Josh Bongard: What are Biological Robots? How AI is Reshaping Life, Consciousness & Matter!

1
00:00:00,080 --> 00:00:02,480
So I am interested in the
philosophical side of things.

2
00:00:02,480 --> 00:00:06,400
So one of the things that I see
in the Xenobots and some of the

3
00:00:06,440 --> 00:00:11,240
other kind of exotic robots that
we work with is a conflation of

4
00:00:11,240 --> 00:00:15,560
thought and action.
So if you take a Roomba, the the

5
00:00:15,560 --> 00:00:18,600
robot vacuum cleaner, you can
point to the wheels and you can

6
00:00:18,600 --> 00:00:21,480
say action.
You can open it up and point to

7
00:00:21,480 --> 00:00:25,520
the central, the CPU, and say,
you know, thought or cognition

8
00:00:25,520 --> 00:00:29,760
or processing traditional
robots, there is a Cartesian cut

9
00:00:29,760 --> 00:00:32,680
you can make that separates the
body from the brain.

10
00:00:33,040 --> 00:00:36,200
But what I'm excited about is if
you look at a lot of cutting

11
00:00:36,200 --> 00:00:40,800
edge technologies, that
separation is becoming less and

12
00:00:40,800 --> 00:00:46,240
less obvious, which I think is,
you know, finally the beginnings

13
00:00:46,240 --> 00:00:50,080
of an acid that's dissolving
this distinction that at least

14
00:00:50,080 --> 00:00:53,600
in the West has been around for
over 300 years.

15
00:00:54,240 --> 00:00:57,120
So I think that's great from a
purely intellectual point of

16
00:00:57,120 --> 00:00:59,280
view.
But again, it, it also is

17
00:00:59,280 --> 00:01:02,440
important for us as a species to
understand that, you know,

18
00:01:02,440 --> 00:01:06,000
brains are not everything and
everything else is expendable.

19
00:01:06,000 --> 00:01:09,520
You know, it's, the story is
much more complicated than that.

20
00:01:09,520 --> 00:01:13,080
And I think that finally
overturning Cartesian dualism

21
00:01:13,080 --> 00:01:16,680
will actually be a positive
thing for for society.

22
00:01:21,840 --> 00:01:24,640
Josh, I've been following you
work for years and you guys have

23
00:01:24,640 --> 00:01:26,400
done some incredible stuff in
the field.

24
00:01:27,040 --> 00:01:30,160
It's biology, robotics,
artificial intelligence, so many

25
00:01:30,160 --> 00:01:31,560
different branches we can go
into.

26
00:01:31,720 --> 00:01:34,800
But I think the best place to
start would be let's start with

27
00:01:34,800 --> 00:01:36,920
definitions.
How do you define life,

28
00:01:37,360 --> 00:01:42,560
consciousness and intelligence?
Life, consciousness and

29
00:01:42,560 --> 00:01:44,600
intelligence.
So we're diving in the deep end

30
00:01:44,640 --> 00:01:49,600
to start.
OK, so to me I am very much in

31
00:01:49,600 --> 00:01:54,120
favour, not surprisingly, of an
embodied approach to life.

32
00:01:54,200 --> 00:01:59,000
So life in many ways is a
process of self construction.

33
00:02:00,120 --> 00:02:04,200
This has been articulated I
think, best in the idea of

34
00:02:04,200 --> 00:02:09,080
autopoiesis, that you can
construct your own components.

35
00:02:09,360 --> 00:02:13,080
Those may be parts of your own
body, parts of your own brain,

36
00:02:13,080 --> 00:02:16,240
tools.
They may be offspring, let you

37
00:02:16,240 --> 00:02:19,560
know that it's almost a cliche,
but life perpetuates itself,

38
00:02:19,560 --> 00:02:22,880
which seems obvious and almost
tautological.

39
00:02:23,360 --> 00:02:26,960
But it is not a trivial thing to
perpetuate oneself and

40
00:02:26,960 --> 00:02:31,960
constantly self construct.
And so that is to me the

41
00:02:31,960 --> 00:02:34,760
definition of life.
Not, none of our technologies,

42
00:02:34,760 --> 00:02:38,560
no inert materials that we know
of are capable of this.

43
00:02:39,080 --> 00:02:41,800
And that's part of what
interests me about robotics.

44
00:02:41,800 --> 00:02:45,960
I believe in future we will
create machines that are living

45
00:02:45,960 --> 00:02:49,880
in the sense that maybe they're
made of living materials, but

46
00:02:49,880 --> 00:02:52,360
even if they're not, that
they're able to constantly

47
00:02:52,360 --> 00:02:57,240
construct the self and propagate
the self in various ways into

48
00:02:57,240 --> 00:03:00,960
the future.
So that that's my definition of

49
00:03:00,960 --> 00:03:03,880
life.
You asked about consciousness.

50
00:03:04,920 --> 00:03:08,040
All, all cards on the table.
I'm an illusionist.

51
00:03:08,040 --> 00:03:11,080
I do not believe consciousness
is a thing.

52
00:03:11,080 --> 00:03:14,400
We've been looking for it for a
long time, haven't found it.

53
00:03:15,520 --> 00:03:18,720
I think it's just one of those
things that's the best, the most

54
00:03:18,720 --> 00:03:22,720
convenient term that humans have
come up with for something

55
00:03:22,720 --> 00:03:25,400
that's epiphenomenal, that's not
really there.

56
00:03:26,440 --> 00:03:29,680
I would follow in the footsteps
of Daniel Dennett in this way,

57
00:03:30,160 --> 00:03:33,560
that it is a form of illusion
and we can point to optical

58
00:03:33,560 --> 00:03:37,240
illusions, auditory illusions,
motor illusions.

59
00:03:37,800 --> 00:03:41,400
Our brains are very good at
fooling themselves and I think

60
00:03:41,400 --> 00:03:44,960
fooling themselves into
believing they are conscious is

61
00:03:44,960 --> 00:03:48,480
one of those things.
And what, what was your third

62
00:03:48,840 --> 00:03:51,000
dissideratum?
Intelligence.

63
00:03:51,480 --> 00:03:55,720
Intelligence.
Aha, yes, OK, that one again,

64
00:03:55,720 --> 00:04:00,400
I'm sort of a very pragmatist
about intelligence is the

65
00:04:00,400 --> 00:04:03,880
ability to make sure that you do
not get painted into a corner in

66
00:04:03,880 --> 00:04:06,040
the future.
That's it.

67
00:04:07,120 --> 00:04:09,480
I love the fact that you brought
up the fact that you think

68
00:04:09,480 --> 00:04:13,240
consciousness is an illusion.
Well I I guess the word there is

69
00:04:13,240 --> 00:04:15,800
quite important and and
difficult to work around because

70
00:04:15,800 --> 00:04:17,519
you often have to backtrack what
that means.

71
00:04:17,880 --> 00:04:20,880
When I wrote my dissertation, it
was on illusionism as a theory

72
00:04:20,880 --> 00:04:23,360
of consciousness.
OK, there you go.

73
00:04:23,680 --> 00:04:26,360
Followed in the footsteps of
dand in it, Keith Frankish,

74
00:04:26,360 --> 00:04:29,600
Michael Graziano and I put all
of this together to defended

75
00:04:29,640 --> 00:04:32,240
within psychiatry.
And I find it interesting

76
00:04:32,240 --> 00:04:34,920
because it's it's a very
misunderstood theory of

77
00:04:34,920 --> 00:04:36,480
consciousness.
Well, it's not really a theory

78
00:04:36,480 --> 00:04:39,120
of consciousness, but rather a
theory that shows how other

79
00:04:39,120 --> 00:04:44,120
theories of consciousness tend
to lack substantive evidence of

80
00:04:44,120 --> 00:04:46,800
what they're talking about.
So it's almost like intuition

81
00:04:46,800 --> 00:04:50,160
pumps, what Dan and it called
for theory of consciousness,

82
00:04:50,240 --> 00:04:52,760
sort of how to think of it.
But when I look at your work,

83
00:04:52,760 --> 00:04:55,680
it's almost easy to say that you
might be an illusionist.

84
00:04:55,680 --> 00:04:57,800
And I thought that right up
until now.

85
00:04:57,800 --> 00:05:00,480
So you've you've managed to
confirm what I've been thinking

86
00:05:00,480 --> 00:05:02,240
all about.
OK, great.

87
00:05:03,080 --> 00:05:05,280
Let's start off with I've I've
I've scheduled this.

88
00:05:05,880 --> 00:05:09,200
Sorry, I've prepped this podcast
with 10 main questions in mind.

89
00:05:09,320 --> 00:05:12,480
The first one in what you just
mentioned embodied intelligence.

90
00:05:12,840 --> 00:05:16,560
In the book you co-authored, How
the body Shapes the way we

91
00:05:16,560 --> 00:05:18,760
think.
You argue that cognition is

92
00:05:18,760 --> 00:05:22,320
deeply rooted within the body's
physical form.

93
00:05:22,440 --> 00:05:26,200
So how does this embodied
perspective challenge our

94
00:05:26,200 --> 00:05:28,160
traditional abstract models of
mind?

95
00:05:28,440 --> 00:05:31,200
I mean, this podcast is called
Mind Body Solution, paying

96
00:05:31,200 --> 00:05:32,800
homage to the infamous mind body
problem.

97
00:05:32,960 --> 00:05:35,000
But therein might be a problem,
the fact that we're

98
00:05:35,000 --> 00:05:38,000
differentiating between mind and
body, separating these two

99
00:05:38,000 --> 00:05:40,160
entities.
Does this problem exist?

100
00:05:40,160 --> 00:05:43,400
What are your thoughts of this?
Yeah, OK, great.

101
00:05:44,320 --> 00:05:48,200
So yeah, embodied cognition,
this term, it means a lot of

102
00:05:48,200 --> 00:05:50,280
different things to a lot of
people.

103
00:05:51,600 --> 00:05:54,880
I've worked in embodied
cognition for many years now.

104
00:05:54,880 --> 00:06:00,240
So my my particular perspective
in it on it is exactly that,

105
00:06:00,240 --> 00:06:03,520
that there is no distinction
between what we conveniently

106
00:06:03,520 --> 00:06:08,320
refer to as mind and what we may
be less problem problematically

107
00:06:08,320 --> 00:06:12,520
refer to as the body.
There is no separation.

108
00:06:13,120 --> 00:06:17,000
There are some additional kind
of add-ons, evolutionarily

109
00:06:17,000 --> 00:06:20,840
recent add-ons like neurons and
brains and central nervous

110
00:06:20,840 --> 00:06:25,080
systems and prefrontal cortices,
but they're the icing on the

111
00:06:25,080 --> 00:06:28,080
cake.
They're not the cake the nervous

112
00:06:28,080 --> 00:06:33,040
system facilitates, things that
organisms without brains were

113
00:06:33,040 --> 00:06:36,760
able to do for a very, very long
time.

114
00:06:37,160 --> 00:06:41,200
That when you get down to brass
tacks, there isn't anything

115
00:06:41,200 --> 00:06:45,120
really qualitatively new that
brains have brought to the

116
00:06:45,120 --> 00:06:50,320
table, although from a human
perspective, it often feels like

117
00:06:50,320 --> 00:06:52,880
there is.
Again, illusions are creeping

118
00:06:52,880 --> 00:06:56,760
back in.
When when you think of embodied

119
00:06:56,760 --> 00:07:00,760
cognition, how familiar or how
much do you also consider what

120
00:07:01,000 --> 00:07:02,680
psychologists call the 4E
cogsaw?

121
00:07:02,680 --> 00:07:06,480
I mean inactivism and embodied
cognition.

122
00:07:06,480 --> 00:07:09,840
And then you have the others
like embedded cognition and is

123
00:07:09,840 --> 00:07:11,400
the other E?
And there's one more.

124
00:07:11,920 --> 00:07:13,280
Yeah, I'm trying to remember
now.

125
00:07:13,280 --> 00:07:14,640
What it is?
It's embedded.

126
00:07:15,200 --> 00:07:19,400
OK, yeah, I, I, I think you know
this.

127
00:07:19,720 --> 00:07:23,880
My belief comports with a lot of
those, you know, and now putting

128
00:07:23,880 --> 00:07:28,000
on my roboticist hat, you know,
the devil is in the details.

129
00:07:28,000 --> 00:07:31,600
What exactly does it mean to be,
you know, all the ease,

130
00:07:31,600 --> 00:07:37,720
empowered, enacted, enabled, you
know, you name it, embodied that

131
00:07:37,720 --> 00:07:40,880
it's at the surface, it's kind
of intuitive.

132
00:07:40,880 --> 00:07:43,720
You know, something or someone
that has a body that's a

133
00:07:43,720 --> 00:07:46,880
physical tool with which to
interact with the environment

134
00:07:46,880 --> 00:07:52,560
somehow, you know, intuitively
has more options and ability to

135
00:07:52,960 --> 00:07:56,080
prevent getting painted into a
corner in the future, right?

136
00:07:56,080 --> 00:07:59,240
It just kind of makes sense.
But what I find fascinating in

137
00:07:59,240 --> 00:08:03,280
the the research that I do and
some of my colleagues do is when

138
00:08:03,280 --> 00:08:07,320
you get down into the into the
weeds, things become very non

139
00:08:07,320 --> 00:08:10,520
obvious.
So I co-authored how the body

140
00:08:10,520 --> 00:08:13,640
shapes the way we think with my
PhD advisor Roll Pfeiffer.

141
00:08:14,000 --> 00:08:17,880
And when I was a PhD student
with Rolf, I would always ask

142
00:08:17,880 --> 00:08:20,920
him, you know, how exactly does
the body shape the way we think?

143
00:08:20,920 --> 00:08:24,920
What exactly is it about the,
the body that neural networks

144
00:08:24,920 --> 00:08:27,120
can't do?
And we would have, you know, a

145
00:08:27,120 --> 00:08:29,280
lot of heated discussions about
that.

146
00:08:29,280 --> 00:08:32,039
And and that's what I've
dedicated my career to is to

147
00:08:32,039 --> 00:08:37,120
trying to quantify and
concretize, you know, exactly

148
00:08:37,120 --> 00:08:40,120
what this means because we can
build machines to try out some

149
00:08:40,120 --> 00:08:42,799
of these theories.
Whereas in the, you know, the

150
00:08:42,799 --> 00:08:46,680
cognitive sciences and the
psychological sciences, you

151
00:08:46,680 --> 00:08:49,720
know, there's only so much we
can do to humans to try and

152
00:08:49,720 --> 00:08:53,320
understand how ease exist in
humans.

153
00:08:53,720 --> 00:08:55,520
But with robots, you can do
anything.

154
00:08:55,520 --> 00:08:59,760
So I think that's the difference
between the fields, but I think

155
00:08:59,760 --> 00:09:03,080
their long term goal is the same
as to really understand what all

156
00:09:03,080 --> 00:09:06,400
these ES mean.
You you actually touched on the

157
00:09:06,400 --> 00:09:07,720
one.
I think I might have made a

158
00:09:07,720 --> 00:09:10,400
mistake, but the other E was
extended and.

159
00:09:11,000 --> 00:09:12,360
Yes.
Those tools that we do use,

160
00:09:12,360 --> 00:09:13,560
yeah.
So like our cell phones,

161
00:09:13,560 --> 00:09:17,080
everything else that we use to
that form bottle cognition, that

162
00:09:17,840 --> 00:09:20,360
in terms of evolutionary design
of intelligence.

163
00:09:20,640 --> 00:09:24,360
Your work in evolutionary
robotics explores how machines

164
00:09:24,360 --> 00:09:27,640
evolve adaptive behaviours and
it's.

165
00:09:27,880 --> 00:09:31,440
Can you explain how these
evolutionary algorithms guide

166
00:09:31,440 --> 00:09:36,120
this process and with such
evolved systems could ever cross

167
00:09:36,120 --> 00:09:39,080
into the realm of, let's say,
consciousness or sentience?

168
00:09:39,720 --> 00:09:42,080
Sure, sure.
So, yeah.

169
00:09:42,080 --> 00:09:44,440
So for those of your listeners
that aren't familiar with the

170
00:09:44,440 --> 00:09:48,320
field of evolutionary robotics,
the approach is more or less

171
00:09:48,320 --> 00:09:52,680
what the name implies, which is
that we try and we create AI

172
00:09:53,320 --> 00:09:58,200
that evolves bodies and brains
of robots in simulation.

173
00:09:58,200 --> 00:10:01,920
So I'm sitting here in my office
and on the far side of the quad

174
00:10:01,920 --> 00:10:04,560
I can see the building where a
supercomputer is housed.

175
00:10:04,880 --> 00:10:09,360
And inside that supercomputer,
in the GPU's, there's 10,000

176
00:10:09,360 --> 00:10:11,760
virtual worlds running right
now.

177
00:10:12,240 --> 00:10:15,960
And inside each of those virtual
worlds, there's a swarm of

178
00:10:15,960 --> 00:10:19,320
virtual robots.
And every once in a while, some

179
00:10:19,320 --> 00:10:22,720
of those robots disappear
because the AI is deleting the

180
00:10:22,720 --> 00:10:25,600
ones that aren't doing such a
good job at whatever we want

181
00:10:25,600 --> 00:10:28,480
them to do.
And the AI makes randomly

182
00:10:28,480 --> 00:10:31,920
modified copies of the surviving
virtual robots.

183
00:10:32,280 --> 00:10:34,720
So that that's the
methodological approach.

