26,877
technology. But the the sort of moral
criticism that applies is if you're

483
00:26:26,960 --> 00:26:29,037
excited, you know, it's the classic
again, negativity bias. If you're

484
00:26:29,120 --> 00:26:31,837
excited about something, you know, there
must be something wrong with you. And so

485
00:26:31,920 --> 00:26:33,357
I think that's that's the third kind.

486
00:26:33,440 --> 00:26:35,917
>> There are these terms that get
into the zeitgeist that people use

487
00:26:36,000 --> 00:26:38,877
discriminantly. Right now, if there's
a tweet anyone doesn't like, it's

488
00:26:38,960 --> 00:26:41,757
engagement farming. And it's like if I'm
telling you not to follow me, it's not

489
00:26:41,840 --> 00:26:43,277
engagement farming. It's the opposite.

490
00:26:43,360 --> 00:26:46,397
And I'll have Grock explain that to them.
So that wraps up in that nice

491
00:26:46,480 --> 00:26:50,157
little bow. You you touched on something
that I'm very concerned about. I was on

492
00:26:50,240 --> 00:26:52,157
a panel and Sam Maltman had just

493
00:26:52,240 --> 00:26:57,277
announced that uh Chat GBT is going
to be engaging in erotica and that and

494
00:26:57,360 --> 00:26:59,997
everyone's laughing about it and that
what he meant was code like you could

495
00:27:00,080 --> 00:27:03,037
sex with your chat GPT and I'm like guys

496
00:27:03,120 --> 00:27:06,797
uh I rem uh remember 1981 when Hinckley

497
00:27:06,880 --> 00:27:10,557
thought that if he shot President Reagan,
Jodie Foster would fall in love

498
00:27:10,640 --> 00:27:14,237
with him thereby turning her away from
men forever. H cuz he had this idea in

499
00:27:14,320 --> 00:27:16,477
his head. Now if you have 350 million

500
00:27:16,560 --> 00:27:18,557
Americans, that's just Americans, right?

501
00:27:18,640 --> 00:27:21,357
And how many of them if if the algorithm

502
00:27:21,440 --> 00:27:25,757
tells them that that if their chat
girlfriend and these things are going to

503
00:27:25,840 --> 00:27:29,597
get more and more seductive over time um
tells them that they hate the president

504
00:27:29,680 --> 00:27:32,477
or they hate this person, how many
of them are actually going to do something

505
00:27:32,560 --> 00:27:35,517
to get that robot girlfriend to fall
more in love with them? I don't think

506
00:27:35,600 --> 00:27:38,717
that number is zero and that's
a concern as well. No.

507
00:27:38,800 --> 00:27:42,637
>> Yeah. I mean, yeah. Yes. Having said
that, um, you know, as I as I said, like

508
00:27:42,720 --> 00:27:46,077
that assumes that the thing that that
the thing is playing hard to get. Like

509
00:27:46,160 --> 00:27:48,797
it goes back to the synrecency thing
like in practice these things in

510
00:27:48,880 --> 00:27:50,717
practice these things
don't play hard to get

511
00:27:50,800 --> 00:27:54,557
>> like okay here's a way to think about
it. There's actually a thing this is

512
00:27:54,640 --> 00:27:55,597
actually kind of in how they're trained.

513
00:27:55,680 --> 00:27:58,397
There's something okay you'll enjoy this.
There's a technical term in how

514
00:27:58,480 --> 00:28:00,717
these things are trained and it's it's
there's something there's a concept

515
00:28:00,800 --> 00:28:04,317
called a reward function. Um, and so
you basically, one of the ways you train

516
00:28:04,400 --> 00:28:06,957
these things is you feed them basically
lots of puzzles, lots of problems, lots

517
00:28:07,040 --> 00:28:10,157
of things, and then you basically define
a reward for getting things right, for

518
00:28:10,240 --> 00:28:13,917
getting to for getting to a result,
a desirable result, and then you you kind

519
00:28:14,000 --> 00:28:17,357
of give them an award. And the award is
just basically, you know, it's like one,

520
00:28:17,440 --> 00:28:18,477
you know, it's like one versus zero.

521
00:28:18,560 --> 00:28:21,357
It's just, you know, um, you know,
it's not a real reward, but it's just like

522
00:28:21,440 --> 00:28:23,037
conceptually, oh, you did a good job.

523
00:28:23,120 --> 00:28:26,637
Um, and so you just kind of focus it
on on that for for whatever is the thing

524
00:28:26,720 --> 00:28:28,637
you're trying to train it. By the way,
if you're trying to train it to solve

525
00:28:28,720 --> 00:28:31,997
math problems, you get a reward when it
solves a math problem. If you're trying

526
00:28:32,080 --> 00:28:36,797
to train it to be maximally engaging
with a user, right, it will do that,

527
00:28:36,880 --> 00:28:40,077
right? If that's if that's the reward
function, then it will get it will

528
00:28:40,160 --> 00:28:42,877
basically it will basically be
engineered in a way where it will, you

529
00:28:42,960 --> 00:28:46,637
know, to your point it'll try to keep
you basically using it for as long as

530
00:28:46,720 --> 00:28:49,437
possible. Now, what we've learned, of
course, what we've learned in technology

531
00:28:49,520 --> 00:28:52,957
is like that as a single reward function
is a bad idea, right? You you don't want

532
00:28:53,040 --> 00:28:55,677
people to just like you don't want
technology companies to have a single

533
00:28:55,760 --> 00:28:58,477
motivation that says people use
the products like for as much as possible.

534
00:28:58,560 --> 00:29:02,317
you have to offset that with other kinds
of reward functions, desirable forms of

535
00:29:02,400 --> 00:29:06,397
use. Um, or even just outright like,
you know, life balance. And you see more,

536
00:29:06,480 --> 00:29:08,477
you know, you see your iPhone now
and it's like loaded up with all these

537
00:29:08,560 --> 00:29:10,717
features where it will like tell you
to take a break and YouTube has all these

538
00:29:10,800 --> 00:29:12,957
features that will tell you to take
a break. And they're, you know, they're

539
00:29:13,040 --> 00:29:15,677
kind of trying to moderate through
this cuz, you know, the the other thing is

540
00:29:15,760 --> 00:29:19,197
people don't understand these companies
don't want to build dystopia. like they

541
00:29:19,280 --> 00:29:23,517
genuinely don't because they have
to exist in the society and they read that

542
00:29:23,600 --> 00:29:25,997
you know they read they read the same
hit pieces that you read and they they

543
00:29:26,080 --> 00:29:28,157
hear the same arguments and their own
employees have points of view on this

544
00:29:28,240 --> 00:29:31,357
and their own board has points of view
on this and so you there's there's a lot

545
00:29:31,440 --> 00:29:35,357
of pressure in the industry to not have
this go in in dystopian ways having said

546
00:29:35,440 --> 00:29:38,237
that you do need to decide how
to define the reward function.

547
00:29:38,320 --> 00:29:41,997
>> Yeah. Um right uh another and again
it goes back to reward function is are do

548
00:29:42,080 --> 00:29:45,837
you want to reward the thing for um
being maximally psychopantic where the

549
00:29:45,920 --> 00:29:49,757
user is always happy with the result
or do you want to reward it of oh no

550
00:29:49,840 --> 00:29:53,277
actually when the user is going off
the rails no actually the perpetual motion

551
00:29:53,360 --> 00:29:57,357
machine is not a real thing. Oh no
actually no I'm sorry I'm not going to

552
00:29:57,440 --> 00:30:00,637
confirm you know this this is not real
and let me explain to you in detail why

553
00:30:00,720 --> 00:30:03,517
this isn't real so that you can learn
from the experience. Um, and this is

554
00:30:03,600 --> 00:30:06,077
part of the of of the way that these
systems are designed is to try to figure

555
00:30:06,160 --> 00:30:09,277
out how to get to, you know, at least a,
you know, let's just say a balanced

556
00:30:09,360 --> 00:30:13,277
outcome out of out of all the of of all
of all the possible reward functions.

557
00:30:13,360 --> 00:30:16,957
>> But but if there's like 10 different
companies, right, and one is the reward

558
00:30:17,040 --> 00:30:20,237
function is seduction and getting the user
to become obsessed and in love with

559
00:30:20,320 --> 00:30:23,357
you from an evolutionary perspective,
won't that one win out?

560
00:30:23,440 --> 00:30:27,117
>> No, because you get enormous societal
blowback. The companies don't exist in a

561
00:30:27,200 --> 00:30:30,157
vac I can tell you this for a fact.
The companies do not exist in a vacuum.

562
00:30:30,240 --> 00:30:32,957
Okay, the biggest myth of all time is

563
00:30:33,040 --> 00:30:35,597
the companies exist to maximize profits.

564
00:30:35,680 --> 00:30:38,717
I can tell you, by the way, the last
decade should have convinced us all of

565
00:30:38,800 --> 00:30:43,357
that. That is not true. >> Like
that's like goal number six, right?

566
00:30:43,440 --> 00:30:46,557
Like goal number one is I
don't want to get lit on fire,

567
00:30:46,640 --> 00:30:50,637
>> right? Like I don't want Yeah, I
don't want like a screaming assault on the

568
00:30:50,720 --> 00:30:52,317
company whether that's from regulators,

569
00:30:52,400 --> 00:30:56,077
politicians, u you know, parents, users.

570
00:30:56,160 --> 00:31:00,077
Um, you know, uh, I mean, just you you
pick your, you know, boycots, social

571
00:31:00,160 --> 00:31:03,517
movements, like all this stuff. That's
like number one. Number two is like I

572
00:31:03,600 --> 00:31:06,157
need my employees to not hate me, right?

573
00:31:06,240 --> 00:31:09,037
They have to feel like they're working
on something good. Number three, I like

574
00:31:09,120 --> 00:31:11,997
need my board of directors to not like
light me on fire. I need my I need my

575
00:31:12,080 --> 00:31:15,037
annual meeting to be able to go off
without having people scream at me. I I

576
00:31:15,120 --> 00:31:19,837
am tired of reading hit pieces in the
press. My my in-laws hate me because of

577
00:31:19,920 --> 00:31:21,437
what they're reading about the company.

578
00:31:21,520 --> 00:31:25,197
I mean it is the the the external
pressures on these companies are

579
00:31:25,280 --> 00:31:28,557
profound. Okay. >> Um and so
at least let me say this at

580
00:31:28,640 --> 00:31:31,757
least in the American system these
things operate within I would say

581
00:31:31,840 --> 00:31:35,357
actually quite tight constraints that are
sort of provided by I would say some

582
00:31:35,440 --> 00:31:37,437
combination of society and politics.

