WEBVTT

1
00:00:00.000 --> 00:00:05.599
Explain what neuralink is and what the
goal of it is. We put a

2
00:00:05.679 --> 00:00:11.400
chip in your brain to control your
mind. Yeah, okay, all right,

3
00:00:11.480 --> 00:00:16.079
concerns, Yeah, jump right in, step right up, who wants

4
00:00:16.120 --> 00:00:23.000
one? So neuralink you'll be able
to see neural link coming from a very

5
00:00:23.000 --> 00:00:28.000
long distance because any device that you
implanted in a human is you have to

6
00:00:28.199 --> 00:00:32.799
go through a million, so many
tests. It moves very slowly. You

7
00:00:32.920 --> 00:00:37.399
just do a few people at a
time and then you go to extreme lengths

8
00:00:37.439 --> 00:00:41.439
to prove safety. You have to
go through the FDA approvals. Like we're

9
00:00:41.479 --> 00:00:47.520
not trying to side step any you
know, regulatory approvals doing everything. You

10
00:00:47.719 --> 00:00:53.280
buy the book and with maximum we're
really actually we're going far beyond what the

11
00:00:53.359 --> 00:00:58.000
requirements are of the FDA from a
safety standpoint, and the initial devices will

12
00:00:58.039 --> 00:01:03.439
really just be pretty basic. It'll
be about restoring functionality to people who've lost

13
00:01:03.479 --> 00:01:07.519
their connection between their their brain and
their body. So you can imagine like

14
00:01:07.519 --> 00:01:14.079
if say Stephen Hawking could talk communicate
as fast as somebody with a fully functioning

15
00:01:14.120 --> 00:01:17.799
body, that'll be amazing. So
that's like the what we're trying to do

16
00:01:18.200 --> 00:01:23.959
that that's our first application is to
restore functionality to quatter fleix, tetraplegics and

17
00:01:23.200 --> 00:01:27.680
people who have just for whatever reason, no longer have a connection between or

18
00:01:27.719 --> 00:01:34.159
if a limited connection between their their
their their brain and their body. And

19
00:01:34.280 --> 00:01:40.239
then the second application would be restoration
of eyesight. So if somebody's gone completely

20
00:01:40.280 --> 00:01:45.719
blind, maybe even just lost the
optic nerve, you can actually still directly

21
00:01:46.680 --> 00:01:53.519
assimilate the neurons in the visual part
of the cortex, so you can give

22
00:01:55.280 --> 00:02:00.680
direct vision to the brain. In
fact, you could actually depending upon what

23
00:02:00.799 --> 00:02:04.480
cameras you use, you can actually
see in different wavelengths. Okay, you

24
00:02:04.560 --> 00:02:07.840
know, like Jory LaForce from like
you know, because I have that,

25
00:02:07.039 --> 00:02:13.800
Like I actually watched like an episode
of Start Take the Next Generation with special

26
00:02:13.800 --> 00:02:17.840
effects compared to what we're used to, like you know, not that great,

27
00:02:17.879 --> 00:02:22.599
but he's got like the wrap round
you know, glasses, and you

28
00:02:22.639 --> 00:02:28.680
can see in different wavelengths, so
you can see like ultra violet, infrared

29
00:02:28.719 --> 00:02:30.639
and that kind of thing. So
you can actually do that. You could

30
00:02:30.639 --> 00:02:35.840
say, like you could see in
radar if you want. And what's the

31
00:02:36.039 --> 00:02:38.280
what's the long term what's the long
term goal for it? Because I think

32
00:02:39.159 --> 00:02:46.360
myself and others I think the first
part is the first part sounds sounds fine,

33
00:02:46.360 --> 00:02:50.080
sounds yeah, that's like card argue
with yeah yeah. But long term,

34
00:02:50.159 --> 00:02:53.919
I think, I mean, the
concern that people have is is this

35
00:02:53.120 --> 00:02:59.080
just leading us into this dystopian transhumanist
future? Where where does it go?