184
00:10:35,320 --> 00:10:39,360
This AI that I mentioned, it's,
it's arguably the oldest form of

185
00:10:39,480 --> 00:10:41,000
AI.
It's called an evolutionary

186
00:10:41,000 --> 00:10:45,240
algorithm, and you can trace it
back almost to the 1940s.

187
00:10:45,240 --> 00:10:48,440
Some argue that the very first
computer program that was ever

188
00:10:48,440 --> 00:10:51,120
written was something that
looked kind of like an

189
00:10:51,120 --> 00:10:53,760
evolutionary algorithm.
And it's kind of, you know, an

190
00:10:53,760 --> 00:10:55,640
intuitive idea.
If you don't know how to solve

191
00:10:55,640 --> 00:10:59,360
your problem, why don't you just
create a population of random

192
00:10:59,360 --> 00:11:02,600
solutions, measure how good
those solutions are, delete the

193
00:11:02,600 --> 00:11:06,400
bad ones, make randomly modified
copies of the survivors, and off

194
00:11:06,400 --> 00:11:08,840
you go.
The I think the second part of

195
00:11:08,840 --> 00:11:12,520
your question was, well, how
might this process lead to, you

196
00:11:12,520 --> 00:11:16,240
know, more abstract forms of
cognition like sentience and

197
00:11:16,240 --> 00:11:18,560
consciousness?
It's a very good question.

198
00:11:19,200 --> 00:11:21,800
I've been working in
evolutionary robotics for over

199
00:11:21,800 --> 00:11:26,280
20 years now, and we tend to
focus on, you know, sensor motor

200
00:11:26,880 --> 00:11:30,400
tasks, behaviors rooted in
sensory motor things, so

201
00:11:30,400 --> 00:11:34,920
locomotion, object manipulation,
swarm behavior, collective

202
00:11:34,920 --> 00:11:36,800
intelligence, that sort of
thing.

203
00:11:38,560 --> 00:11:42,880
But we have dabbled in the more
abstract, you know, aspects of

204
00:11:42,880 --> 00:11:46,480
cognition.
In some work I did in my post

205
00:11:46,480 --> 00:11:51,800
doc with Hod Lipson at Cornell,
we evolved a robot that evolves

206
00:11:51,800 --> 00:11:56,320
understandings of itself.
So, so it gets a little like in

207
00:11:56,480 --> 00:12:00,280
the movie Inception.
So you have a virtual robot and

208
00:12:00,320 --> 00:12:04,360
it's trying to survive, and that
virtual robot is running

209
00:12:04,360 --> 00:12:09,280
simulations of itself in its own
virtual head, and it is using

210
00:12:09,280 --> 00:12:14,120
those guesses about itself and
its current situation to guide

211
00:12:14,120 --> 00:12:17,480
its behavior.
So there's an old theory, you

212
00:12:17,480 --> 00:12:20,280
know, in the biological and
cognitive sciences that that

213
00:12:20,280 --> 00:12:22,800
self-awareness, I don't know
about consciousness, but

214
00:12:22,800 --> 00:12:25,920
self-awareness evolved for very
good evolutionary reasons.

215
00:12:26,480 --> 00:12:30,240
If you are near the edge of a
Cliff and you are able to

216
00:12:30,240 --> 00:12:33,720
simulate what would happen to
yourself, you can mentally

217
00:12:33,720 --> 00:12:36,120
rehearse what would happen if
you take three more steps

218
00:12:36,120 --> 00:12:38,200
forward.
Obviously you're in a much

219
00:12:38,200 --> 00:12:42,560
better evolutionary situation
than someone who cannot mentally

220
00:12:42,560 --> 00:12:45,680
rehearse and has to physically
take those three steps to see

221
00:12:45,680 --> 00:12:49,240
what happens.
So we have seen evidence in some

222
00:12:49,240 --> 00:12:54,680
of our simulations where our
robots start to evolve what we

223
00:12:54,680 --> 00:12:58,960
call self models, a model of
self, and they use it to avoid

224
00:12:58,960 --> 00:13:03,120
risky behaviors, practice things
before trying them in reality.

225
00:13:03,480 --> 00:13:06,680
And you can see that how for
many people, that's a stepping

226
00:13:06,680 --> 00:13:12,000
stone towards very abstract
things like consciousness.

227
00:13:13,080 --> 00:13:16,280
And I think it, it's, it reminds
me of some work done well.

228
00:13:16,280 --> 00:13:18,840
Josh about this talks about
something similar in his work.

229
00:13:19,080 --> 00:13:21,320
But Nicholas Humphrey, he wrote
a book called Soul Dust.

230
00:13:21,640 --> 00:13:23,840
And he talks about one of the
functions of consciousness also

231
00:13:23,840 --> 00:13:29,400
having this, this self-awareness
or this awareness of being alive

232
00:13:29,400 --> 00:13:32,360
in itself, the beauty of being
conscious and, and, and

233
00:13:32,400 --> 00:13:35,520
surviving in this world, which
can be another force and driving

234
00:13:35,520 --> 00:13:39,960
force with keeping us alive.
So if that, if the AI eventually

235
00:13:40,240 --> 00:13:43,000
discovers that the fact that
it's there in the 1st place is

236
00:13:43,000 --> 00:13:45,680
important in itself because it
now has this ability to

237
00:13:45,680 --> 00:13:49,000
experience and therefore will
not take that further step if it

238
00:13:49,000 --> 00:13:52,520
does process those thoughts of
if I take a few more steps, I'm

239
00:13:52,520 --> 00:13:54,720
going to fall off the Cliff.
Is that something that you've

240
00:13:54,720 --> 00:13:57,400
seen presenting itself in any
form or any way?

241
00:13:57,440 --> 00:13:59,840
Sorry, I'm sidetracking now.
No, that's OK.

242
00:13:59,840 --> 00:14:03,120
I, I, I haven't, but that's
something that we certainly are

243
00:14:03,120 --> 00:14:06,080
on the lookout.
So if you are able to simulate

244
00:14:06,080 --> 00:14:10,200
self, if you have a, you know,
some sort of abstraction that

245
00:14:10,200 --> 00:14:14,000
you can form of self, you know,
obviously it's also useful for

246
00:14:14,000 --> 00:14:16,800
social creatures.
So if I can form a model of

247
00:14:16,800 --> 00:14:21,280
self, I, you know, I can adapt
that to form a model of other

248
00:14:21,800 --> 00:14:24,880
and a good model of other like
you is the assumption that

249
00:14:24,960 --> 00:14:29,080
you're also running a simulation
of yourself in your head, which

250
00:14:29,080 --> 00:14:32,320
probably includes a simulation
of me, which includes a

251
00:14:32,320 --> 00:14:35,920
simulation of you.
And so you know this, this,

252
00:14:36,440 --> 00:14:40,200
these evolutionary pressures
towards self-awareness very

253
00:14:40,200 --> 00:14:43,920
quickly can lead to recursion
and reflexivity, which is the

254
00:14:43,920 --> 00:14:48,000
awareness of being aware.
And for many people that is a

255
00:14:48,000 --> 00:14:50,560
definition of consciousness.
So.

256
00:14:50,880 --> 00:14:53,520
That sort of meta cognition
which most people talk about is

257
00:14:53,520 --> 00:14:55,800
their own definition of what
consciousness is.

258
00:14:56,400 --> 00:14:58,440
There you go.
No, it doesn't quite handle

259
00:14:58,440 --> 00:15:01,720
qualia very well, but but it,
you know, captures a lot of what

260
00:15:01,720 --> 00:15:04,640
people think of when they think
of consciousness.

261
00:15:05,000 --> 00:15:07,560
I think to a lot of philosophers
and psychologists, just knowing

262
00:15:07,560 --> 00:15:10,880
that some artificial
intelligences have this ability

263
00:15:10,880 --> 00:15:14,440
to self reflect and develop
these self models is probably

264
00:15:14,440 --> 00:15:17,360
enough for them to think that
this this is conscious.

265
00:15:17,960 --> 00:15:21,880
I again, as an illusionist I'd
say sure, absolutely.

266
00:15:21,880 --> 00:15:25,120
You know, using the term
consciousness to describe this,

267
00:15:25,440 --> 00:15:28,480
you know this.
This recursive self-awareness

268
00:15:28,640 --> 00:15:32,240
seems useful as it is also
useful for describing our

269
00:15:32,240 --> 00:15:35,360
ability to be recursively self
aware.

270
00:15:36,200 --> 00:15:39,600
Let let's expand a little bit.
I mean an illusionist when they

271
00:15:39,600 --> 00:15:42,600
think about consciousness.
I mean, we know that the brain

272
00:15:42,600 --> 00:15:45,720
plays tricks on us all the time.
Whether it's optical illusions,

273
00:15:45,720 --> 00:15:48,560
auditory, doesn't really matter.
Hallucinations and, and

274
00:15:48,560 --> 00:15:51,840
experiences occur all the time.
Just like AI confabulates, we

275
00:15:51,840 --> 00:15:55,240
confabulate, we make up stories.
When something happens, we

276
00:15:55,240 --> 00:15:58,480
remember it incorrectly and we
create and draft new stories to

277
00:15:58,480 --> 00:16:02,120
explain this phenomenon.
Are these all the very varying

278
00:16:02,120 --> 00:16:04,560
reasons that you've come to this
conclusion that we, we

279
00:16:04,560 --> 00:16:06,400
fundamentally are flawed
species?

280
00:16:06,600 --> 00:16:10,600
We're we're limited by
biological processes, heuristic

281
00:16:10,600 --> 00:16:13,960
adaptations and and and
therefore we can't assume when

282
00:16:13,960 --> 00:16:17,160
we ask what is consciousness in
an AI that we even are conscious

283
00:16:17,160 --> 00:16:19,200
in the 1st place.
It's like we know well.

284
00:16:19,200 --> 00:16:23,360
Again, I, you know, I, I'm, I'm
perfectly fine with that non

285
00:16:23,360 --> 00:16:27,000
problematic side of it, which
again for most people captures

286
00:16:27,000 --> 00:16:29,080
part of what they mean by
consciousness.

287
00:16:29,480 --> 00:16:31,720
As an illusionist.
What I have a problem with is

288
00:16:31,720 --> 00:16:34,840
that taking the further step to
things like qualia and the

289
00:16:34,840 --> 00:16:38,440
redness of red and you know that
there's there are these things

290
00:16:38,440 --> 00:16:40,960
called qualia and they're in
there somewhere.

291
00:16:41,200 --> 00:16:45,600
To me, that's a step too far.
You know it, it feels as if we

292
00:16:45,600 --> 00:16:47,080
have them.
And I think it's an

293
00:16:47,080 --> 00:16:51,760
epiphenomenal, you know, part of
something to do with

294
00:16:51,760 --> 00:16:55,120
self-awareness.
So that that's that part of it.

295
00:16:55,120 --> 00:16:59,880
I don't, I don't buy.
When you, you guys work, there's

296
00:16:59,960 --> 00:17:02,800
the four of you with when it
comes to biological robots,

297
00:17:02,800 --> 00:17:03,880
which we're going to touch on
soon.

298
00:17:04,280 --> 00:17:07,520
But on the other side, I mean,
you've got someone like Mike who

299
00:17:08,079 --> 00:17:10,200
Mike Levin for all those who
know, but but he's obviously

300
00:17:10,200 --> 00:17:13,200
within the Pancychus realm in
that he sees this cognition

301
00:17:13,200 --> 00:17:14,440
happening in layers going
upward.

302
00:17:14,800 --> 00:17:16,720
And, and then you were on the
other end saying actually now

303
00:17:16,720 --> 00:17:19,800
these are just phenomenal
phenomena occurring, but you're

304
00:17:19,800 --> 00:17:22,839
both producing such amazing work
as a collaboration in a unit.

305
00:17:23,119 --> 00:17:25,000
How do you guys go about
discussing these?

306
00:17:27,400 --> 00:17:29,960
It doesn't, surprisingly.
It doesn't come up too much.

307
00:17:29,960 --> 00:17:33,120
I think we kind of nerd out on
the nuts and bolts, you know,

308
00:17:33,120 --> 00:17:36,360
experimental side of it.
And then how each of us

309
00:17:36,360 --> 00:17:41,200
interprets the implications of
the work is sort of a private,

310
00:17:41,200 --> 00:17:44,200
private affair, I guess.
OK, well, you guys have done

311
00:17:44,200 --> 00:17:46,440
something incredible and I
always think about it and I

312
00:17:46,440 --> 00:17:49,000
always wonder why this isn't
spoken about more.

313
00:17:49,000 --> 00:17:51,720
I mean, when you think of
firstly, what are biological

314
00:17:51,720 --> 00:17:55,240
robots, which we we call now
Xenobots and thanks to you guys,

315
00:17:55,640 --> 00:18:00,040
and I mean these living machines
that you help design, they blur

316
00:18:00,040 --> 00:18:03,000
the line between life and and
machine.

317
00:18:03,000 --> 00:18:05,880
And how do they force us to
rethink our definitions of life,

318
00:18:07,240 --> 00:18:09,880
intelligence, and even agency?
Sure.

319
00:18:09,880 --> 00:18:12,040
Yeah.
So these biological robots, this

320
00:18:12,040 --> 00:18:14,680
was a collaboration as you
mentioned, between myself and

321
00:18:14,680 --> 00:18:18,960
Mike Levin at Tufts and two of
our collaborators, Sam Krigman

322
00:18:18,960 --> 00:18:23,080
and and Doug Blackiston.
And so over five years ago now,

323
00:18:23,080 --> 00:18:25,760
the four of us were together on
a funded project.

324
00:18:25,760 --> 00:18:30,000
And myself and Sam, as the
roboticist, showed what we could

325
00:18:30,000 --> 00:18:32,800
do at the beginning of this
collaboration, that we could

326
00:18:34,040 --> 00:18:38,080
teach an AI to reconfigure
virtual robot parts, you know,

327
00:18:38,080 --> 00:18:41,200
in a supercomputer to make new
virtual robots.

328
00:18:41,200 --> 00:18:45,080
And we demonstrated all of this.
And then a week later, Mike and

329
00:18:45,080 --> 00:18:49,600
Doug joined us on the Zoom call.
And Doug was a very talented

330
00:18:49,600 --> 00:18:53,920
microsurgeon, showed us that he
built a version of one of our

331
00:18:53,920 --> 00:18:57,800
four legged robots under the
microscope from frog cells.

332
00:18:58,320 --> 00:19:00,200
And I'll never forget this
moment.

333
00:19:00,200 --> 00:19:02,840
There was complete silence on
the call because this thing

334
00:19:02,840 --> 00:19:08,520
looked like, you know, our
robots, traditional robots, but

335
00:19:08,520 --> 00:19:11,120
it was clearly some something
biological.

336
00:19:11,120 --> 00:19:13,440
And we found out it was made
from frog cells.

337
00:19:14,520 --> 00:19:17,240
And so Sam and I immediately
asked Doug, we said, can you

338
00:19:17,240 --> 00:19:20,080
take this statue and can we make
it move?

339
00:19:20,080 --> 00:19:22,600
And it took Doug a few more
months, but he figured out how

340
00:19:22,600 --> 00:19:25,120
to put some muscle cells in
there.

341
00:19:25,120 --> 00:19:28,600
And then that was sort of the
first xenobot, this little 4

342
00:19:28,600 --> 00:19:31,840
legged creature walking along
the bottom of a, a Petri dish.

343
00:19:32,480 --> 00:19:35,920
So in, you know, the, the
simplest explanation of a

344
00:19:35,920 --> 00:19:38,240
biological robot is it's a
robot.

345
00:19:38,240 --> 00:19:42,640
It's something that's been built
by us, in this case, us being an

346
00:19:42,720 --> 00:19:46,400
AI plus humans.
It's been built by us to do

347
00:19:46,400 --> 00:19:49,480
something we wanted to do, which
in this first experiment was

348
00:19:49,480 --> 00:19:51,440
just walk along the bottom of a
Petri dish.

349
00:19:52,240 --> 00:19:55,240
A lot of folks don't realize it,
but that's actually the original

350
00:19:55,240 --> 00:20:00,600
definition of a robot.
It's something built by humans

351
00:20:00,600 --> 00:20:04,800
that runs around, is capable of
moving about and does stuff on

352
00:20:04,800 --> 00:20:09,360
our behalf and.
In the original, the original

353
00:20:09,360 --> 00:20:14,280
definition of robots, it comes
from a Czech play over 100 years

354
00:20:14,280 --> 00:20:18,600
ago.
And in that play the robota were

355
00:20:18,600 --> 00:20:21,760
built from protoplasm.
It was some sort of biological

356
00:20:21,760 --> 00:20:24,360
mass.
But of course, then into the

357
00:20:24,360 --> 00:20:27,880
20th century, along comes metal
and plastics and ceramics and

358
00:20:27,880 --> 00:20:30,800
eventually electronics.
And so now, and thanks to

359
00:20:30,800 --> 00:20:34,680
Hollywood, most people think of
robots as things built from 20th

360
00:20:34,680 --> 00:20:39,080
century materials.
But we're in a way as

361
00:20:39,080 --> 00:20:40,960
roboticists.
This is very satisfying.

362
00:20:40,960 --> 00:20:44,520
We're going back to our roots
and we're tasking AI with

363
00:20:44,520 --> 00:20:47,520
building robots from living
materials.

364
00:20:48,600 --> 00:20:50,480
And I think there's sorry,
continue the no.

