Terrence Deacon & Michael Levin: What is Life? Complexity, Cognition & the Origin of Purpose

1
00:00:05,040 --> 00:00:07,560
Professor Terence Deakin,
Professor Michael Levin.

2
00:00:07,560 --> 00:00:10,440
This is a long time coming.
I think many of the listeners

3
00:00:10,440 --> 00:00:13,000
and viewers of the show have
been waiting to see this

4
00:00:13,000 --> 00:00:16,800
conversation.
Both of you touch distinguished

5
00:00:16,800 --> 00:00:20,320
careers and have massively
shaped and influenced

6
00:00:21,000 --> 00:00:24,240
developmental evolutionary
biology and cognitive science as

7
00:00:24,240 --> 00:00:26,320
a whole.
And you have so many overlapping

8
00:00:26,840 --> 00:00:29,120
areas of expertise and common
interests.

9
00:00:29,120 --> 00:00:31,600
With that being said, Mike,
perhaps I could start with you.

10
00:00:32,400 --> 00:00:35,480
What do you think of Terry's
work over the years and which

11
00:00:35,480 --> 00:00:38,880
aspects of his work do you find
most fascinating?

12
00:00:38,880 --> 00:00:41,160
And then, Terry, I'll let you
answer the same question when

13
00:00:41,160 --> 00:00:46,280
Mike is done.
Yeah, look, I, I, I have read

14
00:00:46,280 --> 00:00:49,320
your stuff since the 90's, the
developmental neuroscience

15
00:00:49,320 --> 00:00:53,120
papers, especially the cross
species neural transplants and

16
00:00:53,120 --> 00:00:55,040
things like that.
I thought it was incredibly

17
00:00:55,040 --> 00:00:56,800
interesting.
Since then, I mean, there's been

18
00:00:56,800 --> 00:01:00,520
some amazing developments that
that you've, that you've LED

19
00:01:01,120 --> 00:01:04,400
the, you know, incomplete nature
and the, the focus on the origin

20
00:01:04,400 --> 00:01:07,240
of and the importance of
semiosis I think is absolutely

21
00:01:07,240 --> 00:01:10,800
critical.
And, you know, all your ideas

22
00:01:10,800 --> 00:01:13,720
about the, the, the tearing down
the binary kind of Cartesian

23
00:01:13,720 --> 00:01:15,720
dichotomies.
I mean, I'm all, I'm all, all,

24
00:01:16,240 --> 00:01:19,160
all on board with that for sure.
The flows between form and

25
00:01:19,160 --> 00:01:20,400
function.
I mean, there's, you know, I

26
00:01:20,400 --> 00:01:23,280
could go on and on, but but all
of this, this, this, this, this

27
00:01:23,280 --> 00:01:26,640
merger of these deep ideas
between philosophy, evolutionary

28
00:01:26,640 --> 00:01:30,200
biology, developmental biology,
I find for absolutely critical

29
00:01:30,200 --> 00:01:31,320
for progress.
So, yeah.

30
00:01:31,320 --> 00:01:35,840
So thank you for doing all that.
In many respects, those are

31
00:01:35,840 --> 00:01:37,960
exactly the things that are
exciting about your work.

32
00:01:38,440 --> 00:01:44,840
And and in particular where we
overlap very strongly is this

33
00:01:44,840 --> 00:01:50,160
recognition of in a sense the
deep commonalities that are

34
00:01:50,160 --> 00:01:54,080
underlying everything is
biological, including the

35
00:01:54,080 --> 00:01:58,640
informational part.
But also the deeper than evo

36
00:01:58,640 --> 00:02:03,040
divo in the sense that that it
actually utilizes sort of the

37
00:02:03,040 --> 00:02:07,280
self organizing logic that we
find in all kinds of systems

38
00:02:07,560 --> 00:02:12,000
that are really critically
critical pieces of development

39
00:02:12,000 --> 00:02:16,360
of evolution of cognition.
You know, having spent a good

40
00:02:16,360 --> 00:02:19,400
part of my career looking at how
brains develop, for example,

41
00:02:20,640 --> 00:02:24,120
seeing that that a lot of
similarities are involved in the

42
00:02:24,120 --> 00:02:26,920
development of, for example,
Planaria.

43
00:02:26,920 --> 00:02:29,720
A lot of your work on Planaria
that shows a lot of that similar

44
00:02:29,720 --> 00:02:32,280
sort of pattern generation stuff
that goes on.

45
00:02:32,520 --> 00:02:37,160
That is not not strictly, you
know, biological.

46
00:02:37,160 --> 00:02:40,080
It's certainly not designed in
any obvious sense.

47
00:02:40,280 --> 00:02:42,920
It's a very different kind of
logic and that that search for

48
00:02:42,920 --> 00:02:46,080
that underlying logic, which is,
I think where we overlap so

49
00:02:46,080 --> 00:02:49,840
strongly.
And of course, the, the very

50
00:02:49,840 --> 00:02:54,200
fact that that that you had
specific time to work with, with

51
00:02:54,200 --> 00:02:56,720
Dan Dennett, who had a
significant influence on me

52
00:02:56,720 --> 00:02:59,680
while I was in Boston.
He and I have have had these

53
00:02:59,680 --> 00:03:02,560
conversations back and forth
over the years on different

54
00:03:02,560 --> 00:03:06,360
sides of the same issue, often
times sort of not agreeing, but

55
00:03:06,360 --> 00:03:09,440
but recognizing, you know,
exactly how these different

56
00:03:09,440 --> 00:03:13,200
perspectives inform each other.
And and I think that's where you

57
00:03:13,200 --> 00:03:16,640
and I will probably fit because
we have a common interest, but

58
00:03:16,640 --> 00:03:19,840
very different backgrounds in
terms of the particular topics

59
00:03:19,840 --> 00:03:22,360
we've worked on specifically.
Yeah, yeah.

60
00:03:22,520 --> 00:03:24,200
And Dan talked about your work a
lot, you know.

61
00:03:25,440 --> 00:03:29,320
I mean, would that be what do
you guys, what are your guys

62
00:03:29,320 --> 00:03:33,040
thoughts on on Dan's work and
his legacy and philosophy and

63
00:03:33,040 --> 00:03:34,280
how has it impacted both of you?
Why?

64
00:03:36,200 --> 00:03:38,280
Don't you go first, Mike?
Since since since you've been

65
00:03:38,280 --> 00:03:40,480
working with him directly or you
worked with him directly?

66
00:03:41,800 --> 00:03:43,840
Yeah.
For for me, there's a few

67
00:03:43,840 --> 00:03:49,480
things.
He was, he was always to, to,

68
00:03:49,560 --> 00:03:53,560
to, to me, He was, he was an
incredible example of somebody

69
00:03:53,560 --> 00:03:58,160
who wanted to understand things
in a, in a fundamental way.

70
00:03:58,160 --> 00:04:02,560
That was, you know, the, the,
the things he said about that

71
00:04:02,560 --> 00:04:04,480
there is no philosophy free
science.

72
00:04:04,480 --> 00:04:06,680
There's only, what do you say?
There's only side with science

73
00:04:06,680 --> 00:04:09,520
where you smuggle in, you know,
philosophical assumptions

74
00:04:09,520 --> 00:04:12,280
without the examination, right?
So, so that I think is very,

75
00:04:12,280 --> 00:04:15,840
was, was very critical and his,
his commitment both to really

76
00:04:15,840 --> 00:04:21,200
clear, concise, rigorous
philosophy and the, the, the

77
00:04:21,200 --> 00:04:23,760
data, you know, the, the
experimental work that that is

78
00:04:23,760 --> 00:04:26,840
necessary to revise your
philosophical outlook is, is

79
00:04:26,840 --> 00:04:29,360
very important.
Because I, I find, you know, in

80
00:04:29,360 --> 00:04:33,120
our field, I find a lot of the
times when, when, when, you

81
00:04:33,120 --> 00:04:34,960
know, I give these talks that
are partially this kind of

82
00:04:34,960 --> 00:04:36,440
philosophical stuff and
partially data.

83
00:04:36,760 --> 00:04:40,000
And, and some people, some
people say, OK, the data are

84
00:04:40,000 --> 00:04:41,320
great, the experiments very
good.

85
00:04:41,320 --> 00:04:43,320
Don't talk like all this
philosophy stuff, you don't need

86
00:04:43,320 --> 00:04:44,360
it.
Don't, don't talk about it.

87
00:04:44,680 --> 00:04:47,560
You know, and I, I, you know,
and I said, well, why do you

88
00:04:47,560 --> 00:04:49,480
think we did those experiments?
You know, we wouldn't have done

89
00:04:49,480 --> 00:04:50,840
them otherwise.
That's the whole, that's the

90
00:04:50,840 --> 00:04:53,640
whole point, right?
And, and, and I think especially

91
00:04:53,640 --> 00:04:57,200
in our, in like some of the
molecular biology approaches and

92
00:04:57,200 --> 00:05:01,520
things like that, people tend to
feel, or at least they act as if

93
00:05:02,080 --> 00:05:05,680
they're really touching reality
directly, that there is no

94
00:05:05,680 --> 00:05:07,360
philosophy underlying their
perspective.

95
00:05:07,600 --> 00:05:10,560
They think that the things I say
are metaphorical, but pathways

96
00:05:10,560 --> 00:05:12,440
and things like that are real.
You know, we got real things.

97
00:05:12,440 --> 00:05:15,720
We got these pathways.
And, and, and just, you know, I,

98
00:05:15,760 --> 00:05:19,760
I always like to dance the kind
of unflinching fusion of, of the

99
00:05:19,760 --> 00:05:23,760
philosophy and the and the data.
I would say that's exactly the

100
00:05:23,760 --> 00:05:27,720
same for me that that that sense
that that he's he's a

101
00:05:27,720 --> 00:05:32,560
philosopher that's totally
focused on the science, the

102
00:05:32,560 --> 00:05:35,720
technology that it always
influenced his thinking.

103
00:05:35,720 --> 00:05:37,840
It just it always is very
important.

104
00:05:37,840 --> 00:05:41,840
And as a result, you know, he
was coming largely and, and

105
00:05:41,840 --> 00:05:46,120
particularly the, the, the 80s
and 90s when we spent a lot of

106
00:05:46,120 --> 00:05:50,200
time interacting his, his inputs
were from, you know, computer

107
00:05:50,200 --> 00:05:53,200
science and cognitive science
has done particularly from the

108
00:05:53,200 --> 00:05:55,480
MIT groups.
And mine of course, was from the

109
00:05:55,480 --> 00:05:57,440
biology side.
And we go back and forth on

110
00:05:57,440 --> 00:06:00,480
this.
I remember he used to have

111
00:06:00,840 --> 00:06:04,600
regular meetings at his home and
we would, we'd sit around and

112
00:06:04,600 --> 00:06:08,480
argue over these things and, and
he was always in the middle.

113
00:06:08,480 --> 00:06:11,160
He was always seen in a sense
recognizing, well, you know,

114
00:06:11,720 --> 00:06:14,440
listen to listen to each other,
that kind of thing.

115
00:06:14,480 --> 00:06:19,320
Just really, really helpful.
And yet, and and yet was after

116
00:06:19,320 --> 00:06:22,600
some of these really basic
philosophical questions, you

117
00:06:22,600 --> 00:06:26,360
know, the the free will question
was, was a big part of his

118
00:06:26,360 --> 00:06:30,360
thinking that the the idea of
experience of consciousness,

119
00:06:30,640 --> 00:06:33,720
that everybody sort of gave him
a hard time for coming up with a

120
00:06:33,720 --> 00:06:37,520
title consciousness explained
that, you know, that wasn't what

121
00:06:37,520 --> 00:06:39,800
he was about.
He was actually about, you know,

122
00:06:39,920 --> 00:06:43,040
interrogating the problem.
You know, what's why do we think

123
00:06:43,040 --> 00:06:46,040
the way we do?
And as a result, even though he

124
00:06:46,040 --> 00:06:48,520
and I came to very different
conclusions about that, at least

125
00:06:48,520 --> 00:06:53,600
during those years, there was it
was just a really constructive

126
00:06:53,800 --> 00:06:56,000
discussion.
And I find like like with

127
00:06:56,000 --> 00:06:59,520
Michael.
So I I began, I should say, just

128
00:06:59,520 --> 00:07:06,280
to be clear, my career was
driven initially by the

129
00:07:06,280 --> 00:07:10,720
discovery of a philosopher in
the late 1970s.

130
00:07:10,720 --> 00:07:13,640
I've read the work of Charles
Sanders Purse mostly by

131
00:07:13,640 --> 00:07:15,160
accident.
I was thinking I was going to go

132
00:07:15,160 --> 00:07:19,240
on and do, you know, just sort
of standard bench science and

133
00:07:19,480 --> 00:07:22,880
the philosophy just sort of
kicked me hard really on that.

134
00:07:22,880 --> 00:07:25,880
He was asking questions at the
end of the 19th century, the

135
00:07:25,880 --> 00:07:29,560
beginning of the 20th century
that nobody was even asking.

136
00:07:29,760 --> 00:07:35,800
Now, 7580 years later, and and
so I approach when I began to do

137
00:07:35,800 --> 00:07:39,320
my neuroscience and evolutionary
work, it was always driven by

138
00:07:39,320 --> 00:07:42,720
this sort of deep philosophical
question that sort of drove it

139
00:07:42,720 --> 00:07:46,760
about, you know, the nature of
of representation of information

140
00:07:47,400 --> 00:07:49,400
and the processes that generated
it.

141
00:07:49,400 --> 00:07:54,680
These are so in that respect, I
always had like Michael, in many

142
00:07:54,680 --> 00:08:00,520
respects, this, this sort of a
philosophical grounding kept,

143
00:08:00,600 --> 00:08:03,280
kept, kept forcing me to ask
these questions.

144
00:08:03,480 --> 00:08:05,120
Now wait a minute.
You know, it's not that simple.

145
00:08:05,120 --> 00:08:07,520
It's not just, it's just
something more.

146
00:08:09,200 --> 00:08:11,960
Well, beautifully said guys.
I mean, rest in peace to Dan.

147
00:08:12,200 --> 00:08:15,360
This conversation has so many
overlapping themes.

148
00:08:15,360 --> 00:08:18,160
While the two of you and your
work that was very difficult to

149
00:08:18,160 --> 00:08:20,640
actually decide on what to what
to ask the both of you.

150
00:08:20,640 --> 00:08:23,320
But I managed to set a few cards
together.

151
00:08:23,320 --> 00:08:24,720
I think let's start off with a
deep one.

152
00:08:24,960 --> 00:08:29,360
Origin of purpose.
At what point, Terry, perhaps

153
00:08:29,360 --> 00:08:32,840
you could start, at what point
do physical systems without

154
00:08:32,840 --> 00:08:37,520
brains, let's say, begin to
exhibit purpose like behavior?

155
00:08:39,240 --> 00:08:43,919
Well, I actually approached this
in a very different sense that

156
00:08:43,919 --> 00:08:48,080
as I realized that that even
simple organisms like bacteria

157
00:08:48,080 --> 00:08:52,280
were way too complex to actually
sort that out.

158
00:08:52,280 --> 00:08:55,400
And so I've actually tried to
simplify it as much as I could,

159
00:08:55,840 --> 00:08:59,560
and I've approached it mostly
from a thermodynamics point of

160
00:08:59,560 --> 00:09:01,920
view.
One of things that I've never

161
00:09:02,320 --> 00:09:05,360
felt comfortable with, and yet
it's sort of the standard

162
00:09:05,360 --> 00:09:09,040
philosophical approach is a list
of properties.

163
00:09:09,280 --> 00:09:11,000
You know, what's the list of
properties that's going to

164
00:09:11,000 --> 00:09:13,920
define life?
And as we've seen over the

165
00:09:13,920 --> 00:09:17,880
course of decades and decades
and decades, that that list of

166
00:09:17,880 --> 00:09:22,120
properties, it's always leads
you to, to not have a definitive

167
00:09:22,120 --> 00:09:24,640
answer.
And So what I wanted to ask was,

168
00:09:24,680 --> 00:09:28,440
you know, what is it about the
thermodynamics of life that's

169
00:09:28,440 --> 00:09:32,760
that's very unusual.
And because when we look at and

170
00:09:32,760 --> 00:09:35,960
I I think what happened is that
over the course, I think it

171
00:09:35,960 --> 00:09:41,480
begins partially with with this
book, What is Life, a series of

172
00:09:41,480 --> 00:09:45,720
lectures that the physicist
Erwin Schroedinger produces in

173
00:09:45,720 --> 00:09:48,360
the 1940s.
He kicks us off in asking the

174
00:09:48,360 --> 00:09:51,880
question, you know, why does
life act so different than other

175
00:09:51,880 --> 00:09:55,720
systems?
And you know, he recognized that

176
00:09:55,720 --> 00:09:58,880
there's an informational issue,
but in particularly recognizes

177
00:09:58,880 --> 00:10:03,360
there's a thermodynamic issue.
And, and I came to believe

178
00:10:03,600 --> 00:10:06,680
partly because of purses work,
but also because of later work

179
00:10:07,560 --> 00:10:10,520
that in effect the thermodynamic
question and the information

180
00:10:10,520 --> 00:10:14,800
question were the same question.
The problem is that I think

181
00:10:15,200 --> 00:10:19,440
really beginning in the early
1980s and and on until the

182
00:10:19,440 --> 00:10:23,400
present, many people in that
particularly approaching this

183
00:10:23,400 --> 00:10:28,120
physically have been satisfied
with the far from equilibrium.

