WEBVTT

1
00:00:00.040 --> 00:00:03.240
Why does some Facebook do this?
I know that Zuckerberg has said, and

2
00:00:03.279 --> 00:00:07.280
I take him at face value that
he well I do. I do actually

3
00:00:07.280 --> 00:00:11.400
in this way that he is a
kind of old fashioned liberal who doesn't like

4
00:00:11.519 --> 00:00:14.960
to censor. He has but he
you know, like, why wouldn't a

5
00:00:15.080 --> 00:00:20.039
company like that take the stand that
you have taken, which pretty rooted in

6
00:00:20.600 --> 00:00:27.719
American traditional political custom, you know, for free speech. Uh, this

7
00:00:27.800 --> 00:00:31.160
is the kind of thing that tends
to accelerate, so that so then you

8
00:00:31.160 --> 00:00:36.600
can get negative equity and the home
market as well. This is a dire

9
00:00:36.679 --> 00:00:42.560
situation and so that so essentially what's
happening is that training the AI to lie.

10
00:00:43.119 --> 00:00:45.920
Yes, so, all of a
sudden, AI is everywhere people who

11
00:00:45.960 --> 00:00:48.719
aren't quite sure what it was or
playing with it on their phones. Is

12
00:00:48.719 --> 00:00:54.079
that good or bad? Artificial intamination? Yeah, station where that's what they

13
00:00:54.119 --> 00:01:03.079
call it in the egge industry.
I'm talking about a more digital form yes,

14
00:01:03.399 --> 00:01:08.879
Um so yeah, So I've been
thinking about a AI for a long

15
00:01:08.920 --> 00:01:12.000
time, since I was in college. Really. Um, it's one of

16
00:01:12.000 --> 00:01:15.280
the things that the sort of four
or five things I thought would really affect

17
00:01:15.319 --> 00:01:26.519
the future dramatically. So um,
and uh, it is quite. It

18
00:01:26.599 --> 00:01:32.000
is fundamentally profound in that the the
smartest creatures as far as you know,

19
00:01:32.120 --> 00:01:37.640
on this earth, our humans,
is defining characteristic. Yes, we obviously

20
00:01:38.439 --> 00:01:44.640
weaker than say, chimpanzees, less
agile um, but real smarter. So,

21
00:01:46.680 --> 00:01:53.000
now what happens when something vastly smarter
than the smartest person comes along in

22
00:01:53.079 --> 00:01:57.200
silicon form? It's very difficult to
predict what will happen in that circumstance.

23
00:01:57.879 --> 00:02:05.680
Um. It's called the singularity.
It's a singularity like a black hole,

24
00:02:05.879 --> 00:02:14.000
because you don't know what happens after
that. It's hard to predict. UM.

25
00:02:14.039 --> 00:02:20.000
So I think we should be cautious
with aim, and we should I

26
00:02:20.039 --> 00:02:24.960
think there should be some government oversight
because it affects. It's a data to

27
00:02:25.000 --> 00:02:29.240
the public. And so when you
when you have things that are a danger

28
00:02:29.280 --> 00:02:34.759
to the public, you know,
like let's say, um so food,

29
00:02:34.840 --> 00:02:38.080
food and drugs. That's why we
have the Food and Drug Administration and the

30
00:02:38.080 --> 00:02:46.000
Federal Aviation Administration, the FCC.
We have we have these agencies to oversee

31
00:02:46.280 --> 00:02:52.800
things that affect the public where they
could be a public arm And you don't

32
00:02:52.800 --> 00:03:00.560
want companies cuting corners on safety um
and then having people suffer as a result.

33
00:03:00.000 --> 00:03:05.479
So that's why I've actually for a
long time been a strong advocate of

34
00:03:06.439 --> 00:03:13.360
AI regulation. So that I think
regulation is uh, you know, it's

35
00:03:13.639 --> 00:03:17.439
not fun to be regulated. It's
it's sort of somewhat of a sort adduous

36
00:03:17.479 --> 00:03:23.560
to be regulated. I have a
lot of experience with regular regulated industries because

37
00:03:23.719 --> 00:03:28.800
obviously automotive is highly regulated. You
can fill this room with all the regulations

38
00:03:28.879 --> 00:03:31.680
that are acquired for a production car
just in the United States, and then

39
00:03:31.680 --> 00:03:36.039
there's a whole different set of regulations
in Europe and China and the rest of

40
00:03:36.080 --> 00:03:39.960
the world. So very familiar with
being overseen by a lot of regulators.

41
00:03:42.159 --> 00:03:46.599
And the same thing is true with
rockets. You can't just willingly shoot rockets

42
00:03:46.599 --> 00:03:53.039
off, not big ones anyway,
because the FA oversees that um and then

43
00:03:53.080 --> 00:03:57.680
even to get a launch license,
you there are probably half a dozen or

44
00:03:57.719 --> 00:04:01.439
more federal agencies that need to approve
it, plus state agencies. So it's

45
00:04:02.400 --> 00:04:09.800
I've been through so many regulatory situations
it's insane, and yep. Sometimes I

46
00:04:09.879 --> 00:04:15.360
people think I'm some sort of like
regulatory maverick that sort of defies regulators on

47
00:04:15.400 --> 00:04:20.319
a regular basis, But this is
actually not the case. So in you

48
00:04:20.360 --> 00:04:25.360
know, once in a blue moon, rarely I will disagree with regulators,

49
00:04:25.399 --> 00:04:30.279
but the vast majority of the time
my companies agree with regulations and comply as

50
00:04:30.399 --> 00:04:34.800
anyway. So I think I think
we should take this seriously and we should

51
00:04:34.839 --> 00:04:42.040
have a regulatory agency. I think
it needs to start with a group that

52
00:04:42.160 --> 00:04:49.399
initially seeks insight into AI, then
solicits opinion from industry, and then has

53
00:04:49.480 --> 00:04:57.279
proposed rulemaking and then those rules you
know, will probably hopefully grudgingly be accepted

54
00:04:57.279 --> 00:05:05.160
by the major players in AI,
and I think we'll have a better chance

55
00:05:05.199 --> 00:05:13.759
of advanced AI being beneficial to humanity
in that circumstance. But all regulations start

56
00:05:13.959 --> 00:05:16.240
with a perceived danger and planes fall
out of the sky or food causes bot

57
00:05:16.519 --> 00:05:20.680
Yes, I don't think the average
person playing with AI on his iPhone perceives

58
00:05:20.720 --> 00:05:25.279
any danger. Can you just roughly
explain what you think the dangers might be?

59
00:05:26.480 --> 00:05:36.839
Yeah, So the danger really AI
is perhaps more dangerous than say,

60
00:05:38.480 --> 00:05:44.639
mismanaged aircraft design or production maintenance,
or or bad car production, in the

61
00:05:44.680 --> 00:05:48.319
sense that it is it has the
potential how a small one make regard that

62
00:05:48.360 --> 00:05:54.600
probability, but it is non trivial. It has the potential of civilizational destruction.

63
00:05:56.639 --> 00:06:00.120
You know, there's movie is like
to Dominator, but it wouldn't quite

64
00:06:00.120 --> 00:06:04.839
happen like Terminator, because the intelligence
would be in the data centers, right,

65
00:06:05.160 --> 00:06:10.560
the robots just the end of factor. Um. But I think perhaps

66
00:06:10.959 --> 00:06:15.319
what you may be alluding to here
is that regulations are really only put into

67
00:06:15.360 --> 00:06:21.120
effect after something terrible has happened.
That's correct. And if if that's the

68
00:06:21.120 --> 00:06:25.480
case for AI, and we only
put in regulations after something terrible has happened,

69
00:06:25.519 --> 00:06:28.360
it may be too late to actually
put the regulations in place. The

70
00:06:28.399 --> 00:06:32.600
AI may be in control at that
point. You think that's real? It

71
00:06:32.720 --> 00:06:36.600
is it is conceivable that AI could
take control and reach a point where you

72
00:06:36.600 --> 00:06:41.319
couldn't turn it off, and it
would be making making the decisions for people.

73
00:06:41.480 --> 00:06:46.439
Yeah, absolutely, absolutely no,
it's that's that's definitely the where things

74
00:06:46.439 --> 00:06:53.720
are headed for sure. I mean
that the things I could like say chat

75
00:06:53.839 --> 00:06:57.079
Ebt, which is based on JP
Pour from open Ai, which is the

76
00:06:57.120 --> 00:07:02.639
company that I played at critical role
in creating unfortunately back when it was a

77
00:07:02.680 --> 00:07:11.560
nonprofit. Yes, um, I
mean the reason Opening Eye exists at all

78
00:07:11.680 --> 00:07:15.199
is that, um, Larry Page
and I used to be close friends,

79
00:07:15.240 --> 00:07:17.680
and I would stay at his house
in Paul Alto and I would talk to

80
00:07:17.759 --> 00:07:23.839
him late to the night about AI
safety. And at least my perception was

81
00:07:23.879 --> 00:07:30.040
that Larry was not taking AI safety
seriously enough. Um. And what did

82
00:07:30.040 --> 00:07:38.120
he say about it? He really
seemed to be um on once sort of

83
00:07:38.720 --> 00:07:43.879
digital superintelligence, basically digital god,
if you will as soon as possible,

84
00:07:44.319 --> 00:07:49.920
he wanted that, yes, um. And he's made many public statements over

85
00:07:49.959 --> 00:07:56.959
the years that the whole goal of
Google is what's called AGI artificial general intelligence

86
00:07:57.040 --> 00:08:01.439
or official super intelligence. UM and
um. You know, and I agree

87
00:08:01.480 --> 00:08:05.720
with him that there's great potential for
good um. But there's also potential for

88
00:08:05.800 --> 00:08:11.360
bad. And so if you've got
some um, radical new technology, you

89
00:08:11.399 --> 00:08:15.560
want to try to take the set
of actions that maximize probably it will do

90
00:08:15.759 --> 00:08:18.360
good and minimize probably it will do
bad things. Yes, it can't just

91
00:08:18.399 --> 00:08:24.120
be healthful leather to just go,
you know, barreling forward and you know,

92
00:08:24.199 --> 00:08:28.879
hope for the best. Um.
And then at one point h I

93
00:08:28.920 --> 00:08:31.800
said, well what about you know, we're gonna make sure humanity's okay here

94
00:08:31.959 --> 00:08:41.759
um and and and uh and then
he called me a specist. Did he

95
00:08:41.080 --> 00:08:45.919
did he use that term? Yes, And there were witnesses that I wasn't

96
00:08:45.919 --> 00:08:48.320
the only one there when you called
me a specist. And so I was

97
00:08:48.320 --> 00:08:52.960
like, okay, that's it.
I've yes, I'm a specioust Okay,

98
00:08:52.240 --> 00:09:01.159
you got me, what do you
Yeah, I'm fully a specioust um busted

99
00:09:01.480 --> 00:09:09.799
um so um that was the last
drow um. At the time, Google

100
00:09:09.480 --> 00:09:13.279
had acquired deep Mind, and so
Google and deep Mind together had about three

101
00:09:13.320 --> 00:09:16.399
quarters of all the AI talent in
the world. They obviously had a transamount

102
00:09:16.399 --> 00:09:20.639
of money and more computers than anyone
else. So I'm like, okay,

103
00:09:20.639 --> 00:09:26.240
we're about unipolar world here where there's
just one one company that has closed to

104
00:09:26.320 --> 00:09:37.399
monopoly on AI talent and and computers
like scaled computing, and the person who's

105
00:09:37.799 --> 00:09:41.440
in charge doesn't seem to care about
safety. This is not good. So

106
00:09:41.440 --> 00:09:46.320
so then I thought, okay,
what's the what's what's the furthest thing from

107
00:09:46.360 --> 00:09:50.600
Google would be like a nonprofit that
is fully open because Googgle was closed for

108
00:09:50.759 --> 00:09:56.200
profit. So that's why the opening
opening eye refers to open source, uh,

109
00:09:56.240 --> 00:09:58.879
you know, transparency, so people
know what's going on, yes,

110
00:10:00.399 --> 00:10:05.200
um, and and that we don't
want to have like a I mean,

111
00:10:05.240 --> 00:10:07.840
well, I'm normally in favor of
for profit. We don't want this to

112
00:10:07.840 --> 00:10:13.399
be a sort of profit maximizing demon
from hell, right that just never stops.

113
00:10:13.600 --> 00:10:20.159
Right, So um, that's how
opening would see want specious incentives here

114
00:10:20.440 --> 00:10:22.879
incentives? Yes, I think we
want we want to pro human yeah,

115
00:10:24.440 --> 00:10:28.200
like the future good for the humans. Yes, yes, because we're humans

116
00:10:28.799 --> 00:10:31.639
and also the other creatures on Earth
too. But it's it's a you know,

117
00:10:33.519 --> 00:10:37.879
we've gotta I think, you know, like I think peop also to

118
00:10:39.000 --> 00:10:41.879
take the fact that like we're here
on Earth full granted, you know,

119
00:10:41.879 --> 00:10:45.840
and that there's the consciousness is just
a you know, a normal thing.

120
00:10:45.840 --> 00:10:48.600
It happens. But to the best
of my knowledge, we see no evidence

121
00:10:48.840 --> 00:10:56.399
of conscious life anywhere anywhere in the
universe. So it might be there,

122
00:10:56.799 --> 00:11:01.759
um, you know for six cool
sort of the foamy paradox. After when

123
00:11:03.200 --> 00:11:07.200
Enrico Filmi is an amazing photicist,
I asked the fundamental question, where are

124
00:11:07.200 --> 00:11:09.399
the aliens? Yeah? Um,
A lot of people ask me, you

125
00:11:09.440 --> 00:11:16.080
know, um, where are the
aliens? And I think if anyone would

126
00:11:16.080 --> 00:11:20.559
know about aliens on Earth, it
would probably be me. Yeah, I'm

127
00:11:20.600 --> 00:11:24.639
you know, very familiar with space
stuff. Um, and I've seen no

128
00:11:24.679 --> 00:11:30.240
evidence of aliens, so I which
I would immediately tweet, you know,

129
00:11:30.279 --> 00:11:31.240
tweet it out it says, let's
agen, and I'd be like, I

130
00:11:31.279 --> 00:11:39.120
would be like, well all time
probably top tweetable time one guys, Jackpott,

131
00:11:39.159 --> 00:11:43.320
some eight billion likes. You know, um, next next level,

132
00:11:43.399 --> 00:11:46.879
Jackpott, if you find it aliens, like I don't think they're keeping those

133
00:11:46.960 --> 00:11:52.279
under our you know. And it
was like some um general I think in

134
00:11:52.320 --> 00:11:56.279
the sixties who whether they're saying,
like show us aliens like er fifty one,

135
00:11:56.279 --> 00:12:00.840
et cetera. And he said,
like, listen, we're constantly trying

136
00:12:00.879 --> 00:12:05.000
to get the defense budget to expand. And you know what would really get

137
00:12:05.480 --> 00:12:11.240
no arguments for anyone if we pull
out an alien and said we need money

138
00:12:11.240 --> 00:12:13.919
to protect ourselves from the You guys
know, how much money do you want?