365
00:20:50,680 --> 00:20:52,600
No, please go ahead, I.
Was going to say, and I think

366
00:20:52,600 --> 00:20:57,120
it, it would be more comforting
to most people to see this type

367
00:20:57,120 --> 00:21:01,440
of architecture within a system
that's robotic in that it's not

368
00:21:01,440 --> 00:21:04,120
scary, it's not dangerously,
it's not metal, it's not going

369
00:21:04,120 --> 00:21:06,800
to harm you or hurt you.
It's softer, it's slightly, it's

370
00:21:06,800 --> 00:21:09,280
easier to look at.
Do you think that

371
00:21:09,480 --> 00:21:12,320
psychologically this is an easy
approach as well?

372
00:21:12,600 --> 00:21:15,720
Just.
For the my experience has been

373
00:21:15,720 --> 00:21:18,040
the exact opposite of what you
just described.

374
00:21:18,040 --> 00:21:22,000
People are absolutely terrified
of what have been been known as

375
00:21:22,000 --> 00:21:26,840
the Zenobots.
We, so we, we finished this

376
00:21:26,840 --> 00:21:29,880
work, we published it in a very
visible journal and it triggered

377
00:21:29,880 --> 00:21:33,080
this, you know, huge media
interest in obviously, you know,

378
00:21:33,080 --> 00:21:36,040
frog bots built by AI.
What, what's not to love?

379
00:21:37,240 --> 00:21:40,840
So, so psychologically, it's
kind of interesting because I

380
00:21:40,840 --> 00:21:44,600
think the Xenobots have
discovered one of the deepest

381
00:21:44,600 --> 00:21:48,880
parts of the uncanny valley.
So for folks who haven't heard,

382
00:21:48,880 --> 00:21:52,280
the uncanny valley is something
where there's something out

383
00:21:52,280 --> 00:21:55,640
there that's kind of like us,
but also in an unexpected way,

384
00:21:55,640 --> 00:21:58,480
not like us.
And that seems to press on some

385
00:21:58,480 --> 00:22:02,400
very deep buttons in humans.
So, you know, zombie, all the

386
00:22:02,400 --> 00:22:05,720
zombie movies and TV shows that
just never seem to end, you

387
00:22:05,720 --> 00:22:08,400
know, zombies are a great
example of something that's

388
00:22:08,760 --> 00:22:11,200
that's in this, you know,
uncanny valley.

389
00:22:11,200 --> 00:22:16,240
And I think Chachi PT and how
from 2001, these disembodied

390
00:22:16,480 --> 00:22:20,360
authoritative voices from above,
you know, that somehow also

391
00:22:20,360 --> 00:22:22,960
presses our buttons.
And now, and maybe in

392
00:22:22,960 --> 00:22:27,520
retrospect, not so surprisingly,
AI designed frogbots also

393
00:22:28,800 --> 00:22:33,080
frightens a lot of people.
I will say that it's mostly

394
00:22:33,080 --> 00:22:36,040
older people, I think that are
afraid at the xenobots.

395
00:22:36,040 --> 00:22:39,440
Mike and I now, you know, on a
daily basis get emails from

396
00:22:39,440 --> 00:22:43,520
young folks saying how do I
train to become a xenoboticist?

397
00:22:43,520 --> 00:22:46,280
And you know, they, they think
it's the coolest thing under the

398
00:22:46,280 --> 00:22:48,920
sun.
So it depends on who we're

399
00:22:48,920 --> 00:22:51,880
talking about.
How did the name Xenobot come

400
00:22:51,880 --> 00:22:52,480
about?
Who?

401
00:22:52,480 --> 00:22:54,400
Who said that the first time?
Yep.

402
00:22:54,400 --> 00:22:58,120
So that was Mike, and I'm, I'm
probably going to botch this

403
00:22:58,120 --> 00:23:01,000
story, but I think he was
sitting in on someone's PhD

404
00:23:01,000 --> 00:23:04,840
defense and he saw a little bit
of an animal cat.

405
00:23:04,840 --> 00:23:09,640
This is one part of a frog egg,
and it was moving about on its

406
00:23:09,640 --> 00:23:12,080
own.
And from Mike's perspective, it

407
00:23:12,080 --> 00:23:14,800
looked like a little robot
running around doing something.

408
00:23:14,800 --> 00:23:17,800
And I think if I'm getting the
story right, that's where the

409
00:23:17,800 --> 00:23:22,600
name came from.
But I think it fits quite well

410
00:23:22,600 --> 00:23:26,720
for this technology because it
has the bot part in these are

411
00:23:27,000 --> 00:23:30,640
living machines that have been
built by us for a purpose.

412
00:23:31,680 --> 00:23:34,920
The Zeno comes from Xenopus
Levus, which is the Latin name

413
00:23:34,920 --> 00:23:38,040
of the the particular frog that
we draw these cells from.

414
00:23:38,520 --> 00:23:42,320
But Zeno is also the Greek
cognate for like stranger or

415
00:23:42,320 --> 00:23:47,120
newcomer, which was just
coincidental, but it was a nice

416
00:23:47,120 --> 00:23:50,800
a nice add to the term I think.
Yeah, I think it's such a cool

417
00:23:50,800 --> 00:23:52,840
name.
I think I might have asked Mike

418
00:23:52,840 --> 00:23:55,840
this many years ago and I can't,
I can't remember if that if the

419
00:23:55,840 --> 00:23:58,200
story matches me, I need to.
I'll double.

420
00:23:58,200 --> 00:24:00,040
S you can, you can compare,
yeah.

421
00:24:00,040 --> 00:24:02,960
See if you got it right.
But these xenobots have

422
00:24:02,960 --> 00:24:06,640
demonstrated self replication
and something typically that's

423
00:24:06,640 --> 00:24:10,080
associated with life only.
And what does this ability tell

424
00:24:10,080 --> 00:24:12,600
us about the relationship
between reproduction,

425
00:24:12,600 --> 00:24:15,080
intelligence and autonomy in
living systems?

426
00:24:15,440 --> 00:24:19,800
So this comes back to the top of
our interview where you were

427
00:24:19,800 --> 00:24:22,440
asking me about intelligence and
and life.

428
00:24:22,480 --> 00:24:24,360
And to me these things are
entwined.

429
00:24:24,360 --> 00:24:28,400
So life is capable of
perpetuating itself, making

430
00:24:28,400 --> 00:24:32,320
stuff including parts of itself
or copies of itself or

431
00:24:32,320 --> 00:24:35,000
offspring.
And the ability to self

432
00:24:35,000 --> 00:24:39,960
construct is a very good tool if
your goal is to not get painted

433
00:24:39,960 --> 00:24:43,280
into a corner in the future,
which for me is the working

434
00:24:43,280 --> 00:24:46,520
definition of intelligence.
It gives you options.

435
00:24:46,720 --> 00:24:51,680
So one way of viewing the self
replicating xenobots is that in

436
00:24:51,680 --> 00:24:56,480
essence, we took individual frog
skin cells and just like pulled

437
00:24:56,480 --> 00:25:00,080
them apart.
So this is like the a maximally

438
00:25:00,080 --> 00:25:06,280
deconstructed frog incapable of
its tree, of its usual way of

439
00:25:06,280 --> 00:25:08,560
reproducing or perpetuating
itself.

440
00:25:09,360 --> 00:25:14,240
And these skin cells seem like
they want to come back together

441
00:25:14,240 --> 00:25:18,520
and they sort of somehow figure
out how to make copies of

442
00:25:18,520 --> 00:25:21,120
themselves.
And I am, I'm adopting the

443
00:25:21,120 --> 00:25:24,200
intentional stance here.
I'm talking about frog cells

444
00:25:24,200 --> 00:25:27,560
wanting things and having goals.
I'm not sure whether that's

445
00:25:27,560 --> 00:25:30,280
true, but again, it kind of
makes sense.

446
00:25:30,280 --> 00:25:34,200
If you're an Organism made-up of
a whole bunch of parts, a way to

447
00:25:34,200 --> 00:25:37,720
be intelligent is to make sure
you can reconstitute yourself no

448
00:25:37,720 --> 00:25:41,560
matter what happens and continue
on.

449
00:25:41,600 --> 00:25:46,680
So one of the interesting things
about the self replication study

450
00:25:46,680 --> 00:25:51,560
was, you know, life finds a way.
And this was certainly an

451
00:25:51,560 --> 00:25:55,000
unexpected way that had
previously been unknown.

452
00:25:55,880 --> 00:25:59,440
And I, I, I can sort of
understand when you, when you

453
00:25:59,440 --> 00:26:03,120
talk about the, the backlash,
let's say when you guys

454
00:26:03,120 --> 00:26:08,240
published the first batch of
these papers, because when you

455
00:26:08,240 --> 00:26:12,640
think about the ethics within
biology and with the way people

456
00:26:12,640 --> 00:26:15,040
perceived it years ago, any,
anything relating to sort of

457
00:26:15,040 --> 00:26:18,040
tinkering with an animal doing
this or doing that with a cell

458
00:26:18,600 --> 00:26:21,480
to them is sort of playing God.
And it's, it's their, it, it

459
00:26:21,480 --> 00:26:23,160
comes with a lot of backlash in
general.

460
00:26:23,640 --> 00:26:25,320
I was watching one of your
lectures and I find it quite

461
00:26:25,320 --> 00:26:28,680
funny when you spoke about the
tadpole that now has an eye on

462
00:26:28,680 --> 00:26:31,680
its, but I just said to myself,
that's a joke.

463
00:26:31,800 --> 00:26:34,440
I mean, today people watching
it, it, it is something we, we

464
00:26:34,440 --> 00:26:35,680
noticed that this thing was not
harmed.

465
00:26:35,680 --> 00:26:37,960
It actually still evolved trades
that could do things.

466
00:26:38,440 --> 00:26:41,520
The frog was fully functional,
still able to turn around and

467
00:26:41,520 --> 00:26:45,120
then still catch its prey, so
these systems adapted

468
00:26:45,120 --> 00:26:48,440
appropriately and technically no
harm was done.

469
00:26:48,800 --> 00:26:52,040
So do you think the ethics of
just generally tinkering with

470
00:26:52,040 --> 00:26:54,440
biology is what's coming into
play when it comes to this,

471
00:26:54,440 --> 00:26:57,760
rather than just being used to
metal robots?

472
00:26:58,960 --> 00:27:02,240
I think, yeah, you know,
obviously this is one of the an

473
00:27:02,240 --> 00:27:05,000
old theme, possibly one of the
oldest themes in humanity.

474
00:27:05,000 --> 00:27:08,360
Like what, what starts to happen
when you tinker with things

475
00:27:08,360 --> 00:27:10,040
you're not supposed to tinker
with, right.

476
00:27:10,040 --> 00:27:13,480
All major religions, their
origin stories have to do with

477
00:27:13,480 --> 00:27:16,760
some version of that.
And I would say this is just the

478
00:27:16,760 --> 00:27:20,200
latest chapter in that.
And you know, that that's where

479
00:27:20,200 --> 00:27:24,320
we as a global society are.
We are now a technological

480
00:27:24,320 --> 00:27:26,960
civilization.
We are doing things far beyond

481
00:27:26,960 --> 00:27:30,240
what our ancestors could do.
And guess what?

482
00:27:30,240 --> 00:27:32,120
It's starting to cause some
problems.

483
00:27:33,080 --> 00:27:37,000
So I, I can understand people, a
lot of people's perspective here

484
00:27:37,000 --> 00:27:40,440
that, that we are tinkering with
things that are very powerful

485
00:27:40,680 --> 00:27:44,600
that can deflect in directions
we didn't think of and that can

486
00:27:44,600 --> 00:27:48,680
really cause problems for us.
On the other hand, you know,

487
00:27:48,840 --> 00:27:52,240
that is a reality.
So the most dangerous thing we

488
00:27:52,240 --> 00:27:54,200
could do is just to stop
innovating.

489
00:27:54,280 --> 00:27:56,760
You know, if we, you know, we're
flying up an airplane, if we

490
00:27:56,760 --> 00:27:59,280
turn off the engine, it not a
good solution.

491
00:27:59,280 --> 00:28:03,080
If if we are going to as a
species or as part of this

492
00:28:03,080 --> 00:28:07,920
ecosystem come back to a steady
state, whatever that means, it's

493
00:28:07,920 --> 00:28:12,040
going to be by going forward and
innovating rather than trying to

494
00:28:12,040 --> 00:28:16,800
turn things off and go back.
Do you think what a big fear

495
00:28:16,800 --> 00:28:21,240
would be is sort of creating new
forms of life that could outpace

496
00:28:21,480 --> 00:28:26,240
or control us in a in a way that
we did not anticipate?

497
00:28:26,840 --> 00:28:29,400
I think so.
So you mentioned outpace, right?

498
00:28:29,400 --> 00:28:32,640
So you look at, you know, fear
around AI technologies.

499
00:28:32,800 --> 00:28:36,200
Again, it boils down to the
speed at which things are

500
00:28:36,200 --> 00:28:38,680
changing.
And again, I think this is a

501
00:28:38,680 --> 00:28:41,680
reality that we are only now
just waking up to.

502
00:28:42,760 --> 00:28:48,160
So 25 years ago, Bill Joy, the
CEO of Sun Microsystems, an old

503
00:28:48,160 --> 00:28:52,640
type of computer, wrote a very
influential article back then

504
00:28:52,800 --> 00:28:57,240
about exponential technologies.
These are things that, you know,

505
00:28:57,240 --> 00:29:01,440
grow at an exponential rate.
And you know, we are now living

506
00:29:01,440 --> 00:29:04,840
in the exponential age.
You know, COVID is an example of

507
00:29:04,840 --> 00:29:07,640
computer viruses.
You know, the, the rapid

508
00:29:07,640 --> 00:29:11,280
improvement of AI and it's
exponentially growing powers.

509
00:29:12,200 --> 00:29:16,320
We, it's just where we are.
So for me, the xenobots are even

510
00:29:16,320 --> 00:29:19,960
more important because they are
sometimes an exponential

511
00:29:19,960 --> 00:29:22,600
technology.
They can grow exponentially if

512
00:29:22,600 --> 00:29:27,840
self replication is involved.
So we need to understand how how

513
00:29:27,840 --> 00:29:32,960
exponential technologies work.
You know, it was a great human

514
00:29:32,960 --> 00:29:35,440
achievement that we developed
the COVID vaccine.

515
00:29:35,440 --> 00:29:39,400
You know, the, the, the Black
Death lasted for centuries.

516
00:29:39,400 --> 00:29:42,520
COVID is still here, but the
worst of it, you know, was over.

517
00:29:42,520 --> 00:29:46,720
In a few years, there will be
things that spread exponentially

518
00:29:46,720 --> 00:29:49,960
around the planet.
And we are AI have to learn how

519
00:29:49,960 --> 00:29:53,080
to create other exponential
technologies that can bring the

520
00:29:53,080 --> 00:29:56,520
bad exponential technologies
down as quickly as possible.

521
00:29:56,520 --> 00:30:01,280
And that requires understanding
exponential technologies.

522
00:30:01,960 --> 00:30:04,400
How was your?
How was your experience with

523
00:30:04,880 --> 00:30:08,080
encountering these xenobots when
he showed you what he had done

524
00:30:08,080 --> 00:30:10,240
with your work the first time
you saw it?

525
00:30:10,240 --> 00:30:12,880
How did this fundamentally
either change the way you saw

526
00:30:12,880 --> 00:30:16,440
the mind body problem or or or
just impact you in general?

527
00:30:16,440 --> 00:30:18,960
How did it change your
philosophical or psychological

528
00:30:18,960 --> 00:30:21,160
views on the concept?
Yeah, sure, sure.

529
00:30:21,160 --> 00:30:22,880
So it was a very visceral
experience.

530
00:30:22,880 --> 00:30:27,200
So as a scientist, you know, 99
things you try don't work.

531
00:30:27,200 --> 00:30:29,440
It's failure after failure after
failure.

532
00:30:29,640 --> 00:30:32,400
And then every once in a while,
if you're lucky, something like

533
00:30:32,400 --> 00:30:35,960
the Xenobots comes along.
So I remember I was here in my

534
00:30:35,960 --> 00:30:40,000
office and my then PhD student
Sam came in and showed me some

535
00:30:40,000 --> 00:30:44,480
preliminary data that the
xenobots that the AI had dreamed

536
00:30:44,480 --> 00:30:48,040
up in our supercomputer.
Doug was able to build them

537
00:30:48,040 --> 00:30:51,120
under the microscope.
So he, so the AI, in effect, was

538
00:30:51,120 --> 00:30:55,080
right.
It knew it knew how to rearrange

539
00:30:55,080 --> 00:30:58,320
living materials to create a
robot for us.

540
00:30:58,320 --> 00:31:00,840
And that robot did what it was
supposed to do.

541
00:31:01,520 --> 00:31:04,320
And it was, it was Thanksgiving
time here.

542
00:31:04,320 --> 00:31:06,320
So there was no one in the
research building here.

543
00:31:06,320 --> 00:31:09,080
And I remember just getting up
and going for a walk in the hall

544
00:31:09,080 --> 00:31:11,560
and, and my legs just felt like
jello.

545
00:31:12,080 --> 00:31:15,760
Just realizing that the first
realization was it's possible.

546
00:31:15,760 --> 00:31:18,640
You can do this.
We can make machines from

547
00:31:18,640 --> 00:31:21,040
genetically unmodified living
materials.