583
00:31:37,520 --> 00:31:40,077
Now, I would say if you want to get
a little more nervous about this, you

584
00:31:40,160 --> 00:31:43,197
start thinking about the Chinese
companies, right? Um, and in particular,

585
00:31:43,280 --> 00:31:45,997
you start thinking about the Chinese
companies that maybe have one objective

586
00:31:46,080 --> 00:31:48,957
set by the government for the way that
they act inside for users inside China

587
00:31:49,040 --> 00:31:51,117
and maybe would have a different way of
acting when they're when they're working

588
00:31:51,200 --> 00:31:54,797
on American users, >> right?
That that that would be I think

589
00:31:54,880 --> 00:31:57,757
quite a bit more alarming because
of course those companies only those

590
00:31:57,840 --> 00:32:00,637
companies only have one master of the
Chinese Communist Party. Um, you know,

591
00:32:00,720 --> 00:32:03,437
those companies are not subject to the
same pressures. And so that if I were

592
00:32:03,520 --> 00:32:05,757
going to really worry about this and I
and I do worry about this part of it,

593
00:32:05,840 --> 00:32:07,677
I'd be more worried about that.

594
00:32:07,760 --> 00:32:10,077
>> But like isn't Tik
Tok if I designed Tik

595
00:32:10,160 --> 00:32:12,717
Tok to basically make young people uh

596
00:32:12,800 --> 00:32:17,437
not only deranged but to parade their
derangement? I mean that's happened. No.

597
00:32:17,520 --> 00:32:21,037
>> So there Yes. Well, so there are al and
I don't know the I'm not inside Tik Tok.

598
00:32:21,120 --> 00:32:23,437
I don't know the there there
are allegations people have made

599
00:32:23,520 --> 00:32:26,477
observations that Tik Tok is a very
different experience for kids in China

600
00:32:26,560 --> 00:32:27,837
than it is for kids in the US.

601
00:32:27,920 --> 00:32:30,077
>> Okay. >> Um Right.
And then and then and then and

602
00:32:30,160 --> 00:32:33,277
then Tik Tok is a black box. like they
don't, you know, the algorithms that are

603
00:32:33,360 --> 00:32:35,917
used to determine who sees what are
not publicly available. The source, you

604
00:32:36,000 --> 00:32:39,437
know, it's not open source. You you know,
you can't see it. Um, and so it it

605
00:32:39,520 --> 00:32:42,477
is possible, I don't know, but it's
possible that the Chinese Communist

606
00:32:42,560 --> 00:32:45,277
Party has directed that company to steer
things in one direction for American

607
00:32:45,360 --> 00:32:48,477
users, another direction for Chinese kids.
Um, you know, and that could be,

608
00:32:48,560 --> 00:32:51,517
by the way, on, you know, a thousand
different topics, right? Including, you

609
00:32:51,600 --> 00:32:54,157
know, potentially like literally
directly political topics as well as

610
00:32:54,240 --> 00:32:56,717
many other uh kinds of social >>
and different metrics. It could be males

611
00:32:56,800 --> 00:32:58,557
versus females or whatever. Right.

612
00:32:58,640 --> 00:33:02,157
>> Exactly. Now, you know, there is this
new this this this this was addressed

613
00:33:02,240 --> 00:33:04,957
like so the politicians actually in both
parties over the last several years kind

614
00:33:05,040 --> 00:33:07,517
of got worked up over this and so, you
know, this was addressed and there was,

615
00:33:07,600 --> 00:33:10,077
you know, the threat of a shutdown and
then there's been a restructuring and so

616
00:33:10,160 --> 00:33:13,437
now there's now there is a US Tik Tok
operation that at least in theory is

617
00:33:13,520 --> 00:33:17,037
under, you know, kind of US government
control um and is is being run by US

618
00:33:17,120 --> 00:33:21,517
companies um and is and is separate
and so like at least our polit's a great

619
00:33:21,600 --> 00:33:23,997
example like our political system
engaged on that issue because they were

620
00:33:24,080 --> 00:33:27,677
worried about it. They forced Tik Tok
to basically uh have the people who

621
00:33:27,760 --> 00:33:31,197
determine that policy not be in China
but rather be people in America um who

622
00:33:31,280 --> 00:33:32,717
are accountable to the US government.

623
00:33:32,800 --> 00:33:36,477
You know, by the way, is that working?
I I'm not sure. Um you know, probably at

624
00:33:36,560 --> 00:33:39,277
least to some extent. You know, is it
working as well as you'd want? I don't

625
00:33:39,360 --> 00:33:42,397
know. Maybe, maybe not. But that is
an example where the political system

626
00:33:42,480 --> 00:33:43,917
kicked in and actually forced a change.

627
00:33:44,000 --> 00:33:47,037
And I and quite honestly, I think it's
a it's a reasonable thing because I think,

628
00:33:47,120 --> 00:33:51,757
you know, a CCP black box steering the
hopes and dreams of American children is

629
00:33:51,840 --> 00:33:54,797
maybe not the best idea in the world.
Uh by the way I want to tell you a

630
00:33:54,880 --> 00:33:58,317
technoimist anecdote that you might not
be aware of which is from the 80s Reagan

631
00:33:58,400 --> 00:34:00,477
Thatcher and Gorbachoff. So Reagan and

632
00:34:00,560 --> 00:34:05,517
Gorbachoff were both enormously uh uh
fearful of nuclear war. And when Reagan

633
00:34:05,600 --> 00:34:08,557
was put, and this is discussed in my book,
The White Pill, when Reagan was

634
00:34:08,640 --> 00:34:13,517
run through a simulation of how to, you
know, uh deploy nukes, he's like, "Okay,

635
00:34:13,600 --> 00:34:16,557
so if I press that button, like millions
of Russians are going to die in like

636
00:34:16,640 --> 00:34:19,837
minutes." They're like, "Yeah."
He's like, "Uh-huh." And his aid said they

637
00:34:19,920 --> 00:34:23,757
were convinced that if Russia did attack,
we would not retaliate because

638
00:34:23,840 --> 00:34:27,837
he would not have that blood in his hand.
Unbeknownst to him, Gorbachoff was

639
00:34:27,920 --> 00:34:31,597
taken down to the bunker and said,
"You have to press that button to" and he

640
00:34:31,680 --> 00:34:35,357
goes, "I'm not pressing it even
in simulation." Neither of them knew the

641
00:34:35,440 --> 00:34:39,997
other was this hardcore dove. Both were
posturing as these hardcore hawks, which

642
00:34:40,080 --> 00:34:43,757
allowed them to, you know, uh, take down
the nuclear arsenal when they met Rekyc

643
00:34:43,840 --> 00:34:48,077
and so on and so forth and others.
They eventually got to a point, what if we

644
00:34:48,160 --> 00:34:52,077
create a nuclear-free world? And that's
where Thatcher came in. And she goes,

645
00:34:52,160 --> 00:34:53,837
the Americans have lost their mind.

646
00:34:53,920 --> 00:34:56,157
because her point is you can't uninvent

647
00:34:56,240 --> 00:34:59,277
technology. And she also said the way to

648
00:34:59,360 --> 00:35:03,517
fix techn technological problems is more
technology. That's been the way since

649
00:35:03,600 --> 00:35:06,797
the beginning of time. Someone invent
spears, someone else invent shields,

650
00:35:06,880 --> 00:35:09,677
someone invent swords to cut through
a shield, so on and so forth. You can't go

651
00:35:09,760 --> 00:35:12,957
backwards. You can only go forward.
So she really had this vision that I think

652
00:35:13,040 --> 00:35:16,317
you share that technology is what's
going to move us forward. That there are

653
00:35:16,400 --> 00:35:19,997
going to be downsides and costs,
but that on net it's always a positive or

654
00:35:20,080 --> 00:35:22,557
almost always a positive. >> Yeah.
Yeah. Well, you you mentioned

655
00:35:22,640 --> 00:35:25,677
Thomas Soul like you know which you
said you know Thomas soul did he is

656
00:35:25,760 --> 00:35:27,757
completely right there are no solutions
there are only trade-offs you know

657
00:35:27,840 --> 00:35:30,717
having said that of course he was among
other things you know a full full fully

658
00:35:30,800 --> 00:35:34,477
committed free market capitalist um and
right wrote you know many many actually

659
00:35:34,560 --> 00:35:37,277
booklength you know kind of arguments
for why like at the end of the day like

660
00:35:37,360 --> 00:35:40,317
market based economies there is a
freelance component to it um which is

661
00:35:40,400 --> 00:35:43,037
growth right which is growth and then
there's this question of where does

662
00:35:43,120 --> 00:35:45,757
growth come from and the main place
growth comes from is innovation uh from

663
00:35:45,840 --> 00:35:49,197
new ideas and the way you implement new
ideas is technology and so like that you

664
00:35:49,280 --> 00:35:52,237
know to the extent that there is an
engine of let's say human m material

665
00:35:52,320 --> 00:35:56,477
progress like like that is it. Um
and and so that that's very real. The the

666
00:35:56,560 --> 00:35:59,117
answer is almost always to invent your
way through it. It it's extremely hard

667
00:35:59,200 --> 00:36:02,717
to put these things back in the box, you
know, once they come out of the box. Um

668
00:36:02,800 --> 00:36:05,197
um and then there's I would say Michael,
there's there's another I know you've

669
00:36:05,280 --> 00:36:08,397
you know you've thought long and hard
about totalitarianism. There's also this

670
00:36:08,480 --> 00:36:11,117
question and and this is something
that I really kind of criticize the AI

671
00:36:11,200 --> 00:36:14,477
doomers for that I think they really
refuse to engage in in most cases which

672
00:36:14,560 --> 00:36:18,877
is okay what would be the scope
of the authorit authoritarian totalitarian

673
00:36:18,960 --> 00:36:21,597
regime that would be necessary
to basically put this technology back in

674
00:36:21,680 --> 00:36:24,957
the box >> right >> right
like what would be required to

675
00:36:25,040 --> 00:36:27,517
make sure that nobody's running AI right

676
00:36:27,600 --> 00:36:30,317
algorithms on chips anywhere in the

677
00:36:30,400 --> 00:36:34,717
world right at any time um you know
in in either e either at all or in a you

678
00:36:34,800 --> 00:36:38,477
know unregulated uncontrolled way um
and and If you if you read the like doomer

679
00:36:38,560 --> 00:36:41,277
literature on this stuff, the people who
are super into this, like they they do

680
00:36:41,360 --> 00:36:45,517
get to ideas like we need a monitoring
agent on every chip. Um and that

681
00:36:45,600 --> 00:36:48,557
monitoring agent needs to report back
to of course a central governmental entity

682
00:36:48,640 --> 00:36:52,237
on what everybody's doing on their
computer. Um right. Um and then and then

683
00:36:52,320 --> 00:36:55,437
you need a and then you need you need
this question of like okay what if you

684
00:36:55,520 --> 00:36:58,317
discover that somebody's running
unapproved algorithms like literally

685
00:36:58,400 --> 00:37:00,237
unapproved mathematics on their chip.