184
00:10:30,000 --> 00:10:32,120
I describe this more of a
dynamic, but a lot of people

185
00:10:32,120 --> 00:10:35,040
call it self organizational
processes that that's all you

186
00:10:35,040 --> 00:10:37,280
have to really cover.
Because according to the

187
00:10:37,280 --> 00:10:40,360
Schrodinger perspective, the
really, the really interesting

188
00:10:40,360 --> 00:10:43,600
question about life was that
it's far from equilibrium and it

189
00:10:43,600 --> 00:10:47,360
keeps itself there and it
produces form rather than

190
00:10:47,360 --> 00:10:50,840
destroying forms, produces
regularities rather than having

191
00:10:50,840 --> 00:10:53,600
regularity spontaneously
breakdown like the second law

192
00:10:53,880 --> 00:10:57,800
would have suggested.
What I came to realize and

193
00:10:57,800 --> 00:11:02,360
really as I was struggling with
how brains develop and, and how

194
00:11:02,560 --> 00:11:06,680
and why language is so unusual
for communication amongst other

195
00:11:06,680 --> 00:11:12,720
species, that in effect, there's
a third transition that we, we

196
00:11:12,720 --> 00:11:15,360
can't just stop and look at how
form gets generated.

197
00:11:15,560 --> 00:11:20,040
We have to think about processes
in which form gets remembered,

198
00:11:20,400 --> 00:11:24,360
gets generated and remembered
and repaired so that in effect,

199
00:11:24,360 --> 00:11:27,760
if it gets modified, it gets
modified back again to where it

200
00:11:27,760 --> 00:11:29,280
was.
It's like a memory that gets

201
00:11:29,280 --> 00:11:32,280
passed on.
And that wasn't answered here.

202
00:11:32,320 --> 00:11:37,240
And so I wrote this book in the
mid 90s, began this book and

203
00:11:37,240 --> 00:11:40,400
then then it doesn't get done
until, you know, a decade later

204
00:11:41,200 --> 00:11:43,960
that came out to be incomplete
nature, which was basically to

205
00:11:43,960 --> 00:11:46,760
say no, the problem of
emergence.

206
00:11:46,760 --> 00:11:48,560
First of all, we had to get rid
of the magic.

207
00:11:48,880 --> 00:11:51,800
It's not magic.
It's about, you know, novelty of

208
00:11:51,800 --> 00:11:53,440
some kind.
But and typically it's

209
00:11:53,440 --> 00:11:56,000
combinatorial novelty in an
interesting way.

210
00:11:56,000 --> 00:11:58,560
And what I realized is that
what's happened is that we have

211
00:11:59,120 --> 00:12:02,400
oversimplified the
thermodynamics story and that we

212
00:12:02,400 --> 00:12:05,800
needed to have a thermodynamics
in which self organizing

213
00:12:05,800 --> 00:12:09,640
processes themselves organize
each other in such a way that

214
00:12:09,640 --> 00:12:12,800
they rebalanced and then had
something like memory.

215
00:12:12,800 --> 00:12:16,800
And so that I really saw the
origins of life question and the

216
00:12:16,800 --> 00:12:20,720
origins of interpretation, not
the origins of information,

217
00:12:20,720 --> 00:12:23,720
because I think we have
oversimplified the concept of

218
00:12:23,720 --> 00:12:26,960
information.
We flattened it so that it we

219
00:12:26,960 --> 00:12:30,000
don't really deal with the
question of evaluation.

220
00:12:30,000 --> 00:12:32,760
We don't really deal with the
question of, of reference, how

221
00:12:32,760 --> 00:12:35,160
something present can be about
something that's not there.

222
00:12:37,640 --> 00:12:40,440
But that that there's an
organization, the thermodynamic

223
00:12:40,440 --> 00:12:42,880
organization that makes that
possible.

224
00:12:43,440 --> 00:12:46,080
And that therefore it's not a
surprise that that some of the

225
00:12:46,080 --> 00:12:49,280
mathematics of thermodynamics
and the mathematics of

226
00:12:49,280 --> 00:12:51,120
information have a lot in
common.

227
00:12:51,800 --> 00:12:53,760
Even though they're talking
about very different things,

228
00:12:53,760 --> 00:12:55,760
they have a lot in common
because there's some similar

229
00:12:55,760 --> 00:13:00,280
processes.
So to just cut to the end of

230
00:13:00,280 --> 00:13:04,400
that story, I realized that the
the problem that everybody has

231
00:13:04,400 --> 00:13:09,200
in biology is with viruses.
Are viruses alive or are they

232
00:13:09,200 --> 00:13:11,400
dead?
And I realized that the the

233
00:13:11,400 --> 00:13:14,480
question about being alive was
not the right question.

234
00:13:15,080 --> 00:13:20,360
It's the question about this
contortion of thermodynamics

235
00:13:21,360 --> 00:13:24,920
that allows something to be
about something, you know,

236
00:13:24,920 --> 00:13:27,920
allows something to make
predictions about the future and

237
00:13:27,920 --> 00:13:29,520
to then maintain those
predictions.

238
00:13:29,760 --> 00:13:31,480
How can a chemical system do
that?

239
00:13:31,480 --> 00:13:33,520
So the question always dogged
me.

240
00:13:33,520 --> 00:13:36,200
I actually wrote a paper with a
sort of title is that how does a

241
00:13:36,200 --> 00:13:40,680
molecule become about something
that it's not that it, it, it

242
00:13:40,680 --> 00:13:44,040
just, it just doesn't have that.
We just can't assume that DNA

243
00:13:44,040 --> 00:13:47,920
and RNA are just information.
They're, they're about

244
00:13:47,920 --> 00:13:49,520
something.
They're not just copying

245
00:13:49,520 --> 00:13:51,440
themselves.
They're about something and

246
00:13:51,440 --> 00:13:54,000
they're about themselves in a,
in an interesting sort of

247
00:13:54,000 --> 00:13:56,960
twisted way.
So I realized that the, the

248
00:13:56,960 --> 00:13:59,880
question of viruses was really
the, the crucial question.

249
00:14:00,600 --> 00:14:03,560
Viruses are at that boundary
where we're not sure we can say

250
00:14:03,560 --> 00:14:06,160
it's alive or not.
And yet it does all the things

251
00:14:06,160 --> 00:14:09,520
that living things do, that is
it evolves, it protects itself.

252
00:14:09,920 --> 00:14:14,160
We can talk about, you know,
killing viruses, all those

253
00:14:14,160 --> 00:14:17,200
things we can do, even though we
know that the, it's not quite

254
00:14:17,200 --> 00:14:20,520
the same language.
So I, I've sort of in the last

255
00:14:20,520 --> 00:14:24,480
decade of my career, I've been
focusing on that question, which

256
00:14:24,480 --> 00:14:27,200
I think is a way to ask the
question about what is

257
00:14:27,200 --> 00:14:30,920
information in all of its
various forms, what I would call

258
00:14:30,920 --> 00:14:34,920
the semihotic concept of
information and, and then how it

259
00:14:34,920 --> 00:14:37,680
comes about and how it evolves.
Mike so.

260
00:14:37,680 --> 00:14:39,200
That's a long answer your
question.

261
00:14:40,320 --> 00:14:42,360
Yeah, that's fine, Terry, Mike,
same question.

262
00:14:42,360 --> 00:14:44,920
I mean, at what point do
physical systems exhibit purpose

263
00:14:44,920 --> 00:14:47,040
like behavior?
But also at this point, you

264
00:14:47,040 --> 00:14:50,400
could also address whatever you
think that Terry said is

265
00:14:50,400 --> 00:14:53,160
important, valuable and whim.
You might disagree or agree.

266
00:14:55,160 --> 00:14:57,400
Yeah, not much to argue with
there.

267
00:14:57,400 --> 00:14:59,440
I, I agree with, with with
almost all of it.

268
00:14:59,640 --> 00:15:01,960
So maybe, maybe I'll just talk
about a, a slightly different

269
00:15:01,960 --> 00:15:04,000
perspective on some of these
things that, that, that I've,

270
00:15:04,080 --> 00:15:07,520
that I've been developing.
So what I'm interested in, I, I,

271
00:15:07,560 --> 00:15:11,400
I think, I think a lot of times
dichotomies lead us astray.

272
00:15:11,640 --> 00:15:13,680
So trying to, you know, is it
alive?

273
00:15:13,680 --> 00:15:15,320
Isn't it alive?
Is it the cognitive?

274
00:15:15,320 --> 00:15:17,640
Isn't it like, I think these
things land us in the pseudo

275
00:15:17,640 --> 00:15:20,000
problems that are probably UN
unresolvable?

276
00:15:20,000 --> 00:15:22,520
And So what I'm really
interested in are models of

277
00:15:22,520 --> 00:15:25,840
scaling and transformation.
So, so as, as Terry said, was,

278
00:15:25,840 --> 00:15:28,600
was, was saying, I want to know,
you know, how, how, how do the

279
00:15:28,600 --> 00:15:31,080
things that we're interested in?
So the ability to generate form

280
00:15:31,080 --> 00:15:33,600
to maintain and to have this
memory to all of these things.

281
00:15:33,600 --> 00:15:36,760
How do they scale up from, from
the tiniest beginnings?

282
00:15:37,240 --> 00:15:40,680
And so, so that's, that's,
that's one, one thing that I'm

283
00:15:40,680 --> 00:15:42,040
really interested in is the
scaling.

284
00:15:42,040 --> 00:15:45,720
And the second thing is this,
this idea, when I think biology

285
00:15:45,720 --> 00:15:47,640
does this really well, is
polycomputing.

286
00:15:47,640 --> 00:15:49,400
So this is something that Josh
Bongard and I have been

287
00:15:49,400 --> 00:15:53,200
developing this idea that what
you really have instead of 1

288
00:15:53,560 --> 00:15:57,280
objective, correct, you know
what this, this set of physical

289
00:15:57,280 --> 00:16:01,080
events is computing this
function or this has information

290
00:16:01,080 --> 00:16:04,200
about this thing is to consider
all the different observers.

291
00:16:04,360 --> 00:16:06,720
And so within a living
structure, you have many nested

292
00:16:06,720 --> 00:16:08,640
observers, right?
You have, you have both

293
00:16:08,640 --> 00:16:11,040
laterally and vertically, you
have the system's parts, you

294
00:16:11,040 --> 00:16:14,600
have the, the the other systems
around it that are all

295
00:16:14,600 --> 00:16:17,640
continuously interpreting each
other and what each other are

296
00:16:17,640 --> 00:16:20,160
doing.
And because they're all hacking

297
00:16:20,160 --> 00:16:21,520
each other, right?
They're all, they're all trying

298
00:16:21,520 --> 00:16:24,160
to figure out where do I end?
Where does the outside world

299
00:16:24,160 --> 00:16:25,360
begin?
How do, what are the things I

300
00:16:25,360 --> 00:16:26,840
have control over?
What can I affect?

301
00:16:26,840 --> 00:16:29,640
How, what are the most efficient
signals I can get it to do other

302
00:16:29,640 --> 00:16:32,800
things and so on.
And so, so this idea of and, and

303
00:16:32,800 --> 00:16:35,160
of course, of course, systems
also, as, as you said, systems

304
00:16:35,160 --> 00:16:37,400
also have to interpret
themselves, say that's the most

305
00:16:37,400 --> 00:16:41,120
interesting thing about
significant agents is that is

306
00:16:41,120 --> 00:16:45,040
that they, they, it's not just
that others apply a sort of

307
00:16:45,040 --> 00:16:47,720
intentional stance to them and
see goal directed behavior, but

308
00:16:47,720 --> 00:16:50,440
they themselves are able to
close that, that strange loop

309
00:16:50,440 --> 00:16:54,520
and then do it for themselves.
So, so this is the sort of thing

310
00:16:54,520 --> 00:16:57,040
that we've been working on is,
is the stories of scaling, which

311
00:16:57,040 --> 00:17:00,440
has many implications for, for,
for cancer and for, and for

312
00:17:00,440 --> 00:17:03,000
other things.
But one, one thing in particular

313
00:17:03,000 --> 00:17:05,599
that I'm really interested in
is, is the concept of mistakes.

314
00:17:05,920 --> 00:17:10,040
So as you, as you observe
chemistry, chemistry doesn't

315
00:17:10,040 --> 00:17:11,960
make mistakes.
Chemistry just sort of does what

316
00:17:11,960 --> 00:17:13,960
chemistry does.
But developmental biology can

317
00:17:13,960 --> 00:17:16,359
certainly make mistakes and
behavior can, can, can make

318
00:17:16,359 --> 00:17:18,760
mistakes.
And so the extent to which you

319
00:17:18,760 --> 00:17:21,680
have a system and, and this is
how I see all of these things as

320
00:17:21,680 --> 00:17:24,480
navigating different spaces.
So of course you have, you have

321
00:17:24,480 --> 00:17:27,319
metabolic spaces and, and
physiological spaces and

322
00:17:27,319 --> 00:17:30,480
transcriptional spaces and then
anatomical amorphous space,

323
00:17:30,480 --> 00:17:32,200
which is, you know, sort of what
I study mostly.

324
00:17:32,960 --> 00:17:35,400
And then you have 3D motion,
space of motion and so on.

325
00:17:36,320 --> 00:17:40,440
As, as systems navigate this
space, it is very natural to

326
00:17:40,440 --> 00:17:43,520
look at that scale that, you
know, the Wiener Rosenbluth,

327
00:17:43,600 --> 00:17:47,000
Bigelow scale from, from passive
matter to, you know, sort of

328
00:17:47,160 --> 00:17:49,080
active matter.
And then, and then the, how much

329
00:17:49,280 --> 00:17:51,480
you know what, what are the,
what are the, what are the tools

330
00:17:51,480 --> 00:17:54,200
that you can bring to understand
the navigation of the system in

331
00:17:54,200 --> 00:17:56,000
that space?
Is it a random walk?

332
00:17:56,000 --> 00:17:57,520
Does it do delayed
gratification?

333
00:17:57,520 --> 00:17:59,520
Does it have memory and forward
planning?

334
00:17:59,520 --> 00:18:02,560
Does it, you know, what are,
what are all the things that

335
00:18:02,560 --> 00:18:05,040
that it does does.
And this site.

336
00:18:05,040 --> 00:18:07,480
And, and So what we've been
doing is taking tools that are

337
00:18:07,480 --> 00:18:13,200
usually deployed all across,
let's say the right side of that

338
00:18:13,200 --> 00:18:15,320
spectrum.
So tools from computational

339
00:18:15,320 --> 00:18:17,720
neuroscience and, and behaviour
science and so on.

340
00:18:18,000 --> 00:18:20,960
And asking whether they apply to
very simple things and simple

341
00:18:21,000 --> 00:18:23,480
in, in other spaces that are
really not easy for us to, to,

342
00:18:23,720 --> 00:18:26,600
to, to sort of visualize.
And and that's been incredibly

343
00:18:26,600 --> 00:18:28,600
enriching because it turns out
that if you're willing to do

344
00:18:28,600 --> 00:18:31,800
that, if you're willing to say
that this is an empirical

345
00:18:31,800 --> 00:18:35,200
project, we can't just assume,
assume that that this is, you

346
00:18:35,200 --> 00:18:36,720
know, how cognitive different
things are.

347
00:18:36,720 --> 00:18:38,080
You have to actually do the
experiments.

348
00:18:38,080 --> 00:18:41,360
When you do the experiments, you
find out some amazing things

349
00:18:41,360 --> 00:18:43,200
even at the very left end of
that spectrum.

350
00:18:43,200 --> 00:18:47,000
So, for example, gene regulatory
networks, not the cell, not that

351
00:18:47,000 --> 00:18:49,760
all the stuff that goes in it,
but just the mathematics of a

352
00:18:49,760 --> 00:18:53,160
few nodes linked to each other
by, you know, these the ordinary

353
00:18:53,160 --> 00:18:55,680
differential equations.
Just that alone can do six

354
00:18:55,680 --> 00:18:57,840
different kinds of learning,
dynamical system learning,

355
00:18:57,840 --> 00:19:00,080
including Pavlovian
conditioning, including

356
00:19:00,200 --> 00:19:03,560
habituation sensitization, They
can count to small numbers,

357
00:19:03,560 --> 00:19:06,280
things like that.
So you get that sort of for

358
00:19:06,280 --> 00:19:09,760
free, so to speak, from the math
right at the very beginning.

359
00:19:10,160 --> 00:19:14,960
And when we talk about purpose,
you can sort of define, you

360
00:19:14,960 --> 00:19:16,880
know, purpose versus goals,
right?

361
00:19:16,880 --> 00:19:18,520
Do you know you have a purpose?
I mean, that's kind of a

362
00:19:18,520 --> 00:19:20,440
metacognitive, you know, sort of
thing.

363
00:19:20,720 --> 00:19:24,760
But but the ability to navigate
that space using the tools that

364
00:19:24,760 --> 00:19:28,760
we associate with some level of
intelligence, I think goes all

365
00:19:28,760 --> 00:19:30,880
the way down.
We we see it in extremely

366
00:19:30,880 --> 00:19:33,920
minimal systems.
We've even done, and you know,

367
00:19:33,920 --> 00:19:36,000
my, my background is, is
originally computer science.

368
00:19:36,000 --> 00:19:38,440
And so I, I like, like very, you
know, very simple kind of

369
00:19:38,440 --> 00:19:40,960
algorithmic things.
And it turns out that even in

370
00:19:40,960 --> 00:19:44,560
very simple deterministic things
like sorting algorithms and so

371
00:19:44,560 --> 00:19:47,560
on, you can find surprises that
look like different kinds of

372
00:19:47,560 --> 00:19:50,920
problem solving that you have no
business to expect from the

373
00:19:50,920 --> 00:19:52,080
algorithm that you thought you
wrote.

374
00:19:52,440 --> 00:19:56,120
So, yeah, you know, to, to, to
me this, these things are start,

375
00:19:56,160 --> 00:20:00,160
start very low down and, and we
just have to be to do it, to do

376
00:20:00,160 --> 00:20:02,120
experiments to, to, to see if we
can find them.

377
00:20:03,960 --> 00:20:05,320
Yeah.
The way I like to think about

378
00:20:05,320 --> 00:20:08,840
that transition is the
transition from chemistry to

379
00:20:08,840 --> 00:20:12,000
normative chemistry.
Chemistry which there is good

380
00:20:12,000 --> 00:20:15,000
chemistry or bad chemistry or
chemistry that works and doesn't

381
00:20:15,000 --> 00:20:17,280
work.
The question is for something to

382
00:20:17,280 --> 00:20:20,320
be normative, there has to be a
beneficiary.