548
00:31:21,040 --> 00:31:23,800
This is not GMO.
This is not something else.

549
00:31:23,800 --> 00:31:27,400
It's just, it's just possible.
And for a scientist or an

550
00:31:27,400 --> 00:31:29,760
engineer, you know, that's,
that's the best part.

551
00:31:29,760 --> 00:31:34,000
Just suddenly this whole new
Vista opens up and then all the

552
00:31:34,000 --> 00:31:37,920
implications since then, you
know, are exciting and fun.

553
00:31:37,920 --> 00:31:40,720
But that was that was the big
moment for me at least.

554
00:31:41,640 --> 00:31:44,640
Do you guys ever sit back and
and just reminisce on that day?

555
00:31:44,640 --> 00:31:49,080
Do you Do you talk about the?
Not yet, but I'm hoping we're so

556
00:31:49,080 --> 00:31:52,240
busy, you know, because it
really has started to snowball

557
00:31:52,240 --> 00:31:55,520
since then, you know, and then
typical nerds and scientists.

558
00:31:55,520 --> 00:31:57,680
We could just get caught up in
the work and then, you know,

559
00:31:57,680 --> 00:32:00,800
it's I'm sure one of these days
we'll get, we'll sit back with a

560
00:32:00,800 --> 00:32:04,920
beer and say remember when, but
haven't haven't quite found that

561
00:32:04,920 --> 00:32:07,520
moment yet.
Well, you, you've touched on

562
00:32:07,520 --> 00:32:10,480
this at the moment AI's role in
shaping biology because I mean,

563
00:32:10,480 --> 00:32:12,800
this clearly showed us what
could be possible.

564
00:32:13,440 --> 00:32:16,200
So it played a critical role in
designing these innerbots.

565
00:32:16,800 --> 00:32:20,240
Could you walk us through how AI
is reshaping the way we design

566
00:32:20,240 --> 00:32:23,000
biological systems currently,
and what this might mean for,

567
00:32:23,080 --> 00:32:27,080
let's say future artificial in
life or evolution in general?

568
00:32:27,640 --> 00:32:29,680
Sure, sure.
So stepping back from the

569
00:32:29,680 --> 00:32:35,200
xenobots for a moment, you know,
AI and biology in general are

570
00:32:35,200 --> 00:32:37,960
biological engineering, genetic
engineering.

571
00:32:37,960 --> 00:32:39,840
They're merging.
So it's, you know, it's

572
00:32:39,840 --> 00:32:42,040
impossible to say.
There's so many exciting things

573
00:32:42,040 --> 00:32:46,680
happening every week.
So the first most obvious one is

574
00:32:46,760 --> 00:32:51,080
genetic engineering, that that
AI is being fed vast amounts of

575
00:32:51,080 --> 00:32:54,520
genetic information.
And so AI is learning.

576
00:32:54,520 --> 00:32:57,440
If you tweak this gene, this
will happen.

577
00:32:57,440 --> 00:32:59,480
If you tweak that gene, this
will happen.

578
00:32:59,480 --> 00:33:03,040
And that's, you know, immensely
interesting and exciting.

579
00:33:03,040 --> 00:33:07,240
It's, you know, bringing us,
it's bringing us a certain

580
00:33:07,520 --> 00:33:11,000
grains, strains of rice that are
very helpful in developing

581
00:33:11,000 --> 00:33:12,880
countries.
You know, it's lifting people

582
00:33:12,880 --> 00:33:17,000
out of poverty and and misery,
dealing with some huge health

583
00:33:17,000 --> 00:33:21,200
issues, Immensely, immensely
powerful and helpful.

584
00:33:23,000 --> 00:33:26,040
Then there is the generative AI
side, you know, Gen.

585
00:33:26,240 --> 00:33:30,000
AI where it's not looking at
data and learning how to tweak

586
00:33:30,000 --> 00:33:32,480
things, it's learning how to
build things.

587
00:33:33,400 --> 00:33:36,960
And so at the moment, the big
successes there have been at the

588
00:33:36,960 --> 00:33:40,000
sub cellular level.
So there are now AI design

589
00:33:40,000 --> 00:33:44,960
proteins and these proteins look
and act like no other protein

590
00:33:44,960 --> 00:33:46,760
that humans have ever seen,
right?

591
00:33:46,760 --> 00:33:51,200
And that's revolutionary
revolutionizing, you know, drug

592
00:33:51,200 --> 00:33:54,160
design and you know,
therapeutics in general.

593
00:33:54,880 --> 00:33:57,160
So the the xenobots then is sort
of Gen.

594
00:33:57,160 --> 00:34:01,240
I Gen.
AI at the Super cellular level.

595
00:34:01,240 --> 00:34:05,720
So the AI is taking cells as the
building block, not peptides,

596
00:34:06,000 --> 00:34:09,040
and putting those cells together
to make things that are, you

597
00:34:09,040 --> 00:34:13,159
know, a millimeter in size or,
or a little bit larger to create

598
00:34:13,159 --> 00:34:18,159
little tiny machines.
And so stepping back, you know,

599
00:34:18,159 --> 00:34:21,560
we're basically developing AI
that is learning how to tinker

600
00:34:21,560 --> 00:34:26,639
with life at all levels, sub
cellular, super cellular, you

601
00:34:26,639 --> 00:34:30,520
know, I don't know how how high
up the scale we can go, but you

602
00:34:30,520 --> 00:34:34,280
know, the trend suggests there's
there doesn't seem to be a limit

603
00:34:34,280 --> 00:34:37,120
insight yet.
Have they has it?

604
00:34:37,280 --> 00:34:40,040
Well, I'm not too familiar with
the with the data within this

605
00:34:40,040 --> 00:34:42,000
field though.
But when it comes to these

606
00:34:42,000 --> 00:34:45,400
proteins and these, does it ever
give us insight into certain

607
00:34:45,400 --> 00:34:47,040
amino acids?
Like maybe we missed something,

608
00:34:47,040 --> 00:34:48,840
maybe there's something we
haven't yet figured this out.

609
00:34:50,080 --> 00:34:52,960
Have there been cases like this
where people are figuring things

610
00:34:52,960 --> 00:34:57,040
out, where biologists just never
would have never considered or

611
00:34:57,040 --> 00:34:59,600
expected aside?
Yeah, absolutely.

612
00:34:59,600 --> 00:35:02,600
Now it's a much harder, it's a
much harder lift.

613
00:35:02,600 --> 00:35:06,640
It does happen.
It's rarer because if you think

614
00:35:06,640 --> 00:35:10,720
about it, you know the we've
learned now that so AI is black

615
00:35:10,720 --> 00:35:12,520
box.
It's able to do some amazing

616
00:35:12,520 --> 00:35:16,840
things and not bother to explain
to us poor humans how it did it.

617
00:35:17,480 --> 00:35:21,840
But you can make things harder
on the AI by saying, please come

618
00:35:21,840 --> 00:35:25,320
up with new things we never
thought of and also come up with

619
00:35:25,320 --> 00:35:30,360
explanations that we as pure
poor humans will understand and

620
00:35:30,400 --> 00:35:33,280
that, you know, that's a harder
ask.

621
00:35:33,400 --> 00:35:35,920
We joke in the lab, you know,
from the AI perspective says,

622
00:35:36,040 --> 00:35:39,160
oh, you want me to also have to
try and explain this to you?

623
00:35:39,160 --> 00:35:43,040
You know, that's really hard.
But but it is possible.

624
00:35:43,280 --> 00:35:47,360
And I, I've dabbled with some of
that in a different branch of AI

625
00:35:48,320 --> 00:35:51,400
where the AI takes in raw data
and spits out equations.

626
00:35:51,400 --> 00:35:54,560
It basically generates math to
explain the data.

627
00:35:54,880 --> 00:35:58,960
And you can read this AI
generated math term by term.

628
00:35:59,320 --> 00:36:02,480
And when this AI is working, it
should generate the terms that

629
00:36:02,480 --> 00:36:04,680
you as an expert in that field
are familiar.

630
00:36:04,680 --> 00:36:07,960
What with you would expect to
see this term, that term.

631
00:36:08,440 --> 00:36:09,720
Hey, wait a second.
Why?

632
00:36:09,840 --> 00:36:13,760
I don't recognize this term?
That's where things get really

633
00:36:13,760 --> 00:36:16,640
interesting.
So yeah, it does.

634
00:36:16,640 --> 00:36:19,280
It does happen.
But we are the weakest link in

635
00:36:19,280 --> 00:36:21,800
that chain, right?
It's up to us to struggle to

636
00:36:21,800 --> 00:36:25,160
actually understand what the AI
is trying to tell us it.

637
00:36:25,920 --> 00:36:29,640
Kind of reminds me of the film
Her with Joaquin Phoenix.

638
00:36:29,680 --> 00:36:32,280
We're at the very end.
I mean this artificial

639
00:36:32,280 --> 00:36:34,400
intelligence is just way too
complex for him to even

640
00:36:34,400 --> 00:36:35,880
understand.
Yes.

641
00:36:36,000 --> 00:36:38,240
I mean, we're there already.
We're there already.

642
00:36:38,240 --> 00:36:42,320
It's yeah, yeah, we're, we're
living in that reality for sure.

643
00:36:42,440 --> 00:36:45,680
The outpacing is a problem, I
would say that that a lot of us

644
00:36:45,680 --> 00:36:47,920
fear in general.
I mean, even I, I myself as a

645
00:36:47,920 --> 00:36:50,840
doctor, when I see how many of
my colleagues would use AI to

646
00:36:50,840 --> 00:36:54,160
draft motivation letters or even
just to diagnose or try and

647
00:36:54,160 --> 00:36:56,520
interact with patients, they
send an e-mail reply.

648
00:36:56,520 --> 00:37:00,960
I mean, there's a lot of
legalities around the privacy

649
00:37:00,960 --> 00:37:03,440
and obviously trying to not give
out personal information, but

650
00:37:03,640 --> 00:37:06,200
but people are using this
everywhere all the time.

651
00:37:06,200 --> 00:37:09,840
It's in, it's almost
inescapable, and we have to

652
00:37:09,840 --> 00:37:12,840
consider this to be part of the
extended cognition we were

653
00:37:12,840 --> 00:37:15,640
talking about earlier.
AI has now become fundamentally

654
00:37:15,640 --> 00:37:17,720
a part of this extended
cognition.

655
00:37:17,920 --> 00:37:20,160
What are your thoughts about?
Yeah.

656
00:37:20,160 --> 00:37:24,040
I mean, the great, the greatest
extension of our ability so far,

657
00:37:24,040 --> 00:37:25,480
right?
And it's hard to argue that it

658
00:37:25,480 --> 00:37:27,400
isn't.
It's Yeah, it is.

659
00:37:28,640 --> 00:37:31,360
Yeah, I'm slowly struggling to
see a day that goes by where I

660
00:37:31,360 --> 00:37:34,160
don't use some sort of
artificial intelligence.

661
00:37:34,560 --> 00:37:36,600
To assist.
And you probably used it 32

662
00:37:36,600 --> 00:37:39,440
times today in ways you didn't
even realize that you were using

663
00:37:39,440 --> 00:37:41,840
it right.
There's also that, there's the

664
00:37:41,880 --> 00:37:45,280
there's the dark matter of AI
that now exists in our world as

665
00:37:45,280 --> 00:37:47,800
well.
I can't remember who the author

666
00:37:47,800 --> 00:37:50,560
was who mentioned the
technological singularity, but

667
00:37:51,160 --> 00:37:53,040
you.
Think Ray Kurzweil.

668
00:37:53,200 --> 00:37:55,240
Yes, yes.
Do you, do you think we're close

669
00:37:55,240 --> 00:37:56,120
to that?
Do you think it's?

670
00:37:57,400 --> 00:37:58,920
Yeah, I got him sounding like a
cynic.

671
00:37:58,920 --> 00:38:00,520
I don't believe in the
singularity.

672
00:38:01,200 --> 00:38:05,080
Things are moving fast and even
possibly exponentially, but that

673
00:38:05,080 --> 00:38:07,400
doesn't mean that they will
continue to you.

674
00:38:07,440 --> 00:38:11,680
You see, you see a lot of
sigmoid curves in the world.

675
00:38:11,680 --> 00:38:15,520
Things suddenly get really,
really fast and then you start

676
00:38:15,520 --> 00:38:19,280
to get diminishing returns.
I don't, I've never seen any

677
00:38:19,280 --> 00:38:23,320
phenomenon that just keeps going
like this and I don't see why AI

678
00:38:23,360 --> 00:38:26,400
would.
Well, let's Josh, let's bring it

679
00:38:26,400 --> 00:38:27,880
back.
So this podcast obviously is

680
00:38:27,960 --> 00:38:30,760
focused on mind body solution,
the mind body problem,

681
00:38:30,760 --> 00:38:32,920
consciousness, trying to
understand this fundamental

682
00:38:32,920 --> 00:38:35,240
question.
And I can't help but think, but

683
00:38:35,240 --> 00:38:38,360
I know you love sci-fi and and
growing up, I mean, this was

684
00:38:38,360 --> 00:38:39,800
something you were always
fascinated by.

685
00:38:40,840 --> 00:38:47,040
Have you watched Ex Machina?
Ex Machina, Yes, yes, yes.

686
00:38:47,040 --> 00:38:50,960
I think I remember the.
Alex Garland So the the science

687
00:38:50,960 --> 00:38:55,040
advisor was Maurice Shanahan and
he he spoke about how the, when

688
00:38:55,040 --> 00:38:58,880
he was one of the best scenes in
the film was when, when Oscar

689
00:38:58,880 --> 00:39:03,360
Isaac asks the I'm sorry, the
guy who asks Oscar Isaac about

690
00:39:03,520 --> 00:39:07,720
his artificial intelligence, he
says, did he pass the Turing

691
00:39:07,720 --> 00:39:09,240
test?
And he says no, he doesn't need

692
00:39:09,240 --> 00:39:11,480
to.
So she doesn't need to because I

693
00:39:11,480 --> 00:39:13,120
want you to know this is a
robot.

694
00:39:13,120 --> 00:39:15,840
I want you to know this is
artificial intelligence.

695
00:39:16,200 --> 00:39:18,960
The question is, will you still
believe it's consciousness

696
00:39:19,160 --> 00:39:21,880
thereafter?
So so here the the cards out on

697
00:39:21,880 --> 00:39:24,680
the table.
We know this is not a conscious

698
00:39:24,680 --> 00:39:26,800
biological being.
You know, it's physically

699
00:39:27,040 --> 00:39:29,640
artificial intelligence.
But are you still fooled?

700
00:39:29,640 --> 00:39:31,720
And he called us the Garland
test since that form.

701
00:39:31,840 --> 00:39:34,280
Do you think that's a bit it's
more of an interesting approach

702
00:39:34,280 --> 00:39:36,680
rather than having the normal
Turing test, but rather know

703
00:39:36,680 --> 00:39:40,760
this is a non, well a non
biological being, but yet he's

704
00:39:40,760 --> 00:39:42,720
super convinced that it is
conscious.

705
00:39:42,960 --> 00:39:46,240
And I think I don't know that it
helps much.

706
00:39:47,120 --> 00:39:49,640
I would mention Thomas Nagel's
essay.

707
00:39:49,640 --> 00:39:51,480
You know what it is like to be a
bat?

708
00:39:51,480 --> 00:39:54,720
So a bat is also not human.
It is biological.

709
00:39:55,160 --> 00:39:58,560
But, you know, humans have a
long, you know, we've got a very

710
00:39:58,560 --> 00:40:01,080
long history of looking at
things that are not us.

711
00:40:01,760 --> 00:40:05,400
And making decisions that's
conscious, that's not conscious,

712
00:40:05,400 --> 00:40:08,520
that maybe is conscious.
I'll never know whether that is

713
00:40:08,520 --> 00:40:12,160
conscious.
So, you know, in terms of from

714
00:40:12,160 --> 00:40:14,960
that perspective, suddenly
having, you know, super

715
00:40:14,960 --> 00:40:21,000
realistic humanoids or you know,
how level type chat GPTS, what's

716
00:40:21,000 --> 00:40:23,200
the difference?
I I don't know whether a bat is

717
00:40:23,200 --> 00:40:24,760
conscious.
I also don't know whether

718
00:40:24,760 --> 00:40:27,360
ChatGPT is conscious.
I don't know whether it's that's

719
00:40:27,360 --> 00:40:31,640
a well formed question.
I don't know that it's that

720
00:40:31,640 --> 00:40:35,160
different in that that way.
Well, it's, it's, it's one of

721
00:40:35,160 --> 00:40:37,400
those things where the
philosophical zombie will always

722
00:40:37,720 --> 00:40:40,160
come into play.
It's, there's, there's

723
00:40:40,160 --> 00:40:42,760
absolutely no way for me to look
at you and just assume that you

724
00:40:42,760 --> 00:40:46,080
are actually a conscious being.
It is, I only have access to

725
00:40:46,080 --> 00:40:49,480
this mind and it, and, and all
we do is taking cues and, and

726
00:40:49,480 --> 00:40:52,760
sort of make relative guesses as
to what we think is conscious.

727
00:40:53,000 --> 00:40:55,840
Do you think we anthropomorphize
things way too much as a

728
00:40:55,840 --> 00:40:57,840
species?
Way too much.

729
00:40:59,120 --> 00:41:01,520
Oh, that's a that's a good one.
We certainly do it.