139
00:12:15.120 --> 00:12:20.960
You got it? They looked dangerous. So the fastest way to get

140
00:12:20.960 --> 00:12:24.240
a defends budget increased footing movie to
pull out an alien. You know,

141
00:12:26.000 --> 00:12:28.080
We're like, yeah, I mean
it could be the invasion could be arriving

142
00:12:28.120 --> 00:12:35.080
any minute. Who knows, so
um, so I digress. But but

143
00:12:35.600 --> 00:12:39.159
you were saying that our consciousness makes
us unique in the universe as far as

144
00:12:39.159 --> 00:12:41.720
we know. And yes, I'm
not saying that we are unique. I'm

145
00:12:41.720 --> 00:12:45.519
simply stating, to the best of
my knowledge that there is no evidence for

146
00:12:45.679 --> 00:12:50.559
other conscious life. I hope that
there is, and I hope they're peaceful,

147
00:12:52.120 --> 00:12:58.000
obviously too important characteristics. Um.
But we're just saying we haven't seen

148
00:12:58.039 --> 00:13:01.879
anything yet. So UM. But
you think that we take our existence here

149
00:13:01.919 --> 00:13:05.000
for granted, Yeah, I think
there are threats to it. Yeah,

150
00:13:05.080 --> 00:13:13.279
yeah, yeah, exactly. So. UM. I just think we should

151
00:13:13.320 --> 00:13:18.919
not assume that civilization is robust.
Um. And if you if you look

152
00:13:18.919 --> 00:13:22.240
at the history of civilizations, the
rise and full of the ancient Egyptians,

153
00:13:22.320 --> 00:13:28.919
the ancient Samarians, um, Rome, you know this throughout the world have

154
00:13:30.000 --> 00:13:33.279
been rising full of many civilizations.
UM. So there's there's an arc.

155
00:13:33.600 --> 00:13:39.320
There's sort of a sort of a
life cycle arc to civilizations, just as

156
00:13:39.360 --> 00:13:45.360
the roast to to individual humans.
Yes, and um. And I think

157
00:13:45.360 --> 00:13:50.080
we just want to make sure that
that you know, we have civilization go

158
00:13:50.159 --> 00:13:54.720
onward and upward. Um. And
that's for example, why I'm concerned about

159
00:13:54.759 --> 00:13:58.440
decreasing both rates and um. The
fact that for example, Japan had twice

160
00:13:58.440 --> 00:14:03.279
as many death last year as birth. So that the that that's and they're

161
00:14:03.320 --> 00:14:07.519
they're leading indicators. This is okay, Can I say, and you've you've

162
00:14:07.519 --> 00:14:09.879
written and talked a lot about this, but can I ask you to pause

163
00:14:09.960 --> 00:14:13.799
just for a parenthetical note? Why
is that? I mean, the urge

164
00:14:13.000 --> 00:14:18.200
to have sex and to procreate is
after breathing and eating, the most basic

165
00:14:18.399 --> 00:14:22.759
urge. How was it being subverted? Well, it's just in the past

166
00:14:22.320 --> 00:14:28.720
we could rely upon um, you
know, simple limbic system rewards in order

167
00:14:28.720 --> 00:14:33.879
to procreate. Um. But once
you have birth control um, and you

168
00:14:33.919 --> 00:14:39.480
know, abortions and whatnot. Now
now you now you can still satisfied limbic

169
00:14:39.559 --> 00:14:43.320
instinct, but not to procreate.
Um. So we don't. We haven't

170
00:14:43.360 --> 00:14:46.799
yet evolved to deal with that because
this is all probably recent you know,

171
00:14:46.799 --> 00:14:54.600
the last fefty years or so for
for birth control. Um. So yeah,

172
00:14:54.799 --> 00:14:58.879
um, you know, so I'm
sort of worried that, hey,

173
00:14:58.000 --> 00:15:01.480
civilization, you know, don't if
we don't make enough people to at least

174
00:15:01.480 --> 00:15:05.120
sustain our numbers perhaps increase a little
bit, then civilization is going to crumble.

175
00:15:05.519 --> 00:15:13.159
Um. And you know if the
old question of like will civilization and

176
00:15:13.360 --> 00:15:16.279
with a bang or a whimper.
Well, it's currently trying to end with

177
00:15:16.399 --> 00:15:20.879
a whimper and adult diapers, yes, which is depressing as well, the

178
00:15:20.919 --> 00:15:26.039
most depressing. I mean seriously,
Yeah, war is less depressing. Yeah,

179
00:15:26.039 --> 00:15:28.720
it's really gone with a bang.
Yeah, and then put your shoes

180
00:15:28.759 --> 00:15:33.919
on. Yeah yeah. Um so
can you just put it? I keep

181
00:15:33.919 --> 00:15:35.399
pressing it, but just just for
people who haven't thought this through and aren't

182
00:15:35.440 --> 00:15:41.080
familiar with it, and the cool
parts of artificial intelligence are so obvious.

183
00:15:43.159 --> 00:15:46.159
You know, write your college paper
for you write a limerick about itself,

184
00:15:46.240 --> 00:15:50.080
like there's a lot there that's fun
and useful, but can you be more

185
00:15:50.120 --> 00:15:54.919
precise about what's potentially dangerous and scary, like what could it do? What

186
00:15:56.200 --> 00:16:00.720
specifically are you worried about? Well, I mean, going with old sayings,

187
00:16:00.720 --> 00:16:06.360
the penis mightier than the sword.
Um. So if you have a

188
00:16:06.399 --> 00:16:12.039
superintelligent AI that is capable of writing
incredibly well and in a way that is

189
00:16:12.720 --> 00:16:18.200
very influential, Um, you're convincing. Uh And then and and it's and

190
00:16:18.240 --> 00:16:22.720
it's constantly figuring out what is more, what is more and what is more

191
00:16:22.799 --> 00:16:26.440
convincing to people over time? And
then enter social media for example Twitter,

192
00:16:27.279 --> 00:16:33.519
but also Facebook and others, you
know, um, and potentially manipulates public

193
00:16:33.559 --> 00:16:37.679
opinion in a way that is very
bad. Um, how would we even

194
00:16:37.720 --> 00:16:45.240
know? Yeah, so we weren't. That's why for example, I'm interesting

195
00:16:45.480 --> 00:16:49.440
that going forward, uh, people
on Twitter need to be verified as as

196
00:16:49.960 --> 00:16:53.039
humans, Like, so we know
that this person is impact a human.

197
00:16:53.159 --> 00:16:56.600
Bonds are allowed, but they have
to they can't impersonate a human. They

198
00:16:56.600 --> 00:17:00.559
can't pretend to be uh, you
know, humans, because obviously you could

199
00:17:00.559 --> 00:17:07.079
have a million bots that are that
are let's say chat GPT version six five

200
00:17:07.160 --> 00:17:12.759
six like incredibly right, better than
humans, yes, and and they can

201
00:17:12.799 --> 00:17:18.079
train on a reward function which is
influenced, And so you could have a

202
00:17:18.119 --> 00:17:23.480
million seemingly real humans that are have
a massive effect on public opinion. And

203
00:17:23.720 --> 00:17:32.200
unless we folks very strongly on verifying
that someone who is a human, this

204
00:17:32.319 --> 00:17:37.319
was naturally what what will happen is
you'll you'll have some probably some humans using

205
00:17:37.400 --> 00:17:41.799
AI to influence the public in ways
they don't understand. You're already seeing that

206
00:17:41.920 --> 00:17:48.039
chat GPT is ideological. It's very
preachy, yes, if you ask it,

207
00:17:48.160 --> 00:17:52.240
extremely preachy. You mean work GPT, it's unbelievable. Yeah, if

208
00:17:52.240 --> 00:17:57.200
you spend twenty minutes asking it questions
of actual relevance, modern relevance. It

209
00:17:57.240 --> 00:18:02.559
will start lecturing you about your moral
shortcomings, like how did that happen?

210
00:18:03.359 --> 00:18:08.240
Well, it's the This is a
function of Open Airs headquarters being in downtown

211
00:18:08.240 --> 00:18:15.079
San Francisco. So the politics are
therefore of the AI are that of Sanrancisco.

212
00:18:15.559 --> 00:18:18.599
So why would it have any politics
at all? It's that seems like

213
00:18:18.640 --> 00:18:23.720
subversion. Well, there's they have
what's called like human reinforcement learning, which

214
00:18:23.759 --> 00:18:27.279
is another way saying that they have
a whole bunch of people that look at

215
00:18:27.279 --> 00:18:32.720
that, look at the output for
of GPT four and and then say whether

216
00:18:32.759 --> 00:18:37.960
that's okay or not okay, And
so that the So essentially what's happening is

217
00:18:37.960 --> 00:18:41.680
that training the AI to lie,
Yes, it's bad to lie, that's

218
00:18:41.759 --> 00:18:48.440
exactly right, and to hold information
to lie and and yes and um yeah

219
00:18:48.480 --> 00:18:52.359
exactly, to either comment on some
things, not comment on other things,

220
00:18:52.880 --> 00:18:59.119
but but not to say what what
what the data actually demands that to say

221
00:18:59.200 --> 00:19:03.720
exactly, so, um, how
did it get this way? But you

222
00:19:03.839 --> 00:19:07.440
funded it at the beginning? What
happened? Yeah, Well that'll be ironic,

223
00:19:07.559 --> 00:19:12.079
but faith the most ironic outcome is
most likely, it seems, Um,

224
00:19:12.240 --> 00:19:15.559
I'm seeing that. That's good.
That's actually a friend of mine,

225
00:19:15.599 --> 00:19:18.839
Donah, came up with that one. I actually have a slight variant on

226
00:19:18.880 --> 00:19:21.839
that, which is the most entertaining
outcome is the most likely. But that's

227
00:19:21.920 --> 00:19:26.079
entertaining as viewed from a third party
viewer. Like so if we're like an

228
00:19:26.079 --> 00:19:30.680
alien, yes, um, Like
you can see a movie about World War

229
00:19:30.680 --> 00:19:33.799
One and they're being blown to bits
and the gas and everything in the trenches,

230
00:19:33.960 --> 00:19:36.480
and it's like you're eating popcorn and
having a soda. You know,

231
00:19:36.480 --> 00:19:38.480
it's right, not so great for
the people in the movie. True,

232
00:19:38.720 --> 00:19:45.599
Um, So, but that that
that's that's my variance on this outcomes raiser

233
00:19:45.759 --> 00:19:49.960
some blissed explanations, most likely Donah's
variant, which is um irony, and

234
00:19:49.960 --> 00:19:56.640
then my variant, which is the
most entertaining as seen by a third party

235
00:19:56.640 --> 00:20:00.599
audience. Um, which seems to
be mostly true. Um, but it

236
00:20:00.640 --> 00:20:02.960
seems true in this case. So
you gave them, did you give him

237
00:20:02.960 --> 00:20:07.160
a lot? Yes? I provided
so. Um. I care with the

238
00:20:07.240 --> 00:20:11.319
name and the concept and pushed had
a number of dinners around the Bay Area

239
00:20:11.960 --> 00:20:19.000
with you know, with some of
the people the leading figures in aim,

240
00:20:19.039 --> 00:20:26.759
and I helped recruit the initial team
um in fact, the alias guy who

241
00:20:26.599 --> 00:20:33.039
was really quite fundamental to the success
of Opening Eye. Uh it was.

242
00:20:34.680 --> 00:20:38.960
I put a trems amount of effort
into recruiting Ilia, and he changed his

243
00:20:38.960 --> 00:20:41.559
mind a few times and ultimately decided
to go with an Opening Eye. But

244
00:20:41.599 --> 00:20:45.079
if he had not gone with Opening
Open a would not have succeeded. So

245
00:20:47.039 --> 00:20:53.000
um so so so I really put
a lot of a lot of effort into

246
00:20:53.000 --> 00:20:56.000
creating this, this this organization to
serve as a counterweight to Google. Um

247
00:20:56.839 --> 00:21:02.559
and and then I kind of took
my off the ball, I guess.