383
00:20:20,600 --> 00:20:26,240
Something has to benefit or, or,
or be harmed in the process to

384
00:20:26,240 --> 00:20:30,400
do better or do worse.
So that it's, it's the critical

385
00:20:30,400 --> 00:20:33,280
transition for me.
So rather than so, you know, I

386
00:20:33,680 --> 00:20:37,720
in, in some respects, I'm at the
very bottom and the top because

387
00:20:37,720 --> 00:20:40,880
I'm, I was very interested in,
in brains and particularly the

388
00:20:41,080 --> 00:20:44,600
uniqueness of human brains and
their anatomy and function.

389
00:20:44,880 --> 00:20:47,920
And then also it forced me to
the very bottom.

390
00:20:47,920 --> 00:20:51,160
Ask the question, well, when can
you say it starts?

391
00:20:51,240 --> 00:20:53,400
When is there finally a
beneficiary?

392
00:20:54,000 --> 00:20:57,400
When can I now say something
actually benefits as opposed to

393
00:20:57,400 --> 00:21:02,440
just being preserved like a rock
is preserved or generated like

394
00:21:02,560 --> 00:21:04,920
like a chemical reaction
generates more of something.

395
00:21:06,240 --> 00:21:10,000
None of those have beneficiaries
in the sense that a beneficiary

396
00:21:10,000 --> 00:21:14,160
is something that in a sense
does work to maintain itself.

397
00:21:14,960 --> 00:21:17,280
And and that's that's an
interesting question.

398
00:21:17,280 --> 00:21:19,960
So work comes part of it.
That's why it's necessarily

399
00:21:19,960 --> 00:21:25,080
thermodynamics, but why I also
force myself to think about the

400
00:21:25,080 --> 00:21:28,600
simplest possible systems,
simpler than a than a virus,

401
00:21:28,600 --> 00:21:31,520
because viruses are, are
parasitic and they're taking

402
00:21:31,520 --> 00:21:34,680
advantage of the lots of stuff.
But but basically virus like

403
00:21:34,680 --> 00:21:38,920
structure has a lot to it that
has this kind of, you can think

404
00:21:38,920 --> 00:21:40,960
about it.
It's just a, you know, I, I, I,

405
00:21:40,960 --> 00:21:43,680
I don't think about it
algorithmically, but certainly

406
00:21:43,920 --> 00:21:46,600
in terms of it has a something
like self.

407
00:21:46,600 --> 00:21:49,480
Whenever we talk about living
systems, we talk about self

408
00:21:49,480 --> 00:21:55,960
regeneration, self repair, self
reproduction, self correction,

409
00:21:56,160 --> 00:21:58,360
you know, all of those things.
We've got, we've got this notion

410
00:21:58,360 --> 00:22:01,080
of self.
I've often thought that this is

411
00:22:01,360 --> 00:22:05,920
one of the critical problems in
philosophy of mind is that we

412
00:22:05,920 --> 00:22:09,440
just sort of assume self.
Self is actually a really

413
00:22:09,440 --> 00:22:11,720
complicated issue.
Even though it might be very

414
00:22:11,720 --> 00:22:14,400
simple.
It's actually something that if

415
00:22:14,400 --> 00:22:18,160
we just leave alone, we just
assume a beneficiary.

416
00:22:18,400 --> 00:22:22,280
We just assume goals and don't
talk about how they're

417
00:22:22,280 --> 00:22:25,400
generated.
I think we get into into

418
00:22:25,400 --> 00:22:28,680
circles.
Yeah, I mean something very

419
00:22:28,680 --> 00:22:32,280
interesting about this concept
of beneficiary is also related

420
00:22:32,280 --> 00:22:34,600
to valence and reward and
punishment.

421
00:22:34,920 --> 00:22:38,040
So, so we can all see that you
can write, you can, you can

422
00:22:38,040 --> 00:22:40,560
reward and punish A paramecium,
simple enough.

423
00:22:41,640 --> 00:22:43,960
But then you can ask yourself
what, what does it actually, you

424
00:22:43,960 --> 00:22:46,400
know, if I started backwards and
I say to people, OK, can I

425
00:22:46,400 --> 00:22:48,480
punish a chemical network?
And they say, well, absolutely

426
00:22:48,480 --> 00:22:50,120
not.
And then and then you show them

427
00:22:50,120 --> 00:22:51,400
paramecium.
Well, well, what's this?

428
00:22:51,400 --> 00:22:52,920
Because, because you can do that
here, right?

429
00:22:53,400 --> 00:22:57,640
And what, what we're finding
that's really interesting and

430
00:22:57,640 --> 00:23:00,040
this will be this, this paper
will be out and I don't know

431
00:23:00,240 --> 00:23:06,640
hopefully, hopefully soon if, if
you, if you have a model of a

432
00:23:06,640 --> 00:23:08,160
pathway or a gene regulatory
network.

433
00:23:08,160 --> 00:23:11,000
So not the rest of the cell, no
evolution that you know nothing.

434
00:23:11,320 --> 00:23:14,600
Just just a small set of a small
set of nodes that are connected

435
00:23:14,600 --> 00:23:17,640
to turn each other on and off.
In addition to the learning that

436
00:23:17,640 --> 00:23:21,000
we find there, you can also find
something very interesting,

437
00:23:21,000 --> 00:23:24,680
which is that if you take
metrics from causal information

438
00:23:24,680 --> 00:23:26,680
theory that people are using to
apply to brain.

439
00:23:26,680 --> 00:23:29,120
So, so people like Giulio Tononi
and Eric Hole and others are

440
00:23:29,320 --> 00:23:31,640
applying these things to try to
understand, do I am I looking at

441
00:23:31,640 --> 00:23:33,640
a pile of neurons or is there
somebody home in there?

442
00:23:33,640 --> 00:23:35,960
You know, is there a person, is
there a human patient that's in

443
00:23:35,960 --> 00:23:39,280
there, right.
So it turns out that not only,

444
00:23:39,280 --> 00:23:43,160
not only do some of these
networks have significant causal

445
00:23:43,160 --> 00:23:46,000
emergence, but that causal
emergence goes up after you

446
00:23:46,000 --> 00:23:48,000
train them so you can calculate
it.

447
00:23:48,000 --> 00:23:50,240
And So what happens is for some
of them, not for all of them,

448
00:23:50,240 --> 00:23:53,040
there are classes, In fact,
there are, I think, 5 distinct

449
00:23:53,040 --> 00:23:54,840
personalities that these
networks fall into.

450
00:23:55,040 --> 00:23:58,160
But for some of them, if you, as
you start to train them, and

451
00:23:58,160 --> 00:24:01,280
when I say train, I mean you,
you stimulate some of the nodes.

452
00:24:01,280 --> 00:24:04,400
So let's say in a, in a
Pavlovian paradigm, you know,

453
00:24:04,400 --> 00:24:06,120
you've got your unconditioned
stimulus, your condition

454
00:24:06,120 --> 00:24:08,320
stimulus, and then there's some
other node that's your response

455
00:24:08,320 --> 00:24:09,760
node.
And so, so you, you start to

456
00:24:09,760 --> 00:24:11,920
pair the stimulations and then
you see what happens to the

457
00:24:11,920 --> 00:24:14,720
response.
So as you do these things for

458
00:24:14,720 --> 00:24:17,120
some of them, the cause of
emergence goes up.

459
00:24:17,680 --> 00:24:21,360
And, and so you've got this
really interesting loop where by

460
00:24:21,360 --> 00:24:24,440
virtue of learning about their
environment, when in this case,

461
00:24:24,440 --> 00:24:26,480
being trained, not so much
learning more, being trained,

462
00:24:27,320 --> 00:24:31,000
they become more integrated as a
tiny little self.

463
00:24:31,000 --> 00:24:33,080
They become right.
They, they, they acquire this

464
00:24:33,080 --> 00:24:35,720
kind of integration.
And for some of these things,

465
00:24:36,080 --> 00:24:40,360
what we, we looked at biological
networks versus random networks.

466
00:24:40,360 --> 00:24:41,720
And so, so it's very
interesting.

467
00:24:41,720 --> 00:24:44,720
The biological networks are
better at this, significantly

468
00:24:44,720 --> 00:24:46,360
better at this.
They do more memories, they do

469
00:24:46,360 --> 00:24:47,440
more, you know, causal
emergence.

470
00:24:47,440 --> 00:24:52,200
So, so, so clearly evolution, I
think like this feature, but

471
00:24:52,200 --> 00:24:54,040
even the random ones do it a
little bit.

472
00:24:54,560 --> 00:24:58,720
And so that suggests to me that
even before you get into any

473
00:24:58,720 --> 00:25:02,600
kind of selection or any kind of
evolutionary cycles, just from

474
00:25:02,600 --> 00:25:05,800
the math of alone, you get this
little free gift that, that,

475
00:25:05,800 --> 00:25:07,400
that, that kick starts some of
the stuff.

476
00:25:07,400 --> 00:25:09,480
And then of course, you can
optimize the heck out of it, you

477
00:25:09,480 --> 00:25:11,560
know, as you as you do these,
these selection cycles.

478
00:25:11,840 --> 00:25:14,720
But, but right at the beginning,
just just the, the, the math

479
00:25:14,720 --> 00:25:17,080
that, that, that governs these
kind of networks.

480
00:25:17,760 --> 00:25:19,520
And of course, some of that, you
know, Stu Kauffman talked about

481
00:25:19,520 --> 00:25:22,160
some of this stuff a long time
ago, but it's, it's not just the

482
00:25:22,160 --> 00:25:24,560
dynamics, it's not just the
dynamical system.

483
00:25:24,760 --> 00:25:28,240
It actually has both learning
ability and this, this kind of

484
00:25:28,480 --> 00:25:31,200
cohesive integration that grows
as a result of experience.

485
00:25:31,400 --> 00:25:33,720
So I think we can trace and as
you said, you know, both both

486
00:25:33,720 --> 00:25:35,240
very, sort of very high and very
low.

487
00:25:35,680 --> 00:25:38,560
There is this kind of weird
cycle where you can take metrics

488
00:25:38,720 --> 00:25:42,000
from the kinds of things that
that that you know,

489
00:25:42,160 --> 00:25:45,160
neuroscientists do with, with
human patients and look at how

490
00:25:45,160 --> 00:25:47,320
this works with very simple
systems.

491
00:25:49,960 --> 00:25:52,920
Or I was going to add to this is
something that we've actually

492
00:25:52,920 --> 00:25:56,040
been done at the other end, that
is the neuroscience and I have a

493
00:25:56,680 --> 00:26:01,080
couple of colleagues at
University of San Francisco and

494
00:26:01,360 --> 00:26:06,920
the in which we're looking at
simultaneous EEG and, and fMRI

495
00:26:07,680 --> 00:26:12,600
and asking the question.
I mean, because my assumption in

496
00:26:12,600 --> 00:26:16,360
all of this is that the, the
dynamics or what you might want

497
00:26:16,360 --> 00:26:19,240
to call it computing, either
called semiosis.

498
00:26:19,240 --> 00:26:23,600
But but basically the, the, the
informational processes are

499
00:26:23,600 --> 00:26:28,800
generating the physical
material, energetic processes

500
00:26:28,960 --> 00:26:31,280
that make the information
processes possible.

501
00:26:31,280 --> 00:26:35,080
That is, there's a what what
Hofstadter would have called a

502
00:26:35,080 --> 00:26:40,400
strange loop in which the
dynamics and the semiotic

503
00:26:40,480 --> 00:26:45,680
activity is generating and
maintaining the physicality that

504
00:26:45,680 --> 00:26:48,280
makes it possible.
So you get this, this loop

505
00:26:49,200 --> 00:26:51,960
between sort of the, you might
say, the ontology and the

506
00:26:52,160 --> 00:26:55,960
epistemology of a system in
which each depend on each other.

507
00:26:56,160 --> 00:26:59,440
Well, we thought that that must
be possible also for how brains

508
00:26:59,440 --> 00:27:02,640
work, that we've been ignoring
this in part because we thought,

509
00:27:02,640 --> 00:27:06,640
we assumed, if MRI was telling
us that when these neurons get

510
00:27:06,640 --> 00:27:10,080
information and they're active,
they change their metabolism.

511
00:27:10,760 --> 00:27:14,680
What we've begun to discover is
that there's depending on the

512
00:27:14,680 --> 00:27:19,240
task, sometimes the metabolism
anticipates and precedes the

513
00:27:19,240 --> 00:27:23,080
neural activity and vice versa.
And you can look at it in terms

514
00:27:23,080 --> 00:27:25,600
of different parts of the
nervous system, you know, in

515
00:27:25,600 --> 00:27:30,320
general, sort of these broad
distinctions like like like a

516
00:27:30,320 --> 00:27:33,320
default network and things like
that, but, but also between

517
00:27:33,320 --> 00:27:36,320
brain stem, midbrain and and
forebrain and so on, so on and

518
00:27:36,320 --> 00:27:37,880
so forth.
There's these, it's interesting

519
00:27:37,880 --> 00:27:41,920
exchange that it looks as though
a lot of what we would, what we

520
00:27:41,920 --> 00:27:45,960
would describe as self generated
activity, intrinsically

521
00:27:45,960 --> 00:27:50,120
generated activity.
The metabolism precedes and it

522
00:27:50,120 --> 00:27:54,720
drives up neural activity which
changes the metabolism, which

523
00:27:54,720 --> 00:27:57,760
drives up the neural activity
and then shifts metabolism to

524
00:27:57,760 --> 00:28:00,800
other parts of the brain.
So that, that in effect, there

525
00:28:00,800 --> 00:28:05,120
is, there is both a leading and
following feature in which the,

526
00:28:05,440 --> 00:28:09,000
you might say the computing or
the, the semiotic activity that

527
00:28:09,000 --> 00:28:14,480
is a signal generation activity
and it's substrate are

528
00:28:14,480 --> 00:28:16,560
critically linked.
And it turns out they're linked

529
00:28:16,800 --> 00:28:19,720
at a number of levels, not just
neurotransmitters, but also

530
00:28:20,400 --> 00:28:25,960
things like nitric oxide and,
and, and various ionic channels

531
00:28:25,960 --> 00:28:29,440
that sort of ionic potentials
that sort of pass through large

532
00:28:29,440 --> 00:28:32,920
areas.
All of these areas are linked so

533
00:28:32,920 --> 00:28:36,600
that there's no clear
distinction between what you

534
00:28:36,720 --> 00:28:40,040
might call the information
processing and its physicality.

535
00:28:40,280 --> 00:28:43,080
It's all part of the same thing.
And and that's, that's sort of

536
00:28:43,080 --> 00:28:45,840
the the other side, the other
extreme, I think of what you've

537
00:28:45,840 --> 00:28:51,880
been talking about, Michael.
You guys were talking about

538
00:28:51,880 --> 00:28:53,880
causal emergence.
I'm going to do my best to just

539
00:28:53,880 --> 00:28:55,960
be a family on the wall of this
conversation, by the way.

540
00:28:55,960 --> 00:28:58,280
So if you guys have any
questions that you want to ask

541
00:28:58,280 --> 00:28:59,840
each other directly, by all
means.

542
00:28:59,840 --> 00:29:03,280
But how do let's let's talk
about constraint and causality.

543
00:29:03,440 --> 00:29:05,200
Michael, let's let's start with
you.

544
00:29:06,440 --> 00:29:09,640
How do constraints shape
outcomes more powerfully than

545
00:29:09,640 --> 00:29:12,440
initial conditions?
Let's say, for example, is

546
00:29:12,440 --> 00:29:15,360
absence more causal than
presence?

547
00:29:19,320 --> 00:29:21,800
Yeah, well.
I mean, the first thing is I'm,

548
00:29:21,840 --> 00:29:25,240
I'm completely on board with
things like absences being

549
00:29:25,240 --> 00:29:29,040
causal functional in inputs into
into everything that happens.

550
00:29:29,040 --> 00:29:31,800
So, so I am on board with that.
And, and I think, you know, I

551
00:29:31,800 --> 00:29:34,720
think I think Terry will, will,
will say probably the, you know,

552
00:29:34,720 --> 00:29:38,360
the most about it.
What I would like to say is that

553
00:29:39,120 --> 00:29:42,680
I, I, I in, in addition to
constraints, I, I tend to track

554
00:29:42,680 --> 00:29:45,400
two other two other things that,
that are kind of in that, in

555
00:29:45,400 --> 00:29:49,720
that Class 1 is this interesting
aspect that was really

556
00:29:49,720 --> 00:29:52,920
highlighted by the, by
morphological computing, by the

557
00:29:52,920 --> 00:29:56,040
way, you know, Pfeiffer and, and
Bonguard and, and, and folks

558
00:29:56,040 --> 00:29:59,720
like that where when you have
embodiment and, and I study

559
00:29:59,720 --> 00:30:01,720
embodiment, not just in the
three-dimensional world, but

560
00:30:01,720 --> 00:30:03,320
embodiment in these other
spaces.

561
00:30:03,320 --> 00:30:05,720
And I think it actually, it
actually applies perfectly well

562
00:30:05,720 --> 00:30:08,160
to, to all these other problems,
spaces that life works in.

563
00:30:09,240 --> 00:30:11,920
There are many constraints that
guide what's going to happen,

564
00:30:12,160 --> 00:30:14,920
but there are also another thing
that's kind of amazing is that

565
00:30:15,080 --> 00:30:19,400
in addition to the constraints,
there are also these ways that

566
00:30:19,400 --> 00:30:22,720
that you can offload an awful
lot of computation onto the

567
00:30:22,720 --> 00:30:24,600
environment.
It's like, and, and this is

568
00:30:24,600 --> 00:30:26,720
what, this is what they show
when they have these robots and,

569
00:30:26,720 --> 00:30:29,200
and other things that basically
there is no controller that has

570
00:30:29,200 --> 00:30:30,720
to handle all the things that
has to happen.

571
00:30:30,880 --> 00:30:33,440
They're basically letting the
environment do the compute for

572
00:30:33,440 --> 00:30:35,960
them.
And this is this is something

573
00:30:35,960 --> 00:30:38,880
very new in our lab that we're
just getting into is asking what

574
00:30:38,880 --> 00:30:40,520
does that look like in these
other spaces?