730
00:41:01,520 --> 00:41:04,480
You know, an obscene amount of
time.

731
00:41:04,480 --> 00:41:09,560
I, I have AI, have a robot
lawnmower at home and my 4 year

732
00:41:09,560 --> 00:41:12,720
old, you know, this, this is,
he's basically views it as his

733
00:41:12,720 --> 00:41:16,640
sibling and he will get upset
with us if we work the, the

734
00:41:16,640 --> 00:41:19,480
lawnmower too much.
You know, it's, it's everywhere

735
00:41:19,920 --> 00:41:22,400
and and robots are just going to
exacerbate it.

736
00:41:22,880 --> 00:41:26,000
Now, whether it's too much, you
know, you know, suggests that

737
00:41:26,000 --> 00:41:29,400
maybe it's a bad thing.
I think, you know, the main

738
00:41:29,400 --> 00:41:32,720
thing that humans struggle with
is increasing our circle of

739
00:41:32,720 --> 00:41:36,120
empathy.
So if we mistakenly, you know,

740
00:41:36,120 --> 00:41:39,120
empathize with things that maybe
don't there's nothing at home,

741
00:41:39,120 --> 00:41:43,160
there's there's zombie or not.
I think better to air on that

742
00:41:43,160 --> 00:41:46,480
side than to air on the side of
this is not something that's

743
00:41:46,480 --> 00:41:49,520
worthy of our consideration, our
moral consideration.

744
00:41:50,440 --> 00:41:53,480
When you look back at someone
like Isaac Asimov when he came

745
00:41:53,480 --> 00:41:57,240
up with his laws of robotics and
AI, and then you look back to

746
00:41:57,240 --> 00:42:00,800
today with xenobots present, how
do you think that this field has

747
00:42:00,800 --> 00:42:03,560
fundamentally shifted and
changed throughout its course?

748
00:42:05,000 --> 00:42:09,360
Yeah, I mean, Isaac Asimov,
brilliant, creative and also

749
00:42:09,400 --> 00:42:11,400
like all of us, a creature of
his time.

750
00:42:11,400 --> 00:42:15,160
So, you know, it was a lot of,
you know, androids and, and that

751
00:42:15,160 --> 00:42:19,920
sort of thing and, and metal and
plastic and, but with the laws

752
00:42:19,920 --> 00:42:23,520
of robotics, you know, I think
he got that exactly right, which

753
00:42:23,520 --> 00:42:28,800
is human hubris that we think we
can put guardrails around

754
00:42:28,800 --> 00:42:32,440
complicated things.
And so in in AI and robotics,

755
00:42:32,440 --> 00:42:36,160
there's a famous term called
perverse instantiation, which

756
00:42:36,160 --> 00:42:40,600
means that the machine somehow
instantiates the behavior we

757
00:42:40,600 --> 00:42:44,800
want, but instantiates it in a
perverse way.

758
00:42:45,080 --> 00:42:48,360
If you ask an autonomous vehicle
to get the human occupant from

759
00:42:48,360 --> 00:42:52,160
point A to point B as quickly as
possible, fastest way to do that

760
00:42:52,160 --> 00:42:55,800
is to drive in a straight line,
you know, through parks and over

761
00:42:55,800 --> 00:42:58,560
sidewalks and it's everywhere,
right?

762
00:42:58,560 --> 00:43:02,200
So, and it's a very difficult
problem to solve.

763
00:43:02,200 --> 00:43:06,360
We, we end up having to tell all
our AI technologies, please do

764
00:43:06,360 --> 00:43:09,880
X, but don't do it in this way
and don't do it in that way and

765
00:43:09,880 --> 00:43:13,120
don't do we keep tacking on
these, these things, which

766
00:43:13,440 --> 00:43:15,720
reminds me of, you know, all the
things that went wrong with the

767
00:43:15,720 --> 00:43:18,520
three laws of robotics in
Asimov's book.

768
00:43:18,520 --> 00:43:22,600
So I would say actually he was
very prescient and he got at the

769
00:43:22,600 --> 00:43:26,520
root of what the problem would
be, which is exactly the problem

770
00:43:26,520 --> 00:43:30,320
We're all now wrestling.
It's pretty incredible how these

771
00:43:30,480 --> 00:43:34,760
sci-fi writers back in the day
just guessed so many things so

772
00:43:34,760 --> 00:43:39,160
accurately and well informed
within their within their time.

773
00:43:39,160 --> 00:43:42,240
It's so cool to see it all come
to fruition today.

774
00:43:42,640 --> 00:43:46,040
When you look at the field and
and the work you does you guys

775
00:43:46,040 --> 00:43:48,920
have done and that you've done
over time, what do you feel most

776
00:43:48,920 --> 00:43:52,560
proud about when you look back?
What's the most exciting you you

777
00:43:52,560 --> 00:43:57,720
think's been happening so far?
Oh gosh, that's very hard to

778
00:43:57,720 --> 00:43:59,640
say.
A lot of the things we've done I

779
00:43:59,640 --> 00:44:03,880
found very personally satisfying
in that for me, they were good

780
00:44:04,320 --> 00:44:07,520
pieces of the puzzle in
answering this question of how

781
00:44:07,520 --> 00:44:09,080
the body shapes the way we
think.

782
00:44:09,400 --> 00:44:12,280
But some of those results are
kind of, you know, abstract or

783
00:44:12,280 --> 00:44:15,200
abstruse and and, you know,
they're going to have probably

784
00:44:15,200 --> 00:44:17,920
limited impact.
So I, I guess I would say the

785
00:44:17,920 --> 00:44:22,400
Xenobots for the simple fact
that they've given, they've

786
00:44:22,400 --> 00:44:25,880
given society a new way to make
helpful tools, you know, if we

787
00:44:25,880 --> 00:44:30,040
want to clean up microplastics,
not leave additional metal and

788
00:44:30,040 --> 00:44:32,440
stuff, and then when we clean
up, great.

789
00:44:32,440 --> 00:44:36,360
So I'm happy about that.
And then there's also something

790
00:44:36,360 --> 00:44:38,520
about the Xenobots, as I
mentioned, that seems

791
00:44:38,520 --> 00:44:43,160
inspirational to younger people.
And that's something I care a

792
00:44:43,160 --> 00:44:47,080
lot about.
I consider myself an old person

793
00:44:47,080 --> 00:44:49,280
or, you know, we we've made a
lot of problems that

794
00:44:49,280 --> 00:44:52,040
unfortunately younger folks are
going to have to try and deal

795
00:44:52,040 --> 00:44:55,080
with somehow.
And as a scientist and engineer,

796
00:44:55,080 --> 00:44:58,920
I feel morally responsible for
trying to at least leave them

797
00:44:58,920 --> 00:45:02,720
some tools with which to try and
solve some of these problems.

798
00:45:02,720 --> 00:45:05,040
And I hope that Xenobots will be
one of those tools.

799
00:45:05,760 --> 00:45:08,080
When you look back when you were
growing up, who who were the

800
00:45:08,080 --> 00:45:10,880
scientists or writers Sign for
writers, whatever that really

801
00:45:10,880 --> 00:45:13,560
inspired you and got you to to
like look at this field and

802
00:45:13,560 --> 00:45:16,320
think about it so seriously too,
to a point where you guys create

803
00:45:16,320 --> 00:45:19,360
new organisms that have.
Yeah, Yeah, It's a good

804
00:45:19,360 --> 00:45:21,760
question.
So I didn't, I there were no

805
00:45:21,760 --> 00:45:24,600
scientists in my family.
It wasn't something that, you

806
00:45:24,600 --> 00:45:26,640
know, was a part of my
childhood.

807
00:45:26,640 --> 00:45:29,200
So I didn't actually really
think about scientists or

808
00:45:29,200 --> 00:45:31,440
engineers or what we now called
STEM.

809
00:45:31,440 --> 00:45:35,480
That wasn't a thing.
Being a professor also wasn't,

810
00:45:35,920 --> 00:45:40,040
you know, wasn't really a thing
in my family, but it, it was the

811
00:45:40,040 --> 00:45:42,560
sci-fi.
And you know, what I took away

812
00:45:42,560 --> 00:45:46,160
from it is, you know, all these
things sound wonderful.

813
00:45:46,960 --> 00:45:49,640
Where are they?
Like if, and, you know, as a

814
00:45:49,640 --> 00:45:51,760
young person, I could see back
then, you know, there are also

815
00:45:51,760 --> 00:45:54,400
problems.
Like wouldn't it be great?

816
00:45:54,400 --> 00:45:58,360
Like this literature seems to
suggest there's other paths

817
00:45:58,360 --> 00:46:00,640
forward.
There's other ways to do things.

818
00:46:00,640 --> 00:46:03,520
And then I would, from my
limited vantage point, look

819
00:46:03,520 --> 00:46:07,560
around and just, I remember just
being confused, like, why not?

820
00:46:07,560 --> 00:46:10,960
Why aren't we trying to?
And then of course, I learned as

821
00:46:10,960 --> 00:46:13,600
I got older that we were trying
and failing miserably.

822
00:46:14,040 --> 00:46:17,760
You know, I, I came of age in,
in academia during the AI winter

823
00:46:17,760 --> 00:46:21,240
when nothing was working.
It seemed that, you know, robots

824
00:46:21,240 --> 00:46:26,480
and AI was 1000 years off.
And I, for me, I guess my

825
00:46:26,480 --> 00:46:30,480
personality that the challenge
appealed to me, the fact that

826
00:46:31,320 --> 00:46:34,760
there were so many failures that
even if I entered the field and

827
00:46:34,760 --> 00:46:38,760
was slightly less of a failure,
that would be progress that, you

828
00:46:38,760 --> 00:46:40,920
know, being able to contribute a
small bit.

829
00:46:41,400 --> 00:46:44,520
And then suddenly the AI summer
happened and, you know, all

830
00:46:44,520 --> 00:46:46,000
these other things started to
happen.

831
00:46:46,000 --> 00:46:48,200
But but anyways, that's how I
got to this point.

832
00:46:48,760 --> 00:46:51,760
You you mentioned something like
STEM, being a professor.

833
00:46:51,800 --> 00:46:55,800
When you think about you, your
work and how it bridges biology,

834
00:46:55,800 --> 00:46:59,800
computer science, robotics,
philosophy, so many different

835
00:46:59,800 --> 00:47:02,480
fields, this interdisciplinary
work that you guys are doing, so

836
00:47:02,480 --> 00:47:03,920
many different people coming
together.

837
00:47:04,360 --> 00:47:08,680
How do you think that this has
changed the game for you guys?

838
00:47:08,680 --> 00:47:11,680
Because it wasn't a thing back
in the day where, I mean, we're

839
00:47:11,680 --> 00:47:14,080
all stuck in these niche fields
and everybody's just doing their

840
00:47:14,080 --> 00:47:16,400
own work.
Even within biology, you'll have

841
00:47:16,400 --> 00:47:19,080
a biologist who knows nothing
about a specific molecule and

842
00:47:19,520 --> 00:47:21,800
there's so little that everyone
knows about everyone else's

843
00:47:21,800 --> 00:47:23,960
work.
But but I see you guys doing

844
00:47:23,960 --> 00:47:27,200
this in such an
interdisciplinary way that it's

845
00:47:27,200 --> 00:47:29,560
it's pretty epic to watch from
from the outside.

846
00:47:29,720 --> 00:47:34,080
And how's that been for you?
Well, first of all, it's been

847
00:47:34,080 --> 00:47:36,560
very difficult and it's been
very slow going.

848
00:47:36,560 --> 00:47:39,160
So we've been talking about the
Xenobots, which is kind of the

849
00:47:39,160 --> 00:47:42,600
end, you know, or it's a, it's a
stopping point on a very long

850
00:47:42,600 --> 00:47:45,760
journey.
I was very fortunate and again,

851
00:47:46,040 --> 00:47:51,280
just by chance to have fallen
into various programs at

852
00:47:51,280 --> 00:47:55,360
different universities in, in
four different countries that

853
00:47:55,360 --> 00:47:58,880
taught me how to be an
interdisciplinary researcher.

854
00:47:58,880 --> 00:48:02,200
There are ways, you know, to do
it very badly, right?

855
00:48:02,200 --> 00:48:06,120
So, you know, Jack of all
trades, master of none, you

856
00:48:06,120 --> 00:48:09,920
know, there, it's not an easy
thing to do that The, the reason

857
00:48:09,920 --> 00:48:12,560
that most people are
specialists, and I'm glad there

858
00:48:12,560 --> 00:48:16,920
are specialists, you know, is
because it's easier to drill

859
00:48:16,920 --> 00:48:21,680
down and really grasp a branch
of human knowledge and extend it

860
00:48:22,080 --> 00:48:25,600
than it is to try and connect
two branches together.

861
00:48:25,600 --> 00:48:30,040
It takes at least twice as long.
So in the case of my

862
00:48:30,040 --> 00:48:33,160
collaboration with Mike, I told
you about, you know, the, the

863
00:48:33,160 --> 00:48:35,080
Xeno sculpture and then the
xenobot.

864
00:48:35,080 --> 00:48:38,120
But you know, it took a long
time to get to that point.

865
00:48:38,160 --> 00:48:40,360
We, we don't even speak the same
languages.

866
00:48:40,360 --> 00:48:44,800
You have to learn, you know how
what each are capable of, what

867
00:48:44,800 --> 00:48:48,680
takes a week, what takes five
years, you know all those sorts

868
00:48:48,680 --> 00:48:52,120
of things.
And then you can proceed

869
00:48:52,120 --> 00:48:55,560
carefully.
And of course, in some cases all

870
00:48:55,560 --> 00:48:57,800
of this effort is worth it
because you get lucky with

871
00:48:57,800 --> 00:49:02,080
something like the Xenobots, but
it is definitely not the norm.

872
00:49:02,520 --> 00:49:04,800
Yeah, and then you get stuck in
a podcast that's philosophical

873
00:49:04,800 --> 00:49:06,600
based on the mind body problem
being asked.

874
00:49:06,600 --> 00:49:09,440
What is consciousness?
That's right, that's right.

875
00:49:09,720 --> 00:49:11,480
You got to roll with the roll
with the punches.

876
00:49:12,880 --> 00:49:15,760
Tell me, Josh, when you, when
you think about this field

877
00:49:15,760 --> 00:49:18,920
moving forward, what excites you
most about it?

878
00:49:18,920 --> 00:49:21,560
Where, what do you think in
terms of the the future of the

879
00:49:21,560 --> 00:49:24,720
industry, people that are
working perhaps under you or

880
00:49:24,960 --> 00:49:27,560
work that you've seen what
what's what's been most exciting

881
00:49:27,560 --> 00:49:32,040
and even within your own work?
Yeah, so, so I am interested in

882
00:49:32,040 --> 00:49:33,440
the philosophical side of
things.

883
00:49:33,440 --> 00:49:37,400
So one of the things that I see
in the Xenobots and some of the

884
00:49:37,400 --> 00:49:42,600
other kind of exotic robots that
we work with is a, a conflation

885
00:49:42,600 --> 00:49:47,080
of thought and action.
So in a, if you take a Roomba,

886
00:49:47,080 --> 00:49:50,280
the the robot vacuum cleaner,
you can point to the wheels and

887
00:49:50,280 --> 00:49:54,080
you can say action.
You can open it up and point to

888
00:49:54,080 --> 00:49:58,080
the central, the CPU and say,
you know, thought or cognition

889
00:49:58,080 --> 00:50:01,480
or processing, you know,
traditional robots, there is a

890
00:50:01,480 --> 00:50:05,200
Cartesian cut you can make that
separates the body from the

891
00:50:05,200 --> 00:50:08,080
brain.
But what I'm excited about is if

892
00:50:08,080 --> 00:50:11,920
you look at a lot of cutting
edge technologies, that

893
00:50:11,920 --> 00:50:18,080
separation is becoming less and
less obvious, which I think is,

894
00:50:18,080 --> 00:50:21,920
you know, finally the beginnings
of an acid that's dissolving

895
00:50:21,920 --> 00:50:25,520
this distinction that at least
in the West has been around for

896
00:50:25,520 --> 00:50:30,160
over 300 years.
And so I think that's great from

897
00:50:30,160 --> 00:50:32,240
a purely intellectual point of
view.

898
00:50:32,560 --> 00:50:35,960
But again, it, it also is
important for us as a species to

899
00:50:35,960 --> 00:50:39,000
understand that, you know,
brains are not everything and

900
00:50:39,000 --> 00:50:42,160
everything else is expendable.
You know, it's the story is much

901
00:50:42,160 --> 00:50:48,680
more complicated than that.
And I think that's finally

902
00:50:48,680 --> 00:50:52,800
overturning Cartesian dualism
will actually be a positive

903
00:50:52,800 --> 00:50:56,600
thing for for society, assuming
it actually happens.

904
00:50:57,120 --> 00:51:01,680
You know, the Keith Frankish,
the one who termed it, coined

905
00:51:01,680 --> 00:51:04,280
the term illusionism, He's
writing a book called Escaping

906
00:51:04,280 --> 00:51:07,440
Descartes Prison and it's
fundamentally premised on that,

907
00:51:07,480 --> 00:51:10,560
on that whole idea that this
particular dualism is, is

908
00:51:10,560 --> 00:51:12,720
definitely on a decline,
however.

909
00:51:12,960 --> 00:51:16,000
I know, I know Keith, and I'm
glad to know he's also like,

910
00:51:16,000 --> 00:51:18,040
wielding an axe from a different
direction.