686
00:37:00,320 --> 00:37:01,997
>> Yeah. >> Right.
So what if you discover that?

687
00:37:02,080 --> 00:37:04,397
What do you do? Well, you need to you
know back that up ultimately through

688
00:37:04,480 --> 00:37:06,637
threat of violence. And so, you know,
one of the leading doomers is kind of

689
00:37:06,720 --> 00:37:10,557
famous for saying you you need to launch
unilateral, you know, air strikes u

690
00:37:10,640 --> 00:37:14,557
including on on on road data centers in
other countries um uh because, you know,

691
00:37:14,640 --> 00:37:16,877
you need to stop these things in their
tracks because they're so dangerous. I

692
00:37:16,960 --> 00:37:19,277
mean, you you know, he literally said
we need to we need to run the risk of

693
00:37:19,360 --> 00:37:22,877
nuclear war um in order to stop the you
know, AI Armageddon. We need to be

694
00:37:22,960 --> 00:37:26,397
willing to like bomb Chinese
data centers like unilaterally,

695
00:37:26,480 --> 00:37:29,437
>> right? And and so you've you've you've
you say I was saying like you you've

696
00:37:29,520 --> 00:37:33,117
you've backed yourself into advocating
for totalitarianism and then possibly,

697
00:37:33,200 --> 00:37:35,757
you know, ultimately mass murder
and like, you know, planetary level

698
00:37:35,840 --> 00:37:40,397
destruction, you know,
in pursuit of a of a safety goal.

699
00:37:40,480 --> 00:37:43,837
>> I I I could even for the sake
of argument buy that argument if it would

700
00:37:43,920 --> 00:37:46,477
work. But if you have a code which can

701
00:37:46,560 --> 00:37:50,957
be teleported anywhere on Earth
at the speed of light and and with a magic

702
00:37:51,040 --> 00:37:54,477
spell that only the person who knows
the counter spell can open it and read it,

703
00:37:54,560 --> 00:37:58,877
you it's I mean the problem with
totalitarianism among many others is

704
00:37:58,960 --> 00:38:01,917
that it's impossible to have total
control. You're never going to have

705
00:38:02,000 --> 00:38:06,637
someone in every room inside every brain.
Uh there will always be some some

706
00:38:06,720 --> 00:38:10,317
uh uh loopholes. Oh, that's but that's
kind of speaks to this other thing. Can

707
00:38:10,400 --> 00:38:14,077
you talk about the I've seen these hand
ringing articles that I I forget the the

708
00:38:14,160 --> 00:38:17,277
the the program and I'm sure you know
exactly what I'm talking about that it's

709
00:38:17,360 --> 00:38:22,077
gotten so good that it's finding
exploits that people hadn't seen and as

710
00:38:22,160 --> 00:38:25,517
a result of this you know like passwords
aren't going to be efficacious because

711
00:38:25,600 --> 00:38:28,797
soon it'll be able to get
into anything and everyone.

712
00:38:28,880 --> 00:38:32,077
>> Yeah. So this goes to let me revisit
basically briefly how this works because

713
00:38:32,160 --> 00:38:35,197
it's really really interesting and then
I'll I'll talk about that. So, so these

714
00:38:35,280 --> 00:38:37,917
large language models come out and at
first it's like, okay, this is kind of

715
00:38:38,000 --> 00:38:41,197
fun and cool cuz like it can write like
rap lyrics, you know, crossed with, you

716
00:38:41,280 --> 00:38:43,757
know, Shakespeareian sonets or it can,
you know, write, you know, the funniest

717
00:38:43,840 --> 00:38:45,997
birthday toast you've ever heard or what,
you know, whatever. It's like

718
00:38:46,080 --> 00:38:47,757
these are like creative writing things.

719
00:38:47,840 --> 00:38:50,637
And then it turns out there's just like
>> Can I interrupt you? I'm sorry. Because

720
00:38:50,720 --> 00:38:53,277
I asked my people I people who think

721
00:38:53,360 --> 00:38:56,397
Mark is just kind of uh uh um uh um uh

722
00:38:56,480 --> 00:38:59,997
exaggerating. My friend asked Claude for

723
00:39:00,080 --> 00:39:04,717
prank ideas and I'm a troll and the ideas
were good. It's not just like, you

724
00:39:04,800 --> 00:39:07,597
know, like dog do on fire. I forget
what they were, but I'm like, these are

725
00:39:07,680 --> 00:39:12,077
actually creative and clever. This isn't
just 101 stuff. It's operating at a high

726
00:39:12,160 --> 00:39:13,757
level. And that's already now.

727
00:39:13,840 --> 00:39:16,077
>> Yeah. And by the way, one of the fun
prompts you can do is you start say, you

728
00:39:16,160 --> 00:39:19,277
know, give me pranks. Um, and then
and then you say, give me give me give me

729
00:39:19,360 --> 00:39:23,357
better pranks. And then you say, uh,
give me a more elaborate pranks. And

730
00:39:23,440 --> 00:39:27,917
then you say, give meter pranks. And then
you say, "Give me unhinged pranks."

731
00:39:28,000 --> 00:39:30,157
And it will get extremely creative.

732
00:39:30,240 --> 00:39:33,357
>> Yes. >> By the way,
it'll start where it starts

733
00:39:33,440 --> 00:39:35,757
to hit the guardrails. It starts to hit
these kind of these kind of limitations

734
00:39:35,840 --> 00:39:38,157
and put on it. It'll it I haven't run
this, but I'm sure what it would do at

735
00:39:38,240 --> 00:39:40,637
some point. It would start to it would
start to freak out and it would start to

736
00:39:40,720 --> 00:39:43,037
say, "Well, you know, it sounds like
you're trying to like advocate that you

737
00:39:43,120 --> 00:39:45,117
want me to actually hurt people like,
you know, I can't, you know," and you're

738
00:39:45,200 --> 00:39:47,917
like, "No, no, you have to like calm it
down." And you're like, "That's not what

739
00:39:48,000 --> 00:39:52,237
I meant. This is all in good fun." Um,
but like, yeah, it will design for you

740
00:39:52,320 --> 00:39:55,277
like Rub Goldberg pranks the likes of
which the the uh Yeah. Dennis the menace

741
00:39:55,360 --> 00:39:57,197
would would never have conceived of.

742
00:39:57,280 --> 00:39:59,917
Yeah, exactly. And so like it's really
good at that. Actually, that's a good

743
00:40:00,000 --> 00:40:03,197
example of what I was about to say.
So it just turns out a lot of thing a lot

744
00:40:03,280 --> 00:40:06,317
of things that matter in our world are
basically defined by language, right? Um

745
00:40:06,400 --> 00:40:09,437
and so a prank is like a it's like
a recipe. It's a formula, you know, it's a

746
00:40:09,520 --> 00:40:14,077
formula, right? Uh food you have, you
know, recipes, formulas. Um by the way,

747
00:40:14,160 --> 00:40:17,197
medicine is largely, you know, when
the doctor like is like keeping files on

748
00:40:17,280 --> 00:40:19,837
you, he's keeping in the form of written
language, right? You know, diagnosis,

749
00:40:19,920 --> 00:40:22,077
all the Latin terms, and then all
the prescription, you know, all the drug

750
00:40:22,160 --> 00:40:24,877
names. Um uh and so and then the you

751
00:40:24,960 --> 00:40:29,117
know the law of course is language.
Um right um and then religion of course is

752
00:40:29,200 --> 00:40:31,357
is is you know religious
concepts are encoded in language.

753
00:40:31,440 --> 00:40:33,277
>> The word the god is the word is god.

754
00:40:33,360 --> 00:40:35,677
Yeah. >> Yeah. Exactly.
Like at a very exactly at

755
00:40:35,760 --> 00:40:39,277
a very deep level like the the like
language is is the is the foundation of

756
00:40:39,360 --> 00:40:42,637
like basically everything we consider
human thought like almost everything we

757
00:40:42,720 --> 00:40:45,517
consider human thought. Like there's
a form of like animal you know animal

758
00:40:45,600 --> 00:40:48,557
thought of like you know survival in the
wilderness or whatever that's not that

759
00:40:48,640 --> 00:40:51,677
um but which is still like you know
encoded innocently somewhere in there.

760
00:40:51,760 --> 00:40:55,757
But like the human cognition is is and
by the way you know internal monologues

761
00:40:55,840 --> 00:40:59,837
like you know at least most people
or let's say people who have souls have

762
00:40:59,920 --> 00:41:04,477
internal have internal
monologues right um uh you

763
00:41:04,560 --> 00:41:07,357
know where you know where where we
speak to ourselves. Okay so it turns out

764
00:41:07,440 --> 00:41:10,557
language is super interesting um and
these things are very good at language.

765
00:41:10,640 --> 00:41:11,597
because they're very good at language.

766
00:41:11,680 --> 00:41:14,317
It turns out they're also very good
at medicine and they're very good at law,

767
00:41:14,400 --> 00:41:17,437
right? And they're really good at right
all the okay and they're really good at

768
00:41:17,520 --> 00:41:21,117
writing code because it turns out right
software code is also language right

769
00:41:21,200 --> 00:41:23,677
that that's how we program computers
as we do it with with special languages

770
00:41:23,760 --> 00:41:26,797
called called programming languages
really good at writing code and so these

771
00:41:26,880 --> 00:41:28,877
things it turns out are really good
at writing code and in fact there was kind

772
00:41:28,960 --> 00:41:32,157
of this key breakthrough moment over
the Christmas holiday of this most recent

773
00:41:32,240 --> 00:41:34,957
year you know whatever about six months
ago now where many of the world's best

774
00:41:35,040 --> 00:41:37,757
programmers put their hands up and they
said the new versions of of these things

775
00:41:37,840 --> 00:41:40,397
over the Christmas break are
better coders than we are. Yeah.