575
00:30:40,640 --> 00:30:43,800
What does it look like to
offload transcriptional problem

576
00:30:43,800 --> 00:30:46,560
solving, physiological problem
solving, anatomical problem

577
00:30:46,560 --> 00:30:49,880
solving to the rules of these
other spaces that absolutely

578
00:30:49,880 --> 00:30:52,400
provide constraints that will
guide subsequent events.

579
00:30:52,880 --> 00:30:55,000
But in addition to the
constraints, they, they're also

580
00:30:55,000 --> 00:30:56,640
helping out.
I mean, that's that's the part

581
00:30:56,640 --> 00:30:59,880
that I think is, is, is, you
know, really, really wild is

582
00:30:59,880 --> 00:31:03,920
that there's actually all kinds
of benefits that that come from

583
00:31:03,920 --> 00:31:07,120
that, that where the environment
is is actually helping you do

584
00:31:07,120 --> 00:31:08,400
the do the thing that you want
it to do.

585
00:31:08,800 --> 00:31:11,840
And, and the, and the other
thing that that I track a lot

586
00:31:11,840 --> 00:31:17,000
are patterns that arise not from
physical facts.

587
00:31:17,000 --> 00:31:19,800
In other words, there are
mathematical structures, there

588
00:31:19,800 --> 00:31:23,600
are laws of computation, laws of
geometry, of number theories,

589
00:31:23,600 --> 00:31:27,160
all these things that evolution,
I think uses a lot.

590
00:31:27,480 --> 00:31:30,280
They, it uses them as, as in
effect free lunches, because you

591
00:31:30,280 --> 00:31:33,880
can save an awful lot of
evolutionary compute time if you

592
00:31:33,880 --> 00:31:37,280
can just be handed to certain
truths from the, you know, from,

593
00:31:37,280 --> 00:31:39,120
from the laws of mathematics and
computation.

594
00:31:39,360 --> 00:31:41,680
And, and I think those are
especially interesting because

595
00:31:41,680 --> 00:31:45,760
they're not, many of them are
not determined by physical

596
00:31:45,760 --> 00:31:47,040
facts.
They wouldn't change if you

597
00:31:47,040 --> 00:31:49,200
shuffle the constants at The Big
Bang and things like that.

598
00:31:49,600 --> 00:31:53,920
And so those kinds of causes
that are not the traditional,

599
00:31:54,080 --> 00:31:56,800
you know, why did this happen?
Oh, because the BMV, you know,

600
00:31:56,800 --> 00:31:59,240
what molecule is well, that that
some of these causes are not

601
00:31:59,240 --> 00:32:00,720
like that at all.
They're like a completely

602
00:32:00,720 --> 00:32:04,320
different class of things.
And what's really critical for

603
00:32:04,320 --> 00:32:07,080
us now is because, because we
try to drive a lot of these

604
00:32:07,080 --> 00:32:09,960
things into very practical
programs in regenerative

605
00:32:09,960 --> 00:32:13,000
medicine and in bioengineering
and, and so on is to how to take

606
00:32:13,000 --> 00:32:15,680
advantage of these things.
So I, you know, we, we would

607
00:32:15,680 --> 00:32:21,160
like to know as we look for
causes that are beyond physics

608
00:32:21,160 --> 00:32:23,520
and history are the things that
biologists love as, as causes.

609
00:32:23,520 --> 00:32:27,720
But then as we, as we see, there
are these other things and being

610
00:32:27,720 --> 00:32:31,640
able to exploit those and
exploit that space of, of, of

611
00:32:31,640 --> 00:32:35,560
these of, of these patterns that
that can be used for both for,

612
00:32:35,560 --> 00:32:38,720
for our construction and can
help us communicate our goals to

613
00:32:38,720 --> 00:32:40,800
cells and tissues in
regenerative context.

614
00:32:41,080 --> 00:32:43,880
So, so yeah, yeah, constraints
and and and various other things

615
00:32:43,880 --> 00:32:48,640
that are like them.
There's so much in this question

616
00:32:49,080 --> 00:32:51,960
and, and, and we'll go back and
forth on it maybe in this

617
00:32:51,960 --> 00:32:55,080
process.
But but one of the, the things

618
00:32:55,480 --> 00:32:59,000
that sort of drove my thinking
about this originally and pushed

619
00:32:59,000 --> 00:33:03,880
me back into the theory was, was
an interest in the Fibonacci

620
00:33:03,920 --> 00:33:09,040
spirals of plants and precise
mathematical relationships.

621
00:33:09,040 --> 00:33:12,480
These wonderful spirals that you
see and for example, the seeds

622
00:33:12,480 --> 00:33:18,320
of a of a sunflower and knowing
that the genetics was not doing

623
00:33:18,320 --> 00:33:22,520
the math that the genetics was
in effect only really

624
00:33:22,520 --> 00:33:27,200
controlling thresholds in which
various regions of the plant was

625
00:33:27,200 --> 00:33:30,400
releasing auxins.
These these these various growth

626
00:33:30,400 --> 00:33:32,840
hormones in the plant.
And yet producing these

627
00:33:32,840 --> 00:33:34,960
remarkable spirals that we find
everywhere.

628
00:33:35,680 --> 00:33:40,040
And recognizing that one of the
advantages of these spirals is

629
00:33:40,040 --> 00:33:44,480
that if it goes up the length of
the stem, for example, for

630
00:33:44,480 --> 00:33:48,600
producing leaves, then the
leaves are maximally out of each

631
00:33:48,600 --> 00:33:51,120
others way.
If you want to capture sunlight

632
00:33:51,120 --> 00:33:54,080
really well, what you want to
have is things that are just,

633
00:33:54,080 --> 00:33:57,520
you know, they're as tight as
they can be, but maximally also

634
00:33:57,520 --> 00:34:01,400
out of each others way.
And it turns out that it's a,

635
00:34:01,520 --> 00:34:05,240
it's wonderful that evolution
would want this, but it didn't

636
00:34:05,240 --> 00:34:08,639
have to in a sense encode it
because it was there in the

637
00:34:08,639 --> 00:34:11,280
geometry itself.
And all you have to do is you

638
00:34:11,280 --> 00:34:15,280
want to, you know, create
different kinds of spirals,

639
00:34:15,280 --> 00:34:19,080
different kinds of Fibonacci's
with tighter or looser spirals.

640
00:34:19,239 --> 00:34:22,320
You just, you just change the
thresholds in which things

641
00:34:22,320 --> 00:34:26,840
respond.
And so at one point in time

642
00:34:27,199 --> 00:34:31,480
somebody said that I should call
this the lazy gene hypothesis

643
00:34:31,760 --> 00:34:35,040
that that that genes will only
control what they have to

644
00:34:35,040 --> 00:34:37,920
control.
That in a sense, if they are

645
00:34:37,920 --> 00:34:41,080
over controlling something, it's
going to eventually degenerate

646
00:34:41,280 --> 00:34:44,760
if there's a redundant capacity
out there, that if that

647
00:34:44,760 --> 00:34:47,639
redundancy is produced by the
constraints in this case of

648
00:34:47,639 --> 00:34:51,520
geometry of space.
That caught my attention.

649
00:34:51,560 --> 00:34:54,960
And in fact, it's now been part
of a recent work that I've been

650
00:34:54,960 --> 00:34:59,560
doing that I that I call inverse
Darwinism as we begin to look at

651
00:34:59,640 --> 00:35:02,080
evolution.
I mean, this was recognized back

652
00:35:02,080 --> 00:35:06,560
as early as 1970.
That is that a lot of genetic

653
00:35:06,560 --> 00:35:10,400
evolution is by virtue of
duplication, but it's not just

654
00:35:10,400 --> 00:35:13,040
gene duplication, it's
duplication at lots of levels.

655
00:35:13,040 --> 00:35:16,880
And one of the things that that
has to happen for evolution, for

656
00:35:16,880 --> 00:35:19,320
life is that you have to always
over produce.

657
00:35:20,280 --> 00:35:22,880
You can't just keep up with the
second law of thermodynamics.

658
00:35:22,880 --> 00:35:24,360
You've got to go a little bit
faster.

659
00:35:24,480 --> 00:35:26,120
So you've got to generate too
much.

660
00:35:26,760 --> 00:35:30,960
And that allows you two things
the redundancy provides.

661
00:35:31,240 --> 00:35:34,920
You know, in information theory,
redundancy is the correct

662
00:35:34,920 --> 00:35:39,000
errors, but redundancy in
biology is to make you error

663
00:35:39,000 --> 00:35:42,000
tolerant.
That is, if you have redundancy,

664
00:35:42,000 --> 00:35:44,360
then things can go wrong and you
can still keep going.

665
00:35:44,840 --> 00:35:48,600
Well, if you got a system that
always overproduces stuff, that

666
00:35:48,600 --> 00:35:52,360
always makes redundant cells,
redundant genes, redundant

667
00:35:52,360 --> 00:35:55,920
proteins and whatever, you've
always got a little error

668
00:35:55,920 --> 00:35:58,400
protection there.
You've got a lot of little

669
00:35:58,400 --> 00:36:02,120
plasticity, which means you've
got exploratory space available

670
00:36:02,120 --> 00:36:05,280
to you.
Well, it turns out that in

671
00:36:05,280 --> 00:36:08,680
evolutionary theory, of course,
we're, we're thinking only of

672
00:36:08,680 --> 00:36:12,800
the eliminative side of things.
But in fact, what we're talking

673
00:36:12,800 --> 00:36:15,800
about is that the generative
side of things is not just

674
00:36:15,800 --> 00:36:19,680
generating random variation,
it's generating redundancy.

675
00:36:19,920 --> 00:36:22,600
And redundancy has a very, very
special role and it plays a

676
00:36:22,600 --> 00:36:25,200
critical role in information
theory.

677
00:36:25,200 --> 00:36:28,280
Now, this will bring me back to
the question of constraint

678
00:36:28,880 --> 00:36:33,720
because a lot when we look at
the original mathematics,

679
00:36:33,800 --> 00:36:35,640
grammatical theory of
communication that Claude

680
00:36:35,640 --> 00:36:41,040
Shannon puts together, there are
two measures of information.

681
00:36:41,240 --> 00:36:44,360
They both involve what he calls
entropy, which is just basically

682
00:36:44,640 --> 00:36:48,720
a way to talk about the possible
variety of something either in

683
00:36:48,720 --> 00:36:53,520
time or in space or or whatever.
But there's the original entropy

684
00:36:53,520 --> 00:36:56,360
of what could have been
produced, and then there's the

685
00:36:56,360 --> 00:36:58,680
entropy of, you know, what
results.

686
00:36:58,960 --> 00:37:01,520
So if you're sending a message,
there might be a bunch of

687
00:37:01,520 --> 00:37:05,920
messages you could send, but the
one that's picked up does not

688
00:37:05,920 --> 00:37:08,400
have all that all that variety
in it.

689
00:37:08,680 --> 00:37:12,320
And so in effect, there's two
entropies for Shannon, and

690
00:37:12,320 --> 00:37:15,480
mostly we focus on one of them,
which is just what he calls the

691
00:37:15,760 --> 00:37:17,560
the original, the entropy of a
channel.

692
00:37:17,760 --> 00:37:21,080
But the entropy of a message is
the difference between what the

693
00:37:21,080 --> 00:37:23,760
channel carries and what was
actually received.

694
00:37:24,160 --> 00:37:26,400
And that tells you that what's
happened is there's been a

695
00:37:26,400 --> 00:37:30,880
constraint imposed upon that
information, and that

696
00:37:30,880 --> 00:37:33,600
information was carried on
something physical.

697
00:37:34,520 --> 00:37:37,400
But because it's something
physical, it's subject to

698
00:37:37,400 --> 00:37:41,040
thermodynamic laws as well.
And this is one of the reasons

699
00:37:41,040 --> 00:37:44,440
why the thermodynamic story and
the information story are

700
00:37:44,440 --> 00:37:47,920
linked, and they're linked
precisely by this concept of

701
00:37:47,920 --> 00:37:51,160
constraint.
That is the the received

702
00:37:51,160 --> 00:37:54,240
information.
That's a constraint on the

703
00:37:54,240 --> 00:37:58,520
possible information.
In effect, it's the result that

704
00:37:58,520 --> 00:38:04,000
there were something had to act
on the signal, something had to

705
00:38:04,000 --> 00:38:07,600
act on it to constrain its
thermodynamics, which then would

706
00:38:07,600 --> 00:38:10,440
constrain its information
entropy.

707
00:38:11,160 --> 00:38:13,920
And that tells you this is how
it could be about something,

708
00:38:14,640 --> 00:38:18,120
because the constraint of the
channel can now be in a sense

709
00:38:18,120 --> 00:38:23,120
about work that was done on it.
Because to shift something to a,

710
00:38:23,280 --> 00:38:25,480
a less probable state takes
work.

711
00:38:26,280 --> 00:38:29,640
And this then helps us
understand the whole concept of

712
00:38:29,640 --> 00:38:32,480
work, which is of course
necessary for, for everything

713
00:38:32,480 --> 00:38:35,960
we're talking about here, that
that it's not just the release

714
00:38:35,960 --> 00:38:39,920
of energy, it's the release of
energy in a constrained context.

715
00:38:40,200 --> 00:38:42,880
You've got to constrain the the
energy release.

716
00:38:43,080 --> 00:38:47,440
So you know, even a, a piston in
an automobile engine that where

717
00:38:47,440 --> 00:38:50,440
you constrain the explosion
allows you to push something far

718
00:38:50,440 --> 00:38:53,360
from equilibrium to push a car
uphill, for example.

719
00:38:54,800 --> 00:38:59,800
So it constraint is unnecessary,
but often times ignored or just

720
00:38:59,800 --> 00:39:03,600
sort of assumed part of work
when we talked about biological

721
00:39:03,600 --> 00:39:05,640
systems.
But it's.

722
00:39:06,520 --> 00:39:09,240
But if you recognize from
information theory that it's the

723
00:39:09,240 --> 00:39:12,240
constraints.
That carry the.

724
00:39:12,240 --> 00:39:16,120
Information.
But constraints are also what

725
00:39:16,120 --> 00:39:20,440
determines the structure of the
work that was done, and the

726
00:39:20,440 --> 00:39:23,440
structure of the work can
generate new constraints.

727
00:39:24,520 --> 00:39:27,360
So if you can generate new kinds
of work with constraints, and

728
00:39:27,360 --> 00:39:30,360
that work can generate new kinds
of constraints, you generate new

729
00:39:30,360 --> 00:39:32,680
kinds of work.
The whole possibility of

730
00:39:32,680 --> 00:39:36,080
evolution is the result of this
interesting relationship between

731
00:39:36,080 --> 00:39:39,080
constraint and the release and
the control of energy.

732
00:39:39,840 --> 00:39:43,000
And that means that if you have
a system that can both generate

733
00:39:43,160 --> 00:39:47,280
and then remember and transmit
constraint, now you have a

734
00:39:47,280 --> 00:39:54,440
system that can evolve.
On let's sticking with that,

735
00:39:54,520 --> 00:39:55,960
let's talk about regeneration
memory.

736
00:39:55,960 --> 00:39:58,160
You'll, you'll notice that the
first question I asked, well

737
00:39:58,160 --> 00:40:01,320
this previous one was more tell
it towards your work, Terrence,

738
00:40:01,320 --> 00:40:03,200
but this next one's more towards
yours.

739
00:40:03,200 --> 00:40:08,720
Mike, perhaps Terry, you could
start with this answer regarding

740
00:40:08,720 --> 00:40:12,400
regeneration and memory.
When a planaria regrows its head

741
00:40:12,560 --> 00:40:15,960
and retains memories, where do
you think those memories are

742
00:40:15,960 --> 00:40:18,480
stored?
And what does this tell us about

743
00:40:18,480 --> 00:40:24,320
the body as memory substrate?
It's interesting because memory

744
00:40:24,320 --> 00:40:28,440
assumes a lot of what we've
talked about already and and it

745
00:40:28,440 --> 00:40:31,200
assumes one thing that we don't
talk about a lot, but I think

746
00:40:31,200 --> 00:40:34,560
it's very characteristic in
Planaria, it's characteristic in

747
00:40:34,560 --> 00:40:38,280
how brains develop in multi cell
bodies and that sort of thing.

748
00:40:38,280 --> 00:40:43,800
And that is that you can
initiate the system started

749
00:40:43,800 --> 00:40:48,560
going with a highly compressed
source of information.

750
00:40:49,520 --> 00:40:53,120
DNA is compression.
And so when we talk about

751
00:40:53,120 --> 00:40:57,080
complexity, one of the problems
we want to talk about is, you

752
00:40:57,080 --> 00:41:01,760
know, you got a list of numbers.
Is that list compressible in the

753
00:41:01,760 --> 00:41:03,440
sense that there's some
redundancy in it?

754
00:41:04,560 --> 00:41:08,800
If it is, you can take advantage
of that redundancy in

755
00:41:08,880 --> 00:41:11,280
reconstructing from a compressed
version.

756
00:41:11,680 --> 00:41:14,800
And this is where algorithms
often times play a critical

757
00:41:14,800 --> 00:41:16,440
role.
And in fact, we've analysed

758
00:41:16,440 --> 00:41:21,000
complexity these days since,
since Chaitin and Kolmogorov and

759
00:41:21,000 --> 00:41:25,440
others in terms of, you know,
what I would call a

760
00:41:25,440 --> 00:41:28,400
decompression rule?
A decompression rule is you've

761
00:41:28,400 --> 00:41:31,200
compressed something and now you
got a decompression rule.

762
00:41:31,200 --> 00:41:35,560
The key is what what is the
decompression process?