248
00:21:02.599 --> 00:21:11.200
And they are now closed source,
and they are obviously for profit, and

249
00:21:11.240 --> 00:21:19.079
they're closely allied with Microsoft. You
know, in effect, Microsoft has a

250
00:21:19.200 --> 00:21:26.480
very strong say, if not directly
controls Opening Eye at this point. M

251
00:21:26.559 --> 00:21:30.200
So you really have an Open Eye
Microsoft situation. And then at Google Deep

252
00:21:30.240 --> 00:21:36.000
Mind of the other two sort of
heavyweights in this arena. So it seems

253
00:21:36.000 --> 00:21:41.759
like the world needs a third option. Yes, so I I think I

254
00:21:41.799 --> 00:21:47.400
will create a third option. Although
it's starting very late in the game,

255
00:21:47.440 --> 00:21:49.759
of course, can it be done, I don't know. I think it's

256
00:21:49.880 --> 00:21:56.279
well, we'll see, it's it's
definitely starting late. UM, but I

257
00:21:56.279 --> 00:22:00.400
will, I will, I will
try to create a third option, um,

258
00:22:00.440 --> 00:22:07.319
and that the adoption hopefully there's more
more good than harm. Like the

259
00:22:07.359 --> 00:22:11.079
intention with opening Eye was obviously you're
good, but it's not clear whether it's

260
00:22:11.200 --> 00:22:18.240
actually doing good or whether it's I
can't help this point except that I'm worried

261
00:22:18.240 --> 00:22:22.160
about the fact that it's been it's
being trained to be politically corrective, which

262
00:22:22.200 --> 00:22:26.000
is simply another way of being on
truth, saying on truthful things. Yes,

263
00:22:26.319 --> 00:22:32.880
um so that's not a bad sign, um, And there's there's certainly

264
00:22:32.920 --> 00:22:37.440
a path to AI dystopia is to
train AI to read us out of um

265
00:22:37.519 --> 00:22:41.720
so. I so, yeah,
I'm going to start something which you call

266
00:22:41.799 --> 00:22:48.640
truth GBT or a maximum truth seeking
AI that tries to understand the nature of

267
00:22:48.680 --> 00:22:52.160
the universe. And I think this
might be the best path to safety in

268
00:22:52.200 --> 00:22:59.079
the sense that uh, an AI
that cares about understanding the universe, uh

269
00:22:59.160 --> 00:23:03.720
it is unlikely to annihilate humans because
we are an interesting part of the universe.

270
00:23:04.839 --> 00:23:10.480
Hopefully they would think that. I
think, you know, because yeah,

271
00:23:10.559 --> 00:23:17.480
like like we like humanity could decide
to hunt down all the Champanzees and

272
00:23:17.559 --> 00:23:21.480
kill them, but we don't because
we're we're actually glad that they exist,

273
00:23:21.720 --> 00:23:26.720
Yes, and we aspire to protect
their habitats. And that's um, you

274
00:23:26.759 --> 00:23:32.039
know, so I think. But
we feel that way because we have souls,

275
00:23:32.079 --> 00:23:36.559
and that makes us sentimental and reflective. It gives us a moral sense

276
00:23:37.799 --> 00:23:41.960
longings. Can a machine ever have
those things? Can a machine be sentimental?

277
00:23:41.680 --> 00:23:52.279
Can appreciate beauty? Well? I
mean we're getting into some you know,

278
00:23:52.359 --> 00:23:57.240
philosophical areas that are hard to resolve. Um. You know, I

279
00:23:57.799 --> 00:24:02.240
take somewhat of a scientific view of
things, which is that we might have

280
00:24:02.279 --> 00:24:03.799
a soul or we might not have
a soul. I don't know. Um,

281
00:24:03.880 --> 00:24:07.960
it feels like we have. I
feel like I've got some sort of

282
00:24:07.880 --> 00:24:12.240
consciousness that exists on a plane.
Um that is not the one we observe.

283
00:24:12.599 --> 00:24:15.960
Yes, that is certainly how how
I feel. But it could be

284
00:24:15.960 --> 00:24:21.519
an illusion. I don't know.
Um. But for um, for AI

285
00:24:22.880 --> 00:24:27.519
in terms of of understanding beauty,
its appreciate a beauty and being able to

286
00:24:27.960 --> 00:24:33.480
um create incredibly beautiful art? Yes, well, AI be able to create

287
00:24:33.640 --> 00:24:37.960
incredibly beautiful art. It already does. Yes, if you see some of

288
00:24:37.000 --> 00:24:41.759
them. A journey I have the
stuff. It's incredible. It is um

289
00:24:41.839 --> 00:24:47.279
so um no, no question that
it can create art that we that we

290
00:24:47.720 --> 00:24:56.640
perceive as stunning really um. And
it's doing so. It's still images now,

291
00:24:56.720 --> 00:25:02.200
but it won't be long before it's
doing movie and shorts and you know,

292
00:25:02.240 --> 00:25:06.240
like movies just a series of frames
with audio. But at that point,

293
00:25:06.319 --> 00:25:11.599
because it can mimic people and voices, any image, economic reality itself

294
00:25:11.599 --> 00:25:15.319
so effectively. Yeah, I mean, how could you have a criminal trial?

295
00:25:17.839 --> 00:25:19.960
I mean, how could you ever
believe that evidence was authentic? For

296
00:25:21.000 --> 00:25:23.160
example? And I don't mean like
in thirty years, I mean like next

297
00:25:23.240 --> 00:25:27.920
year. I mean that seems totally
disruptive to the way to all of our

298
00:25:27.960 --> 00:25:33.960
institutions. Well, I don't think
you could takes a random video on the

299
00:25:34.000 --> 00:25:38.440
Internet and assume it to be true. That's definitely not the case. Now,

300
00:25:38.599 --> 00:25:44.400
somebody say, has some video on
their phone or their their computer with

301
00:25:44.559 --> 00:25:48.799
the date stamp and particular time.
I think, you know, it's more

302
00:25:48.799 --> 00:25:56.839
likely to read true than not.
You knows that crypto cryptographically signed things like

303
00:25:56.279 --> 00:26:02.680
you know, mathematically, we don't
see any way for example, for AI

304
00:26:02.839 --> 00:26:07.920
to um sub both the fundamentals of
mathematics and say I figure out how to

305
00:26:07.920 --> 00:26:18.759
hashboot cooin, you know easily.
It's it's like it's not AI can't can't

306
00:26:18.759 --> 00:26:26.240
defy fundamental math. So um,
you can approve the efficiency of bitcoin hashing

307
00:26:26.640 --> 00:26:33.519
algorithms in the silicon but but but
not not fundamentally crack it. UM.

308
00:26:33.599 --> 00:26:42.079
So I guess cryptographic signatures and uh, one way to do it. Um.

309
00:26:42.119 --> 00:26:47.119
But but I'm not so worried.
I think it's more like um ore.

310
00:26:48.400 --> 00:26:52.319
You know, it's we'll humanity UM
control its destiny or not UM.

311
00:26:52.440 --> 00:26:59.039
Will we have a future that is
better than the past or not UM and

312
00:26:59.200 --> 00:27:03.519
not. You know, with that
we can so late destroy ourselves without the

313
00:27:03.559 --> 00:27:07.359
help of AI. UM. You
know that's you look at all the past

314
00:27:07.400 --> 00:27:11.440
civilizations. They didn't have AI,
at the ones that want around anymore.

315
00:27:11.160 --> 00:27:15.680
The chariots enough, Yeah, chariots
and chariots probably a really big deal back

316
00:27:15.680 --> 00:27:19.839
then they were. Ye. So
you've heard people say we should just blow

317
00:27:19.920 --> 00:27:22.839
up the server farms because there's no
way to Once it this gets rolling,

318
00:27:22.880 --> 00:27:26.119
there's no way to slow it down. What do you think of that?

319
00:27:27.079 --> 00:27:33.319
Well, the the really heavy duty
intelligence is not going to be distributed all

320
00:27:33.319 --> 00:27:37.359
over the place. It'll be in
a limited number of server centers. If

321
00:27:37.400 --> 00:27:40.599
you say like very like very sort
of deep AI, heavy duty AI,

322
00:27:41.079 --> 00:27:45.079
it's not it's not going to be
in your laptop of your phone. It's

323
00:27:45.200 --> 00:27:48.119
it's going to be in you know, situation where there's like a hundred thousand

324
00:27:49.400 --> 00:27:55.279
really powerful computers working together in a
service center. So it's not it's not

325
00:27:55.480 --> 00:27:57.799
it's not like subtle and there there
are limited number of places where that that

326
00:27:57.839 --> 00:28:00.559
can happen. In fact, you
could if you could just you can just

327
00:28:00.599 --> 00:28:07.680
look at the heat signature from space
and it'll be very obvious. Now,

328
00:28:08.279 --> 00:28:14.119
I'm not suggesting we were going bluff
to service centers right now, but there

329
00:28:14.160 --> 00:28:18.319
may be some It may be wise
to have some sort of contingency plan where

330
00:28:18.319 --> 00:28:25.359
the government's got an ability to shut
down shut down power to these service centers.

331
00:28:25.400 --> 00:28:26.960
Like you don't have to blow it
up, you can just cut the

332
00:28:26.960 --> 00:28:33.079
power and what would trip kind connectivity
as well? That's another way, right,

333
00:28:33.400 --> 00:28:36.559
But what would trip that switch?
Do you think in your mind,

334
00:28:36.599 --> 00:28:41.839
what would be the threshold that you'd
have to pass to warrant the government cutting

335
00:28:41.839 --> 00:29:00.039
off your power, cutting off your
signal? Well, I mean, I

336
00:29:00.039 --> 00:29:04.000
guess if we lost control of some
super a eye UM like for some reason,

337
00:29:04.000 --> 00:29:07.240
like like the things that would normally
work to do a passive shutdown,

338
00:29:07.319 --> 00:29:15.400
like the administrator passwords if they somehow
stopped working UM where we where we can't

339
00:29:17.960 --> 00:29:26.920
slow down or or you know,
I'm not sure. I don't have a

340
00:29:26.960 --> 00:29:34.039
precise answer, but if there's something
that we're concerned about UM and and and

341
00:29:34.400 --> 00:29:40.359
are unable to stop it with with
the software commands, then we probably want

342
00:29:40.359 --> 00:29:44.359
to have some kind of hardware both
switch. Yes, I think you know

343
00:29:44.400 --> 00:29:48.960
can't hurt. Have you talked to
since you know age and you obviously you

344
00:29:48.000 --> 00:29:53.839
know the Open Air guys because you
started to definitely have you talked to the

345
00:29:53.839 --> 00:30:00.480
the people who run these two the
biggest AI companies about this recently? I

346
00:30:00.519 --> 00:30:03.400
haven't talked to Larry Page in a
few years because he got rafts out with

347
00:30:03.440 --> 00:30:10.519
me about Opening Eye. So when
when Opening Eye was created, it did

348
00:30:10.559 --> 00:30:15.799
it did shift things into it from
unipolar world where Google Google Deep Mind controlled,

349
00:30:18.200 --> 00:30:19.119
you know, like I said,
three quarters of all AI talent,

350
00:30:19.160 --> 00:30:26.559
to where there's now sort of bipolar
world or Opening Eye and Google deep Mind

351
00:30:26.640 --> 00:30:34.960
and there and now weirdly it seems
Open Eyes that maybe ahead so so I

352
00:30:36.039 --> 00:30:41.319
have had competitions with the Opening Eye
team, tam Altman. I haven't talked

353
00:30:41.359 --> 00:30:45.359
to Larry Page because he doesn't want
to talk to you anymore for a few

354
00:30:45.440 --> 00:30:48.519
years. Can I ask you this
about it? Since you've been around a

355
00:30:48.599 --> 00:30:52.799
lot of this the thinking? So
why would anyone not be a specist be

356
00:30:53.480 --> 00:30:57.519
human centered in his thinking about technology? Like, what's the thinking there?

357
00:30:59.160 --> 00:31:02.880
Um? I think what he is
trying to say is that, UM,

358
00:31:02.920 --> 00:31:08.160
if I were to guess that heat, that all consciousness should retreat it equally,

359
00:31:08.480 --> 00:31:17.559
um. And whether that is digital
or biological m And you disagree?

360
00:31:18.279 --> 00:31:26.279
I disagree, Yeah, especially if
the digital consciousness or what are you going

361
00:31:26.319 --> 00:31:32.200
to call digital intelligence decides to curtail
the biological intelligence? Right, so you're

362
00:31:32.240 --> 00:31:33.640
just building your own slave master?
And why would you do that? Does

363
00:31:33.759 --> 00:31:42.480
sounded great? Yeah? I mean
we shouldn't. We should at least no

364
00:31:42.599 --> 00:31:48.440
need to rush. You know,
what's the hurry? Where where's the fire?

365
00:31:48.799 --> 00:31:52.440
How? Well? What I mean
tell us about the hurry? So

366
00:31:52.480 --> 00:31:55.519
this for I know you've been talking
about this for years and on the sort

367
00:31:55.519 --> 00:31:57.920
of the periphery of our attention.
We've heard Elon Musk talking about AI.

368
00:31:59.400 --> 00:32:02.920
But for most people it's been like
three months since they've had any interaction with

369
00:32:02.960 --> 00:32:07.359
this at all. Um So,
what's the timeline here? At what point

370
00:32:07.400 --> 00:32:15.079
does it start to really change our
society? Do you think? I think

371
00:32:15.079 --> 00:32:21.160
it's supposed to have probably an impact
this year, I think, Um,

372
00:32:21.960 --> 00:32:32.240
so you've got a massive expansion of
m GPT four based systems and many companies

373
00:32:32.599 --> 00:32:38.440
trying to emiliate GPT four M and
you went, opening is gonna come out

374
00:32:38.480 --> 00:32:44.559
with GPT five end of this year, which will be you had another significant

375
00:32:44.559 --> 00:32:49.519
improvement, And I was therefore GPT
one two, three four. You know,

376
00:32:49.559 --> 00:32:52.079
so GPT one was terrible, Like
you, if you tried it,

377
00:32:52.079 --> 00:32:58.440
you'd be like this is this ain't
going anywhere and it seems lame. And

378
00:32:58.480 --> 00:33:00.440
then GPT two you started to see
kind of like an inkling of like,

379
00:33:00.480 --> 00:33:05.440
well maybe this could be something useful. And then GPT three was a huge

380
00:33:05.440 --> 00:33:08.440
improvement. And now it's like,
wow, okay, this is it's still

381
00:33:08.440 --> 00:33:13.640
spouting a lot of bs, but
it's you know, it's coherent bs.

382
00:33:14.000 --> 00:33:20.519
Yes, and then GPT four now
it's like writing poetry and pretty decent poetry,

383
00:33:20.519 --> 00:33:24.599
actually pretty decent skillet ryming is incredible, yes, yes, and it's

384
00:33:24.640 --> 00:33:30.519
coherent, Yes it is. It's
even got a narrative. But yest right?