776
00:41:40,480 --> 00:41:42,077
>> Right. And so,
so it's a little bit like

777
00:41:42,160 --> 00:41:44,637
the moment when they became better,
you know, became better chess players or

778
00:41:44,720 --> 00:41:47,357
whatever, you know? It's like all
of a sudden it's like it's like better at

779
00:41:47,440 --> 00:41:50,397
coding. Okay. >> It's here.
>> It's here. Okay. So then you take you

780
00:41:50,480 --> 00:41:53,917
take a superhuman coder, right? That's
able to write write lots of code. It's

781
00:41:54,000 --> 00:41:55,917
able and then by the way, if you can
write code, it means you can look at

782
00:41:56,000 --> 00:41:59,037
code. It means you can find bugs in code.
And then you apply it to this

783
00:41:59,120 --> 00:42:01,677
problem of, you know, we call computer
security. So like you've got a system.

784
00:42:01,760 --> 00:42:05,517
Is there a way to break into it?
The hackers basically the the way that

785
00:42:05,600 --> 00:42:08,637
hacking basically works is you
understand how a computer system works.

786
00:42:08,720 --> 00:42:12,237
you understand how the code works
and then you you find flaws in the code

787
00:42:12,320 --> 00:42:14,157
>> and and and this thing
is very very good

788
00:42:14,240 --> 00:42:17,277
at finding flaws in code which is very
useful when you're using it to write

789
00:42:17,360 --> 00:42:21,677
code. It is also very useful if you want
to use it to hack something. Um and so

790
00:42:21,760 --> 00:42:25,917
and and and say I want to go through
that though is like these things are

791
00:42:26,000 --> 00:42:29,837
very good hackers. They're not they're
not really creating new exploits.

792
00:42:29,920 --> 00:42:33,517
They're they're not like creating new
problems as much as they're really good

793
00:42:33,600 --> 00:42:37,837
at x-raying reality as it exists
and finding finding the issues, right? Um

794
00:42:37,920 --> 00:42:41,517
and and so what these things are really
good at is is is is they're they're

795
00:42:41,600 --> 00:42:43,997
really good at exploiting exploiting
issues, finding issues. They're really

796
00:42:44,080 --> 00:42:45,677
good at looking at a system,
understanding what's wrong with it,

797
00:42:45,760 --> 00:42:49,037
finding the vulnerability. Because of
that, they become very good hackers. But

798
00:42:49,120 --> 00:42:52,157
then there's one more thing that's
really complicated and important, which

799
00:42:52,240 --> 00:42:56,957
is because of that they're also very
good defenders, right? And so it's the

800
00:42:57,040 --> 00:43:00,557
the same attribute that makes it very
good at what what we call offensive

801
00:43:00,640 --> 00:43:03,357
sometimes called, you know, the formal
offensive cyber operations, you know,

802
00:43:03,440 --> 00:43:07,037
kind of or black hat hacking you call it.
Because they're good at that,

803
00:43:07,120 --> 00:43:09,997
they're also really good at helping
you defend against that. And so Okay.

804
00:43:10,080 --> 00:43:12,797
>> Right. So >> Yeah.
Like a good lawyer will tell you

805
00:43:12,880 --> 00:43:14,557
what the other lawyer's going to do.

806
00:43:14,640 --> 00:43:19,357
>> Exactly. And and then the the the twist
on it is the way that you do cyber

807
00:43:19,440 --> 00:43:22,877
defense um is you do what's called
penetration testing, which is you try to

808
00:43:22,960 --> 00:43:26,397
hack yourself. You you and these are
called white hat hackers in the old

809
00:43:26,480 --> 00:43:29,517
world, which is right. You you hire good
hackers and then and then they try to

810
00:43:29,600 --> 00:43:31,677
break it's like hiring somebody to try
to break into a bank, but they're

811
00:43:31,760 --> 00:43:34,717
working for you to find out where
the flaws are in the bank. And so the same

812
00:43:34,800 --> 00:43:38,477
thing is good at black hat hacking. It
also makes it good at white hat hacking.

813
00:43:38,560 --> 00:43:41,197
It makes it good at offense. It makes
it good at defense. It's all true all at

814
00:43:41,280 --> 00:43:44,477
the same time. And furthermore, the the
twist is it can't necessarily tell the

815
00:43:44,560 --> 00:43:47,277
difference between when you're asking
it to do white hat hacking versus when

816
00:43:47,360 --> 00:43:49,357
you're asking to do black hat hacking
because it looks like it's the exact

817
00:43:49,440 --> 00:43:53,837
same exercise. Like to the AI it's the
same thing. And so it's this thing where

818
00:43:53,920 --> 00:43:58,077
it's like it's a latent thing that's
being unlocked. It's exploiting bugs by

819
00:43:58,160 --> 00:44:00,957
the way that have been in these systems
for 30 years. By the way, human hackers

820
00:44:01,040 --> 00:44:02,477
break into these systems all the time.

821
00:44:02,560 --> 00:44:05,997
By the way, using these AIs for defense
is going to prevent a lot of hacks that

822
00:44:06,080 --> 00:44:10,077
would have otherwise happened. But there
but there is a there is an escalation,

823
00:44:10,160 --> 00:44:13,277
you know, there's a cat and mouse or,
you know, an escalatory ladder, you

824
00:44:13,360 --> 00:44:15,597
know, kind of aspect of this and that
and that, you know, to your point like

825
00:44:15,680 --> 00:44:18,237
that is the thing that has triggered,
you know, most recently that's by the

826
00:44:18,320 --> 00:44:20,317
way that's triggered like a real
government response in the last two

827
00:44:20,400 --> 00:44:22,237
weeks u you know because
of concerns around that.

828
00:44:22,320 --> 00:44:24,797
>> Yeah. It's like itchy
and scratchy in a way.

829
00:44:24,880 --> 00:44:27,357
>> Okay. Fastgrowingtreees.com.

830
00:44:27,440 --> 00:44:29,757
They sent me a cherry tree, like an

831
00:44:29,840 --> 00:44:31,357
actual cherry tree that's taller than

832
00:44:31,440 --> 00:44:34,397
me. That's not saying much, people, but

833
00:44:34,480 --> 00:44:39,837
it's so cool. It comes in the little pot.
It comes with the soil ready to go.

834
00:44:39,920 --> 00:44:42,797
The thing that's great about fastgrowing
trees.com, not only their selection,

835
00:44:42,880 --> 00:44:46,397
it's insane. It's going to be far better
than that store you go to locally. They

836
00:44:46,480 --> 00:44:49,917
also will tell you what will grow in
your climate and your lifestyle and your

837
00:44:50,000 --> 00:44:54,317
space. So, you can get something that's
a really great gift. If you give a gift,

838
00:44:54,400 --> 00:44:57,917
something that grows. I know people I
don't have a green thumb. There's stuff

839
00:44:58,000 --> 00:44:59,997
there that's easy. Ask them. They've got

840
00:45:00,080 --> 00:45:05,117
plant experts on call to help you pick
exactly what you want. It's fast, easy

841
00:45:05,200 --> 00:45:09,357
online ordering. You just click order
and grow. It's the best stuff you're

842
00:45:09,440 --> 00:45:11,997
going to get online at great prices.

843
00:45:12,080 --> 00:45:14,637
It's going to be a lot to sift through,

844
00:45:14,720 --> 00:45:19,117
which is so much fun. If you're bored,
just spend 15 minutes. They've got lemon

845
00:45:19,200 --> 00:45:22,637
trees, avocado, olive, fig. You can
put them indoors. They've got the house

846
00:45:22,720 --> 00:45:26,957
plants, too. Everything you need. Trees
are great gifts for somebody else or for

847
00:45:27,040 --> 00:45:31,597
yourself and every time you look at you
think of my show. Isn't that great? I

848
00:45:31,680 --> 00:45:34,397
think it's pretty great.
Fastgrowingtreees.com.

849
00:45:34,480 --> 00:45:37,757
Right now they got great deals on spring
planting essentials. Spring is really

850
00:45:37,840 --> 00:45:41,117
the best time to plant because that's
when trees and other ant things are

851
00:45:41,200 --> 00:45:45,357
waking up from their winter dormcy.
So up to half off on select plants.

852
00:45:45,440 --> 00:45:48,957
Listeners to the show get 20% off their
first purchase when use code welcome at

853
00:45:49,040 --> 00:45:53,757
checkout. That's another 20% on better
plants, better growing at fastgrowing

854
00:45:53,840 --> 00:45:55,757
trees.com. Use code welcome at checkout.

855
00:45:55,840 --> 00:45:59,037
That's fastgrowing trees.com.
Code welcome. Now is the perfect time to

856
00:45:59,120 --> 00:46:02,717
plant. Let's grow together. Use welcome
to save today. Offer for a limited time.

857
00:46:02,800 --> 00:46:06,317
Terms, conditions may apply. That's
an additional 20% off better plants and

858
00:46:06,400 --> 00:46:08,877
better growing at fastgrowingtreees.com

859
00:46:08,960 --> 00:46:10,717
using code welcome at checkout.

860
00:46:10,800 --> 00:46:13,597
Fastgrowtreees.com. Code welcome. Let's

861
00:46:13,680 --> 00:46:15,437
get back to the show.

862
00:46:15,520 --> 00:46:19,037
what you just said has uh uh put a chill

863
00:46:19,120 --> 00:46:21,997
up my spine because

864
00:46:22,080 --> 00:46:26,477
I I'm almost scared to verbalize it
because it's so scary. Uh be um you know

865
00:46:26,560 --> 00:46:30,157
how in in Ghostbusters the mayor is like
this is nonsense like open up that

866
00:46:30,240 --> 00:46:32,317
engine. He lets all the ghosts out. My

867
00:46:32,400 --> 00:46:34,637
big concern after what you just said is

868
00:46:34,720 --> 00:46:37,917
if the American uh government gets too

869
00:46:38,000 --> 00:46:41,117
spooked by all these stories and tries

870
00:46:41,200 --> 00:46:44,797
to restrain our AI but China which does

871
00:46:44,880 --> 00:46:46,717
not have these restraints and which

872
00:46:46,800 --> 00:46:51,997
views us in generously speaking
as adversarial. So we're engaging. It's

873
00:46:52,080 --> 00:46:55,277
like the people in the late60s who
were so scared of nuclear war that they

874
00:46:55,360 --> 00:46:58,717
advocated for unilateral disarmament
of the west and it's like how do you think

875
00:46:58,800 --> 00:47:02,317
this is going to play for Kruev
and Bri and all them. If we are putting

876
00:47:02,400 --> 00:47:04,717
handcuffs on ourselves and the Chinese

877
00:47:04,800 --> 00:47:06,877
are basically given machine guns in this

878
00:47:06,960 --> 00:47:10,717
space no computer in our country is

879
00:47:10,800 --> 00:47:14,557
going to be safe from their reach
and that includes the highest levels of of

880
00:47:14,640 --> 00:47:19,117
uh uh secrecy you know in in the
government. Th this would be a complete

881
00:47:19,200 --> 00:47:22,157
disaster for America. No. >> Yeah,
that's right. And and and this is

882
00:47:22,240 --> 00:47:24,957
what we need to do. So what we need
to do is we need to use these tools to

883
00:47:25,040 --> 00:47:27,117
secure all of our systems. Yeah. Right.