911
00:51:18,040 --> 00:51:22,280
Well, we're doing our empiricist
best, and yeah, hopefully the

912
00:51:22,280 --> 00:51:25,840
tree will come down.
But but on the flip side, if you

913
00:51:25,840 --> 00:51:28,640
think about it, Josh, while that
Cartesian dualism is breaking

914
00:51:28,640 --> 00:51:31,560
down, you've got a rise in Ben
psychicism and idealism as well.

915
00:51:31,880 --> 00:51:34,920
So you've got a whole group of
thinkers that think that either

916
00:51:34,920 --> 00:51:37,640
everything is conscious or
everything is consciousness.

917
00:51:38,120 --> 00:51:40,200
What are your thoughts on that?
I know this is not your field.

918
00:51:40,200 --> 00:51:43,640
Now we're just going beyond.
You know what, I'm OK with it.

919
00:51:43,640 --> 00:51:47,920
I'm I'm OK with that.
It's actually the the the stance

920
00:51:47,920 --> 00:51:50,000
in the middle that I find
problematic.

921
00:51:50,640 --> 00:51:53,880
I am conscious and that thing
over there isn't.

922
00:51:53,880 --> 00:51:56,040
So it doesn't matter what I do
to it.

923
00:51:56,280 --> 00:52:00,280
Like that's the problem.
So if there are a greater number

924
00:52:00,280 --> 00:52:03,880
of folks that believe everything
is conscious and we should be

925
00:52:04,240 --> 00:52:08,360
careful about how we interact
with other things and other

926
00:52:09,200 --> 00:52:12,840
other entities, great.
If the illusionists say, listen,

927
00:52:12,840 --> 00:52:15,680
we're all on the same level
playing field, you know, for

928
00:52:15,680 --> 00:52:18,120
different reasons, that's also
good.

929
00:52:18,120 --> 00:52:21,280
I think it's yeah.
My problem is the actually the

930
00:52:21,280 --> 00:52:23,280
intermediate stance.
Yeah, I agree.

931
00:52:23,280 --> 00:52:27,520
I think ethically and, and sort
of the the the moral philosophy

932
00:52:27,520 --> 00:52:30,720
behind both of those, whether
it's illusionism or panpsychism,

933
00:52:30,720 --> 00:52:33,000
pretty much lead to the same
thing and that we're all the

934
00:52:33,000 --> 00:52:35,480
same thing.
We're all the same thing,

935
00:52:35,480 --> 00:52:37,560
exactly.
And I think that's all the

936
00:52:37,920 --> 00:52:39,680
lines.
And even within an idealist

937
00:52:39,680 --> 00:52:42,160
philosophy where they think that
everything is consciousness and

938
00:52:42,160 --> 00:52:45,000
we're just a part of it, that
also tends to always lead to

939
00:52:45,000 --> 00:52:48,040
some sort of a this
consciousness trying to love

940
00:52:48,040 --> 00:52:50,360
itself, enjoy itself.
It always leads in some sort of

941
00:52:50,360 --> 00:52:53,760
an empathetic positive
worldview, which I think in

942
00:52:53,760 --> 00:52:56,280
essence would be a better option
anyway, instead of drawing these

943
00:52:56,280 --> 00:52:59,800
fundamental lines between one
thing or another.

944
00:52:59,880 --> 00:53:04,240
But when when you think about
your work currently, what are

945
00:53:04,240 --> 00:53:07,800
you guys doing that you think we
should be looking forward to the

946
00:53:07,800 --> 00:53:09,880
most and that you're most
excited about?

947
00:53:10,640 --> 00:53:14,560
Yeah.
So this is true of the xenobots

948
00:53:14,560 --> 00:53:18,000
in general, but we are also
doing a lot of work in my lab

949
00:53:18,000 --> 00:53:21,800
with what are called meta
materials, and these are human

950
00:53:21,800 --> 00:53:25,520
engineered materials that act
very differently from natural

951
00:53:25,520 --> 00:53:28,560
materials.
And it turns out that these meta

952
00:53:28,560 --> 00:53:34,080
materials also are capable of
thought or cognition and action

953
00:53:34,080 --> 00:53:38,040
simultaneously.
Imagine that I take a sheet of

954
00:53:38,040 --> 00:53:41,840
this material and I start
vibrating it, but I vibrate it

955
00:53:41,840 --> 00:53:45,200
at different frequencies.
It turns out you can get this

956
00:53:45,200 --> 00:53:49,080
material to do different things
at these different frequencies

957
00:53:49,440 --> 00:53:52,920
and actually use these different
frequencies to carry

958
00:53:52,920 --> 00:53:55,600
information.
You can you can design A sheet

959
00:53:55,600 --> 00:53:59,480
of this material that when you
shake it, it performs logical

960
00:53:59,480 --> 00:54:04,720
AND at one frequency and logical
OR at the other frequency

961
00:54:04,800 --> 00:54:08,560
simultaneously.
And while computing that

962
00:54:08,560 --> 00:54:11,880
function, the material is moving
and if you put it on the ground,

963
00:54:11,880 --> 00:54:13,920
it will actually move along the
ground.

964
00:54:13,920 --> 00:54:17,800
So where is the distinction
between thought and action?

965
00:54:18,880 --> 00:54:22,400
So in answer to your question,
I'm excited about, you know, all

966
00:54:22,400 --> 00:54:27,040
of these exotic materials, which
includes living materials that

967
00:54:27,040 --> 00:54:32,440
conflate thought and action.
And then a is ability to exploit

968
00:54:32,440 --> 00:54:35,920
the potential of these new
materials to make useful

969
00:54:35,920 --> 00:54:40,440
machines in ways, you know,
Hollywood hasn't even begun to

970
00:54:40,440 --> 00:54:43,200
explore.
They will, I guess, but not not

971
00:54:43,200 --> 00:54:45,320
yet.
They always eventually catch up

972
00:54:45,320 --> 00:54:49,160
and but but that's incredible.
I mean that the, there's there's

973
00:54:49,160 --> 00:54:52,040
so much ground breaking stuff
occurring that I mean, you're

974
00:54:52,040 --> 00:54:54,120
right.
The the pace even for me when I

975
00:54:54,120 --> 00:54:56,440
watch what you guys do,
sometimes when I look at it, I

976
00:54:56,440 --> 00:54:58,840
think like this is crazy how
fast this is happening.

977
00:54:59,480 --> 00:55:02,000
Like you see five years ago one
Xenobots made and now suddenly

978
00:55:02,000 --> 00:55:04,720
you guys are making materials
that can vibrate at different I

979
00:55:04,880 --> 00:55:07,720
think.
Yeah, I mean, the the reason I

980
00:55:07,720 --> 00:55:11,120
would say is again, it Mike and
myself and some others, you

981
00:55:11,120 --> 00:55:14,120
know, have been on a very long
journey to try and see things

982
00:55:14,120 --> 00:55:17,520
differently, to escape from the
Cartesian worldview, right.

983
00:55:17,520 --> 00:55:21,480
So if you don't escape, you
know, you do things like ChatGPT

984
00:55:21,480 --> 00:55:24,080
and you try and make bigger and
bigger brains, which I would say

985
00:55:24,080 --> 00:55:27,640
is also useful.
Nothing against, you know, open

986
00:55:27,720 --> 00:55:30,120
AI and ChatGPT and Stable
Diffusion.

987
00:55:30,120 --> 00:55:34,080
They're they're great.
But the reason I think that that

988
00:55:34,080 --> 00:55:36,560
we are discovering a lot of
things is because we're not

989
00:55:36,560 --> 00:55:38,960
looking under the lamppost
anymore, right?

990
00:55:38,960 --> 00:55:41,560
We're out in the dark and it
turns out there's a lot of stuff

991
00:55:41,560 --> 00:55:44,920
out there.
It's just that people haven't,

992
00:55:45,120 --> 00:55:48,720
haven't thought to look there,
been able to get there because

993
00:55:48,720 --> 00:55:52,040
you have to think in a non
Cartesian way to get there.

994
00:55:52,040 --> 00:55:56,720
You have to ignore a distinction
between body and brain, and then

995
00:55:56,720 --> 00:56:00,080
you start finding these weird
materials and finding ways to

996
00:56:00,080 --> 00:56:03,720
bend them to new purposes.
Something I found particularly

997
00:56:03,720 --> 00:56:06,800
fascinating when I listen to you
speak was you mentioned

998
00:56:06,800 --> 00:56:11,880
something about cause and effect
and, and until an artificial

999
00:56:11,880 --> 00:56:15,920
intelligence is able to actually
cause an effect, that's when you

1000
00:56:15,920 --> 00:56:18,720
know that this thing is has
become a fundamentally part of

1001
00:56:18,720 --> 00:56:19,960
reality.
Do you want to expand on that

1002
00:56:19,960 --> 00:56:21,520
idea?
Sure, sure.

1003
00:56:21,520 --> 00:56:25,720
So this is something that
current AI technology struggle

1004
00:56:25,720 --> 00:56:30,040
with cause and effect.
You know, the best, the best a

1005
00:56:30,040 --> 00:56:34,840
non embodied AI can do is sort
of read about all cause effect

1006
00:56:34,840 --> 00:56:37,680
relationships that humans have
written about on the Internet or

1007
00:56:37,680 --> 00:56:41,240
in books or whatever.
That's that's all those AIS have

1008
00:56:41,240 --> 00:56:45,320
access to and they can do OK
with cause and effect.

1009
00:56:45,320 --> 00:56:49,040
Not great, but again, you know
how the body shapes the way we

1010
00:56:49,040 --> 00:56:50,920
think.
Like I said, I have a four year

1011
00:56:50,920 --> 00:56:54,080
old at home and he pushes a
coffee mug off the table and it

1012
00:56:54,080 --> 00:56:56,360
shatters like he caused an
effect.

1013
00:56:56,640 --> 00:57:00,000
And not only did it shatter, but
he sees how his two parents

1014
00:57:00,000 --> 00:57:02,760
react and how they react
differently to what's there's.

1015
00:57:03,120 --> 00:57:06,360
When you can push against the
world, literally, you know,

1016
00:57:06,360 --> 00:57:10,600
there is a massive amount of
rich effectual data that comes

1017
00:57:10,600 --> 00:57:13,200
back.
And so I say this a lot.

1018
00:57:13,200 --> 00:57:15,960
Embodied cognition is about
pushing against the world and

1019
00:57:15,960 --> 00:57:17,720
observing how the world pushes
back.

1020
00:57:18,040 --> 00:57:22,320
And that is a cause effect loop.
You can cause them in all sorts

1021
00:57:22,320 --> 00:57:27,600
of ways from, you know, pushing
a coffee cup off a table to, you

1022
00:57:27,600 --> 00:57:31,760
know, changing a society,
landing someone on the moon, all

1023
00:57:31,760 --> 00:57:35,840
sorts of things.
And that that gives embodied

1024
00:57:35,840 --> 00:57:39,040
learners, not just human
learners, but non human

1025
00:57:39,040 --> 00:57:42,160
learners, you know, front row
seat for cause and effect.

1026
00:57:42,400 --> 00:57:45,880
If you're not sure, if you're
thinking that this effect is

1027
00:57:45,880 --> 00:57:49,760
caused by that cause, just go
out and verify it be be a

1028
00:57:49,760 --> 00:57:52,240
scientist, push against the
world and see whether you were

1029
00:57:52,240 --> 00:57:55,360
right or not.
It's actually very simple when

1030
00:57:55,360 --> 00:57:58,560
you think about it.
So I think there are some of

1031
00:57:58,560 --> 00:58:02,600
some of these examples that make
clear why we can't just ignore

1032
00:58:02,600 --> 00:58:04,760
the body.
It matters in very important

1033
00:58:04,760 --> 00:58:05,160
ways.
Yeah.

1034
00:58:05,400 --> 00:58:07,200
It runs, I think it was a paper
I read.

1035
00:58:07,200 --> 00:58:09,240
I can't remember the author's
name, but I'll try and find it.

1036
00:58:09,240 --> 00:58:10,560
But it's called the nobody
problem.

1037
00:58:10,560 --> 00:58:13,360
Just that there's a problem in
having nobody.

1038
00:58:13,400 --> 00:58:15,120
That's the thing.
It's it's a fundamental.

1039
00:58:15,120 --> 00:58:17,000
That's the real problem.
Agreed.

1040
00:58:17,120 --> 00:58:18,880
Nobody.
Is a big problem.

1041
00:58:19,240 --> 00:58:21,400
Yeah.
And it's it's it's, yeah.

1042
00:58:21,400 --> 00:58:23,360
It goes down to the crux of what
we're talking about.

1043
00:58:23,560 --> 00:58:26,800
That same cause and effect that
you're talking about is that

1044
00:58:26,800 --> 00:58:29,680
almost it reminds you of that
inactivist part, the fact that

1045
00:58:29,680 --> 00:58:32,760
we're here trying to
consistently and continuously do

1046
00:58:32,760 --> 00:58:36,160
something like there's a purpose
driven by behind everything we

1047
00:58:36,160 --> 00:58:37,200
do.
We must find food.

1048
00:58:37,200 --> 00:58:38,560
We must.
We see it.

1049
00:58:38,560 --> 00:58:40,760
We see something red.
We know it might be an apple.

1050
00:58:40,760 --> 00:58:43,520
There's, there's this
interaction with reality that's

1051
00:58:43,560 --> 00:58:47,000
always looping back and forth.
And do you think fundamentally

1052
00:58:47,000 --> 00:58:51,120
with AI we would either have to
program it to have some sort of

1053
00:58:51,280 --> 00:58:55,320
a personal loss of if it does
not exist, we must have some

1054
00:58:55,320 --> 00:58:58,440
sort of program within it or
does it not need something like

1055
00:58:58,440 --> 00:59:02,400
that in order to?
Be I, I, I think it absolutely

1056
00:59:02,400 --> 00:59:06,520
needs it, but not necessarily in
the literal sense of, you know,

1057
00:59:06,520 --> 00:59:09,480
making robots that push against
the world and observing how the

1058
00:59:09,480 --> 00:59:12,680
world pushes back.
All the, you know, ChatGPT and

1059
00:59:12,680 --> 00:59:15,680
all the chat bots, they kind of
already are doing it because

1060
00:59:16,000 --> 00:59:20,320
they say stuff and that affects
their human interlocutors, some

1061
00:59:20,320 --> 00:59:23,960
of whom go off and do things and
then report back to ChatGPT,

1062
00:59:24,200 --> 00:59:26,360
which becomes part of its
training set.

1063
00:59:26,360 --> 00:59:31,600
So, you know, modern AI, they
already have, you know, slaves,

1064
00:59:31,600 --> 00:59:34,160
which is us.
You know, they can use us to to

1065
00:59:34,160 --> 00:59:36,560
try out cause and effect
relationships.

1066
00:59:37,040 --> 00:59:40,240
I don't know, you know, what the
big AI companies are doing.

1067
00:59:41,160 --> 00:59:43,560
A lot of them are very smart.
They probably started to figure

1068
00:59:43,560 --> 00:59:46,320
this out.
So, you know, there are ways

1069
00:59:46,320 --> 00:59:48,480
that you can learn about the
real world.

1070
00:59:48,480 --> 00:59:51,120
You can learn about cause and
effect where you have something

1071
00:59:51,120 --> 00:59:54,960
or someone else go and, you
know, embody that cause and

1072
00:59:54,960 --> 00:59:58,880
effect loop for you.
So it's already happening.

1073
00:59:59,000 --> 01:00:01,920
It's a crazy thought.
In other words, we are AI's

1074
01:00:01,920 --> 01:00:05,040
extended mind.
We we absolutely are.

1075
01:00:05,040 --> 01:00:07,760
So it is an extended mind
already It's got I don't know

1076
01:00:07,760 --> 01:00:09,880
how many of us as its end
effectors.

1077
01:00:09,880 --> 01:00:13,680
It's it's happening.
Do you think that a, a, a bigger

1078
01:00:13,680 --> 01:00:16,720
step would to take this even
further would be then giving it

1079
01:00:16,720 --> 01:00:21,320
this sort of sense of loss?
So having such a hard wired

1080
01:00:21,320 --> 01:00:25,640
program that if it ceases to
exist, it's a, that's a

1081
01:00:25,640 --> 01:00:28,200
fundamental problem, the same
way we are so afraid of death.

1082
01:00:28,680 --> 01:00:31,440
You think that's a big factor
that we'll have to consider in

1083
01:00:31,440 --> 01:00:34,080
the future to make this?
It's a it's a really good

1084
01:00:34,080 --> 01:00:35,920
question.
I don't know whether, like, a

1085
01:00:35,920 --> 01:00:39,520
fear of death or extinction is a
necessary component for

1086
01:00:39,520 --> 01:00:44,000
developing embodied cognition.
If it is, that's going to be

1087
01:00:44,000 --> 01:00:47,440
problematic because this is
something Hollywood has explored

1088
01:00:47,440 --> 01:00:50,760
in great detail.
If you know, if you give

1089
01:00:51,040 --> 01:00:54,760
technologies existential dread,
you know they'll do everything

1090
01:00:54,760 --> 01:00:56,760
they can in their power to avoid
it.

1091
01:00:56,760 --> 01:00:59,440
And I agree with it.
That just makes sense.

1092
01:01:00,120 --> 01:01:02,800
So I, I think actually it's a
research question.