763
00:41:37,000 --> 00:41:39,240
It's not that all the
information is stored in the

764
00:41:39,240 --> 00:41:42,400
DNA.
What you've got is a highly

765
00:41:42,400 --> 00:41:46,440
compressed system that if you,
if you pair it with a

766
00:41:46,440 --> 00:41:50,640
decompression system, whether
it's algorithmic or is a set of

767
00:41:50,640 --> 00:41:53,920
constraints that are built into
the geometry of the, of the

768
00:41:53,920 --> 00:42:00,760
world or built into the
recursive activity of some of

769
00:42:00,760 --> 00:42:04,120
some process, then you can take
advantage of this is that lazy

770
00:42:04,120 --> 00:42:08,400
gene idea that I was talking
about that in effect, what we

771
00:42:08,400 --> 00:42:13,560
want to talk about is that
whatever you can rely on in your

772
00:42:13,560 --> 00:42:17,400
decompression process, you can
now leave out of the compressed

773
00:42:17,400 --> 00:42:20,680
source.
And so, you know, this is even

774
00:42:20,680 --> 00:42:25,080
true for something as trivial
as, you know, architectural

775
00:42:25,080 --> 00:42:28,360
design.
And for the architect knows that

776
00:42:28,360 --> 00:42:32,640
there's going to be wood and
metal and tools and people

777
00:42:32,640 --> 00:42:35,720
putting it together knows that
somebody will know how to

778
00:42:35,720 --> 00:42:38,760
decompress this model and do
something with it.

779
00:42:39,200 --> 00:42:42,120
The decompression is ignored
because he can just assume it's

780
00:42:42,120 --> 00:42:45,040
there.
That's the problem is that we

781
00:42:45,040 --> 00:42:47,960
tend to ignore the decompression
process.

782
00:42:49,080 --> 00:42:53,640
And So what I like to like to
say in terms of my favorite

783
00:42:53,640 --> 00:43:00,840
example of this is, you know, is
a frog built out more complex

784
00:43:01,080 --> 00:43:04,080
than if you've taken the body of
the frog and thrown it into a

785
00:43:04,080 --> 00:43:07,360
blender and blend it up into a
mix.

786
00:43:07,640 --> 00:43:10,720
In one sense, the mix is much
more complex.

787
00:43:11,120 --> 00:43:16,240
You would have a hard time
coming up with a systematic set

788
00:43:16,240 --> 00:43:18,840
of algorithms to figure out
where different kinds of

789
00:43:18,840 --> 00:43:22,960
molecules were located.
But in the adult frog, it's much

790
00:43:22,960 --> 00:43:27,600
more simple in one sense.
So it confuses our concept.

791
00:43:27,600 --> 00:43:30,280
It's, it's simple in the fact
that if I know that there's a

792
00:43:30,280 --> 00:43:33,440
bone cell here, I know there's
likely to be a bone cell next to

793
00:43:33,440 --> 00:43:35,440
it.
If there's a skin cell here, I'm

794
00:43:35,440 --> 00:43:38,200
like, you know, you don't have
that in the frog smoothie.

795
00:43:39,040 --> 00:43:41,360
You basically don't have any
predictability.

796
00:43:41,360 --> 00:43:43,960
And yet at that level it's more
complex.

797
00:43:44,400 --> 00:43:47,960
But it's, but it's a
misunderstanding of complexity

798
00:43:48,280 --> 00:43:51,960
because what's actually happened
in the case of the actual frog

799
00:43:52,320 --> 00:43:58,840
is that by taking a highly
compressed representation that

800
00:43:58,840 --> 00:44:05,000
is its genome in a decompression
system, it's cell that now

801
00:44:05,000 --> 00:44:08,920
depends upon other cells, their
position, what they release,

802
00:44:09,040 --> 00:44:12,920
what their electrical ionic
activity is, you know, how

803
00:44:12,920 --> 00:44:16,560
diffusion processes, when they
release molecules will affect

804
00:44:17,120 --> 00:44:20,280
cells close by versus those
different from them and so on.

805
00:44:20,600 --> 00:44:22,880
All of that's part of the
decompression process.

806
00:44:23,080 --> 00:44:27,720
But the decompression process is
not to be confused with the

807
00:44:27,720 --> 00:44:31,960
compressed representation and
and so it's.

808
00:44:32,200 --> 00:44:36,240
I've been very much interested
in two ways.

809
00:44:36,640 --> 00:44:41,280
The decompression process is the
functional part, but in fact

810
00:44:41,920 --> 00:44:45,680
evolution has to work backwards.
It has to figure out

811
00:44:45,680 --> 00:44:49,280
compressions.
And I think one of the real

812
00:44:49,280 --> 00:44:52,880
problems we have is that we tend
to think of it only in one way.

813
00:44:52,880 --> 00:44:55,360
We look at the function of
something and say, now let me

814
00:44:55,360 --> 00:44:57,880
just now sort of break it down
into his parts and his function.

815
00:44:58,200 --> 00:45:00,920
Now it evolved backwards in some
sense.

816
00:45:01,800 --> 00:45:05,200
The compression comes second.
So I've always told people that,

817
00:45:05,200 --> 00:45:09,160
you know, this view that DNA or
RNA came first is a mistake.

818
00:45:10,440 --> 00:45:14,280
They that was a compression of
some other dynamical process.

819
00:45:15,240 --> 00:45:18,320
And once we understand the
decompression process and how

820
00:45:18,320 --> 00:45:21,880
the two have to, in a sense,
evolve together, we realize that

821
00:45:21,880 --> 00:45:24,680
things get better and better at
compressing and decompressing.

822
00:45:24,920 --> 00:45:28,720
And that's the complexity issue.
But it's a pairing of those two

823
00:45:28,720 --> 00:45:32,320
processes.
Yeah, yeah, So, so super.

824
00:45:32,320 --> 00:45:34,480
Super important this, this
compression decompression thing

825
00:45:34,480 --> 00:45:39,600
very is, is is critical because
one of the kind of meta

826
00:45:39,600 --> 00:45:42,280
constraints that biology has is
that the substrate is

827
00:45:42,280 --> 00:45:44,680
unreliable.
In other words, on a large

828
00:45:44,680 --> 00:45:46,920
scale, you're guaranteed you're
going to be mutated.

829
00:45:46,920 --> 00:45:49,360
On the small scale, you don't
know how many copies of your

830
00:45:49,360 --> 00:45:51,520
proteins you have.
Things are being degraded.

831
00:45:51,920 --> 00:45:54,280
And that's, that's quite
different than all of the

832
00:45:54,280 --> 00:45:56,880
computational architectures we
make now where you work very

833
00:45:56,880 --> 00:45:59,480
hard to have abstraction layers,
where if you work at a high

834
00:45:59,480 --> 00:46:02,200
level language, you don't, don't
worry about your, your, you

835
00:46:02,200 --> 00:46:05,720
know, a copper of your, your,
your registers floating off and

836
00:46:05,720 --> 00:46:07,240
so on.
And so, so I think one of the

837
00:46:07,240 --> 00:46:11,040
implications of that if, if, and
we've done a lot of simulations

838
00:46:11,040 --> 00:46:15,080
on this now that evolving over
an unreliable medium.

839
00:46:15,080 --> 00:46:17,040
And so, so there's two
components that, that, that I'm

840
00:46:17,040 --> 00:46:19,360
really interested in.
One is the unreliable medium and

841
00:46:19,360 --> 00:46:24,040
the other is the fact that in,
in living systems, you have

842
00:46:24,240 --> 00:46:25,880
problem solvers all the way
down.

843
00:46:26,040 --> 00:46:28,800
In other words, there are
systems all the way down that

844
00:46:28,800 --> 00:46:31,440
are not just mechanically doing
whatever they do, but they

845
00:46:31,440 --> 00:46:35,400
actually do have a homeo dynamic
and other other problem solving

846
00:46:35,440 --> 00:46:37,760
capacities.
And, and that works.

847
00:46:37,800 --> 00:46:42,400
What, what, what that ends up
doing is first of all, the, the,

848
00:46:42,440 --> 00:46:45,120
the, what One of the amazing
things about the decompression

849
00:46:45,120 --> 00:46:48,000
side is that because you've lost
information during the

850
00:46:48,000 --> 00:46:51,480
compression, it means that the,
the right side of that bow tie

851
00:46:51,560 --> 00:46:54,120
is, has to be creative.
I, I tend to think that the left

852
00:46:54,120 --> 00:46:56,280
side, the compression can be out
the rhythmic, but the right side

853
00:46:56,480 --> 00:47:00,000
has to be creative because it's
kind of like in, in, in the, on

854
00:47:00,000 --> 00:47:01,280
the cognition and it's the same
thing.

855
00:47:01,280 --> 00:47:03,200
We don't have access to the
past.

856
00:47:03,200 --> 00:47:05,920
What we have access to are the N
grams that was, that were made

857
00:47:05,920 --> 00:47:08,600
by previous versions of you in
your brain and body.

858
00:47:09,000 --> 00:47:12,000
And then it's your job every, I
don't know, 300 milliseconds or

859
00:47:12,000 --> 00:47:15,440
whatever it is to continuously
interpret those things and build

860
00:47:15,440 --> 00:47:19,680
up an ongoing, A labile story of
what you are, what, what you're

861
00:47:19,680 --> 00:47:22,080
doing, what's the outside world?
What do your memories mean?

862
00:47:22,600 --> 00:47:27,040
And in doing that, it's
basically a semiosis because

863
00:47:27,360 --> 00:47:29,880
these are messages left for you
from your past self.

864
00:47:29,880 --> 00:47:32,280
They're just like messages you
get laterally from, from your,

865
00:47:32,360 --> 00:47:34,160
you know, from things that are
going on at the same time.

866
00:47:34,680 --> 00:47:38,360
And it's, it's this process of,
of building the best story that

867
00:47:38,360 --> 00:47:41,440
you can at any given moment, not
necessarily the story you used

868
00:47:41,440 --> 00:47:43,080
to have.
And that becomes really

869
00:47:43,080 --> 00:47:46,480
interesting in, in, in
development because if you don't

870
00:47:46,480 --> 00:47:49,200
have, if you, if your allegiance
is not to the fidelity of the

871
00:47:49,200 --> 00:47:51,840
interpretation, meaning I need
to do it exactly the way my

872
00:47:51,840 --> 00:47:53,920
ancestors did.
But instead what I think it's

873
00:47:53,920 --> 00:47:57,320
doing is I think it's a problem
solving agent that says, here

874
00:47:57,320 --> 00:47:58,880
are the prompts that I've been
given.

875
00:47:59,040 --> 00:48:01,560
What is the best thing I can put
together right now?

876
00:48:01,560 --> 00:48:04,200
And that results in some amazing
plasticity in biology.

877
00:48:04,200 --> 00:48:06,480
I'll just give you a couple of
my, my favorite examples.

878
00:48:08,160 --> 00:48:11,520
This this, this Fankhauser work
was, was, was was incredible.

879
00:48:11,520 --> 00:48:14,400
So you got this Newt and you
take a cross section through the

880
00:48:14,400 --> 00:48:17,600
kidney tubules and you see
there's like 8 to 10 cells that

881
00:48:17,600 --> 00:48:19,560
work together and they give you
this thing with a lumen in the

882
00:48:19,560 --> 00:48:20,920
middle.
And it's this nice, nice tubule.

883
00:48:21,320 --> 00:48:23,040
So you can make these polyploid
nudes.

884
00:48:23,040 --> 00:48:25,600
If you make polyploid nudes,
they have, you know, extra

885
00:48:25,600 --> 00:48:27,680
copies of of their, the genetic
material.

886
00:48:27,960 --> 00:48:30,400
And as a result, what they do is
they make their cells bigger to

887
00:48:30,400 --> 00:48:33,680
accommodate the bigger nucleus.
OK, so that's cool, but the Newt

888
00:48:33,680 --> 00:48:36,160
stays the same size.
So you say, well, how did that

889
00:48:36,160 --> 00:48:37,520
happen?
Then you take the cross section

890
00:48:37,520 --> 00:48:39,440
and you see that actually.
Well then there are fewer of

891
00:48:39,440 --> 00:48:41,280
these larger cells to build the
same thing.

892
00:48:42,080 --> 00:48:44,960
If you really go all the way,
you know, all the way up and,

893
00:48:45,080 --> 00:48:47,520
and you make I think there were
like 6 and newts or something

894
00:48:47,520 --> 00:48:50,160
like that.
You have one gigantic cell that

895
00:48:50,160 --> 00:48:53,520
bends around itself and leaves
and leaves a hole in the middle.

896
00:48:53,720 --> 00:48:56,280
Now, now that's a different
mechanism of cytoskeletal

897
00:48:56,280 --> 00:48:58,040
bending instead of cell to cell
communication.

898
00:48:58,320 --> 00:48:59,720
And just just think about what
it means.

899
00:48:59,720 --> 00:49:01,960
If you're a new coming into the
world, what can you count on?

900
00:49:02,120 --> 00:49:03,880
I mean, you know, you can't
really know what your

901
00:49:03,880 --> 00:49:05,760
environment's going to be, but
you can't even, you can count on

902
00:49:05,760 --> 00:49:07,280
your own parts.
You don't know how many copies

903
00:49:07,280 --> 00:49:08,960
of your genetic material you're
going to have.

904
00:49:09,200 --> 00:49:11,080
You don't know how big your
cells are going to be, how many

905
00:49:11,080 --> 00:49:14,240
cells as, as, as Cook showed you
can, you know, take away cells,

906
00:49:14,240 --> 00:49:16,040
add cells, It doesn't matter.
It'll still rescale.

907
00:49:16,720 --> 00:49:20,520
You have to be able to do the
best you can under whatever

908
00:49:20,520 --> 00:49:23,560
circumstances you find yourself.
And in bioengineering, we see

909
00:49:23,560 --> 00:49:25,200
this all the time.
Life is incredibly

910
00:49:25,200 --> 00:49:27,080
interoperable.
You can do these crazy things as

911
00:49:27,080 --> 00:49:29,800
you know, chimeras and, and and
you know, you can engineer all

912
00:49:29,800 --> 00:49:31,440
kinds of novel materials and
scaffolds.

913
00:49:31,880 --> 00:49:34,800
This is one reason I love so so
we make xenobots, we make

914
00:49:34,800 --> 00:49:37,200
anthrobots, which are these
human cell derived biobots.

915
00:49:37,440 --> 00:49:40,640
They have they have these
incredible features because it

916
00:49:40,640 --> 00:49:44,680
will work like hell to get to
the, the species specific target

917
00:49:44,680 --> 00:49:47,400
morphology that that that's the
default.

918
00:49:47,600 --> 00:49:50,840
But if it can't do that, it will
certainly put together something

919
00:49:50,840 --> 00:49:53,760
else that's coherent and that
has, you know, these anthrobots.

920
00:49:54,840 --> 00:49:57,680
No, no change.
We don't touch the genome.

921
00:49:57,680 --> 00:49:59,480
So exactly normal wild type
genetics.

922
00:49:59,680 --> 00:50:02,960
They have 9000 differently
expressed genes than their

923
00:50:02,960 --> 00:50:05,320
tissue of, of origin.
They're in the same kind of

924
00:50:05,320 --> 00:50:06,520
familiar.
They're in the same, but they

925
00:50:06,520 --> 00:50:08,400
have a different lifestyle.
There are these little silly

926
00:50:08,400 --> 00:50:10,320
things that run around and do
interesting things.

927
00:50:11,000 --> 00:50:13,040
Half their genome is now
expressed differently.

928
00:50:13,840 --> 00:50:16,640
Xenobots express about 900 genes
differently.

929
00:50:16,640 --> 00:50:18,680
They, they express a cluster for
hearing.

930
00:50:18,680 --> 00:50:19,920
And so we've actually tested
that.

931
00:50:19,920 --> 00:50:22,880
And it turns out that that
unlike for unlike frog embryos,

932
00:50:22,880 --> 00:50:25,120
these, these xenobots can
actually respond to sound.

933
00:50:25,440 --> 00:50:27,520
And so they, they have this new,
you know, they have this new

934
00:50:27,520 --> 00:50:29,640
lifestyle.
So, so I think this creative

935
00:50:29,640 --> 00:50:32,760
aspect, the, the constraint is,
it's, it's kind of like a meta

936
00:50:32,760 --> 00:50:34,680
constraint.
The constraint is you can't

937
00:50:34,680 --> 00:50:36,680
count on the past, then you
can't take literally.

938
00:50:36,680 --> 00:50:39,640
You need to take the messages
you get from the past, both your

939
00:50:39,640 --> 00:50:42,800
own memories and the genetic,
you know, experience of your, of

940
00:50:42,800 --> 00:50:44,760
your ancestry.
You take it seriously, but you

941
00:50:44,760 --> 00:50:47,200
don't take it literally.
And as a result of that, because

942
00:50:47,200 --> 00:50:52,200
you can't count on the, the,
the, the, the, the, the details

943
00:50:52,200 --> 00:50:55,760
of the past, you have to commit
to a creative interpretation,

944
00:50:55,880 --> 00:50:58,720
which makes it makes for
incredible plasticity.

945
00:50:58,920 --> 00:51:01,400
It's why these tadpoles that the
that the, that the Doug

946
00:51:01,400 --> 00:51:03,680
Blackiston in my lab made that
have eyes on their tails.

947
00:51:03,840 --> 00:51:06,720
They can see out-of-the-box.
You don't need new the the eye

948
00:51:06,720 --> 00:51:08,720
doesn't connect to the brain.
It connects to the optic nerve

949
00:51:08,920 --> 00:51:10,880
until the the optic nerve
connects to the spinal cord.

950
00:51:11,040 --> 00:51:12,920
No problem.
No new rounds of selection and

951
00:51:12,920 --> 00:51:14,720
and mutation.
No, they just they can, they can

952
00:51:14,720 --> 00:51:16,760
see out-of-the-box.
So all of this, this this

953
00:51:16,760 --> 00:51:21,400
amazing plasticity is I think is
because they could never depend

954
00:51:21,400 --> 00:51:23,440
on doing it the same.
Maybe C elegans can, I don't

955
00:51:23,440 --> 00:51:25,320
know, you know, maybe nematodes
liquid or something.