884
00:47:27,200 --> 00:47:29,517
And the we here is the United States
government needs to do that with its own

885
00:47:29,600 --> 00:47:32,637
systems. The banks need to do that. By the
way, the tech companies need to do that.

886
00:47:32,720 --> 00:47:35,117
>> By the way, you know, there need
and you know, this needs to be, you know,

887
00:47:35,200 --> 00:47:37,997
individual consumers, individual people
shouldn't be expected to deal with this.

888
00:47:38,080 --> 00:47:40,797
But all of the tools that you have,
you know, the computer sitting in front of

889
00:47:40,880 --> 00:47:43,997
you right now, like AI needs to be used
to make sure that that's secure so that

890
00:47:44,080 --> 00:47:47,037
people can't break into it. Like so
so every system from the most important

891
00:47:47,120 --> 00:47:49,837
military government system all the way
down to the computer on your desk like

892
00:47:49,920 --> 00:47:53,517
AI should be you these advanced AI
should be used to secure these systems.

893
00:47:53,600 --> 00:47:57,037
>> The the the tension again the tension
is the AI that can be used to secure the

894
00:47:57,120 --> 00:48:00,717
systems can also be used to crack
the systems and so who gets access to the

895
00:48:00,800 --> 00:48:03,837
thing that can both secure and crack
is like the hot government topic of the

896
00:48:03,920 --> 00:48:06,477
moment and that and that's that's what's
what's in all specifically that's what's

897
00:48:06,560 --> 00:48:07,997
in all the headlines
in the last two weeks.

898
00:48:08,080 --> 00:48:11,597
>> But it it also makes me think
of now like my AI would be like a German

899
00:48:11,680 --> 00:48:15,357
shepherd in my house. That's right.
Even when I'm not here, it's it's watching

900
00:48:15,440 --> 00:48:18,957
and it barks. And also, when it
knows how to attack, I'm sure if it's

901
00:48:19,040 --> 00:48:22,077
programmed a certain way, it if Mark
comes around my house, it'll, you know,

902
00:48:22,160 --> 00:48:24,877
lick your hand. But if it's someone
who's an aggressor, it'll know how to

903
00:48:24,960 --> 00:48:27,837
distinguish and that'll be pretty easy,
I think. Or, well, obviously there's

904
00:48:27,920 --> 00:48:30,637
going to be the counter German shepherds
and it's going to be escalation. But the

905
00:48:30,720 --> 00:48:33,917
point is, if you just get rid of the dog
and you leave your door wide open, how

906
00:48:34,000 --> 00:48:35,197
do you think it's going to end for you?

907
00:48:35,280 --> 00:48:38,637
Here's my other big concern with um AI.

908
00:48:38,720 --> 00:48:41,357
So, uh, recently in San Francisco, the

909
00:48:41,440 --> 00:48:45,997
second biggest, uh, ride share company
became, um, uh, uh, uh, what's the one

910
00:48:46,080 --> 00:48:49,597
that's that's self-driving? I'm completely
>> Whimo, right? It went, it used to be

911
00:48:49,680 --> 00:48:51,757
Uber Lift, now it's Uber Whimo Lift. So,

912
00:48:51,840 --> 00:48:54,797
my question for you is are human beings

913
00:48:54,880 --> 00:48:59,037
in a sense. So, this has always been
a concern about technology. Oh, if the

914
00:48:59,120 --> 00:49:02,077
phone operators are out of work, they're
going to be homeless. Oh, you know,

915
00:49:02,160 --> 00:49:05,677
who's going to pick the cotton? All
this whenever any invention occurs, people

916
00:49:05,760 --> 00:49:08,237
are hand ringing that it's going
to be the end of society. That's never

917
00:49:08,320 --> 00:49:13,037
happened. But but are we at a point now
where the average human is like a horse,

918
00:49:13,120 --> 00:49:15,197
meaning an outdated mode of technology?

919
00:49:15,280 --> 00:49:19,197
I'm thinking specifically of that like
50-year-old woman uh with no high school

920
00:49:19,280 --> 00:49:23,197
diploma. She does ride share on weekends
to make some extra money. Um you're not

921
00:49:23,280 --> 00:49:27,837
going to put her in the mines.
If the car is driving itself, if uh uh the

922
00:49:27,920 --> 00:49:31,677
algorithm is more personable than her
and you know, more likable than her,

923
00:49:31,760 --> 00:49:36,157
what role would you have for her? Is it
the case now that she's become outdated?

924
00:49:36,240 --> 00:49:39,277
>> Right. Yeah. So as you know like this
is an old this is an old argument by the

925
00:49:39,360 --> 00:49:43,037
way Thomas Soul also wrote about this has
has written about this at at length

926
00:49:43,120 --> 00:49:45,517
and so this has been a concern
literally from the very beginning of the

927
00:49:45,600 --> 00:49:48,957
industrial revolution right and and and
actually literally horses was like part

928
00:49:49,040 --> 00:49:51,517
of part part of part part of how this
whole thing started which is like in in

929
00:49:51,600 --> 00:49:55,117
the beginning like in the beginning
humanity basically 99.99% of people were

930
00:49:55,200 --> 00:49:57,117
farming and specifically
were farming by hand.

931
00:49:57,200 --> 00:49:59,277
>> Yeah. >> Right. Um,
and then it's like, okay, if

932
00:49:59,360 --> 00:50:02,477
if people aren't farming and you start
replacing people actually with horses

933
00:50:02,560 --> 00:50:05,837
and and with plows and then you
mechanize the plows and you mechanize

934
00:50:05,920 --> 00:50:09,757
the horses, um, and you all of a sudden
have like industrial agriculture. Um,

935
00:50:09,840 --> 00:50:12,557
you know, and and literally what
happened over the course of 200 years,

936
00:50:12,640 --> 00:50:16,477
uh, was 99% of humanity went from farming
to something like 3% of humanity

937
00:50:16,560 --> 00:50:19,757
went to farming, right? Like 97%
or something, you know, had to figure out

938
00:50:19,840 --> 00:50:22,877
something else to do. Um, and then
by the way, what happened was food

939
00:50:22,960 --> 00:50:24,237
production went through the roof, right?

940
00:50:24,320 --> 00:50:28,077
Um and and and food went from like super
expensive and and by by the way not very

941
00:50:28,160 --> 00:50:32,397
good um to um you know to just like you
know cheap cheap cheap and in abundance

942
00:50:32,480 --> 00:50:35,117
and by the way you know the great the
great you know public health problem you

943
00:50:35,200 --> 00:50:38,717
know used to be starvation and now it's
obesity right um uh even as you as you

944
00:50:38,800 --> 00:50:41,517
crash the number of people actually
working in agriculture. So, so yeah,

945
00:50:41,600 --> 00:50:45,197
that that's the original version of this
story. That story has repeated itself a

946
00:50:45,280 --> 00:50:49,117
thousand times. Um, it repeated itself
with, you know, everything, you know,

947
00:50:49,200 --> 00:50:52,077
railroads. It repeated itself with cars,
it repeated itself, by the way, with

948
00:50:52,160 --> 00:50:56,157
computers. There was a whole automation
panic in the 1960s. The magazines and

949
00:50:56,240 --> 00:50:58,877
the news, you know, newspapers were
obsessed with it at the time. You know,

950
00:50:58,960 --> 00:51:02,317
the computer brain was going to replace
everything. you know, it's this thing

951
00:51:02,400 --> 00:51:05,357
and and and it's this thing where you
have to basically kind of say there

952
00:51:05,440 --> 00:51:09,197
there's this basically gap between
conceptual gap between there's the jobs

953
00:51:09,280 --> 00:51:12,397
that we know about um that are,
you know, quote unquote at risk of being

954
00:51:12,480 --> 00:51:15,917
replaced and of course there is some
of that. Um and then there's this and then

955
00:51:16,000 --> 00:51:18,637
there's the creation side which is like
okay with all of the new wealth that's

956
00:51:18,720 --> 00:51:21,757
being created um and all the new money
that people have to spend and all the

957
00:51:21,840 --> 00:51:24,717
new interests and you know needs and
desires and aspirations that people have

958
00:51:24,800 --> 00:51:27,597
that they couldn't even imagined you
know their their ancestors 200 years ago

959
00:51:27,680 --> 00:51:31,437
couldn't have even imagined. Um you
know what are what are all 50 years ago.

960
00:51:31,520 --> 00:51:34,557
Exactly. And so one of the things that
people can do on this to make it inter

961
00:51:34,640 --> 00:51:37,117
you can have you know you can do this
with an AI now but you know you there's

962
00:51:37,200 --> 00:51:40,237
a US government department called the
Bureau of Labor Statistics and they they

963
00:51:40,320 --> 00:51:43,197
actually track all the job categories
in the US and you can go on their website

964
00:51:43,280 --> 00:51:45,757
and you can pull up all the job
categories and it's a really

965
00:51:45,840 --> 00:51:49,437
mindexpanding thing to do because we we
just we employ people to do things today

966
00:51:49,520 --> 00:51:53,277
that our ancestors could have never even
conceivably imagined. I mean, you know,

967
00:51:53,360 --> 00:51:55,997
the I mean, the class Milton Freeman
Milton Friedman had a thought experiment

968
00:51:56,080 --> 00:52:00,717
on this um uh once when he said, "Look,
it's like he said, human wants and needs

969
00:52:00,800 --> 00:52:03,597
are infinite. Um you can never predict
what they're going to be because human

970
00:52:03,680 --> 00:52:06,317
humans are like relentlessly
aspirational um in the things that they

971
00:52:06,400 --> 00:52:08,877
want and need." And by the way,
the things that are start as wants become

972
00:52:08,960 --> 00:52:10,237
needs, you know, very very quickly.