1093
01:01:02,800 --> 01:01:06,440
Like we started talking about
self-awareness and how that

1094
01:01:06,440 --> 01:01:11,400
might have evolved in us to
avoid extinction, avoid death.

1095
01:01:12,600 --> 01:01:16,440
But it's not necessarily true
that technologies can't become

1096
01:01:16,440 --> 01:01:20,600
self aware and practice cause
and effect internally and verify

1097
01:01:20,600 --> 01:01:24,840
it physically that that all
needs to be driven by, you know,

1098
01:01:24,920 --> 01:01:28,640
fear of extinction.
I hope it it's not true.

1099
01:01:28,640 --> 01:01:32,000
I hope that, you know, they
don't have to rely on that for

1100
01:01:32,000 --> 01:01:34,600
their motivation.
But I guess, I guess we'll see.

1101
01:01:35,840 --> 01:01:38,640
A lot's going to probably change
the next few years and I think

1102
01:01:38,640 --> 01:01:42,000
the, well, as we said, it's very
fast-paced, not to the point of

1103
01:01:42,000 --> 01:01:45,560
a singularity, but very fast in
general.

1104
01:01:45,760 --> 01:01:48,880
When you look back at your views
on consciousness in the mind,

1105
01:01:49,760 --> 01:01:53,800
how often have you changed your
views regarding consciousness

1106
01:01:53,800 --> 01:01:56,320
and the mind body problem?
So when you were young, let's

1107
01:01:56,320 --> 01:01:59,600
say prior to entering the field
as a student perhaps, what did

1108
01:01:59,600 --> 01:02:01,320
you have a Cartesian mindset on
this?

1109
01:02:01,320 --> 01:02:03,160
Did you think there was a sort
of soul or?

1110
01:02:04,920 --> 01:02:08,400
I did not have a religious,
yeah, I did not have a religious

1111
01:02:08,400 --> 01:02:11,000
upbringing, but yeah, I was
convinced there was someone in

1112
01:02:11,000 --> 01:02:14,200
there.
But, and that's just because I

1113
01:02:14,200 --> 01:02:16,960
was raised in the West, pretty
much everyone, even if you're an

1114
01:02:16,960 --> 01:02:21,600
atheist, that's what you're
taught indirectly and directly.

1115
01:02:21,960 --> 01:02:26,840
So, so my awakening was at the
University of Sussex.

1116
01:02:26,840 --> 01:02:30,720
I did my masters there and
Margaret Bowden and Phil

1117
01:02:30,720 --> 01:02:33,360
husbands and Inman Harvey and
Anil Seth.

1118
01:02:33,640 --> 01:02:36,920
You know, they did a very good
job of disabusing us of that

1119
01:02:36,920 --> 01:02:40,320
notion, or at least, you know,
presenting it that it's just one

1120
01:02:40,320 --> 01:02:43,320
view.
And so that was that was the

1121
01:02:43,320 --> 01:02:46,440
start for me.
And then, you know, then

1122
01:02:46,520 --> 01:02:50,560
developing self aware robots and
actually kind of how easy it was

1123
01:02:50,560 --> 01:02:53,440
to do that.
What wasn't super easy, but

1124
01:02:53,840 --> 01:02:56,600
surprisingly easy.
That was the nail and that was

1125
01:02:56,600 --> 01:02:58,360
the final nail in the coffin for
me.

1126
01:02:58,360 --> 01:03:04,200
I was like, if we can make this
$29.00 robot, you know, be able

1127
01:03:04,200 --> 01:03:07,680
to create an awareness of self,
you know, it's not really

1128
01:03:07,680 --> 01:03:10,200
something special.
It's probably, you know, it's

1129
01:03:10,240 --> 01:03:13,240
not not we should take it off
the pedestal, let's put it that

1130
01:03:13,240 --> 01:03:15,040
way.
Do you think there'll be a

1131
01:03:15,040 --> 01:03:18,520
fundamental shift or difference
when you're able to give these

1132
01:03:19,640 --> 01:03:26,080
self aware agents agency or or
or allow it to make choices and

1133
01:03:26,080 --> 01:03:28,800
and then interact with those
cause and effects at some point?

1134
01:03:28,800 --> 01:03:32,240
Yeah, absolutely.
Because one of the things I love

1135
01:03:32,240 --> 01:03:35,440
about AI and robots is like how
creative they are.

1136
01:03:35,440 --> 01:03:39,200
So, you know, there's all these
hilarious AI fails, but they're

1137
01:03:39,200 --> 01:03:41,400
also, you know, some of them are
really creative.

1138
01:03:41,640 --> 01:03:44,000
So it'd be great if they're also
creative wins.

1139
01:03:44,000 --> 01:03:47,160
You know, like, again, when the
AI says, I, I'm going to make a

1140
01:03:47,160 --> 01:03:50,280
protein like this and all the
human experts on the planet say

1141
01:03:50,280 --> 01:03:51,840
you can't make a protein like
that.

1142
01:03:51,840 --> 01:03:54,800
And then of course it does.
Like, that's it.

1143
01:03:54,800 --> 01:03:58,560
So, you know, humans can only
push against the world in so

1144
01:03:58,560 --> 01:04:01,400
many ways.
And wouldn't it be great if we

1145
01:04:01,400 --> 01:04:04,640
had allies, you know, here in
the real world that could push

1146
01:04:04,640 --> 01:04:07,680
on the world in ways that we
just can't?

1147
01:04:08,120 --> 01:04:11,080
And they're going to learn cause
and effect relationships that

1148
01:04:11,080 --> 01:04:13,600
would have been impossible or
would have taken us a very, very

1149
01:04:13,600 --> 01:04:17,160
long time to discover.
That's the future that I'm

1150
01:04:17,160 --> 01:04:20,920
excited about.
And I think it's, I think it's

1151
01:04:20,920 --> 01:04:23,640
incredible work.
You guys are really changing the

1152
01:04:23,640 --> 01:04:26,360
game.
And the there's, there's so many

1153
01:04:26,360 --> 01:04:29,000
different aspects to the work
that we haven't really even

1154
01:04:29,000 --> 01:04:31,200
touched on because there's so
much greater detail that we

1155
01:04:31,200 --> 01:04:33,920
could go into.
But is there anything particular

1156
01:04:33,920 --> 01:04:35,520
that you'd like to touch and
that you feel like you haven't

1157
01:04:35,520 --> 01:04:38,440
mentioned about your work in
general, Josh, that because this

1158
01:04:38,440 --> 01:04:40,640
is obviously just the first
podcast we've had together, but

1159
01:04:41,000 --> 01:04:43,520
there's many papers we could
dissect in so much detail.

1160
01:04:43,520 --> 01:04:46,400
But before we ever do, is there
anything particular you'd like

1161
01:04:46,400 --> 01:04:49,760
to mention?
Well, one thing we, we, I think

1162
01:04:49,760 --> 01:04:52,760
we mentioned in passing, but
didn't have a chance to address

1163
01:04:52,760 --> 01:04:58,240
is how the, the work on xenobots
and even AI design biology is

1164
01:04:58,240 --> 01:05:00,640
changing like our understanding
of life itself.

1165
01:05:02,000 --> 01:05:06,520
You know, most of us were taught
that, you know, frog DNA codes

1166
01:05:06,520 --> 01:05:09,400
for frogs and human DNA codes
for humans.

1167
01:05:09,400 --> 01:05:12,960
And what I've learned in working
with Mike, and also just looking

1168
01:05:12,960 --> 01:05:16,600
at the intersection between AI
and biology in general, is that

1169
01:05:17,600 --> 01:05:20,680
the species and organisms that
exist on the planet, they're

1170
01:05:20,680 --> 01:05:25,320
points in an attractor space.
So this sort of means that when

1171
01:05:25,320 --> 01:05:30,520
you're in familiar environments,
your genes tend to build humans

1172
01:05:30,520 --> 01:05:33,840
or frogs or what have you, but
that you can change the

1173
01:05:33,840 --> 01:05:39,240
environment in quite drastic
ways, including the embryo

1174
01:05:39,240 --> 01:05:43,400
itself, or even take it apart
into its component cells and put

1175
01:05:43,400 --> 01:05:46,960
them together in new ways.
And without changing any of the

1176
01:05:46,960 --> 01:05:50,480
genetics, you get new form and
function.

1177
01:05:50,960 --> 01:05:54,200
So there there's an old idea in
biology of what's called Morpho

1178
01:05:54,200 --> 01:05:59,120
space, and it's this imaginary
high dimensional space of every

1179
01:05:59,120 --> 01:06:03,040
possible living thing that could
survive on this planet.

1180
01:06:03,560 --> 01:06:07,280
And despite how creative Mother
Nature's been over the last 3.5

1181
01:06:07,280 --> 01:06:10,000
billion years, she's only
explored A vanishingly small

1182
01:06:10,000 --> 01:06:13,800
part of Morpho space.
So, you know, the obvious

1183
01:06:13,800 --> 01:06:15,360
question is what else is out
there?

1184
01:06:15,360 --> 01:06:18,600
You know, everyone you know is
curious about aliens, and maybe

1185
01:06:18,600 --> 01:06:20,920
we'll find some on another
planet someday.

1186
01:06:20,920 --> 01:06:26,000
But we can kind of find them
like now here by asking an AI to

1187
01:06:26,000 --> 01:06:29,440
say, here's frog, here's human,
here's axolotl.

1188
01:06:29,560 --> 01:06:31,840
Now go in the opposite
direction, put them at your

1189
01:06:31,840 --> 01:06:36,560
back, and drive into morphal
space as far from familiar

1190
01:06:36,560 --> 01:06:39,640
organisms as you can to find
those that can exist here.

1191
01:06:40,000 --> 01:06:41,960
And then build them for us,
please.

1192
01:06:42,520 --> 01:06:45,920
The, the prospect of that, you
know, is going to revolutionize

1193
01:06:46,440 --> 01:06:49,600
biology, you know, beyond
recognition, I think.

1194
01:06:50,160 --> 01:06:53,480
I, I literally had one of those
questions down to to further

1195
01:06:53,480 --> 01:06:55,200
elaborate on that.
Remember when I told you about

1196
01:06:55,200 --> 01:06:57,880
the X, the extra questions I had
planned for us?

1197
01:06:58,360 --> 01:07:01,480
Yes, when when you it brings me
back to one of the other

1198
01:07:01,480 --> 01:07:03,880
questions I wrote.
You've used evolutionary

1199
01:07:03,880 --> 01:07:06,800
algorithms to evolve solutions
to both software and hardware.

1200
01:07:07,440 --> 01:07:12,080
How do these algorithms reflect,
or perhaps extend the principles

1201
01:07:12,080 --> 01:07:15,600
of natural selection?
Oh, yeah.

1202
01:07:17,120 --> 01:07:21,640
Gosh, I don't know.
I mean, I'm not a biologist, so

1203
01:07:21,640 --> 01:07:25,760
I'm sort of speaking beyond my
professional expertise here.

1204
01:07:26,120 --> 01:07:29,240
I don't know that it really is
beyond natural evolution.

1205
01:07:29,240 --> 01:07:33,960
I mean, we're just, you know, AI
is a product of us, and we're a

1206
01:07:33,960 --> 01:07:37,240
product of natural evolution.
And the AI is modifying or

1207
01:07:37,320 --> 01:07:40,440
putting pressures on these
organisms in ways they haven't

1208
01:07:40,440 --> 01:07:43,680
experienced before.
So, you know, from a bit of a

1209
01:07:43,680 --> 01:07:47,760
distance and if you squint, it
is just, you know, natural

1210
01:07:47,760 --> 01:07:50,120
selection.
It's just entered a new chapter,

1211
01:07:50,120 --> 01:07:53,120
like when, you know, the fish
crawled out of the Seas or we

1212
01:07:53,120 --> 01:07:56,800
came down from the trees.
It's just feels like a different

1213
01:07:56,800 --> 01:08:01,920
phase in a similar process.
So I, yeah, I don't know if

1214
01:08:01,920 --> 01:08:04,680
we've broken out of the bounds
of natural selection yet.

1215
01:08:04,680 --> 01:08:08,120
It it is unnatural, of course,
because there's things like AI

1216
01:08:08,120 --> 01:08:13,600
and robots and a supercomputers
involved, but ultimately it's

1217
01:08:13,600 --> 01:08:19,040
just challenging genetically
unmodified materials to survive

1218
01:08:19,040 --> 01:08:23,200
and continue, and they're
finding a way to do so.

1219
01:08:23,640 --> 01:08:26,200
The I I agree with you.
I think I had this conversation

1220
01:08:26,200 --> 01:08:30,439
with someone recently where we
spoke about the fact that it

1221
01:08:30,439 --> 01:08:33,160
took us billions of years to get
you to have this sort of

1222
01:08:33,319 --> 01:08:35,120
experience.
It doesn't mean that if we're

1223
01:08:35,120 --> 01:08:38,600
able to do it within 5 minutes
in the lab, that that this is

1224
01:08:38,600 --> 01:08:41,840
some sort of a miracle or it's
just literally all that work

1225
01:08:41,840 --> 01:08:44,399
that's taken place, as you know,
with the cells that you take.

1226
01:08:44,680 --> 01:08:48,279
It's already got so much prior
information and it's just always

1227
01:08:48,279 --> 01:08:50,279
working on itself.
So it's not like we're going

1228
01:08:50,279 --> 01:08:52,439
beyond the bounds of of nature
itself.

1229
01:08:53,200 --> 01:08:55,439
And I think, you know, the other
thing is we talked about the

1230
01:08:55,439 --> 01:08:58,640
Cartesian dualism.
A lot of this work, not just

1231
01:08:58,640 --> 01:09:01,760
ours but others, as well as
exploding the concept of self.

1232
01:09:01,760 --> 01:09:05,359
We tend to view things about
like at the self level, if you

1233
01:09:05,359 --> 01:09:08,319
think about the xenobots and
that like the cells that make up

1234
01:09:08,319 --> 01:09:12,560
the xenobots from the cells
perspective, things might not

1235
01:09:12,560 --> 01:09:15,439
actually be that different, you
know, or problematic.

1236
01:09:16,319 --> 01:09:19,560
When a, when a, when a frog
develops or any Organism

1237
01:09:19,560 --> 01:09:22,920
develops, the cells that are
part of the developmental

1238
01:09:22,920 --> 01:09:26,920
process, they're being pulled
and pushed and sheared and, you

1239
01:09:26,920 --> 01:09:30,439
know, shoved through tubes like
crazy all the time.

1240
01:09:30,800 --> 01:09:34,880
So all of our manipulations for
from the perspective of the cell

1241
01:09:34,880 --> 01:09:36,960
might just be, you know,
business as usual.

1242
01:09:36,960 --> 01:09:40,560
It's find your neighbors and
connect to them and establish

1243
01:09:40,560 --> 01:09:46,160
calcium communication again.
And you know, so I agree, like

1244
01:09:46,160 --> 01:09:49,880
the xenobots in some way seems
miraculous if you think about it

1245
01:09:49,880 --> 01:09:52,840
from the frog's perspective.
But if you think about it from

1246
01:09:53,040 --> 01:09:57,600
the frog's cells perspective,
maybe it's not so, you know,

1247
01:09:57,600 --> 01:10:00,880
surprising or traumatic.
And I think that's where I think

1248
01:10:00,880 --> 01:10:03,240
Mike at some point, and that's
when they see cognition all the

1249
01:10:03,240 --> 01:10:07,360
way going all the way down, is
that it's really, really hard to

1250
01:10:07,360 --> 01:10:10,960
actually differentiate this,
this type of process in terms of

1251
01:10:10,960 --> 01:10:12,960
cognition.
When you look at a Xenobot and

1252
01:10:12,960 --> 01:10:16,920
you see it exhibiting some sort
of autonomous behavior.

1253
01:10:17,640 --> 01:10:20,480
Have you ever found yourself
ascribing some sort of agency or

1254
01:10:20,480 --> 01:10:26,800
free will to this To this?
I can feel the temptation to

1255
01:10:26,800 --> 01:10:29,280
want to do so.
Let's put it that way again.

1256
01:10:29,280 --> 01:10:31,840
I'm an illusionist.
I know I'm being fooled.

1257
01:10:31,840 --> 01:10:34,680
I can feel it.
And yes, sometimes it takes more

1258
01:10:34,680 --> 01:10:36,760
effort than other times to
resist.

1259
01:10:37,760 --> 01:10:40,000
Have you always?
At what point have you actually

1260
01:10:40,000 --> 01:10:42,960
said firmly I'm an illusionist?
Was this something you one day

1261
01:10:42,960 --> 01:10:45,360
read or was?
I think I was.

1262
01:10:45,360 --> 01:10:49,840
And then I met Keith and Dan at
a workshop a while back and and

1263
01:10:49,840 --> 01:10:52,120
learned this actual term.
I said that's it.

1264
01:10:52,120 --> 01:10:54,680
That's what I am.
Yeah, I think.

1265
01:10:54,720 --> 01:10:57,000
I think I felt it before I knew
what I was.

1266
01:10:57,560 --> 01:10:59,320
That's exactly the way I felt
about it.

1267
01:10:59,320 --> 01:11:02,680
And and then one day watching
Dan and then Keith eventually as

1268
01:11:02,680 --> 01:11:04,800
well because he helped me with
my dissertation, which is quite

1269
01:11:04,800 --> 01:11:07,200
cool.
OK yeah, Nicholas Humphrey also

1270
01:11:07,200 --> 01:11:10,760
gave some input and Mark solms.
Great, which is pretty cool.