956
00:51:25,560 --> 00:51:29,120
But but most things are are are
they, they're going to find

957
00:51:29,120 --> 00:51:31,920
themselves in all sorts of
novelty and they never assume it

958
00:51:31,920 --> 00:51:34,240
was going to be the same.
They have to resolve the problem

959
00:51:34,240 --> 00:51:37,000
each time.
And that and that when, when we

960
00:51:37,080 --> 00:51:39,960
actually we see this in the in
the evolutionary simulations

961
00:51:39,960 --> 00:51:43,760
that we do, because what happens
is as you start simulating this

962
00:51:43,760 --> 00:51:47,680
kind of unreliable and yet
competent material, what happens

963
00:51:47,680 --> 00:51:50,600
is.
You, you, you.

964
00:51:50,600 --> 00:51:53,920
Have some, some mutations, but
the material tends to fix it.

965
00:51:54,320 --> 00:51:57,400
And then selection can't see the
genome as well because because

966
00:51:57,400 --> 00:51:59,400
the phenotype doesn't, doesn't
show evidence.

967
00:51:59,440 --> 00:52:01,480
You know you fixed it.
And so when, when selection

968
00:52:01,480 --> 00:52:04,800
can't tell whether your genome
was great or actually your, your

969
00:52:04,800 --> 00:52:07,600
competence was good, all the
work of that evolutionary cycle

970
00:52:07,600 --> 00:52:10,400
starts being put into making
the, the algorithm better.

971
00:52:10,720 --> 00:52:12,840
And the more you do that well,
the harder it is to see the

972
00:52:12,840 --> 00:52:14,360
genome.
And so you've got this feedback

973
00:52:14,360 --> 00:52:15,840
loop.
It's just like positive feedback

974
00:52:15,840 --> 00:52:18,760
loop where over and over again,
what you're getting really good

975
00:52:18,760 --> 00:52:23,040
at is doing something coherent,
even if the starting material

976
00:52:23,040 --> 00:52:25,600
and the starting information is
different, degraded, you know,

977
00:52:25,600 --> 00:52:28,520
unreadable, whatever.
And then and then that's how you

978
00:52:28,520 --> 00:52:31,840
end up with things like Planaria
that have a ridiculously noisy,

979
00:52:32,800 --> 00:52:36,480
you know, genome that's
mixaployed where every cell has

980
00:52:36,480 --> 00:52:37,720
a different number of
chromosomes.

981
00:52:38,000 --> 00:52:40,280
And, and yet they're the ones
that have this like incredible

982
00:52:40,280 --> 00:52:45,040
cancer resistant, no, no aging,
highly regenerative animal, even

983
00:52:45,040 --> 00:52:47,360
though, even though the genome
is, is, is basically junk.

984
00:52:47,880 --> 00:52:51,000
So yeah, I, I, I love the, I
love the, the interpretation of,

985
00:52:51,120 --> 00:52:53,840
of that, that aspect of it, you
know, about how to, how to

986
00:52:53,840 --> 00:52:57,720
interpret that information.
Right.

987
00:52:57,760 --> 00:53:01,080
So I like to think about this
also as the as looking at the

988
00:53:01,080 --> 00:53:04,800
hypothesis generation side of
things that in fact that's

989
00:53:04,800 --> 00:53:08,520
what's going on is that you use
the same old information but in

990
00:53:08,520 --> 00:53:11,360
a different context.
You you generate A hypothesis.

991
00:53:11,560 --> 00:53:15,440
The hypothesis already starts
with bias.

992
00:53:16,040 --> 00:53:19,640
It's like this, this process we
call Bayesian, but the

993
00:53:19,640 --> 00:53:23,600
hypothesis are actually quite
sophisticated in biology because

994
00:53:23,600 --> 00:53:26,680
you've already got a system that
evolved to be confident in

995
00:53:26,680 --> 00:53:30,840
multiple ways that involve
evolved to be able to interpret

996
00:53:31,080 --> 00:53:32,840
under a lots of diverse
conditions.

997
00:53:33,800 --> 00:53:36,400
I used to tell my students when
we're talking about development

998
00:53:36,400 --> 00:53:41,160
of the nervous system, that that
nervous systems always in a

999
00:53:41,160 --> 00:53:44,720
sense adapt to the bodies they
find themselves in during

1000
00:53:44,720 --> 00:53:48,080
development.
You add an extra eye, it adapts

1001
00:53:48,080 --> 00:53:50,000
to it.
You, you move the eyes.

1002
00:53:50,000 --> 00:53:55,000
But you know, in in those of us
who are mammals that have eyes,

1003
00:53:55,600 --> 00:53:58,200
sometimes the side of our head,
sometimes towards the front of

1004
00:53:58,200 --> 00:54:01,400
our head.
It turns out that that the way

1005
00:54:01,400 --> 00:54:05,760
that that system develops in the
in the visual cortex in effect,

1006
00:54:05,760 --> 00:54:10,520
is an adaptive system.
It's taking information, saying

1007
00:54:10,640 --> 00:54:15,440
use that information to guide
how it's wired up, adapt to the

1008
00:54:15,440 --> 00:54:19,520
fact that that No2 individuals
will have eyes in the same

1009
00:54:19,720 --> 00:54:25,440
relative position.
And so the system couldn't work.

1010
00:54:25,440 --> 00:54:30,360
It couldn't evolve if it had to
be fine, fine-tuned by virtue of

1011
00:54:30,560 --> 00:54:32,880
a of a kind of deterministic
wiring.

1012
00:54:33,200 --> 00:54:37,480
It has to be adaptive in order
to work for precisely these

1013
00:54:37,480 --> 00:54:39,880
reasons.
But that's the in a sense, the

1014
00:54:40,000 --> 00:54:42,800
the the decompression side of
things, the decompression side

1015
00:54:42,800 --> 00:54:47,080
of things has to itself be a
kind of evolutionary process,

1016
00:54:47,760 --> 00:54:49,160
the.
Hypothesis.

1017
00:54:49,320 --> 00:54:53,040
Right, in testing hypothesis,
Yeah, something something.

1018
00:54:53,040 --> 00:54:56,920
That, that we, we wanted, we
recently did.

1019
00:54:56,920 --> 00:54:58,800
And so this is this preprint
will go up.

1020
00:54:58,800 --> 00:55:02,240
I think this week is we made
some xenobots and we stuck some

1021
00:55:02,240 --> 00:55:05,160
neural cells in it such that
what, what is the structure of a

1022
00:55:05,160 --> 00:55:08,520
nervous system that doesn't have
a, you know, a selection of

1023
00:55:08,680 --> 00:55:10,560
history of selection for a
specific purpose?

1024
00:55:10,720 --> 00:55:12,680
So they're finding themselves in
a completely new context.

1025
00:55:12,680 --> 00:55:14,960
They're sitting excess cells,
you know, that, that, that that

1026
00:55:14,960 --> 00:55:16,360
shouldn't be there.
What you know, what are the

1027
00:55:16,360 --> 00:55:17,840
nervous systems going to look
like?

1028
00:55:17,840 --> 00:55:20,440
And these things are incredibly
plastic.

1029
00:55:20,480 --> 00:55:24,200
And because of this, because of
the, the material has this

1030
00:55:25,560 --> 00:55:28,720
ability to interpret and
reinterpret the information you

1031
00:55:28,720 --> 00:55:33,760
get, you get these really
interesting competitions between

1032
00:55:33,760 --> 00:55:36,360
interpretations.
So one of my favorite examples

1033
00:55:36,360 --> 00:55:40,920
is in the early embryo, one of
the things that tells the eyes

1034
00:55:40,920 --> 00:55:42,920
where they should be.
There's a particular

1035
00:55:42,920 --> 00:55:45,560
bioelectrical pattern in the
nascent face that has these

1036
00:55:45,560 --> 00:55:47,000
spots and see, this is where the
eye goes.

1037
00:55:47,280 --> 00:55:51,960
So if, if, if you, if you inject
one of the number of channels

1038
00:55:51,960 --> 00:55:54,800
into some other location, you
set up a similar bioelectrical

1039
00:55:54,960 --> 00:55:58,040
pattern, those cells will get
the message and they'll and

1040
00:55:58,040 --> 00:56:00,240
they'll make an eye.
But what's really interesting is

1041
00:56:00,240 --> 00:56:03,400
The thing is that there's
actually a competition because

1042
00:56:04,280 --> 00:56:07,080
if if you if you section those
eyes, what you find is that

1043
00:56:07,280 --> 00:56:09,560
there's a few cells that we
injected and then they

1044
00:56:09,560 --> 00:56:12,280
secondarily recruited a bunch of
their neighbors to participate

1045
00:56:12,280 --> 00:56:13,720
in making that eye.
We didn't touch them.

1046
00:56:13,720 --> 00:56:17,600
It's, it's a, you know, the
cells we did inject broadcast,

1047
00:56:17,640 --> 00:56:20,120
there needs to be an eye here.
And then so, so many cells need

1048
00:56:20,120 --> 00:56:23,040
to sort of cooperate.
But sometimes you do that and

1049
00:56:23,040 --> 00:56:25,840
sometimes you get no eye at all
because what's happening is

1050
00:56:25,840 --> 00:56:28,640
there's a competing
interpretation where there's a,

1051
00:56:28,640 --> 00:56:31,280
which is a cancer suppression
mechanism where the surrounding

1052
00:56:31,280 --> 00:56:33,920
cells are saying your voltage is
wrong and we're going to, we're

1053
00:56:33,920 --> 00:56:36,480
going to equalize you out to us.
They make gap junctions and they

1054
00:56:36,480 --> 00:56:38,120
try to equal.
So you get this, you get this

1055
00:56:38,120 --> 00:56:39,640
battle.
So these cells are saying no,

1056
00:56:39,640 --> 00:56:42,120
we're gut and these other cell
cells are saying no, you should

1057
00:56:42,120 --> 00:56:45,720
be, you should be I.
And it's the degree to which we

1058
00:56:45,720 --> 00:56:48,680
are good or not good at
providing convincing

1059
00:56:49,240 --> 00:56:52,160
interpretations via these, these
voltage gradients that we can

1060
00:56:52,160 --> 00:56:55,480
specify a successful eye or
something that you can even see.

1061
00:56:56,040 --> 00:56:59,720
Sometimes you can use animals
that are transgenic for like

1062
00:56:59,720 --> 00:57:02,280
early eye genes, like Rick's one
or pack six or something.

1063
00:57:02,480 --> 00:57:06,200
And so you inject your, your RNA
and you see like there's 88

1064
00:57:06,200 --> 00:57:07,760
ectopic spots.
You say, oh, we're going to have

1065
00:57:07,760 --> 00:57:10,120
a tadpole with eight eyes.
Well, what you find out is that

1066
00:57:10,120 --> 00:57:12,720
a bunch of them get, get winked
out by their neighbors and, and

1067
00:57:12,720 --> 00:57:14,840
you, and you don't get those.
And then some, if you, you know,

1068
00:57:14,840 --> 00:57:15,920
if you do it right, some of them
will.

1069
00:57:16,360 --> 00:57:19,320
And so, so, so that battle of
interpretations, I, I think is

1070
00:57:19,320 --> 00:57:21,880
also really interesting because
that, that material, it's not a,

1071
00:57:21,960 --> 00:57:24,280
it's not passive matter.
It can support things like this,

1072
00:57:24,280 --> 00:57:26,800
like a, like a, like a debate
between different, different

1073
00:57:26,800 --> 00:57:29,280
anatomical, you know, world
views about what are, what are

1074
00:57:29,280 --> 00:57:34,240
we building here, you know.
Jerry, anything you'd like to

1075
00:57:34,240 --> 00:57:39,400
add to?
That so much but but I'm not

1076
00:57:39,400 --> 00:57:42,760
sure I want to I want to keep
going on this and there's

1077
00:57:43,040 --> 00:57:45,040
there's so many pieces.
One of the, one of the things

1078
00:57:45,040 --> 00:57:51,240
that, you know, when I, when I
think about cognition, there's

1079
00:57:51,240 --> 00:57:54,080
a, a, a piece that I think that
Michael was involved in, in

1080
00:57:54,080 --> 00:57:57,600
terms of, of talking about, you
know, getting the biology back

1081
00:57:57,600 --> 00:58:03,520
into the notion of cognition.
I I have often thought that the

1082
00:58:03,520 --> 00:58:08,280
way that that we misunderstand
brains is that we think of them

1083
00:58:08,840 --> 00:58:12,240
in too much of A computational
like sense.

1084
00:58:12,840 --> 00:58:16,320
And I think that brains are
actually doing what embryos do,

1085
00:58:16,640 --> 00:58:19,240
that brains are doing what
evolution does, that, that we

1086
00:58:19,240 --> 00:58:23,600
want to look at a thought
process as in a sense what

1087
00:58:23,600 --> 00:58:25,760
sometimes described as
microgenesis.

1088
00:58:26,160 --> 00:58:30,320
That is, it's in a sense you're
differentiating something as

1089
00:58:30,320 --> 00:58:33,760
opposed to activating it or
inactivating it.

1090
00:58:34,200 --> 00:58:37,120
And the differentiation process,
we experience it, you know, as,

1091
00:58:37,240 --> 00:58:41,280
you know, new ideas come to
fruition or when we're sitting

1092
00:58:41,280 --> 00:58:44,440
here talking and a new idea pops
up, we differentiate it.

1093
00:58:44,440 --> 00:58:46,840
It starts out pretty
undifferentiated.

1094
00:58:47,480 --> 00:58:51,800
And I, and one of my approaches
to language is to think about in

1095
00:58:51,800 --> 00:58:55,000
effect, you know, what's the
sentence before I've produced

1096
00:58:55,000 --> 00:58:57,040
it?
Well, it's undifferentiated.

1097
00:58:57,440 --> 00:59:00,200
It's not that I go about and I
say, OK, it's a bunch of rules.

1098
00:59:00,360 --> 00:59:04,160
And I want to pick out a bunch
of, you know, rules and

1099
00:59:04,160 --> 00:59:07,120
positions and words.
No, the words don't come first.

1100
00:59:07,120 --> 00:59:09,040
The words are at the very end of
the process.

1101
00:59:09,240 --> 00:59:11,280
The sequence is the very end of
the process.

1102
00:59:11,280 --> 00:59:14,240
And it's, and it's flexible.
Thank goodness that I can, you

1103
00:59:14,240 --> 00:59:16,360
know, move things around and it
still works.

1104
00:59:17,680 --> 00:59:22,720
In effect, it's we think about
cognition too much like we think

1105
00:59:22,720 --> 00:59:26,920
about reasoning, you know, in a
formal sense, when it's in fact,

1106
00:59:27,000 --> 00:59:30,800
you know, biology uses the same
strategy wherever we look at it.

1107
00:59:31,920 --> 00:59:35,440
So one of the ways I like to
study different functional

1108
00:59:35,440 --> 00:59:38,960
regions of the brain is to,
first of all, look at how that

1109
00:59:38,960 --> 00:59:44,280
region developed its circuitry.
Now a lot of it's heavy and in

1110
00:59:44,280 --> 00:59:46,960
the sense that there's a
statistical logic to it.

1111
00:59:47,680 --> 00:59:51,920
But actually if you look at how
different regions utilize the

1112
00:59:51,920 --> 00:59:54,920
information that they're
getting, the the over connection

1113
00:59:54,920 --> 00:59:58,000
that they received and so on,
and the difference in connection

1114
00:59:58,000 --> 01:00:02,200
from different areas and how
they settle it, neurons don't

1115
01:00:02,200 --> 01:00:04,880
change their strategy when they
become mature.

1116
01:00:05,760 --> 01:00:07,960
They're doing the same thing at
all levels.

1117
01:00:08,160 --> 01:00:10,720
So they're, you know, whether
it's a nervous system just

1118
01:00:10,720 --> 01:00:16,560
getting started and starting to
wire itself up as a result.

1119
01:00:16,560 --> 01:00:19,800
The best way from my perspective
to look at what a particular

1120
01:00:19,800 --> 01:00:22,600
region of the brain is doing,
what this class of neurons are

1121
01:00:22,600 --> 01:00:25,360
doing is to ask the question,
how did they get that way in the

1122
01:00:25,360 --> 01:00:28,280
1st place?
You know, what developmental

1123
01:00:28,280 --> 01:00:30,760
process were that was each
neuron involved in?

1124
01:00:30,880 --> 01:00:35,960
How did it amplify and eliminate
certain connection patterns?

1125
01:00:36,200 --> 01:00:40,440
How did it, you know, adapt to
the the diversity of time frames

1126
01:00:40,440 --> 01:00:42,400
that are coming in?
All of those things in

1127
01:00:42,400 --> 01:00:45,200
development determine how its
circuit is developed.

1128
01:00:45,200 --> 01:00:48,080
But I think that's what's going
on, you know, in you and I right

1129
01:00:48,080 --> 01:00:52,240
now that each of these is sort
of a, you might say a micro

1130
01:00:52,240 --> 01:00:56,960
Embryology, a micro evolution,
each thought that we produce.

1131
01:00:57,440 --> 01:01:00,800
But then then what we were
talking about in terms of this

1132
01:01:00,800 --> 01:01:04,760
sort of compression,
decompression, creating memories

1133
01:01:05,120 --> 01:01:09,880
is, is the compression process.
We, you know, we strengthen

1134
01:01:09,880 --> 01:01:15,120
certain synapses and we can
certain others in order that

1135
01:01:15,120 --> 01:01:19,400
when we throw information back
into the system, we put

1136
01:01:19,520 --> 01:01:24,040
metabolism back into the system,
it will bias the dynamical

1137
01:01:24,040 --> 01:01:25,760
pattern.
That's that's the result.

1138
01:01:27,680 --> 01:01:32,080
But what's going on is that the
interpretation or you might say

1139
01:01:32,080 --> 01:01:34,920
the decompression is going on
dynamically.

1140
01:01:35,760 --> 01:01:38,920
So I think of it very much like
I think about the evolution of

1141
01:01:38,920 --> 01:01:43,880
compressed signals in evolution,
the compressing it into DNA,

1142
01:01:44,520 --> 01:01:48,440
creating the what's sometimes
been called the engram is

1143
01:01:48,440 --> 01:01:53,440
basically this highly compressed
difference in synaptic strengths

1144
01:01:53,440 --> 01:01:58,400
over a network that's very, very
different than how we do

1145
01:01:58,400 --> 01:02:00,400
computing.
And yet it's of course

1146
01:02:00,400 --> 01:02:04,240
accomplishing something similar.
But it's doing it in a sense,

1147
01:02:04,240 --> 01:02:10,960
it's biologizing cognition.
And so so I know Michael likes

1148
01:02:10,960 --> 01:02:13,600
to talk about cognition all the
way down.