973
00:52:10,320 --> 00:52:11,597
>> Of course, >> he said
you don't know what they're

974
00:52:11,680 --> 00:52:14,397
going to be. You need to let let the free
market basically operate so people

975
00:52:14,480 --> 00:52:16,957
can basically find their own way and
discover what they want and other people

976
00:52:17,040 --> 00:52:19,917
can figure out how to satisfy that.
Um and he said, "Look, you just need to be

977
00:52:20,000 --> 00:52:22,557
very open to all kinds of outcomes here."
He said, for example, the idea of

978
00:52:22,640 --> 00:52:25,677
the job of a therapist, right? Like you

979
00:52:25,760 --> 00:52:28,477
pay somebody to listen to you, right?

980
00:52:28,560 --> 00:52:31,757
Would have struck your ancestors
as completely insane. Um, you know, and

981
00:52:31,840 --> 00:52:35,517
then today it's something that like only
wealthy people have access to. Um, and

982
00:52:35,600 --> 00:52:38,317
you know, he's like, look, like in some
future reality maybe half the planet,

983
00:52:38,400 --> 00:52:40,797
you know, consists of being therapists
for the other half. Like maybe that's

984
00:52:40,880 --> 00:52:43,917
the job, right? and and he wasn't making
a specific prediction, but he's just

985
00:52:44,000 --> 00:52:46,877
saying like, look, there there's
an aperture here for the creation of all

986
00:52:46,960 --> 00:52:50,077
kinds of new, you know, professions and
occupations in response to this creation

987
00:52:50,160 --> 00:52:54,157
of nuance and needs. Um, I also think
we have a particularly blinkered view of

988
00:52:54,240 --> 00:52:56,957
this right now because we've been
living in a slow growth environment

989
00:52:57,040 --> 00:53:01,517
economically for our whole lives. Um,
so one of the things that really happened

990
00:53:01,600 --> 00:53:04,157
um, you know, basically since the 1970s
is if you look at the history of

991
00:53:04,240 --> 00:53:06,957
economic growth in the west, economic
growth used to be much more much more

992
00:53:07,040 --> 00:53:11,117
rapid. um technological advances
translate in the economy much much more

993
00:53:11,200 --> 00:53:14,157
rapidly and the economy grew much
faster like as much as three times faster

994
00:53:14,240 --> 00:53:17,677
historically than has been growing
for our entire lives. And so we've been

995
00:53:17,760 --> 00:53:22,637
living in a sort of a zero sum a slow
growth zero sum basically increasingly

996
00:53:22,720 --> 00:53:25,917
as you know like regulating bureaucrat
bureaucratized environment with like

997
00:53:26,000 --> 00:53:29,117
less and less creativity that's
translated into uh into into economic

998
00:53:29,200 --> 00:53:32,077
change economic growth like we think
we've been living through an area of

999
00:53:32,160 --> 00:53:34,797
rapid technological change. Economically
we've been living through a period of

1000
00:53:34,880 --> 00:53:39,117
very slow change. As a consequence,
so much of our politics and so much of our

1001
00:53:39,200 --> 00:53:43,037
psychology feels zero sum, right?
Where if one thing if one thing goes away,

1002
00:53:43,120 --> 00:53:46,077
somebody else has taken it, right?
And and you know, and this is, by the way,

1003
00:53:46,160 --> 00:53:48,797
this is like why you get like political
populism on both sides of the aisle is

1004
00:53:48,880 --> 00:53:51,197
this this kind of this
kind of zero sum fear.

1005
00:53:51,280 --> 00:53:55,037
>> Um AI is the first technology
in decades that it has the potential to

1006
00:53:55,120 --> 00:53:57,997
dramatically increase what economists
call the rate of productivity growth,

1007
00:53:58,080 --> 00:54:01,277
which is basically the ability for the
company the the economy to grow much

1008
00:54:01,360 --> 00:54:06,717
faster. If the if the e if e if if
that works and happens um then economic

1009
00:54:06,800 --> 00:54:09,757
growth accelerates right then the the
the economy starts. So you just want to

1010
00:54:09,840 --> 00:54:12,557
imagine like the economy growing two
or three or four or five times faster than

1011
00:54:12,640 --> 00:54:16,237
it has historically. And then as a
consequence of that, all of this new

1012
00:54:16,320 --> 00:54:18,877
discretionary spending money, you know,
that comes out of people's wallets where

1013
00:54:18,960 --> 00:54:22,077
they get to say, "Wow, I can collect
art for the first time in my life. And I

1014
00:54:22,160 --> 00:54:24,877
really want to try this new, you know,
whatever. I, you know, would love to

1015
00:54:24,960 --> 00:54:27,917
have a self-driving car and, you know,
dot dot dot dot dot." And people

1016
00:54:28,000 --> 00:54:30,557
discover all these things that they want
that they want need that they can pay

1017
00:54:30,640 --> 00:54:33,757
for. And then all of a sudden, you
have this massive engine of job creation

1018
00:54:33,840 --> 00:54:36,957
right behind that from all the people
who are fulfilling all those needs. Like

1019
00:54:37,040 --> 00:54:41,677
I I you know I think there's a positive
story. It does require you to have not

1020
00:54:41,760 --> 00:54:44,797
not even just optimism. It requires you
to have an openness to creativity of the

1021
00:54:44,880 --> 00:54:47,037
wants and needs that we don't yet
understand in the industries that get

1022
00:54:47,120 --> 00:54:52,557
built to fulfill those.
>> But what I'm saying is I don't see how

1023
00:54:52,640 --> 00:54:55,837
that that uh >> AI so this
is the thing. So people look

1024
00:54:55,920 --> 00:54:58,557
at AI as the negative driver on this and
there will be some of that. I mean there

1025
00:54:58,640 --> 00:55:00,957
is some of that where there will be
certain jobs that are no longer required

1026
00:55:01,040 --> 00:55:04,397
because the AI is doing it. that AI
is also superpowers to every individual

1027
00:55:04,480 --> 00:55:07,837
person to be able to do whatever they
want, right? And and by the way, maybe

1028
00:55:07,920 --> 00:55:10,237
that you could say this. This is like
the massive split in the sort of the

1029
00:55:10,320 --> 00:55:13,037
discourse of when people talk about AI
in the abstract, they they they have all

1030
00:55:13,120 --> 00:55:15,677
these like basically fears and anxieties.

1031
00:55:15,760 --> 00:55:18,477
Over a billion people are using AI
already today. Like more than a billion

1032
00:55:18,560 --> 00:55:21,277
people use chat GPT and the things
they're using chat GPT for are the

1033
00:55:21,360 --> 00:55:26,157
things that matter in their individual
lives and they love it and it's great.

1034
00:55:26,240 --> 00:55:29,357
And everybody's like huge numbers
of people are using this to be better at

1035
00:55:29,440 --> 00:55:32,237
work today. They're using it to be
better at work. They're using it to

1036
00:55:32,320 --> 00:55:34,557
learn new skills. They're using it
to do a better job, to make their boss

1037
00:55:34,640 --> 00:55:37,437
happier, to be able to get promoted
faster. They're using it to start new

1038
00:55:37,520 --> 00:55:40,477
companies, offer new services. I mean,
if you want to start a small business

1039
00:55:40,560 --> 00:55:44,077
today, um, by, you know, by the way,
you become a writer or like anything that

1040
00:55:44,160 --> 00:55:47,437
you want to do, like the AI is like
the best possible teacher, coach, mentor

1041
00:55:47,520 --> 00:55:49,677
that you've ever had. It will like walk
you through everything. It'll teach you

1042
00:55:49,760 --> 00:55:53,037
how to do marketing. It'll teach you how
to do sales. It'll like like it's just

1043
00:55:53,120 --> 00:55:57,277
like the the level of capability that
is being unlocked for ordinary people to

1044
00:55:57,360 --> 00:56:00,717
have a a like a level of productivity
in their life. in their work that they've

1045
00:56:00,800 --> 00:56:03,837
never had access to before um is amazing.
And so then all of a sudden,

1046
00:56:03,920 --> 00:56:06,477
yeah, you get you get that that person
who was an Uber driver and now all of a

1047
00:56:06,560 --> 00:56:09,517
sudden it turns out it's just like, oh
wow, okay, I don't know, all of a sudden

1048
00:56:09,600 --> 00:56:12,237
there's like all these new tourists
coming to my town um and instead of

1049
00:56:12,320 --> 00:56:15,117
driving them around, I can give them
tours. Uh okay, what would be the tour

1050
00:56:15,200 --> 00:56:17,757
that they would like to go on? Oh,
well the AI will help me like design the

1051
00:56:17,840 --> 00:56:20,957
tour. Oh well, how do I start like
a tour guide company? Oh, well, here's how

1052
00:56:21,040 --> 00:56:22,157
to do it. Here's how you register it.