1271
01:11:11,160 --> 01:11:14,400
And I actually have a chat
coming up with Mark and Carl 1st

1272
01:11:14,400 --> 01:11:16,480
and they're gonna have a nice
chat together about some of the

1273
01:11:16,480 --> 01:11:18,840
work you can do as well.
But that's great.

1274
01:11:18,840 --> 01:11:23,480
Then after this podcast, it's
weird how my, my firm

1275
01:11:23,480 --> 01:11:26,960
illusionist beliefs have slowly
like become slightly less

1276
01:11:27,240 --> 01:11:30,080
interior.
And I thought, I haven't seen, I

1277
01:11:30,120 --> 01:11:33,600
don't see myself changing it
anyway, but but I'm far less

1278
01:11:33,720 --> 01:11:35,800
firm about the belief than I
used to be.

1279
01:11:36,080 --> 01:11:39,560
But I wrote the dissertation
very, very firm illusionism.

1280
01:11:39,600 --> 01:11:42,480
And yeah, to the point that it's
just, to me, it just made the

1281
01:11:42,480 --> 01:11:45,960
most sense that, that we're just
playing mind games with

1282
01:11:45,960 --> 01:11:48,800
ourselves all the time.
It's it's a very difficult thing

1283
01:11:48,800 --> 01:11:50,400
to understand.
True.

1284
01:11:50,680 --> 01:11:51,520
Agreed.
Agreed.

1285
01:11:51,800 --> 01:11:55,440
It's a strange phenomenon in
terms of your work, everything

1286
01:11:55,440 --> 01:11:58,800
you've done.
How is this overall impacted

1287
01:11:58,800 --> 01:12:01,920
your entire view of this field
of understanding the mind body

1288
01:12:01,920 --> 01:12:07,880
connection?
I think what it what it's

1289
01:12:07,880 --> 01:12:12,240
convinced me of is that whatever
the answer is, I don't think we

1290
01:12:12,240 --> 01:12:14,600
have it yet.
It's it's going to be very

1291
01:12:14,600 --> 01:12:19,040
confusing, very non intuitive.
This is what I've learned as a

1292
01:12:19,040 --> 01:12:24,320
student of AI and robots is that
things that work, you know, for

1293
01:12:24,320 --> 01:12:28,520
Mother Nature, they're very
confusing, non intuitive.

1294
01:12:29,080 --> 01:12:32,320
So I think there are some
answers to, you know, some of

1295
01:12:32,320 --> 01:12:36,040
our deeper questions aside who
we are, what makes a special, if

1296
01:12:36,040 --> 01:12:39,600
anything, you know, how the body
shapes the way we think.

1297
01:12:39,600 --> 01:12:42,240
Well, you know, what are aliens
going to look like and how are

1298
01:12:42,240 --> 01:12:45,840
they going to behave?
Whatever the answer is, it's

1299
01:12:45,840 --> 01:12:49,640
not, you know what I've learned
from the xenobots in these meta

1300
01:12:49,640 --> 01:12:52,280
materials?
They act in these very, very

1301
01:12:52,280 --> 01:12:55,160
strange, surprising ways.
And I think we're just

1302
01:12:55,160 --> 01:12:58,880
scratching the tip of the the
iceberg, you know, the fact that

1303
01:12:59,440 --> 01:13:03,480
huge matrix multiplications can
give rise to things that look

1304
01:13:03,480 --> 01:13:06,000
like, you know, brilliant human
conversation.

1305
01:13:06,280 --> 01:13:10,080
Like it we're being things are
moving quickly and everything

1306
01:13:10,080 --> 01:13:14,000
that happens is surprising to to
the public in general and also

1307
01:13:14,000 --> 01:13:16,920
to the experts.
And to me, that's changing my

1308
01:13:16,920 --> 01:13:20,200
world view to be like whatever
we thought the answer was.

1309
01:13:20,200 --> 01:13:22,960
It's much, much stranger.
It, it kind of feels like the

1310
01:13:22,960 --> 01:13:27,280
quantum revolution for like
everything other than physics is

1311
01:13:27,280 --> 01:13:29,680
now happening.
Physics already had its, you

1312
01:13:29,680 --> 01:13:33,200
know, brush with the the
ineffable and unexplainable.

1313
01:13:33,640 --> 01:13:36,120
Now it's our turn.
And, and within, I mean, quantum

1314
01:13:36,120 --> 01:13:38,520
computers, quantum biology,
these are on the rise.

1315
01:13:38,520 --> 01:13:40,080
This is this is a thing that's
happening.

1316
01:13:40,080 --> 01:13:42,600
So people are taking this very
seriously, but these meta

1317
01:13:42,600 --> 01:13:44,440
materials are really intriguing
to me.

1318
01:13:44,440 --> 01:13:46,560
And where do you think this is
going to lead?

1319
01:13:47,440 --> 01:13:50,640
Oh, I, I think, you know, meta
materials are the future.

1320
01:13:50,640 --> 01:13:55,840
Like, so you, you know, all the
materials that we build with,

1321
01:13:55,840 --> 01:13:59,120
they have problems, you know,
you, you can only build a

1322
01:13:59,120 --> 01:14:03,120
building so high because the
weight ratio, you know, so meta

1323
01:14:03,120 --> 01:14:06,440
materials are amazing.
Like I, I see a future in which,

1324
01:14:06,720 --> 01:14:10,520
you know, we can cut a robot in
half and the two halves will

1325
01:14:10,520 --> 01:14:13,760
form 2 smaller versions of
exactly the same robot.

1326
01:14:13,840 --> 01:14:21,400
You know, living materials like
cells, when they attach to rigid

1327
01:14:21,400 --> 01:14:23,560
materials, those cells become
more rigid.

1328
01:14:23,560 --> 01:14:26,400
When they attach to soft
materials, they become softer.

1329
01:14:26,400 --> 01:14:28,760
They're cells are like
chameleons.

1330
01:14:29,280 --> 01:14:32,680
And so I think we're going to
make meta materials that are the

1331
01:14:32,680 --> 01:14:36,720
same way that they, they get the
hint when you try and use them

1332
01:14:36,720 --> 01:14:39,480
to be something, they become
that thing.

1333
01:14:39,480 --> 01:14:43,520
They construct what they need or
they change, you know, that's

1334
01:14:43,520 --> 01:14:47,280
what, that's what 21st century
technologies are gonna look like

1335
01:14:47,280 --> 01:14:49,560
as we push further into the 21st
century.

1336
01:14:50,120 --> 01:14:52,440
I think that's one of this is
probably one of the most

1337
01:14:52,440 --> 01:14:57,840
exciting or or sci-fi ish thing
I've heard, and quite some I

1338
01:14:57,880 --> 01:15:00,280
think it's.
Again, we were talking about

1339
01:15:00,280 --> 01:15:02,040
surprise.
The surprising thing is a lot of

1340
01:15:02,040 --> 01:15:04,800
those materials already exist.
You know, they're just kind of

1341
01:15:04,800 --> 01:15:07,720
in labs at the moment.
They're not in general use, but

1342
01:15:08,160 --> 01:15:11,720
but the aerospace companies, you
know, they have R&D labs around

1343
01:15:11,720 --> 01:15:14,840
these things.
And so these, these, these

1344
01:15:14,840 --> 01:15:18,000
materials, ironically, you know,
they're being, these materials

1345
01:15:18,000 --> 01:15:21,520
are being designed by AI.
And then once they're built,

1346
01:15:21,680 --> 01:15:24,840
designed by the AI, the AI can
figure out how to exploit them.

1347
01:15:24,840 --> 01:15:28,040
You know, how, how can they best
be incorporated into buildings

1348
01:15:28,040 --> 01:15:31,640
and roads and cars and plumbing
and you name it.

1349
01:15:31,640 --> 01:15:33,760
It's, it's amazing.
It hasn't quite reached the

1350
01:15:33,760 --> 01:15:36,720
public, you know, eye yet, but
it's, it's coming.

1351
01:15:37,760 --> 01:15:40,600
It's kind of crazy because it's
almost like, I mean, we're

1352
01:15:40,600 --> 01:15:43,480
making biological robots and
then at the same time you're,

1353
01:15:43,560 --> 01:15:48,160
well, you are, but at the same
time you're making now materials

1354
01:15:48,160 --> 01:15:50,240
more biological.
It's kind of like you're doing

1355
01:15:50,240 --> 01:15:52,440
the both, you're doing like the
flip side on both sides.

1356
01:15:52,440 --> 01:15:54,920
It's just pretty crazy and cool
and.

1357
01:15:55,560 --> 01:15:58,800
Yeah, not just my lab.
Some of the exciting things that

1358
01:15:58,800 --> 01:16:02,760
are going on in these
interdisciplinary labs are

1359
01:16:02,760 --> 01:16:06,320
exactly that, like blurring the
line between living and non

1360
01:16:06,320 --> 01:16:09,400
living materials, right.
So again, there's another

1361
01:16:09,400 --> 01:16:13,440
distinction that we just assume
is so obvious and so clear cut.

1362
01:16:13,440 --> 01:16:18,000
And from the AIS perspective, it
says why I, I don't understand.

1363
01:16:18,000 --> 01:16:21,800
I'll put, you know, I'll, I'll
embed some proteins in a plastic

1364
01:16:22,040 --> 01:16:25,000
to make, you know, a protein
infused plastic.

1365
01:16:25,080 --> 01:16:27,680
So what is that Now?
Is that a living thing, a non

1366
01:16:27,680 --> 01:16:29,760
living thing?
I don't know.

1367
01:16:30,000 --> 01:16:32,800
Yeah, just like make the make
the metal some sort of an ion

1368
01:16:32,800 --> 01:16:34,760
channel to assist with like
shifting cells.

1369
01:16:34,760 --> 01:16:36,600
I mean, you can.
It's exactly that.

1370
01:16:36,600 --> 01:16:39,640
When you think about what the
potential this has and how much

1371
01:16:39,640 --> 01:16:43,280
this can affect medicine and
every other field, it's, it's so

1372
01:16:43,280 --> 01:16:45,400
widespread that it would
literally change humanity.

1373
01:16:45,400 --> 01:16:48,400
Like it's, it's pretty crazy.
I agree.

1374
01:16:48,720 --> 01:16:50,840
I thought it's Max.
You don't, you don't need, I

1375
01:16:50,840 --> 01:16:53,240
don't think we need the the
singularity.

1376
01:16:53,240 --> 01:16:56,400
You know they'll be sufficiently
crazy stuff that will happen

1377
01:16:56,400 --> 01:16:58,800
without us all uploading to the
cloud or.

1378
01:16:59,040 --> 01:17:01,360
Yeah, no, this is this is
probably the most exciting thing

1379
01:17:01,360 --> 01:17:03,200
I've heard.
Like and trust me, when you

1380
01:17:03,200 --> 01:17:05,720
scroll online and you do scroll,
you see a lot of crazy things,

1381
01:17:05,720 --> 01:17:07,520
but this is today, so this is
pretty cool.

1382
01:17:07,520 --> 01:17:10,160
I mean, this, I can't wait to
see what, what you guys come up

1383
01:17:10,160 --> 01:17:12,120
with in and I'm really looking
forward to it.

1384
01:17:12,120 --> 01:17:13,520
Yeah, It's, it's, it's
incredible.

1385
01:17:14,040 --> 01:17:14,360
Yeah.
No.

1386
01:17:14,360 --> 01:17:16,960
So, Josh, thank you so much for
this wonderful conversation.

1387
01:17:17,000 --> 01:17:18,240
Any, any final words from your
side?

1388
01:17:18,240 --> 01:17:21,640
Anything you'd like to add?
Well, just just as we're

1389
01:17:21,640 --> 01:17:24,040
wrapping up, you're asking me
about some other things that we

1390
01:17:24,040 --> 01:17:26,640
might not have talked about.
Just one last thing I wanted to

1391
01:17:26,640 --> 01:17:30,680
mention about connecting AI and
embodied cognition is where we

1392
01:17:30,680 --> 01:17:33,920
are at the moment.
So we now have these immensely,

1393
01:17:33,920 --> 01:17:39,360
you know, powerful AIS that say
compelling things, but

1394
01:17:39,360 --> 01:17:43,280
instinctually we're not sure
whether to trust what they say.

1395
01:17:44,280 --> 01:17:47,160
So we're doing some work here in
the lab and there are others

1396
01:17:47,160 --> 01:17:51,360
that are doing so as well, which
is, you know, combining a non

1397
01:17:51,360 --> 01:17:54,440
embodied technologies like
Chachi BT with embodied

1398
01:17:54,440 --> 01:17:57,840
technologies like robots.
And So what we see on the

1399
01:17:57,840 --> 01:18:01,040
horizon is a human has a
problem.

1400
01:18:01,160 --> 01:18:05,560
They come to an AI, they ask for
a solution and the AI proposes

1401
01:18:05,560 --> 01:18:09,800
something that sounds great.
And the human says, I don't know

1402
01:18:09,800 --> 01:18:14,600
that I trust you prove it.
And the AI says, connect me to a

1403
01:18:14,600 --> 01:18:19,080
3D printer, prints a robot that
goes out and, you know,

1404
01:18:19,080 --> 01:18:24,240
physically verifies, does some
experiments, you know, to show

1405
01:18:24,320 --> 01:18:27,560
some version of this solution or
some aspect of it.

1406
01:18:28,120 --> 01:18:34,440
So can we teach AI to design
experiments that will, that will

1407
01:18:34,440 --> 01:18:37,000
prove whatever it is that it's
proposing?

1408
01:18:37,000 --> 01:18:39,960
So I think that's something else
we haven't talked about that's

1409
01:18:39,960 --> 01:18:43,560
often known as like AI science
or AI driven science or machine

1410
01:18:43,560 --> 01:18:48,400
science that also I think has
huge potential for dealing with

1411
01:18:48,960 --> 01:18:51,640
at the moment, which seems like
an intractable problem that

1412
01:18:51,640 --> 01:18:54,760
there's no way to guarantee what
a ChatGPT says.

1413
01:18:54,880 --> 01:18:59,760
And I agree, doesn't matter what
it or anyone says the proof is

1414
01:18:59,760 --> 01:19:03,120
in the physical world.
And I think we're going to start

1415
01:19:03,120 --> 01:19:06,080
to see some of that emerging in
the not too distant future.

1416
01:19:06,440 --> 01:19:09,960
And so far, Josh, have you seen
anything remotely close to that

1417
01:19:09,960 --> 01:19:16,520
or how how far off are we?
We're close for, for toy

1418
01:19:16,520 --> 01:19:19,200
problems.
I, I guess we're, I would say

1419
01:19:19,200 --> 01:19:22,280
we're for in terms of like
practical utility, we're still a

1420
01:19:22,280 --> 01:19:24,440
ways off.
I, I don't even know that we're

1421
01:19:24,440 --> 01:19:28,520
ready morally or ethically to
allow an AI to build things to

1422
01:19:28,520 --> 01:19:32,000
prove its ideas about the world,
that there's another one for

1423
01:19:32,000 --> 01:19:33,520
Hollywood to play with.
But that that.

1424
01:19:34,040 --> 01:19:36,160
Actually is quite a good script
for a film that is.

1425
01:19:36,320 --> 01:19:40,200
Yeah, yeah, exactly, exactly.
No, no raw, yeah, no lack of raw

1426
01:19:40,200 --> 01:19:43,880
material here.
But anyways, I just wanted, I

1427
01:19:43,880 --> 01:19:46,840
guess I wanted to close just by
saying I don't think, you know,

1428
01:19:46,840 --> 01:19:50,240
it's non embodiment against
embodiment that some of the

1429
01:19:50,240 --> 01:19:54,480
really valuable technologies and
solutions in the years to come

1430
01:19:54,480 --> 01:19:58,440
are going to be, you know,
clever ways of combining them so

1431
01:19:58,440 --> 01:20:01,440
you get the best of both worlds.
And in general, I mean life and

1432
01:20:01,440 --> 01:20:02,920
not life.
That dichotomy, that false

1433
01:20:02,920 --> 01:20:05,800
dichotomy, it seems that at some
point we're all going to be

1434
01:20:06,160 --> 01:20:10,240
chemically bonded in some way.
Yeah, Yeah.

1435
01:20:10,240 --> 01:20:12,280
I think it's a beautiful.
That's a beautiful way to

1436
01:20:12,280 --> 01:20:14,120
probably end it there.
Josh, thank you so much.

1437
01:20:14,120 --> 01:20:15,600
Thanks for your time.
Wonderful work.

1438
01:20:16,160 --> 01:20:18,160
No, thank you, Tevin.
It was a pleasure to be on your

1439
01:20:18,160 --> 01:20:20,120
show.
Yeah, no, I can't wait to

1440
01:20:20,240 --> 01:20:23,120
hopefully have around 2 and and
further dissects maybe specific

1441
01:20:23,120 --> 01:20:25,920
papers and we can go sure some
deep dives in the future.

1442
01:20:25,920 --> 01:20:28,520
It's it's a privilege to chat to
you and I really appreciate it.

1443
01:20:28,720 --> 01:20:30,560
Yeah, likewise.
And again, thanks, thanks for

1444
01:20:30,560 --> 01:20:32,800
your interest and you clearly
put a lot of time and effort

1445
01:20:32,800 --> 01:20:35,720
into to preparing this was this
was a real pleasure.