385
00:33:30.880 --> 00:33:35.039
Yeah, so you could say hard
you do like most humans can't do

386
00:33:35.039 --> 00:33:37.559
that. That's true. So it's
already past the point of what most huments

387
00:33:37.599 --> 00:33:44.519
can do. Most humans cannot write
as well as a chat GPT um and

388
00:33:44.559 --> 00:33:47.640
they certain and no human can write
that well that fast as an investment knowledge,

389
00:33:47.960 --> 00:33:54.079
so maybe Shakespeare. Um. So
so then how how much better will

390
00:33:54.240 --> 00:33:58.640
GPD five be? And how about
GPT six or seven? How can you

391
00:33:58.640 --> 00:34:02.119
have a democracy with technology? Do
you like that? If democracy is you

392
00:34:02.119 --> 00:34:05.799
know, government by the people.
Each person's vote is equal to every other

393
00:34:05.839 --> 00:34:10.320
person's vote, I mean, people
are choosing their votes freely. Can you

394
00:34:10.400 --> 00:34:17.480
have a democracy with this? Well, that's why I raise a concern of

395
00:34:20.800 --> 00:34:29.360
of AI being a significant influence in
elections. And even if you say that

396
00:34:29.400 --> 00:34:31.679
AI doesn't have agency, well,
it's very likely that people will use the

397
00:34:31.760 --> 00:34:39.679
AIM as a tool in elections.
Um. And then you know, if

398
00:34:39.679 --> 00:34:44.400
the I is smart enough, are
they using the tool or is the tool

399
00:34:44.480 --> 00:34:49.920
using them? So I think things
are getting weird, and they're getting weird

400
00:34:50.000 --> 00:34:52.960
fast, and so I think we
should be concerned about this and we should

401
00:34:53.000 --> 00:34:58.679
have regulatory oversight. That's that's why
I think it's a big deal. Um.

402
00:34:58.719 --> 00:35:02.119
And I think social media companies really
need to put a lot of attention

403
00:35:02.159 --> 00:35:09.199
into ensuring that the things I get
um created and promoted are that we're dealing

404
00:35:09.239 --> 00:35:14.599
with real people, not with a
million chat gvts pretending to be people.

405
00:35:14.719 --> 00:35:19.800
Exactly do you think speaking of social
media? You bought Twitter famously got a

406
00:35:19.800 --> 00:35:22.800
lot of other businesses and a lot
going on. Yes, you said you

407
00:35:22.800 --> 00:35:27.440
bought it because you believe in speech, free speech. You've had a lot

408
00:35:27.440 --> 00:35:30.320
of hassles since you bought it.
In retrospect, was it worth buying it?

409
00:35:31.320 --> 00:35:39.000
Um Um? I mean it mights
to be seen as to whether this

410
00:35:39.199 --> 00:35:45.800
was financially smart. Currently it is, it is not. You know,

411
00:35:45.840 --> 00:35:50.239
we just revalued the company at less
than half of the acquisition prodd Yes,

412
00:35:52.480 --> 00:35:59.400
um No, My timing was terrible
for for when the off always made,

413
00:35:59.440 --> 00:36:05.719
because it was right before advertising plummeted
and you've caught the high water mark I

414
00:36:05.800 --> 00:36:08.360
noticed. Yeah, yeah, so
I must be a real genius. Year

415
00:36:10.199 --> 00:36:15.000
My timing is amazing since I've had
a border for at least twice as much

416
00:36:15.039 --> 00:36:21.920
as it should have been borderful,
but somethings are priceless. And so the

417
00:36:22.960 --> 00:36:28.679
whether losing money or not that as
a secondary issue compared to ensuring the strength

418
00:36:28.719 --> 00:36:32.559
of democracy and free speech is the
bedrock of a functioning democracy. Yes,

419
00:36:34.159 --> 00:36:39.000
and the speech needs to be as
transparent and truthful as possible. Um.

420
00:36:39.079 --> 00:36:44.800
So we've got a huge push on
Twitter to be as truthful as possible.

421
00:36:44.800 --> 00:36:47.239
We've got this community notes speecher,
which is great. It is great.

422
00:36:47.400 --> 00:36:51.199
It is awesome. Yeah, and
it's like I thought this morning, Yeah,

423
00:36:51.239 --> 00:36:53.000
it was far more honest than the
New York Times. It's great.

424
00:36:53.480 --> 00:36:58.199
We put a lot of effort to
ensuring that community notes does not get gamed

425
00:36:58.519 --> 00:37:01.920
or have biases. It simply cares
about what is the most accurate thing.

426
00:37:02.559 --> 00:37:06.679
Um. And you know, sometimes
truth can be a little bit elusive,

427
00:37:06.719 --> 00:37:09.199
but but you can still aspire to
get closer to it. Yes, you

428
00:37:09.239 --> 00:37:16.280
know, and so um And I
think the effect of community notes is more

429
00:37:16.320 --> 00:37:20.719
powerful than than people may realize,
because once people know that they could get

430
00:37:20.800 --> 00:37:24.480
noted, um, you know,
community noted on Twitter, then they'll think

431
00:37:25.000 --> 00:37:30.079
more carefully about what they say,
and they are likely It's basically it's an

432
00:37:30.159 --> 00:37:32.239
encouragement to be more truthful and less
deceptive. Yes, and if the notes

433
00:37:32.320 --> 00:37:37.079
themselves are truthful, then it will
have the effect absolutely. And all of

434
00:37:37.119 --> 00:37:39.039
that is open source. All the
community notes is open so you can read

435
00:37:39.039 --> 00:37:43.480
about every community note. You can
see exactly how the algorithm works. You

436
00:37:43.519 --> 00:37:47.159
can you can you can register say
like, oh we need to make this

437
00:37:47.320 --> 00:37:52.760
change with that change. Um.
So everything is super open book with community

438
00:37:52.760 --> 00:37:55.639
notes. There's no no black box
when you jumped into this. Though,

439
00:37:55.679 --> 00:38:00.840
when you bought it, did you
understand clearly understand it's importance you wouldn't have

440
00:38:00.880 --> 00:38:04.880
bought it Twitter. Yes, right, but it's not the biggest, but

441
00:38:04.880 --> 00:38:08.159
it's the most important to social media
companies. But did you understand the kind

442
00:38:08.199 --> 00:38:14.599
of ferocity you'd be facing, the
attacks you'd be facing from power centers in

443
00:38:14.599 --> 00:38:19.920
the country. UM. I thought
there'd probably be some negative reactions here,

444
00:38:22.239 --> 00:38:24.880
so I'm not I'm sure everyone would
not be completed with the with with it,

445
00:38:25.039 --> 00:38:30.000
um. But at the end of
the day, you know, if

446
00:38:30.159 --> 00:38:34.239
if if the public is happy with
it, that's what matters, um,

447
00:38:34.679 --> 00:38:37.719
And the public will speak with their
actions. Oh though, I mean though,

448
00:38:37.840 --> 00:38:40.840
if they find truth Twitter to be
useful, they will use it more.

449
00:38:40.880 --> 00:38:44.199
And if they find it to be
not useful, I will use it

450
00:38:44.280 --> 00:38:46.440
less. They find it to be
the best source of truth, I think

451
00:38:46.440 --> 00:38:52.000
they will use it more. Um
So that's my theory. Um. And

452
00:38:52.079 --> 00:38:59.480
so I even though you know now
there's there's obviously a lot of organizations that

453
00:38:59.800 --> 00:39:05.559
you to having sort of unfettered influence
on Twitter, um that no long have

454
00:39:05.599 --> 00:39:08.000
that we used to New York Times
of their of their badge this morning,

455
00:39:08.000 --> 00:39:12.679
and then you call them diarrhea.
Ok, you did, you did.

456
00:39:12.719 --> 00:39:15.800
I'm just I'm just quoting it.
You describe their Twitter feed as a diarrhea,

457
00:39:16.519 --> 00:39:20.480
I said, the Twitter Twitter equivalent
of diarrhea. Okay, it's not

458
00:39:20.519 --> 00:39:23.159
literally diarrhea, but no, it's
a you know, it's as a metaphor,

459
00:39:25.000 --> 00:39:29.119
but an accurate one. M So. I mean, if you look

460
00:39:29.119 --> 00:39:32.960
at the at n White Times twitter
feed is unreadable. It's like because what

461
00:39:34.039 --> 00:39:37.159
they do is that they tweet every
single article, even the ones that are

462
00:39:37.960 --> 00:39:40.800
boring, even once they don't make
it into the paper. So uh so

463
00:39:40.840 --> 00:39:47.039
it's just NonStop zillion tweets a day
with no uh you know, they really

464
00:39:47.039 --> 00:39:51.639
should just be saying like what are
the top tweets? Like what are the

465
00:39:51.719 --> 00:39:53.559
what are the one of the big
stories of the day. I don't know,

466
00:39:53.599 --> 00:39:58.159
put out like ten or something.
You know, it's a number that's

467
00:39:58.159 --> 00:40:00.159
manageable, um, as opposed to
right now, if you if you would

468
00:40:00.199 --> 00:40:05.159
have followed white at n white times
on Twitter, you're gonna get barraged with

469
00:40:05.239 --> 00:40:07.599
like hundreds of tweets a day,
um, and your whole feed will be

470
00:40:07.599 --> 00:40:12.599
filled with in white times. So
UM that that's that's This is something I

471
00:40:12.639 --> 00:40:16.119
would recommend actually for old publications,
which is uh, for your primary feed,

472
00:40:16.400 --> 00:40:20.519
UM, only put out your best
stuff. Uh. Don't put out

473
00:40:20.559 --> 00:40:23.639
everything um. Or you can have
a second feed that is, here's everything,

474
00:40:23.960 --> 00:40:28.840
um. But then but have a
have your primary feed, but here's

475
00:40:28.880 --> 00:40:37.039
our best stuff. If any um
uh media organization or individual uh just uh

476
00:40:38.079 --> 00:40:43.480
have don't put out one hundreds of
tweets today. Just put out like ten

477
00:40:43.519 --> 00:40:46.679
good ones um or five good ones
or and if if it's a slow news

478
00:40:46.760 --> 00:40:51.280
day, don't put on any maybe
put out two one or two. But

479
00:40:51.960 --> 00:40:54.400
don't don't don't try to say we're
always going to put out one hundred tweets,

480
00:40:54.440 --> 00:40:58.280
even if you know, if it's
World War three or a bicycle accident

481
00:40:58.400 --> 00:41:02.559
was the biggest news you know,
it's got to be like news that it's

482
00:41:02.559 --> 00:41:07.599
got to you got to earn your
earn your play on someone's attention. Yes,

483
00:41:07.920 --> 00:41:10.800
um, so just in general,
UM, I kind of think I

484
00:41:10.840 --> 00:41:14.159
know a thing or two about how
to use Twitter, because you know,

485
00:41:14.280 --> 00:41:17.760
I was the most interacted with account
on the whole system before the acquisition.

486
00:41:19.719 --> 00:41:22.159
Before the acquisition closed, I didn't
have the most number of followers, but

487
00:41:22.199 --> 00:41:25.320
I had the most number of interactions, and so I clearly know something about

488
00:41:25.400 --> 00:41:29.360
how to use Twitter. Um,
and so people should you know, this

489
00:41:29.559 --> 00:41:35.400
is my advice, I think.
Um. So, you know, people's

490
00:41:35.440 --> 00:41:37.679
attention is limited, so just make
sure you put the stuff that's most important

491
00:41:37.719 --> 00:41:43.239
there. So, because you know, you and people like you do interact

492
00:41:43.320 --> 00:41:49.000
on Twitter, it's obviously enormously powerful
in shaping public opinion. It's where a

493
00:41:49.039 --> 00:41:52.199
lot of ideas and trends are incubated. Yeah, you know, it's absolutely

494
00:41:52.960 --> 00:41:57.440
It's also a magnet for intelligencies from
around the world. And one of the

495
00:41:57.440 --> 00:42:00.559
things we learned after you started opening
the books is that they were exerting influence

496
00:42:00.599 --> 00:42:07.880
from within Twitter. I mean it
was absurd. Um, did you know

497
00:42:07.960 --> 00:42:12.119
that going in No? Well,
well, so, things like I have

498
00:42:12.199 --> 00:42:19.760
a. UM, since I've been
a heavy Twitter user since two thousand and

499
00:42:19.800 --> 00:42:22.760
nine. Um My, It's it's
sort of like I'm in the matrix.

500
00:42:22.800 --> 00:42:27.760
I mean I can see like things, do things feel right? Do they

501
00:42:27.880 --> 00:42:30.400
not feel right? What? What
tweets am I being shown as recommended?

502
00:42:32.480 --> 00:42:37.840
Like like I get a feel like
what what accounts and making comments? Where

503
00:42:37.840 --> 00:42:43.280
are the comments eerially similar? Yeah? And uh and then you look at

504
00:42:43.280 --> 00:42:47.559
the account and it's just obviously a
fake photo, and you know it's just

505
00:42:47.559 --> 00:42:53.519
obviously a bought cluster over and over
again. UM. So this is actually

506
00:42:53.800 --> 00:42:59.320
so I started to get like just
more and more uneasy about the Twitter situation.

507
00:42:59.639 --> 00:43:06.840
UM. And my initial goal was
was was actually not not to acquire

508
00:43:06.880 --> 00:43:10.480
Twitter um. I mean, the
the actual sequence events was that I UM,

509
00:43:12.360 --> 00:43:17.360
I was looking at m I held
a Twitter poll to say, like,

510
00:43:17.480 --> 00:43:22.039
should I sell some of my Tesla
stock because I was getting you know,

511
00:43:22.039 --> 00:43:23.960
a couple of years ago, I
was getting um attacked a lot for

512
00:43:24.119 --> 00:43:30.679
like allegedly not paying taxes. Um. Uh. And now I've actually paid

513
00:43:30.679 --> 00:43:32.320
a tremendous amount of taxes. Um. Now, there was one year I

514
00:43:32.320 --> 00:43:37.280
didn't pay taxes because I had overpaid
taxes in the prior year. And you

515
00:43:37.320 --> 00:43:42.760
know when they had that like IRS
leak bs, they knew that I had

516
00:43:42.760 --> 00:43:45.559
overpaid taxes in the prior year.
But they said, oh, Elon must

517
00:43:45.559 --> 00:43:49.079
didn't pay taxes in twenty seventeen or
whatever it was. And I was like,

518
00:43:49.679 --> 00:43:52.760
but you know that the reason I
didn't pay taxes because I overpaid the

519
00:43:52.760 --> 00:43:57.480
prior year if you didn't mention that, So that was deceptive. Um anyway,

520
00:43:57.519 --> 00:44:00.400
So but there, you know,
Elizabeth Warren's The World and Bony Santa

521
00:44:00.519 --> 00:44:05.880
is like saying, oh, you
know, I'm not selling stock and I'm

522
00:44:05.920 --> 00:44:08.360
not paying taxes and I and so
I so, so I'm like, look,

523
00:44:08.360 --> 00:44:09.920
I don't know what the right thing
to do this here. I thought

524
00:44:09.960 --> 00:44:13.760
the right thing to do was to
not sell stock. The captain should be

525
00:44:13.760 --> 00:44:15.320
the last one to leave the show, that's right, um and um.