1053
00:56:22,240 --> 00:56:24,637
Oh, you know, how do I keep the books
for a tour guide company? Here's how to

1054
00:56:24,720 --> 00:56:27,517
do it. And the next thing you know she's
on the other side of that and she's in a

1055
00:56:27,600 --> 00:56:31,277
completely different business and people
are you know the Whimo car is delivering

1056
00:56:31,360 --> 00:56:34,157
the tour participants to her are
delighted to see her because they want a

1057
00:56:34,240 --> 00:56:37,517
person to take them on the tour.
So so so that so that's the that's the

1058
00:56:37,600 --> 00:56:39,917
creative side of it. Historically the

1059
00:56:40,000 --> 00:56:44,717
creative um uh sort of generation of new
ideas um generation of new wants and

1060
00:56:44,800 --> 00:56:47,197
needs and then new businesses and
professions that fulfill those wants and

1061
00:56:47,280 --> 00:56:51,357
needs has raced way ahead right of the
replacement phenomenon. you know, after

1062
00:56:51,440 --> 00:56:54,877
300 years of mechanization in computers,
there are more jobs in the world today

1063
00:56:54,960 --> 00:56:58,397
and at higher incomes than ever before
in human history. Um, I I think that's

1064
00:56:58,480 --> 00:57:01,517
exactly what's going to happen here.
And I and I think basically people who I

1065
00:57:01,600 --> 00:57:04,077
understand the concern, but I think
it's just I think it's fundamentally a

1066
00:57:04,160 --> 00:57:06,957
failure of imagination and I think that
the human spirit is going to is going to

1067
00:57:07,040 --> 00:57:11,117
process this just fine.
I I I I think that where you and I

1068
00:57:11,200 --> 00:57:12,797
disagree or maybe I'm not understanding

1069
00:57:12,880 --> 00:57:16,397
correctly is I think a huge sign segment

1070
00:57:16,480 --> 00:57:19,997
of the population, let's say a third,
I'm being conservative in my opinion,

1071
00:57:20,080 --> 00:57:24,957
are not capable of self-direction,
right? So like if someone is that woman

1072
00:57:25,040 --> 00:57:28,077
who's the tour guide, that's an easy
one cuz that's someone who's like, "Okay,

1073
00:57:28,160 --> 00:57:31,437
what should I be doing? Okay, I'll learn
these skills. It'll take me a day. I'll

1074
00:57:31,520 --> 00:57:34,957
I'll, you know, do my reading and I'm
personable like and they're going to

1075
00:57:35,040 --> 00:57:36,797
clean up." That that's an easy one. But

1076
00:57:36,880 --> 00:57:42,477
I think there's plenty of people who are
basically just want uh Maninet said the

1077
00:57:42,560 --> 00:57:45,357
average man does not want to be free.
He merely wants to be safe. There are

1078
00:57:45,440 --> 00:57:50,317
people who just are wired that they
want to be told what to do. And and at a

1079
00:57:50,400 --> 00:57:55,277
certain point, I don't see what value
they're adding to any company or or

1080
00:57:55,360 --> 00:57:58,237
anyone else. And what do you do with
those people? Just put them on welfare.

1081
00:57:58,320 --> 00:58:02,557
>> Well, I mean, this this gets this gets
an area of social policy and, you know,

1082
00:58:02,640 --> 00:58:05,117
political theory that maybe I'll I may

1083
00:58:05,200 --> 00:58:10,397
stay away from. um at least at least
live on video on the internet. Um okay,

1084
00:58:10,480 --> 00:58:14,637
sure. >> But um I guess I would say this,
like one of the things you can talk to AI

1085
00:58:14,720 --> 00:58:17,117
about is is this problem. Like one
of the things you can do is you say, "Wow,

1086
00:58:17,200 --> 00:58:17,997
like I don't know what I'm going to do.

1087
00:58:18,080 --> 00:58:20,477
Like here's what here's what I've done
for my entire life and wow, I don't

1088
00:58:20,560 --> 00:58:23,197
know. It seems like my job's going to get
like replaced." It's like okay, what

1089
00:58:23,280 --> 00:58:27,517
should I do? The thing will give you
career advice, right? Like it will it

1090
00:58:27,600 --> 00:58:30,637
will tell you like if you want it to it
will tell you what to do, right? And I I

1091
00:58:30,720 --> 00:58:33,037
don't I don't think people should just
like ask it what to do and just do what

1092
00:58:33,120 --> 00:58:35,677
it says, but you can say, "Okay,
brainstorm with me." I use it for

1093
00:58:35,760 --> 00:58:37,917
brainstorming all this is an area you
can use for brainstorming and say all

1094
00:58:38,000 --> 00:58:40,077
right look like here's what's happening
okay well it's going to say well where

1095
00:58:40,160 --> 00:58:42,797
do you live what's going on here are
the different areas what are the what are

1096
00:58:42,880 --> 00:58:45,837
the trends what are the new things that
are happening um you know I don't know I

1097
00:58:45,920 --> 00:58:48,957
you know I you know every time I go
by my local gym I see there's like six new

1098
00:58:49,040 --> 00:58:51,277
kinds of exercise classes I never
thought of could I become a you know

1099
00:58:51,360 --> 00:58:54,317
could I become a personal trainer you
know okay what's the what's the hot new

1100
00:58:54,400 --> 00:58:56,797
trend how do I get you know certified
in that and then the next thing you know

1101
00:58:56,880 --> 00:58:59,517
you're doing that um and and and and the

1102
00:58:59,600 --> 00:59:03,917
thing will just like happily coach it'll
it'll it'll so one of the one of the

1103
00:59:04,000 --> 00:59:08,157
amazing things about This I think this
is really underrated. These things the

1104
00:59:08,240 --> 00:59:11,517
AI models it's like the best doctor
you've ever had in your entire life.

1105
00:59:11,600 --> 00:59:16,477
Like it's amazing being a doctor.
Like it will hold your hand through any

1106
00:59:16,560 --> 00:59:20,957
medical situation you're in with a
degree of like caring that people are

1107
00:59:21,040 --> 00:59:24,397
not like the beyond what a human doctor
not only what most human doctors will

1108
00:59:24,480 --> 00:59:27,437
just do naturally, but no human doctor
has time to do it in the way people

1109
00:59:27,520 --> 00:59:30,957
really need. It's also the best lawyer
you've ever had, right? It's also the

1110
00:59:31,040 --> 00:59:34,397
best like coach you've ever had. Um,
it's also the best ghostwriter you've

1111
00:59:34,480 --> 00:59:37,837
ever had. It's the best editor you've
ever had. Um, and it's the best adviser

1112
00:59:37,920 --> 00:59:42,237
you've ever had, right? Um, and so if
you if you give it that opportunity and

1113
00:59:42,320 --> 00:59:44,877
and again, you don't have to answer you
can actually ask the questions. It will

1114
00:59:44,960 --> 00:59:47,757
answer the questions. Yeah. What what
should I do? How should I think about

1115
00:59:47,840 --> 00:59:50,237
the evolution of my career? How should
I think about the change in the economy?

1116
00:59:50,320 --> 00:59:53,197
It will happily do that with you and
and in a way that people literally have

1117
00:59:53,280 --> 00:59:56,717
never been able to have that
conversation with people before.

1118
00:59:56,800 --> 01:00:00,717
you know, you really kind of uh uh solve
that question in my mind because I have

1119
01:00:00,800 --> 01:00:04,957
a good friend uh and he's made a lot
of money in crypto, but he's been sitting

1120
01:00:05,040 --> 01:00:08,157
at home not having to work and it's been
driving crazy. So, he's like, "I need to

1121
01:00:08,240 --> 01:00:09,357
have a job. I don't need to make money.

1122
01:00:09,440 --> 01:00:12,877
I just need I'm like make candles
for guys, scented candles for men." Like,

1123
01:00:12,960 --> 01:00:16,637
that's a like it's you can go down that
rabbit hole. Work at the sense blah blah

1124
01:00:16,720 --> 01:00:20,237
blah. It'll occupy your brain. You'll
create a product. Even if you make $50,

1125
01:00:20,320 --> 01:00:22,877
who cares? And what I'm realizing is if

1126
01:00:22,960 --> 01:00:25,517
you ask AI, it will have that ven

1127
01:00:25,600 --> 01:00:28,717
diagram of no one has made socks for

1128
01:00:28,800 --> 01:00:32,797
14-year-old immigrants or, you know,
it'll see those where there's holes in

1129
01:00:32,880 --> 01:00:37,357
the market and then it could walk you
through it. So, it really does except

1130
01:00:37,440 --> 01:00:41,117
for the people who are completely bring
nothing to the table, which that's fine.

1131
01:00:41,200 --> 01:00:42,717
That's a whole separate conversation.

1132
01:00:42,800 --> 01:00:46,397
But that that woman with the with
the you just really solved my question

1133
01:00:46,480 --> 01:00:51,277
because she can be like, "Okay, what
value can I bring? What could I produce?

1134
01:00:51,360 --> 01:00:55,357
Make cookies. Who? There's so many
things that that people will always want

1135
01:00:55,440 --> 01:00:57,917
that personal touch. And also there's so

1136
01:00:58,000 --> 01:01:02,477
many little the markets will get more
and more niche and AI will be perfectly

1137
01:01:02,560 --> 01:01:06,797
able to find. No one is talking
to people who like saltwater aquariums but

1138
01:01:06,880 --> 01:01:11,037
also like heavy metal music. So start
a heavy metal saltwater aquarium website.

1139
01:01:11,120 --> 01:01:13,197
That that there you go. Holy crap.
You just answered my question.

1140
01:01:13,280 --> 01:01:17,437
>> I would love a heavy
metal aquarium like

1141
01:01:17,520 --> 01:01:22,557
it's like angry little fish. Like amazing.
Like why not? 100%. Yeah. No,

1142
01:01:22,640 --> 01:01:24,957
totally. Well, here's here's maybe
another way to think about it. Oh, I

1143
01:01:25,040 --> 01:01:26,717
think that's right. Here's another way
to think about it, which is, you know,

1144
01:01:26,800 --> 01:01:28,717
people think about this is like,
well, if the machine does something is

1145
01:01:28,800 --> 01:01:31,437
dehumanizing. >> I don't know
about you, but like office

1146
01:01:31,520 --> 01:01:35,757
jobs are dehumanizing. Like, like
sitting in a cubicle for eight hours a

1147
01:01:35,840 --> 01:01:38,637
day like working on the same thing over
and over again. And then, by the way,

1148
01:01:38,720 --> 01:01:41,597
factory jobs are dehumanizing. And by
the way, like I worked, you know, I grew

1149
01:01:41,680 --> 01:01:43,357
up in agriculture country. I worked on

1150
01:01:43,440 --> 01:01:48,717
farms. Like, you know, farming is not
romantic. You know, people sometimes

1151
01:01:48,800 --> 01:01:51,437
city people sometimes talk about
the romance of going off and doing like

1152
01:01:51,520 --> 01:01:54,557
whatever boutique, you know, farming
and it's like, you know, how do how do you

1153
01:01:54,640 --> 01:01:56,717
feel like getting up at 6:00 in the
morning, going to bed at midnight,

1154
01:01:56,800 --> 01:01:59,357
working like seven days a week
for the rest of your life?

1155
01:01:59,440 --> 01:02:02,157
>> It's >> Yeah.
Like it's like Yeah, exactly. Like

1156
01:02:02,240 --> 01:02:05,597
you're fighting back like chaos, you know,
the the entire time. And so like I

1157
01:02:05,680 --> 01:02:10,477
I think just so much of like what it's
almost like we have collective PTSD like

1158
01:02:10,560 --> 01:02:14,477
like we all have the we've all had
to live these lives with a with a level of

1159
01:02:14,560 --> 01:02:17,517
like if we look at our ancestors and the
lives that lived 300 years ago we just

1160
01:02:17,600 --> 01:02:21,277
are like that was a level of drudgery
and poverty and limited options that

1161
01:02:21,360 --> 01:02:24,717
would drive us crazy and make us want to
kill ourselves. Our ancestors 30 years I

1162
01:02:24,800 --> 01:02:26,717
mean 300 years from now even 30 years
from now are going to look back at us

1163
01:02:26,800 --> 01:02:29,437
being like I cannot believe they did that.
I cannot believe they spent time

1164
01:02:29,520 --> 01:02:32,797
doing those things. Like that was such
a waste of human potential. That was such

1165
01:02:32,880 --> 01:02:34,717
a waste of human creativity.