36
00:02:59.400 --> 00:03:01.840
What are the what are the ethical
boundaries of it? Well, I mean

37
00:03:02.639 --> 00:03:06.680
the thing I wanted to empasize is
that it's not going to like sort of

38
00:03:06.719 --> 00:03:09.400
pounce on us overnight. You'll see
it coming. It's going to be very

39
00:03:09.400 --> 00:03:15.360
slow. In fact, I really
think that artificial general intelligence or digital superintelligence

40
00:03:15.719 --> 00:03:22.759
is likely to arrive before we have
really advanced in your links at least that's

41
00:03:22.759 --> 00:03:30.159
where the trend is right now.
But ultimately the idea would be to achieve

42
00:03:30.599 --> 00:03:38.840
symbiosis between our biological mind and our
kind of digital mind. So we're already

43
00:03:38.919 --> 00:03:42.719
kind of a cyborg. If you
think of like your phone and your computers

44
00:03:42.759 --> 00:03:46.919
as an extension of yourself. In
fact, like if you leave your phone

45
00:03:46.919 --> 00:03:49.759
behind, it's like you have missing
limbs syndrome. You're like, where do

46
00:03:49.800 --> 00:03:54.039
it go? You know, And
so the phone is kind of an extension

47
00:03:54.039 --> 00:03:59.919
of ourselves, like computer is.
The various applications that we use are an

48
00:04:00.000 --> 00:04:01.960
extention of self. So we are
already a cyborg. It's just that the

49
00:04:02.000 --> 00:04:10.560
interfaces with our eyes and our fingers
and and that that interface especially output.

50
00:04:10.639 --> 00:04:15.240
The rate at which we can type
words into a phone or a computer just

51
00:04:15.400 --> 00:04:19.560
it's very slow. Our input is
much better because with the data rate from

52
00:04:19.639 --> 00:04:26.040
vision is you know, I don't
know, many thousands of times maybe a

53
00:04:26.040 --> 00:04:30.079
million times better than the rate of
which we can output. So input is

54
00:04:30.120 --> 00:04:38.319
like maybe roughly a million times better
than output. And so so what what

55
00:04:38.439 --> 00:04:44.079
a neural device can do is improve
that bandwidth, allow allow you to be

56
00:04:45.079 --> 00:04:50.000
sort of much more symbiotic with your
the AI extension of yourself. So you

57
00:04:50.040 --> 00:04:54.759
can think of like like a human
brain really is could be augubly divided into

58
00:04:54.879 --> 00:05:00.800
two parts. One is kind of
like the primitive brain, the reptile brain

59
00:05:00.839 --> 00:05:02.399
it's sometimes called, you know,
it's like got a sort of basic instinct.

60
00:05:02.560 --> 00:05:08.319
And then we've got the cortex,
like the higher level thinking, planning

61
00:05:08.360 --> 00:05:13.079
and that kind of thing. The
two operate symbiotically. So I haven't yet

62
00:05:13.160 --> 00:05:16.560
met anyone who wants to delete their
olympic system or deleat their cortex. Everyone's

63
00:05:16.639 --> 00:05:19.439
quite happy having both they're like,
I like that the way there is,

64
00:05:19.519 --> 00:05:24.600
you know, but your cortex is
way smarter than your Olympic system. So

65
00:05:24.959 --> 00:05:27.720
but the irony is that even though
the cortex is way smarter than the Limbic

66
00:05:27.800 --> 00:05:30.839
system, most of what it's doing, you're trying to make the Olympic system

67
00:05:30.879 --> 00:05:33.399
happy. It's like Limbic systems hunky
hungry. Okay, let's get some food.

68
00:05:34.800 --> 00:05:39.079
The Olympic system is horny. Okay, let's you know, have sex

69
00:05:39.199 --> 00:05:43.839
or whatever. See, I mean
the sheer amount of of of effort,

70
00:05:43.920 --> 00:05:49.199
the you know, the cortexes have
of all the humans have put into trying

71
00:05:49.240 --> 00:05:56.319
to get laid is insane. It's
got to keep the species gone. Yeah,

72
00:05:56.360 --> 00:05:59.959
but it's not even with even if
you know, it's not for appropriate

73
00:06:00.240 --> 00:06:03.399
like the like the Limbic system is. It's it's really like too simple to