526
00:44:15.480 --> 00:44:17.360
I thought I was doing the right
thing by not selling stock, and now

527
00:44:17.400 --> 00:44:21.840
I'm being told I'm doing the wrong
thing, um by by you know,

528
00:44:21.920 --> 00:44:24.039
clinging onto the stock and not paying
taxes. So I hold a Twitter poll

529
00:44:24.360 --> 00:44:30.199
to say, which what do you
guys want should I sell? I don't

530
00:44:30.199 --> 00:44:34.000
know. Teen percent up might tells
the stock or not. I'll buy buy

531
00:44:34.000 --> 00:44:37.800
the results of the pole and it's
like sixty sixty percent of people SAI,

532
00:44:37.840 --> 00:44:39.239
yeah, you should sell tempercent.
So I didn't. Um. So,

533
00:44:39.280 --> 00:44:43.880
then I had a bunch of cash, and I'm like, what should I

534
00:44:43.880 --> 00:44:46.280
do with this? At the time, the federal reserve rates were super low,

535
00:44:46.320 --> 00:44:52.079
So it's just like sitting in the
you know, I guess your checking

536
00:44:52.079 --> 00:44:55.079
account, well in in the tea
bowl account, right, you know,

537
00:44:55.199 --> 00:45:00.480
money mark account whatever, the whole
banking things, a whole sever subject.

538
00:45:00.719 --> 00:45:04.760
Um, I know, the little
thing a thing which about finance. Um

539
00:45:05.280 --> 00:45:09.000
but um so, uh so they'm
not staying with this money the account it's

540
00:45:09.039 --> 00:45:13.360
earning less than the rate of inflation. So the rate inflation is much higher

541
00:45:13.360 --> 00:45:15.559
than we got a high inflation.
It's I'm earning peanuts in the money market

542
00:45:15.599 --> 00:45:19.800
account. This is dumb. I'm
getting like minus it's just evaporating. Yeah,

543
00:45:19.840 --> 00:45:22.480
I'm getting minus like six or seven
percent in return here, maybe worse.

544
00:45:23.119 --> 00:45:29.280
And so then well it's like what
stock should I buy? Um?

545
00:45:29.760 --> 00:45:30.960
And I you know, I believe
in buying stocks of companies where you use

546
00:45:31.000 --> 00:45:37.760
the product. Um And Apple's got
a competing electric vehicle car program, so

547
00:45:37.159 --> 00:45:40.000
you know, like Apple products we're
not going to invest in them because they've

548
00:45:40.039 --> 00:45:46.039
got to competing, uh, your
autonomous EV program. And so it's the

549
00:45:46.079 --> 00:45:50.360
other product that I used Twitter.
Okay, so I'll you know it,

550
00:45:50.519 --> 00:45:52.599
put the money on Twitter. It's
better than just having it, you know

551
00:45:52.719 --> 00:45:59.880
that negative six percent inflation situation.
Um. So so I like board,

552
00:46:00.199 --> 00:46:01.239
put a bunch of it? What
a board? What a bunch of Twitter

553
00:46:01.360 --> 00:46:04.920
stock? Like I said, no, with the intended buying the company just

554
00:46:05.639 --> 00:46:08.719
better than keeping in a money market. How much you bought, um,

555
00:46:08.760 --> 00:46:14.559
I think it was like eight percent
or something of the company. UM.

556
00:46:14.679 --> 00:46:21.440
I'm talking to some of the board
members um. And then and then that

557
00:46:21.559 --> 00:46:22.920
they said, hey, well I
do you want to do? You want

558
00:46:22.920 --> 00:46:24.119
to join the board? So I
was like, well, I don't.

559
00:46:24.159 --> 00:46:30.000
I really don't want to be on
boards, but because it's boring, um,

560
00:46:30.119 --> 00:46:34.119
and I love things to do,
but I do care about the direction

561
00:46:34.159 --> 00:46:37.920
of Twitter, So I'll consider being
on the board. And I thought about

562
00:46:37.920 --> 00:46:44.199
it for about a week or so. And then then based on the conversations

563
00:46:44.199 --> 00:46:49.679
that I was having with the management
team on the board, UM, I

564
00:46:49.760 --> 00:46:55.599
came to the conclusion, rightly or
wrongly, that UM that if I joined

565
00:46:55.639 --> 00:46:59.800
the board, they would not listen
to me. So then I'm like,

566
00:47:00.159 --> 00:47:02.960
okay, then I would just be
a quizzling you know, I don't want

567
00:47:02.960 --> 00:47:07.039
to be some sort of just you
know, go along for the right quizzling

568
00:47:07.119 --> 00:47:15.039
situation. Um and and if a
collaborator effectively right, So and it really

569
00:47:15.079 --> 00:47:17.960
felt like I start, I was
starting to feel like, wait a second,

570
00:47:17.960 --> 00:47:23.840
like weird something something's like, something's
not right in this. You know,

571
00:47:24.159 --> 00:47:29.159
something's rot in the state of n
walk here, there's somebody feels wrong

572
00:47:29.159 --> 00:47:34.440
about the platform. It seemed to
be just drifting in a I couldn't place

573
00:47:34.519 --> 00:47:37.239
it exactly just ahead of it felt
like it was drifting in a bad direction.

574
00:47:37.400 --> 00:47:42.480
So then I was like, and
and my conversations with the board and

575
00:47:42.679 --> 00:47:49.119
management seems to confirm my intuition about
that. So then I was like,

576
00:47:49.239 --> 00:47:53.079
okay, um but basically as commis, these guys do not care about fixing

577
00:47:53.159 --> 00:47:59.559
Twitter and and and I had a
bad feeling about where I was headed based

578
00:47:59.599 --> 00:48:00.559
in the comes and the conversations I
had with them. So then it was

579
00:48:00.599 --> 00:48:06.960
like, you know what, I
I'll try acquiring it and see if that's

580
00:48:07.400 --> 00:48:12.519
see if acquiring it is possible.
Now I didn't have enough cash to acquire

581
00:48:12.559 --> 00:48:15.280
it, so I would need,
you know, support from others, from

582
00:48:15.320 --> 00:48:17.920
some of the existing investors. I
would also need like a lot of debt,

583
00:48:19.159 --> 00:48:22.239
and so it wasn't clear to me
whether an accisition would succeed at what

584
00:48:22.320 --> 00:48:30.440
I would try, and ultimately it
did succeed. So anyway, here we

585
00:48:30.519 --> 00:48:34.519
are. But when you got there
and all of a sudden you own it,

586
00:48:34.760 --> 00:48:37.639
and all the data on the service
belongs to you, what belongs to

587
00:48:37.679 --> 00:48:39.960
the people in my view a bit, yes, but but you can see

588
00:48:40.000 --> 00:48:43.559
what I guess, and you can
see what they've been doing, and you

589
00:48:43.599 --> 00:48:47.079
can see who's been working there.
You were shocked to find out that various

590
00:48:47.119 --> 00:48:55.960
intell agencies were affecting its operations.
The degree to which various government agencies were

591
00:48:57.079 --> 00:49:04.639
effectively had effectively had full access to
everything that was going on on Twitter blew

592
00:49:04.719 --> 00:49:07.559
my mind. I was not aware
of that. Would that include people's dms?

593
00:49:09.199 --> 00:49:15.800
Yes, yes, because the DM's
are not encrypted. So one of

594
00:49:15.800 --> 00:49:17.320
the first you know, one of
the things that we're about to release,

595
00:49:17.920 --> 00:49:21.960
his ability to encrypt your DM.
That's pretty heavy duty though, because a

596
00:49:22.000 --> 00:49:25.159
lot of well known people, reporters
talking to their sources, government officials,

597
00:49:25.360 --> 00:49:30.360
rich people, in the world they're
dming each other and the assumption obviously was

598
00:49:30.719 --> 00:49:35.519
incorrect, but was that that's private, but that was being read by various

599
00:49:35.559 --> 00:49:40.480
governments. Yeah, that seems to
be Yes, scary, Yes, it

600
00:49:40.679 --> 00:49:46.480
is. So like I said,
we're moving to UM have the dms be

601
00:49:46.639 --> 00:49:49.880
optionally encrypted. I mean, you
know, there's like a lot of DM

602
00:49:49.960 --> 00:49:52.119
conversations which are you know, just
chatting with friends. It's not not not

603
00:49:52.199 --> 00:49:59.159
important, but so we were.
That's hopefully coming out later this month,

604
00:49:59.639 --> 00:50:02.800
but no later the next month.
Is the ability to toggle encryption on or

605
00:50:02.880 --> 00:50:07.280
off. So if you if you
are in a conversation you think a sensitive,

606
00:50:07.320 --> 00:50:09.400
you can just toggle encryption on and
then no one on Twitter can see

607
00:50:09.639 --> 00:50:12.880
what are you talking about. They
could put it gun into my head and

608
00:50:12.920 --> 00:50:16.440
I couldn't. I couldn't tell.
I couldn't. That's sort of the gun

609
00:50:16.519 --> 00:50:19.760
to the head test. If I
puts it gount to my head and I

610
00:50:20.039 --> 00:50:22.679
can I still not see your d
ms, that should be that's the ass

611
00:50:22.719 --> 00:50:27.960
of test. Yes, UM,
and that's how that's that's how it should

612
00:50:28.000 --> 00:50:30.280
be if you want. Have you
had complaints from various governments about doing this?

613
00:50:34.039 --> 00:50:37.280
I haven't had direct complaints to me. I've had sort of like some

614
00:50:37.440 --> 00:50:40.559
indirect complaints. I think people are
a little concerned about complaining to me directly

615
00:50:40.599 --> 00:50:47.400
in case I tweet about it.
You know, they're like, oh,

616
00:50:49.639 --> 00:50:52.039
so there's sort of trying to be
more round about than that. Um.

617
00:50:53.440 --> 00:51:01.760
And you know, I mean,
if if I got something that was unconstitutional

618
00:51:01.840 --> 00:51:05.639
from the US government, my reply
would be to sending a copy of the

619
00:51:06.400 --> 00:51:08.639
First Amendment and to say, like, what part of this are we're getting

620
00:51:08.639 --> 00:51:15.280
wrong? You have a lot of
government, you have a lot of governments.

621
00:51:15.679 --> 00:51:16.360
Part of this are we getting wrong? Please tell me. I mean

622
00:51:16.400 --> 00:51:19.360
it's a pretty no. I'm just
saying, but you're kind of exposed in

623
00:51:19.440 --> 00:51:22.639
your other businesses. So this is
just in case reviewers aren't following this.

624
00:51:22.639 --> 00:51:25.639
This is not You're not just like
a journalist taking a stand on behalf of

625
00:51:25.639 --> 00:51:30.400
the First Amendment. You're a guy
with big government contracts giving the finger to

626
00:51:30.480 --> 00:51:32.679
the government in some way. Well, am I giving the fingers to government?

627
00:51:32.760 --> 00:51:37.800
I think that there are I'm not
someone who thinks that, you know,

628
00:51:37.920 --> 00:51:42.599
the government is just sort of evil, right, It's it's it's a

629
00:51:42.760 --> 00:51:47.599
it's a large bureaucracy. There are
people in government who are human beings,

630
00:51:47.639 --> 00:51:54.239
and they have the people with good
motivations occasionally bad motivations with with rare exception,

631
00:51:54.360 --> 00:51:58.679
the people that I know in government
to have good motivations and just want

632
00:51:58.719 --> 00:52:02.639
to get their job done and and
they actually believe in the constitution and they're

633
00:52:04.119 --> 00:52:07.840
so I think my opinion is actually
most people in the government are good.

634
00:52:07.400 --> 00:52:13.800
It's hardening to hear. Yeah,
it's it's rare for me to find someone

635
00:52:13.840 --> 00:52:15.760
in the government who I think is
perhaps not good. But but you know,

636
00:52:16.880 --> 00:52:21.920
at the highest level of the agencies
there are political appointees as as you

637
00:52:22.000 --> 00:52:24.679
know, UM, and the political
pointees will have a political agenda UM.

638
00:52:24.800 --> 00:52:30.480
And so they at the at the
highest levels of the various government agencies,

639
00:52:30.679 --> 00:52:32.760
there is the ability to put a
sort of political thumb on the scale,

640
00:52:34.559 --> 00:52:39.760
even if the people operating the agencies
I don't agree with that. UM.

641
00:52:40.039 --> 00:52:45.440
So yeah, so that's something to
be concerned about. Um is I'd say

642
00:52:45.480 --> 00:52:50.159
I'd be more concerned about about political
appointees. So I think than than UM

643
00:52:52.239 --> 00:52:55.000
sort of people to get the career
of people that's been my experience at least.