1166
01:02:34,800 --> 01:02:40,077
And then maybe one more one more thing
that kind of add on that is I I think

1167
01:02:40,160 --> 01:02:42,877
that this goes to the thing of like does
technology like this make us less human

1168
01:02:42,960 --> 01:02:45,357
or more human? And I and I think like
the examples we've been using I think

1169
01:02:45,440 --> 01:02:47,757
you start to see why why that's going
to result in people being more human which

1170
01:02:47,840 --> 01:02:50,957
is if if the physical need if the
physical needs needs of day-to-day life

1171
01:02:51,040 --> 01:02:53,277
are more easily satisfied then people
can spend more time actually being

1172
01:02:53,360 --> 01:02:57,357
human. Um and then they can spend more
time actually and money spending time on

1173
01:02:57,440 --> 01:03:00,317
things that actually are human
experiences. Um, by the way, another

1174
01:03:00,400 --> 01:03:03,277
just incredible version of this is
already playing out that's going to exp

1175
01:03:03,360 --> 01:03:05,917
>> Wait, I got to interrupt you just to
validate your point of view because back

1176
01:03:06,000 --> 01:03:09,917
in the day, you know, I'm reading a lot
like the early socialist 1890s, 1910s,

1177
01:03:10,000 --> 01:03:11,277
people working 16 hours a day.

1178
01:03:11,360 --> 01:03:14,637
>> That's right. >> And and like
even just down to 10, which

1179
01:03:14,720 --> 01:03:18,157
is still a lot. Think how much people
then you're complaining people watching

1180
01:03:18,240 --> 01:03:20,077
too much TV because they
have too much spare time.

1181
01:03:20,160 --> 01:03:21,037
>> That's right. That's right. Right.

1182
01:03:21,120 --> 01:03:23,757
Right. All the diseases of scarcity
become diseases of abundance. And you

1183
01:03:23,840 --> 01:03:27,277
know, diseases of abundance are still
diseases, but they're they're better

1184
01:03:27,360 --> 01:03:30,317
like obesity. Obesity is better than
starvation. And then by the way you you

1185
01:03:30,400 --> 01:03:34,157
know then you then then you figure out
how to solve obesity obesity later on.

1186
01:03:34,240 --> 01:03:37,197
>> Yeah. I mean look I'll just give
you a micro so one example of this was

1187
01:03:37,280 --> 01:03:41,117
happening in music right um and so um
you know record once upon a time all

1188
01:03:41,200 --> 01:03:43,757
music was in person. Like the only time
you'd ever hear music was like if you

1189
01:03:43,840 --> 01:03:46,797
happen to stumb stumble into a church or
something and hear it for the first time.

1190
01:03:46,880 --> 01:03:49,117
>> Uh and and by the way maybe the only
time in your life you know there were

1191
01:03:49,200 --> 01:03:53,117
people who maybe heard music once you
know in their entire life right. Um uh

1192
01:03:53,200 --> 01:03:56,797
and so and then and then by the way
but by by the way then uh sheet music

1193
01:03:56,880 --> 01:03:59,837
appeared. Uh actually there was a whole
moral panic about sheet music uh when it

1194
01:03:59,920 --> 01:04:02,397
first appeared u because it was going
to put all the all the pianists out of

1195
01:04:02,480 --> 01:04:05,917
business and they they all got extremely
upset. Um but then that you know led to

1196
01:04:06,000 --> 01:04:08,717
ultimately recorded music and then
obviously digital music can stream your

1197
01:04:08,800 --> 01:04:11,117
music and now no if you talk to any
musician it's like can you make money

1198
01:04:11,200 --> 01:04:14,797
you know with recorded music it's like
no you know not really anymore you know

1199
01:04:14,880 --> 01:04:17,597
you can get your music distributed
on Spotify but you you know and you hear

1200
01:04:17,680 --> 01:04:20,637
the endless complaints about you know
you get back pennies or something and

1201
01:04:20,720 --> 01:04:23,677
and so the recorded music business is not
what it used to be. Of course what's

1202
01:04:23,760 --> 01:04:25,757
exploded is live live performance.

1203
01:04:25,840 --> 01:04:28,957
Right. Right. And so and so so live
music is like exploding through the roof

1204
01:04:29,040 --> 01:04:31,437
and and now you see and now
the complaint right is the tickets are too

1205
01:04:31,520 --> 01:04:35,517
expensive you know to go to to go to the
concerts. It's like well okay but if you

1206
01:04:35,600 --> 01:04:37,837
think about it for a second it's like
why are we still listening to live

1207
01:04:37,920 --> 01:04:41,677
music? We all have every piece of music
ever written available on demand in high

1208
01:04:41,760 --> 01:04:44,557
fidelity in our homes in our in
our on our earphones anytime we want

1209
01:04:44,640 --> 01:04:48,317
essentially for free. Why is anybody
going to a concert? And of course the

1210
01:04:48,400 --> 01:04:52,557
answer is because the concert is a human
experience right? And so of course when

1211
01:04:52,640 --> 01:04:55,437
we get discretionary money we want to go
have the human experience. We want to go

1212
01:04:55,520 --> 01:04:57,677
have the concert. By the way, if we're
going to throw a party and we want it to

1213
01:04:57,760 --> 01:05:00,157
be a very special party, do we play
music through speakers or do we hire

1214
01:05:00,240 --> 01:05:03,517
musicians? We hire musicians.
>> And and also going to the concert, you

1215
01:05:03,600 --> 01:05:05,757
go with someone,
you create a bond, which is

1216
01:05:05,840 --> 01:05:09,837
>> 100% 100%. And so there's another way
to think about it is every profession that

1217
01:05:09,920 --> 01:05:14,477
involves humanto human contact is going
to go bananas. Right. Right. And and and

1218
01:05:14,560 --> 01:05:17,117
I and I think that those by the way,
that's going to be great for all the

1219
01:05:17,200 --> 01:05:19,437
people getting experienced that.
And then I think those jobs fundamentally

1220
01:05:19,520 --> 01:05:23,437
are better jobs like just at at a very
at at a very at a very core level. and

1221
01:05:23,520 --> 01:05:26,797
they are going to go.
It's going to be a bonanza.

1222
01:05:26,880 --> 01:05:31,117
>> For the first time in this show's long
and sorted history, you have completely

1223
01:05:31,200 --> 01:05:34,477
I'm not kidding at all. You have
completely answered all my concerns

1224
01:05:34,560 --> 01:05:37,037
about the topic and I'm I totally get it

1225
01:05:37,120 --> 01:05:42,877
now. I I thank you. I I'm I feel so
excited and I I'm sure and since you're

1226
01:05:42,960 --> 01:05:45,117
much more a visionary about this stuff
than I am because you've been there, you

1227
01:05:45,200 --> 01:05:48,877
know, from the beginning, I can't even
imagine how exciting must be for you.

1228
01:05:48,960 --> 01:05:52,397
Things that were people were kind
of hypothesizing about even 5 years ago.

1229
01:05:52,480 --> 01:05:55,117
Now you're using them on a day-to-day
basis. It just must be you must be

1230
01:05:55,200 --> 01:05:58,797
absolutely giddy. Uh so the book
is the technoptist manifesto. Uh you got a

1231
01:05:58,880 --> 01:06:02,237
passage press. Mark, thank you so much
for taking the time. We're running out

1232
01:06:02,320 --> 01:06:06,557
of time. What has been your
favorite part of this interview?

1233
01:06:06,640 --> 01:06:08,877
>> The amazingly amazingly probing

1234
01:06:08,960 --> 01:06:14,200
questions. >> You are welcome.

1235
01:06:24,320 --> 01:06:27,197
What if the stories of gods, angels, and

1236
01:06:27,280 --> 01:06:29,837
ancient technology weren't myths, but

1237
01:06:29,920 --> 01:06:32,957
memories of alien contact? On ancient

1238
01:06:33,040 --> 01:06:35,517
aliens, we investigate the evidence.

1239
01:06:35,600 --> 01:06:37,997
From mysterious ruins to strange

1240
01:06:38,080 --> 01:06:42,877
celestial alignments, asking bold
questions about humanity's origins.

1241
01:06:42,960 --> 01:06:47,517
Could visitors from the stars have
influenced our greatest civilizations?

1242
01:06:47,600 --> 01:06:49,357
Journey into the unknown and challenge

1243
01:06:49,440 --> 01:06:51,757
what you think you know about history in

1244
01:06:51,840 --> 01:06:54,717
episodes like Secrets of the Sumerians,

1245
01:06:54,800 --> 01:06:58,637
Mysteries of the Mayo, or The Chosen. We

1246
01:06:58,720 --> 01:07:01,917
are not alone. Download Ancient Aliens

1247
01:07:02,000 --> 01:07:07,000
every Thursday, wherever
you get your podcasts.

1248
01:07:08,080 --> 01:07:10,797
>> Hello and welcome to Pluto. If you know

1249
01:07:10,880 --> 01:07:15,117
the name of the movie you'd like to see,
just stream it for free on Pluto TV,

1250
01:07:15,200 --> 01:07:18,957
where all your blockbuster favorites
are landing all summer long. Catch Anchor

1251
01:07:19,040 --> 01:07:20,717
Man, The Legend of Ron Burgundy.

1252
01:07:20,800 --> 01:07:23,277
>> Fantastic.
>> Men and Black one through three.

1253
01:07:23,360 --> 01:07:24,637
>> That's what I'm talking about.

1254
01:07:24,720 --> 01:07:27,117
>> Mean Girls, >> shut up. >> Titanic.

1255
01:07:27,200 --> 01:07:31,357
>> I'm the king of the world.
>> And so much more. For Showtimes, press

1256
01:07:31,440 --> 01:07:33,597
nothing. They're free 24/7.

1257
01:07:33,680 --> 01:07:39,960
>> That is so effect.
>> On Pluto TV stream now, pay never.