1149
01:02:13,600 --> 01:02:16,880
I like to talk about biology,
biologizing all the way up.

1150
01:02:20,520 --> 01:02:22,240
Mike, anything you?
Want to say about that?

1151
01:02:24,880 --> 01:02:26,560
Yeah.
I'm, I'm, I'm on board it.

1152
01:02:26,600 --> 01:02:28,120
It makes sense.
These are these are kind of

1153
01:02:28,120 --> 01:02:30,720
symmetrical views that that we
have here.

1154
01:02:31,600 --> 01:02:33,760
Terry is there.
Any part of Mike's work that

1155
01:02:33,760 --> 01:02:37,840
when you first read, you found
particularly different from your

1156
01:02:37,840 --> 01:02:40,560
own and that you questioned and
that you'd like to perhaps ask

1157
01:02:40,560 --> 01:02:42,680
him about?
Wow, good.

1158
01:02:42,920 --> 01:02:45,560
That's a good question.
Because some of the earlier work

1159
01:02:46,280 --> 01:02:48,600
troubled me, and particularly
with planaria.

1160
01:02:49,480 --> 01:02:52,920
Planaria are a very unusual kind
of animal because they're,

1161
01:02:53,160 --> 01:02:57,840
they're, they're weird genetics
and but also because of all of

1162
01:02:57,840 --> 01:03:00,960
these things that we see in many
plants where a part of a plant

1163
01:03:00,960 --> 01:03:04,320
can generate the whole plant.
But we don't see that very often

1164
01:03:04,320 --> 01:03:08,160
in animals, mostly simple
animals, you know, hydra and

1165
01:03:08,160 --> 01:03:10,240
stuff like that have these
capacities.

1166
01:03:10,760 --> 01:03:13,280
But as we get more complex
animals, it becomes harder and

1167
01:03:13,280 --> 01:03:16,640
harder, except for parts, you
know, parts of salamanders,

1168
01:03:16,800 --> 01:03:21,120
parts of some frogs and so on
there that they can do this.

1169
01:03:21,320 --> 01:03:25,200
And of course even some, some of
our parts, but minimally and

1170
01:03:25,200 --> 01:03:28,200
they've become minimal.
But I was worried that

1171
01:03:28,200 --> 01:03:32,640
initially, particularly talking
about the electrical and ionic

1172
01:03:32,640 --> 01:03:36,680
systems, that it initially
sounded a little bit like what I

1173
01:03:36,680 --> 01:03:40,360
would call preformationism.
That has initially sounded like

1174
01:03:40,520 --> 01:03:45,040
something that a man who I never
liked his work, a guy named

1175
01:03:45,040 --> 01:03:48,000
Rupert Sheldrake.
He had this idea that there was

1176
01:03:48,000 --> 01:03:52,360
sort of you had a field for a
final map that everything just

1177
01:03:52,360 --> 01:03:54,880
sort of, you know, if it
disappeared, there's an

1178
01:03:54,880 --> 01:03:57,320
electrical map of how things
were supposed to go.

1179
01:03:59,640 --> 01:04:02,120
At first, when I read this
stuff, I was troubled by that

1180
01:04:02,120 --> 01:04:04,880
and I thought, no, this the pre
formation ISM is not what we're

1181
01:04:04,880 --> 01:04:08,640
talking about here.
But but in a sense reading

1182
01:04:08,640 --> 01:04:11,400
further and seeing how this
stuff has developed in Michelle

1183
01:04:11,400 --> 01:04:16,280
work, it's it's very clear that
it came back towards this sort

1184
01:04:16,280 --> 01:04:20,520
of decompression
reinterpretation logic, that it

1185
01:04:20,520 --> 01:04:23,400
was not a freeform map, that
there was not an electrical

1186
01:04:23,400 --> 01:04:25,880
preform map here.
I of course, in working with the

1187
01:04:25,880 --> 01:04:29,400
development of the nervous
system, initially did not have

1188
01:04:29,400 --> 01:04:32,840
any of the ionic information to
deal with and studying it

1189
01:04:32,840 --> 01:04:35,240
because I was studying, you
know, the nervous systems of

1190
01:04:35,240 --> 01:04:38,120
rats and mice and monkeys and
eventually people.

1191
01:04:39,680 --> 01:04:43,040
And so there I was looking at
sort of macroscopic features

1192
01:04:43,040 --> 01:04:46,200
having to do with, you know, how
axons find their targets and

1193
01:04:46,200 --> 01:04:49,960
things like that, mostly
chemical and, and material

1194
01:04:49,960 --> 01:04:53,520
processes.
And from that it was very clear

1195
01:04:53,520 --> 01:04:58,200
that there was not, although
things found their targets went

1196
01:04:58,200 --> 01:05:01,680
through and, you know,
interpreted signals and in a

1197
01:05:01,680 --> 01:05:04,520
sense developed 3 dimensional
complicated structures.

1198
01:05:04,680 --> 01:05:09,040
There was not an initial map.
There was this compressed system

1199
01:05:09,040 --> 01:05:13,240
of biases that set up, set up
another higher order physical

1200
01:05:13,240 --> 01:05:16,560
process that was interpreting
those biases and decompressing

1201
01:05:16,560 --> 01:05:19,880
it.
But, but in hindsight, when I

1202
01:05:19,880 --> 01:05:23,280
look back at this, this early
work with Planaria, I realized

1203
01:05:23,440 --> 01:05:25,840
that it there wasn't quite this
map.

1204
01:05:25,840 --> 01:05:29,080
And I particularly like the the
later stuff you've done,

1205
01:05:29,080 --> 01:05:31,680
Michael, where you've, you've
cut them in interesting ways and

1206
01:05:31,680 --> 01:05:37,680
produced, you know, sort of star
headed Plenaria and and do 22

1207
01:05:37,680 --> 01:05:41,200
headed Plenaria where you can
see that in fact, it's it's not

1208
01:05:41,200 --> 01:05:46,480
a free form map, but it's a, a
complicated interpretive process

1209
01:05:46,760 --> 01:05:49,320
that you can bias the
interpretation in interesting

1210
01:05:49,320 --> 01:05:51,320
ways by sort of critically doing
it.

1211
01:05:51,320 --> 01:05:55,720
So in this respect, I was wary
of the early work and I've

1212
01:05:55,720 --> 01:05:59,400
recognized that it converts
converges really well with how I

1213
01:05:59,400 --> 01:06:00,560
like to think about these
processes.

1214
01:06:00,560 --> 01:06:05,240
But but I'm particularly curious
about the relationship of

1215
01:06:05,560 --> 01:06:11,760
Lanaria and Hydra to plants and
how they generate from a part

1216
01:06:12,120 --> 01:06:16,360
the whole.
Yeah, yeah, there's a very, very

1217
01:06:16,360 --> 01:06:18,240
good point.
So, so let me let me talk about

1218
01:06:18,240 --> 01:06:20,040
this this this business a little
bit.

1219
01:06:20,040 --> 01:06:23,440
And I'll just preface, preface
it by saying that it is

1220
01:06:23,440 --> 01:06:26,120
absolutely an interpretation and
decoding process.

1221
01:06:26,120 --> 01:06:29,320
I'm not saying that there is a
pre formed map that sits there

1222
01:06:29,320 --> 01:06:33,240
as it is.
However, if you zoom into the

1223
01:06:33,240 --> 01:06:36,160
center of that bow tie, if you
take if you take a slice of it,

1224
01:06:36,360 --> 01:06:41,120
there are time periods where the
bioelectric pattern acts as as

1225
01:06:41,120 --> 01:06:43,120
if it was a map.
And I'll just give you a simple

1226
01:06:43,120 --> 01:06:44,680
example, a couple of simple
examples.

1227
01:06:45,560 --> 01:06:48,600
And it's not just Planaria.
We've done this in, in, in, in,

1228
01:06:48,800 --> 01:06:50,640
in frog, and we're now doing
this in Misa.

1229
01:06:50,640 --> 01:06:52,800
This is, you know, this is, this
is, I think a general thing.

1230
01:06:53,040 --> 01:06:56,480
So, so let me, let me just
describe what, what we were

1231
01:06:56,480 --> 01:07:01,120
looking for in the, the, what,
what, what I was looking for is

1232
01:07:01,760 --> 01:07:05,160
a physical encoding of a set
point for this homeostatic

1233
01:07:05,160 --> 01:07:07,240
process.
So, so the thing about planaria

1234
01:07:07,480 --> 01:07:10,880
and, and, and you know, maybe
sometimes I think that all

1235
01:07:10,880 --> 01:07:13,800
development, like even mammals,
you know, half of us can

1236
01:07:13,800 --> 01:07:16,080
regenerate an entire body from 1
cell, right?

1237
01:07:16,080 --> 01:07:18,680
That's a, that's a kind of
regenerative process.

1238
01:07:18,680 --> 01:07:21,000
You're, you're down to 1 cell
and you can sort of re inflate

1239
01:07:21,000 --> 01:07:24,080
the back into the body, right?
So, and we know that all of

1240
01:07:24,080 --> 01:07:26,320
these systems will work like
hell to get to where they're

1241
01:07:26,320 --> 01:07:27,680
going.
Even if you perturb them, you

1242
01:07:27,680 --> 01:07:29,800
can cut them into pieces, you
can move things around.

1243
01:07:29,800 --> 01:07:32,800
We can, we can make a Picasso,
the frogs where everything's in

1244
01:07:32,800 --> 01:07:35,440
the wrong place and and they'll
still kind of come go back to

1245
01:07:35,440 --> 01:07:38,560
where they need to go.
What I was looking for is the

1246
01:07:38,560 --> 01:07:41,680
encoding of the set point.
Not, not magic, but

1247
01:07:41,680 --> 01:07:44,800
nevertheless, any homeostatic
system has to have somewhere you

1248
01:07:44,800 --> 01:07:47,280
have to store, you have to have
a memory of, of what, what, what

1249
01:07:47,280 --> 01:07:49,320
it is that you're reducing error
against, right.

1250
01:07:49,320 --> 01:07:51,400
There's some sort of error
minimization schemes I was

1251
01:07:51,400 --> 01:07:54,920
looking for.
So it turns out that both in

1252
01:07:54,920 --> 01:08:00,160
Planaria head tail decisions and
in the frog face, if you look

1253
01:08:00,160 --> 01:08:03,120
very early before all the genes
come on that are going to

1254
01:08:03,120 --> 01:08:05,560
regionalize the face and all
that, there's a bi.

1255
01:08:05,560 --> 01:08:07,320
So, so let's just talk about the
frog face for a second.

1256
01:08:07,320 --> 01:08:11,440
There's a bioelectrical pattern
which is readily decodable and

1257
01:08:11,440 --> 01:08:13,560
there are many other patterns
that we are working hard to

1258
01:08:13,560 --> 01:08:17,520
decode and some are very hard.
But the frog face one, it looks

1259
01:08:17,520 --> 01:08:19,560
like a face.
You can tell exactly.

1260
01:08:19,560 --> 01:08:20,960
Here's where the eyes are going
to go.

1261
01:08:20,960 --> 01:08:23,120
Here's where the mouth is going
to go here where the plaque goes

1262
01:08:23,120 --> 01:08:26,080
out to the side going to be.
And, and sure enough, if you

1263
01:08:26,080 --> 01:08:29,120
move any of those voltage
states, not the cells, the

1264
01:08:29,120 --> 01:08:31,200
voltage states, and we can do
that without the genetics and

1265
01:08:31,200 --> 01:08:33,399
with other tools.
If you move any of those things,

1266
01:08:33,560 --> 01:08:35,880
the gene expression follows and
the anatomy follows.

1267
01:08:35,880 --> 01:08:38,279
And so you can and in the
planarian, it's the same thing.

1268
01:08:38,279 --> 01:08:40,840
We if you cut a then you say,
how does this thing know how

1269
01:08:40,840 --> 01:08:41,960
many heads it's supposed to
have?

1270
01:08:42,240 --> 01:08:44,520
There's an actual voltage
pattern that says one head and

1271
01:08:44,520 --> 01:08:46,479
one tail.
And I mean now we know that

1272
01:08:46,479 --> 01:08:49,120
that's what it says.
And if you artificially change

1273
01:08:49,120 --> 01:08:51,960
it to say 2 heads, sure enough,
you get a worm with with two

1274
01:08:51,960 --> 01:08:55,240
heads and and and and
parenthetically, if you then

1275
01:08:55,240 --> 01:08:58,240
start chopping those worms, you
forever more will get 2 headed

1276
01:08:58,240 --> 01:09:00,560
worms, even though the genetics
are still completely wild.

1277
01:09:00,560 --> 01:09:06,840
That so So for the duration of
that homeostatic process that

1278
01:09:06,840 --> 01:09:10,080
says what am I supposed to be
reducing error against?

1279
01:09:10,359 --> 01:09:13,840
There is in fact a pattern that
has a very important feature.

1280
01:09:13,840 --> 01:09:17,640
The feature is I don't have to
worry about like with any good

1281
01:09:17,720 --> 01:09:20,800
homeostat, I don't have to worry
about the hardware of how it

1282
01:09:20,800 --> 01:09:23,000
does it.
If I understand the encoding of

1283
01:09:23,000 --> 01:09:25,359
the set point, I just change the
set point and then I take my

1284
01:09:25,359 --> 01:09:28,640
hands off the wheel and the
system autonomously builds to

1285
01:09:28,640 --> 01:09:32,800
that set point.
So for that period of time, I

1286
01:09:32,800 --> 01:09:37,200
think you are looking at a, a, a
map that isn't some, some

1287
01:09:37,240 --> 01:09:40,800
eternal, you know, static thing
that sits there, but it is the

1288
01:09:40,800 --> 01:09:44,399
set point against which that
that drives error minimization.

1289
01:09:44,760 --> 01:09:49,319
Now that pattern has to, has to,
has to emerge from, from, from

1290
01:09:49,319 --> 01:09:51,520
prior States and then eventually
disappears.

1291
01:09:51,680 --> 01:09:54,480
So it is not a permanent thing
that just sits there, but it's

1292
01:09:54,480 --> 01:09:56,080
sort of that center of the bow
tie.

1293
01:09:56,080 --> 01:09:59,400
It's it's the previous events
have given rise to at this

1294
01:09:59,400 --> 01:10:01,520
moment, we're building the face.
Later, we're going to build

1295
01:10:01,520 --> 01:10:05,520
some, some, some other thing.
It, it gives you the set point

1296
01:10:05,520 --> 01:10:08,400
that guides the behavior.
And that set point is really

1297
01:10:08,400 --> 01:10:11,440
convenient.
It's convenient because if you

1298
01:10:11,440 --> 01:10:14,760
try to stick with, I mean, the
standard story of development

1299
01:10:14,760 --> 01:10:18,880
that, that, that we were all
taught is that it's a feat.

1300
01:10:18,880 --> 01:10:22,040
It's an open loop complexity
emergence process.

1301
01:10:22,040 --> 01:10:24,760
There's a bunch of local rules
that everything goes according

1302
01:10:24,760 --> 01:10:27,120
to the local rules.
And eventually it's like voila,

1303
01:10:27,160 --> 01:10:29,440
you know, emergence and, and you
get this complicated thing.

1304
01:10:29,720 --> 01:10:33,160
And yes, that can happen.
And we know like a cellular

1305
01:10:33,160 --> 01:10:35,360
automata and other things, you
will always, you can get complex

1306
01:10:35,360 --> 01:10:38,360
things from simple rules, but
that's not what this process is,

1307
01:10:38,400 --> 01:10:40,440
is actually doing.
If you start to interfere with

1308
01:10:40,440 --> 01:10:42,760
it, you will see that it works
actually really hard.

1309
01:10:42,760 --> 01:10:46,160
It's not an open loop process.
It, it tries really hard to get

1310
01:10:46,280 --> 01:10:50,040
to a particular outcome.
So that means and, and, and the

1311
01:10:50,040 --> 01:10:52,160
thing that's really hard about
the standard story is that if

1312
01:10:52,160 --> 01:10:54,280
you wanted to do, let's say
regenerative medicine and you

1313
01:10:54,280 --> 01:10:56,840
wanted to I, I, you know, this
is the, this is the wrong

1314
01:10:56,840 --> 01:10:57,920
outcome.
I want a different outcome.

1315
01:10:58,200 --> 01:11:00,400
Or like what, what Terry was
just saying about about

1316
01:11:00,400 --> 01:11:01,960
evolution, like going backwards,
right?

1317
01:11:01,960 --> 01:11:03,760
And, and, and figuring out how
to encode it.

1318
01:11:04,200 --> 01:11:07,200
It becomes really hard if all
you have is this feed forward

1319
01:11:07,200 --> 01:11:10,480
that like that, that inverse
problem is, is is not solvable

1320
01:11:10,480 --> 01:11:12,280
in general.
But if you have a component

1321
01:11:12,280 --> 01:11:15,840
where what I have is a self
organizing pattern, it's self

1322
01:11:15,840 --> 01:11:18,560
organizing because of some math
and also because of the kinds of

1323
01:11:18,560 --> 01:11:20,560
ion channels that evolution has
provided us with.

1324
01:11:20,720 --> 01:11:22,920
So it's an excitable medium.
And then of course also, you

1325
01:11:22,920 --> 01:11:25,200
know, Turing patterns of
chemical signals and bio

1326
01:11:25,200 --> 01:11:29,280
mechanics and everything else,
the math will give us a, a, a

1327
01:11:29,280 --> 01:11:32,440
set point that for a period of
time will allow us to get to a

1328
01:11:32,440 --> 01:11:36,120
particular point.
And serves as a really tractable

1329
01:11:36,120 --> 01:11:39,200
control knob for when you want
to make 2 headed worms or fixed

1330
01:11:39,200 --> 01:11:40,760
birth defects.
So that's one of the things we

1331
01:11:40,920 --> 01:11:45,280
we, we, we work out now is a
very simple stimuli that that

1332
01:11:45,280 --> 01:11:48,880
that repair very complex defect.
For example, notch mutations,

1333
01:11:48,880 --> 01:11:50,640
right?
If you mutate notch in the in

1334
01:11:50,640 --> 01:11:52,160
the brain is just completely
wrecked.