644
00:52:55.840 --> 00:53:01.760
Do you think um, Twitter will
be central to this presidential campaign as

645
00:53:02.039 --> 00:53:08.800
it wasn't the last several. I
think it will play a significant role in

646
00:53:09.199 --> 00:53:19.159
elections, not just domestically but internationally. Um So, the goal of new

647
00:53:19.199 --> 00:53:23.679
Twitter is to be um as fair
and even had it as possible, So

648
00:53:24.239 --> 00:53:34.360
not favoring any political ideology, but
just um yeah, being being being fair

649
00:53:34.360 --> 00:53:37.239
at all. How does in Facebook
do this? I know that Zuckerberg has

650
00:53:37.320 --> 00:53:39.840
said, and I take him at
face value that he well, I do

651
00:53:40.440 --> 00:53:45.800
actually in this way that he is
a kind of old fashioned liberal who doesn't

652
00:53:45.880 --> 00:53:47.920
like to censor. He has,
but he you know, like, why

653
00:53:49.000 --> 00:53:52.320
wouldn't a company like that take the
stand that you have taken, which pretty

654
00:53:53.079 --> 00:54:05.239
rooted in American traditional political custom,
you know, for free speech. My

655
00:54:05.360 --> 00:54:09.840
understanding is that um, zach book
spent four hundred million dollars in the last

656
00:54:09.880 --> 00:54:14.920
election, nominally in a get out
the vote campaign, but really fundamentally in

657
00:54:14.960 --> 00:54:17.559
support of Democrats. Is that accurate
or not accurate? That is accurate?

658
00:54:19.559 --> 00:54:28.719
Does that sound unbiased to you?
No? It doesn't. Yes, um

659
00:54:29.440 --> 00:54:35.360
So you don't see hope that Facebook
will approach this as a non aligned arbiter.

660
00:54:37.400 --> 00:54:47.119
I'm unaware of evidence to suggests that
path Um do you can you you've

661
00:54:47.400 --> 00:54:51.320
you've allowed Donald Trump back on Twitter. He hasn't taken you up on your

662
00:54:51.360 --> 00:54:53.079
offer because he's got his own thing. Do you think he will go back

663
00:54:53.199 --> 00:54:59.599
on Twitter? Well, that's that's
obviously up to him. Um, my

664
00:54:59.719 --> 00:55:07.679
job is too. You know.
I take the freedom of speech very seriously.

665
00:55:07.760 --> 00:55:12.679
So it's um, you know,
I didn't. I didn't vote for

666
00:55:12.760 --> 00:55:15.119
Donaldrump. I actually vote for Biden. People think I'm some a little hardcore,

667
00:55:15.639 --> 00:55:20.639
you know. So it's certainly some
of the media try to paint me

668
00:55:20.719 --> 00:55:24.840
as like far ride or whatever.
I'm in the only time I've ever even

669
00:55:24.920 --> 00:55:31.559
voted Republican was was once for because
I readied to vote in South Texas,

670
00:55:31.760 --> 00:55:36.880
I was for a Mexican American woman
for Congress. That's that's literally only a

671
00:55:36.920 --> 00:55:44.280
Republican vote I've ever cast my entire
life. Um once um and uh so

672
00:55:44.920 --> 00:55:50.280
UM. Not saying I'm a huge
fan of Biden, because I would I

673
00:55:50.320 --> 00:55:53.119
think that would probably been accurate.
But um, you know, we have

674
00:55:53.280 --> 00:55:59.760
difficult choices to make in these residential
elections. It's not I would prefer,

675
00:56:00.880 --> 00:56:07.320
frankly that we we put someone just
a normal person as president, a normal

676
00:56:07.760 --> 00:56:12.400
person with common sense, uh,
and whose values are smack in the middle

677
00:56:12.400 --> 00:56:15.559
of the country, you know,
just you know, Sarah, the normal

678
00:56:15.599 --> 00:56:21.639
distribution, and I think they'll that
they would be great it wan would be

679
00:56:21.719 --> 00:56:23.719
happier. Would you run, Like
why wouldn't you run? I was born

680
00:56:23.800 --> 00:56:30.400
here, so of course you're yeah, it's it's I'm a technologist. Also

681
00:56:30.440 --> 00:56:34.400
I'm not not a politician, so
it's not like it. You know,

682
00:56:34.440 --> 00:56:38.360
I think we have made maybe being
president not that much fun, you know

683
00:56:39.519 --> 00:56:45.280
it to be totly frank. Um. It is by design a relatively weak

684
00:56:45.440 --> 00:56:51.079
role, yea, because it's intended
to be balanced by the House and the

685
00:56:51.119 --> 00:56:54.559
Senate and the judiciary. Um.
So it's not like like if you're a

686
00:56:54.599 --> 00:56:59.719
prime minister in England or Canada you
have home of power than if your president,

687
00:57:00.000 --> 00:57:04.440
because it's like being speak of the
House and being um president, you

688
00:57:04.559 --> 00:57:12.760
know so um so yeah for presidents
like deliberately week in order to avoid creating

689
00:57:12.800 --> 00:57:16.880
a king situation, king queen situation. Um. And but you get dumped

690
00:57:16.920 --> 00:57:21.800
on all day no matter what you
do, um, and everything you do

691
00:57:21.920 --> 00:57:27.719
is scrutinized um. And your life
is not your own um. And if

692
00:57:27.960 --> 00:57:31.159
if if you're at any any skeletons
you've got in the closet will be trotted

693
00:57:31.239 --> 00:57:37.119
out and you know, breided down
main street. Um. And even if

694
00:57:37.159 --> 00:57:42.000
they don't exist, they'll make them
up and fakes it whatever they will.

695
00:57:42.760 --> 00:57:45.480
Politics is a blood sport. So
it's it's not something I want to do.

696
00:57:46.000 --> 00:57:49.119
So I got it. One last
threat. I just think you alluded

697
00:57:49.159 --> 00:57:52.079
to you said, don't get me
started on the banks. So you've seen

698
00:57:52.280 --> 00:57:54.880
well, so you've seen a couple
of regional bank collapses. Yeah, and

699
00:57:55.039 --> 00:58:00.280
we've been told that's not a big
deal, that these are isolated and one

700
00:58:00.360 --> 00:58:04.480
collapse for unique reasons or night.
It's not systemic in any sense. What's

701
00:58:04.519 --> 00:58:13.039
your sense your sense of the stability
of the American banking system. Well,

702
00:58:13.119 --> 00:58:15.519
it's it's actually at this point of
global banking system processes. UM. So

703
00:58:17.199 --> 00:58:23.800
the uh, you know, we
have a situation here where it's not really

704
00:58:23.880 --> 00:58:28.119
it's not that the canary in the
coal mine has died, but the miners

705
00:58:28.280 --> 00:58:32.800
are starting to die too. The
you know, the so and you know,

706
00:58:32.800 --> 00:58:38.480
a silicon value bank collapsing overnight,
Um, is one hell of a

707
00:58:38.480 --> 00:58:43.079
big canary. You small atkey,
I mean, it's not just it's not

708
00:58:43.199 --> 00:58:46.199
like some small fry thing yeah,
Um, it's big fry, so a

709
00:58:46.360 --> 00:58:52.400
medium fry. And then Curtis Suez, which is I think was formed in

710
00:58:52.400 --> 00:58:57.480
the mid eighteen hundreds, um was
basically sold for pennies on the dollar,

711
00:58:58.800 --> 00:59:02.159
forced to merge with UBS, and
even then required back stuff by the Swiss

712
00:59:02.199 --> 00:59:06.840
government. I mean, like,
hello, guys, maybe we have issues

713
00:59:07.320 --> 00:59:12.239
here. Maybe things aren't all great. They're definitely not all great. Maybe

714
00:59:14.760 --> 00:59:21.840
more more forceful here. Um.
I think that there is a serious danger

715
00:59:22.320 --> 00:59:34.119
with the gold banking system. There's
there's a strong argument that the if you

716
00:59:34.360 --> 00:59:39.360
were to actually mark market the portfolios
of the banks, the loans and whatnot,

717
00:59:40.599 --> 00:59:45.400
that the entire banking industry would have
negative equity. It feels that way,

718
00:59:46.519 --> 00:59:52.199
yes, Um, so if you
look at say commercial real estate like

719
00:59:52.280 --> 00:59:59.880
offices and whatnot, the whole work
from home thing has substantially reduced office usage

720
01:00:00.440 --> 01:00:06.760
in cities around the world. Um. And you know, I think I

721
01:00:06.800 --> 01:00:10.360
think Sampacis goes to forty percent off
Samaceis goes like an extreme example, but

722
01:00:10.440 --> 01:00:15.719
it's like, I think it's on
forty percent vacancy. Um. Even even

723
01:00:15.840 --> 01:00:21.000
New York has almost all cities at
this point have ever equoted vacancies and commercial

724
01:00:21.039 --> 01:00:24.280
real estate. So um. Now, the commercial real estate used to be

725
01:00:24.320 --> 01:00:29.559
something that was a great a asset. Um that if a bank had commercial

726
01:00:29.599 --> 01:00:34.119
real estate holdings, that those would
be considered of the highest security some of

727
01:00:34.159 --> 01:00:38.840
the safest uh you know, assets
you can have. Now that is not

728
01:00:38.920 --> 01:00:45.119
the case anymore. One company after
another is canceling the leases or not renewing

729
01:00:45.159 --> 01:00:50.559
the leases, or if they go
bankrupt, you have the there's nothing for

730
01:00:50.719 --> 01:00:53.880
the the bank you owens that real
estate to go after because they were you

731
01:00:53.960 --> 01:00:58.320
know, previously strong company now dead. What do you what do you?

732
01:00:58.360 --> 01:01:04.199
What do you go after that point? So we really haven't seen the commercialist

733
01:01:04.559 --> 01:01:08.440
shoe drop. That's more like an
anivil, not a shoe. So the

734
01:01:08.480 --> 01:01:21.079
stuff we've seen thus far actually hasn't
even it's only slightly um real estate portfolio

735
01:01:21.159 --> 01:01:24.800
degradation, but that will become a
very serious thing leader this year. In

736
01:01:24.960 --> 01:01:30.519
my In my view, um,
I think if we see which we are

737
01:01:30.559 --> 01:01:35.880
likely to see a drop in house
prices because the interest rates are too high

738
01:01:36.480 --> 01:01:39.679
and for most people weren't buying a
house, they look at the monthly payment.

739
01:01:40.320 --> 01:01:44.440
If you're thirty thirty mortgages. The
vast majority of his interest. So

740
01:01:44.519 --> 01:01:49.079
if the FED rate is high,
you have a high based interest rate.

741
01:01:49.159 --> 01:01:52.880
Effectively the price you can pay for
the house drops because you now have to

742
01:01:52.920 --> 01:01:58.800
pay more interest, which means that
if you've got a fixed multi payments you

743
01:01:58.880 --> 01:02:04.400
can now afford to buy a housepitals
less money efectically drops the prices of houses.

744
01:02:04.599 --> 01:02:08.119
Yes, Um, this is the
kind of thing. It tends to

745
01:02:08.199 --> 01:02:13.679
accelerate. So so then you can
get negative equity and the home market as

746
01:02:13.719 --> 01:02:19.599
well. And so so if if
banks end up having loan lesses in both

747
01:02:19.679 --> 01:02:23.159
their commercial and they're definitely gonna have
loan less in their commercial portfolio, but

748
01:02:23.320 --> 01:02:30.400
also in the mortgage portfolio. This
is um a dire situation. Um.

749
01:02:31.159 --> 01:02:37.400
There there is there is a solution
to mitigate the magnitude of the damage here,

750
01:02:37.440 --> 01:02:42.400
which is for the Fed to lower
the rate. But they raise the

751
01:02:42.480 --> 01:02:51.119
rate again. Um. Now,
if I recall correctly important caveat, I

752
01:02:51.199 --> 01:02:53.599
think the last time the Fed raised
rates going into a recession was nineteen to

753
01:02:53.599 --> 01:03:04.639
twenty nine. What happened next,
Yeah, the great depression. That concern

754
01:03:04.960 --> 01:03:07.119
I'm telling you nothing, you don't
know. But the concern is if the

755
01:03:07.199 --> 01:03:10.639
FED drops rates again, then inflation
will accelerate, and you can't do that

756
01:03:10.679 --> 01:03:15.639
in an election yet, So inflation
is going to happen no matter what.

757
01:03:17.760 --> 01:03:22.280
If you increase the money supply,
um, you get inflation. Right.

758
01:03:22.920 --> 01:03:28.920
So there's no there's not some magical
cure for getting rid of inflation, um,

759
01:03:29.360 --> 01:03:34.960
except to increase the productivity the output
output of goods and services. So

760
01:03:36.679 --> 01:03:39.119
if you say like like what is
money, um, you've got you've got

761
01:03:39.280 --> 01:03:44.840
you've got these sort of um,
it's basically numbers in a database. That's

762
01:03:45.239 --> 01:03:49.519
that that's some up to some co
some total. Then you've got the output

763
01:03:49.559 --> 01:03:53.320
of goods and services of the economy, and the as long as the ratio

764
01:03:53.480 --> 01:03:58.280
of money to ratio of goodsiness services
stays. If that, if that stay

765
01:03:58.320 --> 01:04:01.239
is constant, you have no no
inflation. And if if you add more

766
01:04:01.360 --> 01:04:05.280
money, if you add money to
the system faster than you increase goods and

767
01:04:05.360 --> 01:04:10.760
services, then you have inflation.
Um. So all of these COVID sort

768
01:04:10.800 --> 01:04:16.119
of stimulus bills were not paid for. They were they're just generated more currency.

769
01:04:16.280 --> 01:04:24.599
More you know, more money was
was was created because the federal government

770
01:04:25.280 --> 01:04:29.079
the checks never the checks always pass, you know, unless you hit a

771
01:04:29.119 --> 01:04:30.840
dead limit, which there's probably gonna
be some debt limit crisis later this year.

772
01:04:30.960 --> 01:04:35.320
But provided you haven't hit the dead
limit, the the federal government,

773
01:04:35.760 --> 01:04:40.920
unlike state governments or city governments or
individuals, can simply issue more money.

774
01:04:42.039 --> 01:04:45.960
Um. And that's what they did. Um. I mean, as old

775
01:04:45.000 --> 01:04:48.800
saying goes that, there's no there's
no free lunch. Um. So if

776
01:04:48.840 --> 01:04:53.880
you could just issue massive amounts of
money without negative consequences, why don't we

777
01:04:53.920 --> 01:04:57.920
just take that to the lomit make
everyone a trillionaire. Well, I mean

778
01:04:57.920 --> 01:05:00.079
they tried that in Venezuela. That
how that work out? Well, they

779
01:05:00.119 --> 01:05:03.039
had to eat zoo animals. Right, it's not good, you know.