74
00:06:03.560 --> 00:06:09.879
understand that, like sex does not
resulting in procreation because for almost all of

75
00:06:09.959 --> 00:06:15.160
human existence that it did. Both
control is a very recent thing. So

76
00:06:15.399 --> 00:06:23.319
the Limbic system is like trying to
insent procreation. And but but now we

77
00:06:23.399 --> 00:06:27.480
can effectively hack Thelympic system by doing
procreation, but by having sex without a

78
00:06:27.519 --> 00:06:35.000
procreation. So which is I guess
maybe part of the issue with you know,

79
00:06:35.040 --> 00:06:38.279
why do we have more kids is
like in the past, that would

80
00:06:38.279 --> 00:06:42.040
just happened because you know, in
order to make the limbic system happy,

81
00:06:42.040 --> 00:06:45.879
you'd have sex. You didn't have
both control, so you pop out some

82
00:06:45.959 --> 00:06:50.720
goods. Yes, So that's that's
how it used to be. So so

83
00:06:51.000 --> 00:06:55.160
so so even though you've got cortex
that's way smarter than the limbic system,

84
00:06:55.519 --> 00:06:59.720
the cortex is still basically just trying
to make the limbic system happy. And

85
00:07:00.920 --> 00:07:02.519
and then if you think of like
the computers as a sort of a third

86
00:07:02.639 --> 00:07:08.279
layer, the AI is a third
layer. It's not necessarily the case that

87
00:07:08.439 --> 00:07:15.439
the AI would be acting contrary to
our interests. I think if it's closely

88
00:07:15.519 --> 00:07:21.759
linked with a biological intelligence, I
think it could actually be just again trying

89
00:07:21.800 --> 00:07:25.319
to make the cortex happy, which
is trying to make the limit system happy.

90
00:07:25.680 --> 00:07:28.360
So I think we'll put even more
computing power to try to get laid.

91
00:07:28.759 --> 00:07:36.240
Basically, now the AI is going
to help you get laid. I

92
00:07:36.319 --> 00:07:40.279
think. I think what it is
that that concerns people about it number one,

93
00:07:40.480 --> 00:07:42.600
you know number one being new,
But I think the main thing is

94
00:07:43.680 --> 00:07:47.519
is the the fact that this is
something that's like inside inside your body.

95
00:07:48.480 --> 00:07:53.959
Right, I think that the idea
of putting a chip ins the idea of

96
00:07:54.000 --> 00:07:56.639
someone putting a chip in my brain. I just have a bit, I

97
00:07:56.720 --> 00:08:03.399
have a visceral. No, I
have I have a very an optional thing.

98
00:08:03.480 --> 00:08:05.480
I mean it's going to you know, would it would it be the

99
00:08:05.639 --> 00:08:09.959
long term? I mean, I
hope so I think it should be optional

100
00:08:09.519 --> 00:08:13.439
too bizarre if it was not.
No, I think it's kind of like

101
00:08:13.920 --> 00:08:18.839
kind What was the reason I ask? Is because smartphones right now are sort

102
00:08:18.879 --> 00:08:24.439
of optional. But if you don't
have a smartphone the way everything is sort

103
00:08:24.439 --> 00:08:30.160
of designed around society, you,
yes, you cannot have a smartphone,

104
00:08:30.639 --> 00:08:35.600
but it's a massive hindrance and disadvantage
in many ways. True, a smartphone

105
00:08:35.679 --> 00:08:39.320
is almost essentral and a mout in
society to do things. Yeah, so

106
00:08:39.480 --> 00:08:43.320
could a neuralink end up in that
same sort of place. It's possible,

107
00:08:43.360 --> 00:08:48.080
But I'm just saying if it does, it's it's it's many decades from now,

108
00:08:48.159 --> 00:08:50.919
so it's not like, you know, it's not today's problem. Yeah,

109
00:08:52.039 --> 00:08:56.960
I'd worry about more about digital superintelligence. I'd worry about let's try to

110
00:08:56.960 --> 00:09:01.960
avoid world War three, let's sure
we're at least having enough kids to sustain

111
00:09:03.039 --> 00:09:07.639
our population, you know, basic
stuff like that. Yeah, what's your

112
00:09:07.720 --> 00:09:13.279
greatest concern with artificial intelligence? Because
I know that you have worked with it,

113
00:09:13.480 --> 00:09:20.360
but also I know it's something that
you personally have massive concerns about as

114
00:09:20.440 --> 00:09:26.360
well, which you've voiced in this
podcast and elsewhere. Yeah. I think

115
00:09:26.399 --> 00:09:33.000
that actually the most likely outcome for
artificial intelligence is that it is good.

116
00:09:33.080 --> 00:09:37.720
That it will improve our lives most
likely, but there's some chance that it

117
00:09:37.000 --> 00:09:41.720
will not, And I think we
just need to be cognizant of that and

118
00:09:41.000 --> 00:09:46.360
understand that it's a powerful technology.
It's a double edged sword, and we

119
00:09:46.480 --> 00:09:50.559
need to put a lot of effort
into ensuring that we have a good AI

120
00:09:50.679 --> 00:09:54.840
outcome and not a bad one.
What does a good one look like?

121
00:09:56.919 --> 00:10:00.639
Well, if I'd recommend people to
read the in Banks books, the culture

122
00:10:00.679 --> 00:10:07.799
books, that's the best representation of
a positive AI human future that I've seen.

123
00:10:09.399 --> 00:10:16.679
Okay, so that's my recommendation.
I mean it effectively a I I

124
00:10:16.720 --> 00:10:24.240
think it will massively enhance human ability. It's like just a massive amplifier of

125
00:10:26.000 --> 00:10:33.639
human ability, just like the computer
was. So really is just a question

126
00:10:33.679 --> 00:10:39.960
of like does does it the some
risk that it doesn't merely amplify human activity,

127
00:10:39.000 --> 00:10:43.879
but it starts basically just being in
charge. And you know there's some

128
00:10:45.120 --> 00:10:52.120
risk at me view humanity negatively and
decide that we're a blied on the earth.

129
00:10:52.159 --> 00:10:54.480
Like like I think, like a
very dangerous thing would be if if

130
00:10:54.519 --> 00:11:00.519
the if the sort of human extinctionist
philosophy somehow got into AI, that would

131
00:11:00.559 --> 00:11:03.000
be bad. Like that guy that
was part page in The York Times saying,

132
00:11:03.399 --> 00:11:05.440
you know, the eight billion people
on Earth, we don't want him

133
00:11:05.480 --> 00:11:09.720
coding it. We don't want him
coding you. Yeah, nice to say

134
00:11:09.399 --> 00:11:11.840
what. And there's quite a few
people in the cicular words who have that

135
00:11:13.120 --> 00:11:18.120
either explicit or implicit extinctionis sphere.
Yes, whether if your humanity is a

136
00:11:18.159 --> 00:11:20.879
blight on the face the Yeah,
there's a lot of anti natalists in the

137
00:11:22.080 --> 00:11:26.879
in the world and people who actually
anti human yes, yes, like you

138
00:11:26.919 --> 00:11:30.279
said, people who don't have the
pride in being human, right, these

139
00:11:30.279 --> 00:11:35.000
people guilty or even say that that's
a speciesism you know that's yeah. Yeah,

140
00:11:35.159 --> 00:11:39.000
I mean I'm a human supremacist.
I'll be open about that one.

141
00:11:39.159 --> 00:11:41.200
Yeah. I mean I think we've
got to fight for team Human, Like,

142
00:11:41.360 --> 00:11:43.720
if we don't fight for as humans, who who will? Yeah?

143
00:11:43.759 --> 00:11:48.000
I mean if the Earth exists and
humans do not, then it's like we're

144
00:11:48.039 --> 00:11:52.600
saving the planet for who at that
point. Yeah. What do you think

145
00:11:52.639 --> 00:11:54.960
about AI replacing jobs though, because
that's a that's a big concern. Even

146
00:11:54.960 --> 00:12:01.720
if AI doesn't sort of go rogue
and start doing something directly horrible, I

147
00:12:01.759 --> 00:12:05.720
think a lot of people's concerns is
just, hey, what about everyone's jobs.