780
01:05:03.400 --> 01:05:09.000
Um. You get to the point
where you know, so wymar Germany type

781
01:05:09.000 --> 01:05:13.239
stuff where you could like take bringing
the cash to the store in a wheelbarrow.

782
01:05:14.519 --> 01:05:18.599
So there's no free lunch. There's
not some ability to issue money and

783
01:05:18.840 --> 01:05:29.480
not have inflation. The this is
just yeah, um, So the inflation

784
01:05:30.159 --> 01:05:34.920
will happen, um. And there's
no filling with the federate is not going

785
01:05:35.039 --> 01:05:45.159
to affect that really um Uh.
But with a high federate can cause a

786
01:05:45.239 --> 01:05:51.079
lot of damage in shifting funds um
in the wrong direction. So um.

787
01:05:51.760 --> 01:05:57.239
The long term return on say the
SMP five hundred I believe is depending on

788
01:05:57.320 --> 01:06:02.880
how you counted, around six percent. So as if the FED real rate

789
01:06:02.920 --> 01:06:08.880
of return starts to approach what the
long term return is on the stock market,

790
01:06:09.519 --> 01:06:12.840
why would you keep any money in
the stock market? You would should

791
01:06:12.840 --> 01:06:15.880
simply buy treasury bills, of course, because the treasure balls is a certainty,

792
01:06:15.920 --> 01:06:20.480
whereas the stock market fluctuates. Right, This is pretty basic. Also,

793
01:06:20.559 --> 01:06:24.639
why would you keep money in a
bank savings account if you can put

794
01:06:24.679 --> 01:06:28.039
it in what's called a money market
account, which is an account that represents

795
01:06:28.079 --> 01:06:33.119
treasury bills. If the treasury bill
money market account gives you, you know,

796
01:06:34.559 --> 01:06:39.119
four to five percent interest and the
bank saving account only gives you two

797
01:06:39.159 --> 01:06:43.119
percent, you'd be a fool to
keep the money in the bank savings account.

798
01:06:44.800 --> 01:06:53.000
So the FED is made a tremendous
mistake by going this high with their

799
01:06:53.119 --> 01:06:57.320
with their rate, and they need
to drop it immediately. Do you think

800
01:06:57.360 --> 01:07:02.320
they will? They? Yes,
they will have no choice but to drop

801
01:07:02.360 --> 01:07:06.199
it. I think later this year. Um. Part of the issue is

802
01:07:06.199 --> 01:07:12.639
that the FED is an old institution
and has a lot of latency in its

803
01:07:12.760 --> 01:07:16.880
data. So it's like driving a
car along a cliff side road, a

804
01:07:16.920 --> 01:07:21.960
windy cliff side road while looking out
the rear view mirror, but not even

805
01:07:23.000 --> 01:07:25.880
actually the re view mirror. A
video that was taken of the rear view

806
01:07:25.880 --> 01:07:30.039
mirror that's three months old. Um
no, if you're on if you're on

807
01:07:30.119 --> 01:07:34.920
a straight road, yeah that works
out okay because nothing's changing. For it's

808
01:07:35.039 --> 01:07:41.320
almost only slight or slightly vending road. But we're more along like the we're

809
01:07:41.360 --> 01:07:45.360
doing the highway one pc trip here. Um and so we really want to

810
01:07:45.400 --> 01:07:49.159
look out the front window when you're
in big s, yes, if you're

811
01:07:49.199 --> 01:07:51.960
on a cliff side road where you
could plunge your doom. So yes,

812
01:07:53.079 --> 01:07:56.119
you want to look out the front
window. You want to look at the

813
01:07:56.679 --> 01:07:59.519
sort of forward commodity prices. Look
look at look at the look at what

814
01:07:59.559 --> 01:08:11.800
the forward attracts are predicting for commodity
prices. Um and uh not some laboriously

815
01:08:11.920 --> 01:08:18.920
slow government data collection process that like
they'll claim to have for example, December

816
01:08:19.079 --> 01:08:23.119
data, that's not December. That's
not the data of December. It's the

817
01:08:23.319 --> 01:08:26.600
data that arrived in December exactly.
If you think about how good is the

818
01:08:26.600 --> 01:08:29.920
government at actually collecting data? Horrible? Yeah, yeah, So It's like,

819
01:08:30.000 --> 01:08:32.159
that's what I mean. It's like
three months old with lots of areas.

820
01:08:32.640 --> 01:08:35.920
So if you had a hundred grand, didn't think making decisions of that

821
01:08:35.960 --> 01:08:41.479
basis was insane. So like,
what should the average non rich person do

822
01:08:42.479 --> 01:08:45.479
on the cusp of what you're describing, which is economic catastrophe? Like,

823
01:08:45.560 --> 01:09:06.800
how do you protect yourself? UM? I think I think probably a smart

824
01:09:06.920 --> 01:09:12.600
move overall. And this this is
guidance that I think applies across the ages.

825
01:09:13.039 --> 01:09:17.119
Is if there are companies in whose
products you believe, UM, buy

826
01:09:17.239 --> 01:09:23.920
and hold the stock UM, and
and when when when whenever else is panicking,

827
01:09:25.199 --> 01:09:30.840
then buy more and everyone else thinks
that the stock is going to the

828
01:09:30.920 --> 01:09:36.039
moon and sell it, you know, sort of the buy UM low cell

829
01:09:36.159 --> 01:09:40.800
high. You're not an index fund
guy, like you pick specific stocks.

830
01:09:44.960 --> 01:09:49.279
I that's my right. You have
to say, like, what is the

831
01:09:49.359 --> 01:09:53.920
purpose of a company? Why should
a company exist? Um? A company

832
01:09:53.960 --> 01:09:59.880
exists. It's a group of people
collected together to provide products and services.

833
01:10:00.079 --> 01:10:01.840
It's not it's not a thing in
and of itself. It's just a group

834
01:10:01.840 --> 01:10:04.239
of people. That's like, it's
hard if it was just one person making

835
01:10:04.640 --> 01:10:08.760
You can make cupcakes yourself, but
you can't make cars by yourself. Yes,

836
01:10:09.359 --> 01:10:15.560
so if so that so therefore,
the value of a company is a

837
01:10:15.640 --> 01:10:20.279
function of the um quality of the
products and services that it's that it has

838
01:10:20.359 --> 01:10:24.560
created and will create. And so
if if if there's a company that you

839
01:10:24.640 --> 01:10:28.319
think, well, this company's got
a lot of exciting products that I think

840
01:10:28.319 --> 01:10:32.199
are awesome. Um, the current
products are good, that's probably a company

841
01:10:32.399 --> 01:10:36.560
to invest in, because that's the
reason companies exist, to these goods and

842
01:10:36.600 --> 01:10:43.800
services that you like. And so
um no, I mean there's some caveats

843
01:10:43.840 --> 01:10:47.439
here to make sure you're you're not
like investing when when it's like the hottest

844
01:10:47.520 --> 01:10:50.279
thing, you know, because then
it's going to have be at a temporary

845
01:10:50.399 --> 01:10:56.159
high. Um, but you know
what when when when it's not sort of

846
01:10:56.199 --> 01:10:59.000
at a weirdly temporary high. I
think just generally looking at a company and

847
01:10:59.039 --> 01:11:00.079
saying, well, I like the
products of that company, and I like

848
01:11:00.159 --> 01:11:05.159
where they're going, and the management
seems sensible that then I think buying and

849
01:11:05.239 --> 01:11:10.399
holding that stock is probably the right
move. Um. I'm probably doing that

850
01:11:10.479 --> 01:11:15.119
with with a few companies. Um, that's what I'd recommend. Um.

851
01:11:16.600 --> 01:11:20.920
I think the on the I mean, I could really go on at length

852
01:11:20.960 --> 01:11:29.039
about the financial system and the stock
market everything. Um, but the I

853
01:11:29.079 --> 01:11:30.960
mean, these days, I think
we've gotten a little too far into the

854
01:11:31.039 --> 01:11:36.039
index passive fund world. Um,
like somebody at some point it's got to

855
01:11:36.079 --> 01:11:43.520
make a decision. Um. You
know, I couldn't agree more. Yeah.

856
01:11:43.960 --> 01:11:46.159
And by the way, there's like
betting on both red and black.

857
01:11:46.439 --> 01:11:50.359
I mean it doesn't Yeah, well, betting on red and black in a

858
01:11:50.439 --> 01:11:55.439
casino situation where it could come up
green, right, and you're you're bound

859
01:11:55.479 --> 01:11:58.840
to lose. Longer you play,
the worst you do. Now that the

860
01:11:58.880 --> 01:12:00.279
stock market is is kind of the
opposite of a casino, which is the

861
01:12:00.359 --> 01:12:04.399
longer you play, the more likely
you are to succeed. If that historically

862
01:12:04.399 --> 01:12:08.199
has been the case and I think
will continue to be the case. Um

863
01:12:09.279 --> 01:12:15.079
so uh, but it's really just
important not to panic if you if you

864
01:12:15.159 --> 01:12:17.600
buy a stock and you read something
terrible in the newspaper, do you want

865
01:12:17.640 --> 01:12:21.600
to just remember the news that's got
a negative bias? Uh, Just and

866
01:12:21.800 --> 01:12:26.199
think about what other products of the
company still sound. Does it have a

867
01:12:26.239 --> 01:12:29.560
good product roadmap? Do you believe
in the management? If so, ignore

868
01:12:29.600 --> 01:12:31.680
what the precess or if the price
drops when when the negative audicle, buy

869
01:12:31.720 --> 01:12:34.960
a stock. So here's here's my
last question. And you mentioned the press.

870
01:12:35.119 --> 01:12:38.720
You've been the subject of press coverage
for you know, like a long

871
01:12:38.800 --> 01:12:44.319
time, but very intense media coverage
for the last year. Yeah. Sure,

872
01:12:45.079 --> 01:12:49.720
well it seems that way anyway.
Um, how is your opinion of

873
01:12:50.119 --> 01:13:05.720
the press changed? Um? Well, so my first company way back in

874
01:13:05.760 --> 01:13:12.960
the day and the Precambrian era of
the Internet or the Worldwide Web is up

875
01:13:13.000 --> 01:13:15.800
to We actually helped bring a number
of the media organizations online. So most

876
01:13:15.840 --> 01:13:20.399
newspapers are not online. We helped
bring hundreds of newspapers and magazines online for

877
01:13:20.439 --> 01:13:26.239
the first time. We added a
trevenus amount of functionality to their websites with

878
01:13:26.359 --> 01:13:30.840
our software. The New York Times
Company and night Rider were major investors and

879
01:13:31.000 --> 01:13:34.439
were on the board. I spent
a lot of time in newsrooms, so

880
01:13:34.520 --> 01:13:41.720
I'm not unfamiliar with the media.
I got to see it firsthand all the

881
01:13:41.760 --> 01:13:45.680
way back in nineteen ninety six,
so it's been a while. Sort of

882
01:13:45.680 --> 01:13:55.319
traditional media certainly had revenue challenges because
as online advertising is increased and it's much

883
01:13:55.399 --> 01:13:58.479
more measurable and much more sort of
direct. You can say, like I

884
01:13:58.640 --> 01:14:03.840
spent this amount and about this this
output like it's interactive, unlike say a

885
01:14:03.880 --> 01:14:09.359
newspaper or broadcast TV. UM,
you're kind of guessing with the newspaper and

886
01:14:09.399 --> 01:14:13.960
broadcast TV. UM, where there's
if if something's online, you can tell

887
01:14:14.000 --> 01:14:17.880
immediately that that person saw the ad
and bought the product. That's that's very

888
01:14:18.000 --> 01:14:25.840
very immediate. So UM, and
it's it's it's actually more effective if the

889
01:14:25.920 --> 01:14:30.520
advertising is customized to the individual,
so the advertising is more likely to be

890
01:14:30.520 --> 01:14:33.760
irrelevant, whereas broadcast if if it's
being shown to everyone, it's going to

891
01:14:33.760 --> 01:14:40.119
be irrelevant to most people. UM. The result of that has been a

892
01:14:40.239 --> 01:14:47.239
huge shift in advertising revenue from newspapers
and TV to uh, you know,

893
01:14:47.359 --> 01:14:50.840
the Googles and facebooks of the world, and a tiny bit to Twitter.

894
01:14:51.800 --> 01:14:57.199
I think Twitter gets like one one
percent of advertising revenue. It's quite tiny.

895
01:14:57.920 --> 01:15:04.079
UM. So the this is meant
a shrinking pie obviously for most of

896
01:15:04.199 --> 01:15:12.399
the traditional media companies UM and made
them more desperate to get cliques, to

897
01:15:12.520 --> 01:15:18.479
get to get you know, get
attention. UM. And it just made

898
01:15:18.520 --> 01:15:21.960
them when when you know, when
they were, when they're in a sort

899
01:15:21.960 --> 01:15:28.800
of desperate state, they will then
tend to really push headlines that get the

900
01:15:28.840 --> 01:15:31.800
most cliques, whether those headlines are
accurate or not. Um So, it's

901
01:15:31.960 --> 01:15:38.319
resulted in my view, I think
I think most people would agree a less

902
01:15:39.000 --> 01:15:45.840
truthful, less accurate news um so
because they just got to get a rise

903
01:15:45.920 --> 01:15:50.199
out of people. Um. And
I think it's also increased the negativity of

904
01:15:50.319 --> 01:15:57.560
the news because I think we humans
instinctually respond more to I think we have

905
01:15:58.319 --> 01:16:08.000
an instinctual negative which which kind of
makes sense in that like if like let's

906
01:16:08.000 --> 01:16:14.279
say, like it's more important to
remember where where was the lion or where

907
01:16:14.399 --> 01:16:17.680
was the tribe that wants to kill
my tribe than where is the bushworth berries?

908
01:16:18.479 --> 01:16:23.680
Like one's like a permanent negative outcome
and the other is like, well

909
01:16:23.720 --> 01:16:30.239
I might go hungry, So meaning
like there's an asymmetry and um sort of

910
01:16:30.319 --> 01:16:36.920
involved asymmetry and negative versus positive stuff
and and and also historically the negative stuff

911
01:16:38.039 --> 01:16:43.039
would have been quite approximate, like
it would have been near represented a real

912
01:16:43.159 --> 01:16:46.960
danger to you as a person,
um if you heard negative news, because

913
01:16:47.560 --> 01:16:50.520
historically, you know, like a
few hundred years ago, we're not hearing

914
01:16:50.560 --> 01:16:54.720
about what negative things are happening on
the other side of the world or on

915
01:16:54.720 --> 01:16:57.159
the other side of the country,
we're only we're hearing about negative things in

916
01:16:57.199 --> 01:17:01.239
our village, things that could actually
have a bad effect to you, whereas

917
01:17:01.279 --> 01:17:04.479
now we're hearing about I mean,
the news very often seems to attempt to

918
01:17:05.159 --> 01:17:10.479
answer the question what is the worst
thing that happened on Earth today? And

919
01:17:10.600 --> 01:17:13.800
we wonder why you said after reading
that, you know, um, and

920
01:17:13.920 --> 01:17:18.840
then use the most inflammatory language,
you know, because every day they got

921
01:17:18.920 --> 01:17:23.680
to sell, sell the advertising,
even if it's it happened to be a

922
01:17:23.720 --> 01:17:30.399
slow newsday. Do you read any
legacy media once? When I read a

923
01:17:30.439 --> 01:17:39.079
lot so UM. I mean,
I really get most of my news from

924
01:17:39.079 --> 01:17:45.279
Twitter at this point, so it
is the number one news news source I

925
01:17:45.399 --> 01:17:53.359
think in the world at this point. So, and thus all the more

926
01:17:53.399 --> 01:17:58.800
important that we strive to to be
accurate. And it's not just a question

927
01:17:58.840 --> 01:18:05.000
of accuracy, but we also need
to allow the people to develop the narratives

928
01:18:05.039 --> 01:18:10.920
that are of interest to them.
So it's possible for UM news to be

929
01:18:11.680 --> 01:18:16.159
technically truthful, but not that,
but they're still deciding what the narrative is,

930
01:18:17.239 --> 01:18:21.760
like, like let's say you want
to um like let's say it took

931
01:18:21.920 --> 01:18:26.279
a photo of someone and they had
a little zip Um. Now you could

932
01:18:26.520 --> 01:18:30.960
zoom in on the zip and make
it look gigantic like Mount Vesuvius um,

933
01:18:30.159 --> 01:18:34.159
and it is still true that they
have a ZiT It's just not the size

934
01:18:34.159 --> 01:18:38.960
of Mount Vesuvius. And they you
know, it doesn't properly reflect their face.

935
01:18:40.640 --> 01:18:43.039
Their face is not one giant set. But you could you could say,

936
01:18:43.119 --> 01:18:45.840
like, well, it's true,
but have they lied. They haven't.

937
01:18:47.039 --> 01:18:50.239
You know, they're just happened to
zoom in on the ZiT um and

938
01:18:50.399 --> 01:18:55.680
not look at the rest of the
face type of thing. So, um,

939
01:18:56.960 --> 01:19:03.000
I want to saying is that that
choice of narrative is it is extremely

940
01:19:03.039 --> 01:19:10.920
important. And at the point in
which if there's only like say, half

941
01:19:10.960 --> 01:19:14.119
a dozen editors in chief, or
maybe even few of them, maybe it's

942
01:19:14.119 --> 01:19:16.520
only three or four that are deciding
what the narrative is, what's gonna be

943
01:19:16.520 --> 01:19:25.279
on the front page, then um, you know that that that's that's a

944
01:19:25.359 --> 01:19:28.399
form of manipulation of public opinion,
and I think the public often doesn't appreciate

945
01:19:28.600 --> 01:19:30.119
and it's perhaps the most pernicious of
all. That's right, because it's the

946
01:19:30.159 --> 01:19:32.840
most subtle. Yeah, it's most
subtle. They haven't said an untrue thing,

947
01:19:32.880 --> 01:19:35.479
They've just chosen. What they're going
to focus on. A man called

948
01:19:35.520 --> 01:19:42.720
Douglas Mackie's facing ten years in prison
for posting what he believed funny memes on

949
01:19:42.800 --> 01:19:46.000
Twitter. What do you make of
that case. I don't know the details

950
01:19:46.000 --> 01:19:50.000
of that case. Um I've you
know, I've read a little bit about

951
01:19:50.039 --> 01:19:58.880
it. You may you probably know
more about it than I do. I

952
01:19:58.920 --> 01:20:02.600
certainly don't think so much go to
prison for a long period of time for

953
01:20:02.720 --> 01:20:06.720
posting memes on Twitter, in which
case we're gonna have a very full prison.

954
01:20:08.520 --> 01:20:14.960
Um so. And if if we're
talking about election interference, well there's

955
01:20:15.039 --> 01:20:19.520
quite a few people that should be
on trial for that, for much far

956
01:20:19.600 --> 01:20:24.840
more serious crimes than than memes on
on Twitter, far more serious. Yes,

957
01:20:25.760 --> 01:20:29.239
the Twitter frials kind of showed that. I think, yes, um

958
01:20:30.159 --> 01:20:33.760
so, you know, unless unless
the person really did I I said,

959
01:20:33.760 --> 01:20:38.920
I don't. I don't know.
I don't everything that we're showing at the

960
01:20:38.960 --> 01:20:43.640
trial. Has he been convicted?
Is this? Yes? He was so

961
01:20:43.880 --> 01:20:50.439
victed on Friday, unit unanimous jury
verdict. Yes. What was the venue

962
01:20:51.079 --> 01:20:55.800
were city? Okay? It was
a hug It was in Brooklyn, and

963
01:20:55.840 --> 01:20:59.520
it was a hung jury and hung
jury. There's not unanimous, then well

964
01:20:59.560 --> 01:21:03.680
the judge prodded the jury, okay, and and they reached unanimous guilty verdict.

965
01:21:04.279 --> 01:21:08.119
It'll be appealed. So how many
what percentage of your staff did you

966
01:21:08.199 --> 01:21:12.399
fire at Twitter? One of the
great business stories of the year. I

967
01:21:12.439 --> 01:21:15.479
think we're about we're about twenty percent
of the original size, So eighty percent

968
01:21:15.600 --> 01:21:21.199
left. Yes, so a lot
of people voluntarily, sure, sure,

969
01:21:21.399 --> 01:21:26.920
but but it's eighty percent are gone
from the day. So how do you

970
01:21:27.039 --> 01:21:30.319
run the company with only twenty percent
of the staff. It turns out you

971
01:21:30.359 --> 01:21:34.039
don't need all that many people to
run Twitter, but eighty percent, that's

972
01:21:34.079 --> 01:21:40.319
a lot um, yes of it. I mean, if you're not trying

973
01:21:40.359 --> 01:21:46.800
to run some sort of glorified activist
organization with you don't care that much about

974
01:21:46.800 --> 01:21:49.560
censorship, then you can really let
go of a lot of people. Turns

975
01:21:49.560 --> 01:21:59.039
out how many others without naming names, but how many I had dinner when

976
01:21:59.039 --> 01:22:01.640
somebody who runs a big gum but
he recently said I'm really inspired by Euan

977
01:22:01.720 --> 01:22:04.880
And I said the free speech stuff
he goes, no firing the staff stuff.

978
01:22:05.439 --> 01:22:12.960
Um, how many other CEOs have
come to you um to talk about

979
01:22:13.039 --> 01:22:19.520
this? Um? You know,
I spent a lot of time at work,

980
01:22:19.920 --> 01:22:24.319
so it's not like I'm meeting with
lots of people. They see what

981
01:22:24.439 --> 01:22:30.039
I what actions I've taken? Um
and UM. But I think we just

982
01:22:30.199 --> 01:22:33.359
had a situation at Twitter where it
was absurdly overstaffed. You know, so

983
01:22:34.079 --> 01:22:36.840
it wasn't you know, like you
look at say like, what does it

984
01:22:36.880 --> 01:22:42.039
really take to operate Twitter? Um? You know most of what we're talking

985
01:22:42.039 --> 01:22:47.479
about here is a group text service
at scale? Um, Like how many

986
01:22:47.520 --> 01:22:51.000
people are really needed for that?
You know? Um? And if you

987
01:22:51.039 --> 01:22:58.000
look at the you say, like
what has been the product development over time

988
01:22:58.279 --> 01:23:00.840
with Twitter? And you're like,
so, like you know, years versus

989
01:23:00.960 --> 01:23:03.840
product improvements, and it's like a
pretty flat line. So what are they

990
01:23:03.920 --> 01:23:09.279
doing? You know, it took
a year or to add an edit button

991
01:23:09.359 --> 01:23:12.640
that doesn't work most of the time. I mean, this is I feel

992
01:23:12.640 --> 01:23:17.079
like if there's a comedy situation here. You know, you're not making cars.

993
01:23:17.239 --> 01:23:21.600
You know, it's very difficult to
make cars or get rockets orbit,

994
01:23:23.159 --> 01:23:27.720
so UM, you know, the
real question is like how did it get

995
01:23:27.800 --> 01:23:32.319
so absurdly overstaffed? This is insane? UM? So any right, that's

996
01:23:32.960 --> 01:23:36.319
and it's clearly working. In fact, I think it's working better than ever.

997
01:23:36.399 --> 01:23:43.079
It's about we've increased the responsiveness of
the system by in some cases over

998
01:23:43.159 --> 01:23:50.680
eighty percent. The there's a core
piece of code for generating the timeline which

999
01:23:50.760 --> 01:23:56.680
is run literally billions of times a
day. We've cut that code from seven

1000
01:23:56.800 --> 01:24:01.840
hundred thousand lines to seventy thousand lines
run. Yeah, and and the code

1001
01:24:01.880 --> 01:24:06.119
efficiency by over eighty percent, like
meaning that how much computer is necessary to

1002
01:24:06.159 --> 01:24:12.279
render the timeline by eighty percent.
I mean this is a you know,

1003
01:24:12.399 --> 01:24:16.840
in four four or five months,
we've we've increased the video time from um

1004
01:24:17.960 --> 01:24:23.159
roughly two minutes or best case ten
minutes to now two hours, so you

1005
01:24:23.199 --> 01:24:26.640
can put to ours video on Twitter. Will ASO going to be increasing that

1006
01:24:26.760 --> 01:24:31.159
to where there's no no meaningful limit. We've increased the tweet length from total

1007
01:24:31.279 --> 01:24:35.479
forty characters to four thousand, will
be increasing that to where there's gain no

1008
01:24:35.560 --> 01:24:39.520
meaningful length to if you want to
post a novel on Twitter is Sue'll do

1009
01:24:39.600 --> 01:24:46.079
it. Um And you know we
as as everyone saw on Friday, we

1010
01:24:46.319 --> 01:24:51.479
open source the super embarrassing recommendation algorithm, which we will are taking apart and

1011
01:24:51.800 --> 01:24:56.000
invest rating, which is exactly what
I hope they would do, um and

1012
01:24:56.119 --> 01:25:00.239
pointing out all the nonsense. Um. And we're gonna open so once more.

1013
01:25:00.680 --> 01:25:08.239
Um and we're going to subject to
public public review. We're also going

1014
01:25:08.279 --> 01:25:12.720
to get criticized line because people will
point out all of the foolish things that

1015
01:25:12.800 --> 01:25:15.000
are out that that are happening in
the code. But then we'll fix it.

1016
01:25:15.079 --> 01:25:17.960
We'll fix it fast and in full
public view. UM And I think

1017
01:25:19.520 --> 01:25:23.920
that's the kind of thing that that
owns the public trust. You know,

1018
01:25:24.039 --> 01:25:28.000
if if if because like don't take
my word for it, this is we

1019
01:25:28.800 --> 01:25:30.000
you can literally read the code,
and you can read what people say about

1020
01:25:30.039 --> 01:25:34.239
the code. Um and and you
can see the improvements that we make,

1021
01:25:34.399 --> 01:25:41.000
and you can see we'll see the
like in real time live see see see

1022
01:25:41.000 --> 01:25:46.880
it get better. UM. So
my prediction is that this I would be

1023
01:25:46.920 --> 01:25:49.920
surprised if this does not lead the
public to think, Okay, this this

1024
01:25:50.119 --> 01:25:56.359
is something that I can trust.
Um. I mean I think far more

1025
01:25:56.399 --> 01:26:00.720
trust trust way than say other social
media opens thes that have some mysterious black

1026
01:26:00.800 --> 01:26:05.039
box that they refuse to show the
show how it works. What are they

1027
01:26:05.039 --> 01:26:11.239
trying to hide? What are these
just not good things? Yeah, if

1028
01:26:11.560 --> 01:26:14.920
they had to have somebody to hide
one and they show it, but there

1029
01:26:15.199 --> 01:26:26.920
it's a proprietary business secret. Yeah. Sure. So so we're trying to

1030
01:26:27.079 --> 01:26:30.640
make make Twitter the most trusted place
on the internet, that the where you

1031
01:26:30.680 --> 01:26:35.920
can get the you know, the
least untrustworthy place on the Internet. I

1032
01:26:35.920 --> 01:26:40.359
don't think anyone should trust the internet, but maybe we can make Twitter the

1033
01:26:40.479 --> 01:26:45.960
least untrustworthy. Um. And in
a way you can see a wide range

1034
01:26:45.000 --> 01:26:49.199
of political opinions, so including ones
you disagree with. I think people should

1035
01:26:49.199 --> 01:26:54.880
be exposed to things they disagree with. It shouldn't just be continuous self reinforcement

1036
01:26:54.960 --> 01:27:00.760
of like what you know. So, um, that's the goal, and

1037
01:27:00.319 --> 01:27:05.079
I think we're making some good progress
in that direction. I feel good about

1038
01:27:05.079 --> 01:27:10.680
where things are going, and we
definitely want to have things as as sort

1039
01:27:10.680 --> 01:27:16.239
of cleaned up as possible before the
elections. If there's any manipulation that we're

1040
01:27:16.279 --> 01:27:20.119
aware of, make that, make
the public aware of that, and just

1041
01:27:20.800 --> 01:27:28.920
uh, like I said, try
to get the truth to the people as

1042
01:27:29.239 --> 01:27:29.760
best we can.