148
00:12:05.759 --> 00:12:11.399
I will certainly be very disruptive because
jobs will be different. So now

149
00:12:11.480 --> 00:12:16.120
I think the rate of change,
the rate the change coursed by AI is

150
00:12:16.200 --> 00:12:20.840
it's going to be pretty radical,
so that there will you know, a

151
00:12:20.879 --> 00:12:24.000
lot of jobs that are that currently
exist won't exist in the future. But

152
00:12:24.039 --> 00:12:26.799
I think there will be new jobs. And I do think like in a

153
00:12:26.840 --> 00:12:35.080
benign AI scenario, we will really
have an age of abundance the I think

154
00:12:35.159 --> 00:12:39.240
really goods and services unless they are
artificially made scarce, like like it's specialized

155
00:12:39.360 --> 00:12:43.399
artwork, or you want that particular
house in that neighborhood, like it has

156
00:12:43.440 --> 00:12:48.919
been artificial scarcity. But anything that
has not made artificially scarce will be plentiful,

157
00:12:48.960 --> 00:12:52.360
as in a benign AI scenario,
you'll be able to have any products

158
00:12:52.360 --> 00:13:00.919
and services you want that anyone will. Is that not start to cause a

159
00:13:00.919 --> 00:13:03.519
whole new level of problems? Yeah, it does. It's like it's it's

160
00:13:03.919 --> 00:13:09.519
like, you know, there's there's
some things which sound like like a a

161
00:13:09.600 --> 00:13:11.840
blessing but may in fact be a
curse. Like I think you will live

162
00:13:11.879 --> 00:13:16.240
forever. It sounds like a blessing, it's actually a curse. You would

163
00:13:16.279 --> 00:13:18.320
not want to live forever. You
don't know how long forever it is.

164
00:13:20.840 --> 00:13:26.000
And if you say, another potential
curse is you can have anything you want

165
00:13:26.639 --> 00:13:31.360
effortlessly. Are you sure you want
that? Yeah? Does anything matter at

166
00:13:31.399 --> 00:13:35.039
that point? Yeah? And what
about the value? I mean, I

167
00:13:35.120 --> 00:13:39.519
think a lot of the value of
work isn't just getting money to buy things.

168
00:13:39.799 --> 00:13:45.559
It's the meaning, it's the purpose, it's knowing that you worked for

169
00:13:45.759 --> 00:13:48.720
this thing. Right, It's different. It's different to get ten million dollars

170
00:13:48.720 --> 00:13:52.200
because someone just gave you ten million
dollars, or you want a lottery versus

171
00:13:52.360 --> 00:13:58.639
you created a business and sold products
and services that helped you to earn that

172
00:13:58.759 --> 00:14:03.759
money exactly. So it's like,
how do we find meaning and relevance?

173
00:14:05.799 --> 00:14:09.279
If you have an age of abundance
where the computer you just ask for anything

174
00:14:09.279 --> 00:14:13.360
and get it, it is something
we'll have struggle with that that'll that'll be

175
00:14:15.519 --> 00:14:18.480
I don't know, it's I think
that's kind of most likely when we're headed

176
00:14:18.559 --> 00:14:26.720
is an age of abundance, but
it will definitely cause some existential angst.

177
00:14:31.519 --> 00:14:33.519
What are you most optimistic about?
In fact, if you just say,

178
00:14:33.600 --> 00:14:37.159
like, like for a neural link
device, I mean, part of what

179
00:14:37.200 --> 00:14:39.919
I'm hoping is here in a benign
AI scenario, like, how do we

180
00:14:41.000 --> 00:14:43.279
even go along for the ride?
How do we even understand the AI?

181
00:14:43.919 --> 00:14:48.960
Like if you've got some sort of
you know, incredible intelligence unless we do

182
00:14:50.159 --> 00:14:56.600
sort of effectively have better symbiosis,
increase the bandwidth to our the AI extension

183
00:14:56.639 --> 00:15:03.000
of ourselves, effectively augment our human
intelligence, and substantially we may not even

184
00:15:03.039 --> 00:15:07.320
be able to appreciate the wonders that
will exist in the future.