1335
01:11:52,880 --> 01:11:54,960
But the first thing that gets
wrecked from that is the

1336
01:11:54,960 --> 01:11:57,200
bioelectrical pattern that
dictates the shape and size of

1337
01:11:57,200 --> 01:11:59,160
the brain.
And if you force the correct

1338
01:11:59,160 --> 01:12:02,160
pattern, you can have perfectly
normal tadpoles with behavior

1339
01:12:02,160 --> 01:12:04,440
indistinguishable from control,
like learning, they have

1340
01:12:04,440 --> 01:12:07,160
learning and distinguishable
from controls, even though that

1341
01:12:07,160 --> 01:12:10,080
dominant notch overactive notch
mutant is, is still there,

1342
01:12:10,280 --> 01:12:12,440
right?
And so, so for a period of time,

1343
01:12:12,800 --> 01:12:16,080
you get access to that set point
that, that is, is very

1344
01:12:16,080 --> 01:12:18,120
convenient.
But of course, it's not just,

1345
01:12:18,160 --> 01:12:19,280
you know, sitting there the
whole time.

1346
01:12:19,280 --> 01:12:22,080
It comes and it goes and, and
very much it's this inflation

1347
01:12:22,080 --> 01:12:24,440
and deflation process.
And what you're looking at is

1348
01:12:24,440 --> 01:12:27,520
that, that, that bow tie, no,
that snapshot of that of that

1349
01:12:27,520 --> 01:12:30,840
process.
Yeah, I like, I like that

1350
01:12:30,840 --> 01:12:34,400
because and this is that that's
the differentiation logic that

1351
01:12:34,400 --> 01:12:37,640
is you, you start from, you
know, I'm thinking about the

1352
01:12:37,640 --> 01:12:40,400
development of of vertebrate
embryos.

1353
01:12:40,680 --> 01:12:43,960
You know, you start with a head
tailed dorsal ventral

1354
01:12:44,320 --> 01:12:48,560
distinction.
Then once that's set, now you

1355
01:12:48,560 --> 01:12:53,680
can partition it into segments.
Once those segments are set, you

1356
01:12:53,680 --> 01:12:58,480
can now take segment A and
resegment recut, cut it in a

1357
01:12:58,480 --> 01:13:03,400
different way, whether it's in
gene expression in terms of

1358
01:13:03,720 --> 01:13:05,960
polarities of things and so on
and so forth.

1359
01:13:06,280 --> 01:13:08,120
But it's, it allows
differentiation.

1360
01:13:08,120 --> 01:13:12,480
Differentiation always involves
this sort of sequence of steps,

1361
01:13:12,480 --> 01:13:15,240
like you say, these, these sort
of the center of the bow tie

1362
01:13:15,240 --> 01:13:20,280
where or basically it's got a,
it's got a point where now this

1363
01:13:20,280 --> 01:13:22,400
has been accomplished, I can
move on to the next

1364
01:13:22,560 --> 01:13:25,800
differentiation, shut down the
old system.

1365
01:13:26,000 --> 01:13:30,640
And now once it's set there and
I've got all of these divisions,

1366
01:13:30,640 --> 01:13:33,680
dorsal ventral, for example,
division set, I can now do

1367
01:13:33,680 --> 01:13:36,440
something else.
But it's a differentiation

1368
01:13:36,440 --> 01:13:39,720
process that begins
undifferentiated and then

1369
01:13:39,720 --> 01:13:43,120
becomes progressively step by
step differentiated.

1370
01:13:43,320 --> 01:13:46,840
And each of those is a decision
like like you're talking about

1371
01:13:46,840 --> 01:13:50,520
it sort of a note of decision.
Now I can do this because that's

1372
01:13:50,520 --> 01:13:53,880
been accomplished, but this is
also the case that then if you

1373
01:13:53,880 --> 01:13:57,960
mess up one of the early ones,
the later ones have to do

1374
01:13:57,960 --> 01:14:02,000
compensation to deal with what
was modified in the first.

1375
01:14:02,000 --> 01:14:05,440
But as a result, you can get
because the differentiation

1376
01:14:05,440 --> 01:14:08,320
process as you, you're
describing it sort of like a

1377
01:14:08,360 --> 01:14:12,800
telescoping process that that
that amplifies something, it can

1378
01:14:12,800 --> 01:14:15,760
amplify something in a very
different direction as well.

1379
01:14:15,960 --> 01:14:17,640
And I think that's one of the
other things that we're finding

1380
01:14:17,640 --> 01:14:22,240
interesting about all of these
processes because this is that

1381
01:14:22,240 --> 01:14:25,840
generative side.
This is why I think it's been so

1382
01:14:25,840 --> 01:14:29,760
exciting in evolutionary theory
to begin to focus on the

1383
01:14:29,760 --> 01:14:33,840
generative side of things.
It was originally described as

1384
01:14:33,840 --> 01:14:36,640
evo devo, but I think it's a,
it's a much more subtle

1385
01:14:36,920 --> 01:14:41,240
distinction than that.
Evolution is taking advantage of

1386
01:14:41,520 --> 01:14:44,200
both the compression,
decompression side of things and

1387
01:14:44,360 --> 01:14:47,560
how the differentiation can sort
of be built upon prior

1388
01:14:47,560 --> 01:14:50,920
differentiation.
I can start from like a single

1389
01:14:50,920 --> 01:14:54,840
cell and develop a body.
You, you can start with a, you

1390
01:14:54,840 --> 01:14:58,440
know, that's why I like to say
that, you know, in thinking I

1391
01:14:58,440 --> 01:15:01,120
have what's a thought before
I've differentiated it.

1392
01:15:01,440 --> 01:15:04,280
You know, it's, it's a sort of,
I sort of know where it's going,

1393
01:15:05,080 --> 01:15:07,000
but it's not done yet.
And like this sentence I'm

1394
01:15:07,000 --> 01:15:09,400
producing right now, you know,
I, I sort of know where it's

1395
01:15:09,400 --> 01:15:11,320
going to go.
You sort of know where it's

1396
01:15:11,320 --> 01:15:14,240
going to go, but it's, it's
undifferentiated.

1397
01:15:14,440 --> 01:15:17,400
And the words just sort of pop
in at the right time.

1398
01:15:17,400 --> 01:15:22,360
Thank goodness, Mike.
Is there anything, any part of

1399
01:15:22,360 --> 01:15:24,720
Terry's work that you can
articulate that you found

1400
01:15:24,720 --> 01:15:28,560
particularly fascinating but yet
also disagreed with and think

1401
01:15:28,560 --> 01:15:29,920
you could work on and build
upon?

1402
01:15:29,920 --> 01:15:32,320
I.
Didn't I?

1403
01:15:32,320 --> 01:15:33,560
Didn't.
Disagree with anything?

1404
01:15:33,560 --> 01:15:36,160
The only thing if I so I've been
sitting here trying to remember

1405
01:15:36,160 --> 01:15:40,440
what the good example of it.
The, the only thing I guess we

1406
01:15:40,440 --> 01:15:43,360
could talk about somewhere, and
I don't remember where it was,

1407
01:15:44,440 --> 01:15:48,480
you had talked about teleonomy
as a kind of watered down

1408
01:15:48,480 --> 01:15:52,080
version of, of teleology, you
know, and right.

1409
01:15:52,560 --> 01:15:54,040
That's, I'm remembering that
correctly.

1410
01:15:54,280 --> 01:15:57,480
So, so the only thing, the only
thing I'll I can say to that.

1411
01:15:57,480 --> 01:16:00,560
And, and I, I think I take
teleology very, very seriously.

1412
01:16:00,560 --> 01:16:02,880
I think, I think it's, it's
absolutely a real thing.

1413
01:16:03,760 --> 01:16:06,160
I use teleonomy slightly
differently.

1414
01:16:06,440 --> 01:16:10,640
I don't use it as a way to, to,
to sort of placate the, the

1415
01:16:10,640 --> 01:16:12,440
folks who don't like actual
teleology.

1416
01:16:12,440 --> 01:16:14,240
Like I'm full on teleology, I
think.

1417
01:16:14,320 --> 01:16:17,000
I think what's what's, what's
perhaps useful.

1418
01:16:17,000 --> 01:16:19,120
And I don't know if this is one
of these cases where I should I

1419
01:16:19,120 --> 01:16:21,680
should use a different word or
whether we should rehabilitate

1420
01:16:21,680 --> 01:16:24,760
the existing word.
But what I like about teleonomy

1421
01:16:24,760 --> 01:16:29,200
is that it reminds us that what
we observe is from the

1422
01:16:29,200 --> 01:16:32,360
perspective of an observer.
So when people say it's apparent

1423
01:16:32,360 --> 01:16:36,720
teleology, I don't mean it's a
lesser version or not real or

1424
01:16:36,720 --> 01:16:39,040
some kind of like, you know, not
the real thing.

1425
01:16:39,280 --> 01:16:42,120
What I mean is what I think it
does for us, which I, which I

1426
01:16:42,120 --> 01:16:45,680
think is important is to remind
us of, of this basically of this

1427
01:16:45,680 --> 01:16:48,400
kind of intentional stance idea
that it's not that it's not

1428
01:16:48,400 --> 01:16:51,720
real, it's that you have to be
an observer who's capable of, of

1429
01:16:51,720 --> 01:16:54,280
seeing it.
And finding teleology and things

1430
01:16:54,280 --> 01:16:57,240
is an IQ test for all of us
because sometimes you don't see

1431
01:16:57,240 --> 01:16:58,320
it.
It's not because it's not there,

1432
01:16:58,320 --> 01:17:00,240
it's because we didn't know how
to look or what space we were

1433
01:17:00,240 --> 01:17:01,720
looking at or what the system
was doing.

1434
01:17:02,160 --> 01:17:06,760
And, and, and, and again, the
system of a good agent will have

1435
01:17:06,760 --> 01:17:09,120
its own perspective.
It doesn't need our perspective.

1436
01:17:09,120 --> 01:17:14,040
But nevertheless, we, when all
of these things like, I think I,

1437
01:17:14,120 --> 01:17:16,880
I tend to think of a lot of
these some cognitive terms that

1438
01:17:16,880 --> 01:17:20,840
people use as kind of as, as
basically interaction protocols.

1439
01:17:21,080 --> 01:17:24,280
What we're really saying is not
that there's some objective, but

1440
01:17:24,280 --> 01:17:27,240
unique perspective and we should
fight about which, which one is

1441
01:17:27,240 --> 01:17:29,080
right.
But but what you're, what we're

1442
01:17:29,080 --> 01:17:33,880
saying is this is the formalism
that I see that I claim is going

1443
01:17:33,880 --> 01:17:36,520
to be applicable to the system.
And maybe it's some sort of high

1444
01:17:36,520 --> 01:17:39,560
level planning, or maybe it's a
simple, you know, homeostasis or

1445
01:17:39,560 --> 01:17:42,320
something in between.
And, and then I'm going to

1446
01:17:42,320 --> 01:17:44,400
interact with that system using
this as a guide.

1447
01:17:44,400 --> 01:17:45,920
And we'll all find out how well
I did.

1448
01:17:45,920 --> 01:17:48,360
And, and, and if you have a
different estimate, we'll find

1449
01:17:48,360 --> 01:17:50,880
out how well you did.
So, so that's kind of the only

1450
01:17:50,880 --> 01:17:53,680
thing that I found to, you know,
even quibble about.

1451
01:17:53,680 --> 01:17:57,160
It's just that I think teleonomy
has another use besides trying

1452
01:17:57,160 --> 01:17:59,880
to water down teleology.
It's just to just to remind us

1453
01:17:59,880 --> 01:18:02,640
of this observer relevant to the
relativity of some of these

1454
01:18:02,640 --> 01:18:05,480
things.
That's an interesting way of

1455
01:18:05,480 --> 01:18:10,440
putting it because my my work
has actually gone in a different

1456
01:18:10,440 --> 01:18:13,880
direction in that respect.
So, so thinking about the

1457
01:18:13,880 --> 01:18:17,320
history of the term teleonomy,
of course it is an observer

1458
01:18:17,320 --> 01:18:19,600
perspective.
And the idea says that, you

1459
01:18:19,600 --> 01:18:25,400
know, they go back to Bigelow
and Wiener in that group and and

1460
01:18:25,480 --> 01:18:29,000
fit and rich when he comes up
with a term, basically, he said,

1461
01:18:29,040 --> 01:18:32,360
you know, we don't know.
And at the time it was in a

1462
01:18:32,360 --> 01:18:35,000
sense, even a stronger claim
that there is no such thing as

1463
01:18:35,000 --> 01:18:40,840
real teleology, that that it's
just sort of feedback and and

1464
01:18:40,840 --> 01:18:43,000
deviation minimization
processes.

1465
01:18:43,800 --> 01:18:50,040
And and that rather than make a
commitment to some metaphysical

1466
01:18:50,040 --> 01:18:54,120
concept like teleology, the idea
was that, you know, come up with

1467
01:18:54,120 --> 01:18:58,360
a description.
Teleonomy is describing

1468
01:18:58,360 --> 01:19:05,040
processes that are and directed
accomplishing the same end from

1469
01:19:05,040 --> 01:19:08,400
different origins, you know, all
the ways that people used it.

1470
01:19:09,120 --> 01:19:12,720
Ernst Meyer did an interesting
job of saying, but, you know,

1471
01:19:12,720 --> 01:19:15,360
there's different kinds of end
directedness in nature that

1472
01:19:15,360 --> 01:19:18,040
maybe we should distinguish.
So he came up with the idea of

1473
01:19:18,160 --> 01:19:24,240
teleometry, teleometric
processes and he said the second

1474
01:19:24,240 --> 01:19:28,360
law of thermodynamics, it's a
teleometric process in a sense

1475
01:19:28,360 --> 01:19:32,640
that it it, it goes towards an
end, it has end directedness, it

1476
01:19:32,640 --> 01:19:35,520
looks like it's indirectedness.
It's not trying to get there.

1477
01:19:35,640 --> 01:19:37,520
Things just tend to go in that
direction.

1478
01:19:38,520 --> 01:19:43,680
And he said that, but telianomic
processes again descriptive.

1479
01:19:43,680 --> 01:19:48,520
And This is why your perspective
way of thinking about it is

1480
01:19:48,520 --> 01:19:50,240
appropriate.
There's from the outside

1481
01:19:50,240 --> 01:19:53,440
looking, looking at something,
it seems to be going in a

1482
01:19:53,440 --> 01:19:55,160
direction, but it's a different
kind of direction.

1483
01:19:55,160 --> 01:19:58,760
He said this is a direction
that's that's error minimizing

1484
01:19:58,760 --> 01:20:02,120
or deviation minimizing and that
we can see all that.

1485
01:20:02,120 --> 01:20:05,400
And he said, but you know it, it
may be.

1486
01:20:05,400 --> 01:20:08,240
And he he was not willing to
make a commitment that there are

1487
01:20:08,240 --> 01:20:12,400
some things that actually are
not just not just deviation

1488
01:20:12,400 --> 01:20:16,440
minimizing, they use deviation
minimizing, but they also.

1489
01:20:17,400 --> 01:20:21,880
Have goals and ends.
And the question is that when we

1490
01:20:21,880 --> 01:20:27,480
design a system like a like a
thermostat or a heat seeking

1491
01:20:27,480 --> 01:20:31,000
missile or something like that,
the the end is divide is

1492
01:20:31,000 --> 01:20:34,720
designed into the mechanism.
But of course, with life, we're

1493
01:20:34,720 --> 01:20:38,920
dealing with systems that that
create their own ends, design

1494
01:20:38,920 --> 01:20:42,480
their own ends, design their own
thermostats, you know, sort of

1495
01:20:42,480 --> 01:20:46,360
keeping my body warm.
You know, my, my body, the

1496
01:20:46,360 --> 01:20:51,240
process is very homeostatic in
the sense that it's, you know,

1497
01:20:52,040 --> 01:20:54,840
it's constructed like a really
complicated thermostat system.

1498
01:20:55,200 --> 01:20:57,760
Now it's, it's, you know, people
who could even describe it.

1499
01:20:58,440 --> 01:21:01,800
It's not accurate, but roughly
mechanistically in the trivial

1500
01:21:01,800 --> 01:21:05,360
sense.
But but the fact that my body

1501
01:21:05,600 --> 01:21:10,640
has evolved to have that
mechanism to set that value is

1502
01:21:10,640 --> 01:21:14,760
is like me building the
thermostat to have a certain

1503
01:21:14,760 --> 01:21:19,240
behavior, to have a certain end.
The question is where does that

1504
01:21:19,240 --> 01:21:21,320
come from?
And so this is one of the

1505
01:21:21,320 --> 01:21:24,560
reasons why I think we are
really forced to go back and

1506
01:21:24,560 --> 01:21:31,920
talk about information about
things, because if teleology or

1507
01:21:31,920 --> 01:21:37,360
teleonomy or teleometric, all of
those refer to ends.

1508
01:21:37,640 --> 01:21:40,440
But the question is, how do ends
get established?

1509
01:21:41,080 --> 01:21:44,120
How do they come into the world?
That's been the problem I've

1510
01:21:44,120 --> 01:21:46,720
been struggling with.
And the answer is that you no

1511
01:21:46,720 --> 01:21:49,680
longer can do this by the
observer perspective.

1512
01:21:50,520 --> 01:21:53,880
You want to ask the question,
what kind of a system generates

1513
01:21:53,880 --> 01:21:56,320
them?
What's the architecture of a

1514
01:21:56,320 --> 01:21:58,760
system that generates its own
ends?

1515
01:21:59,680 --> 01:22:02,760
And so that's been how I've
approached it.

1516
01:22:02,760 --> 01:22:06,280
And as a result, I've, I've in a
sense, as I said early on,

1517
01:22:06,560 --> 01:22:09,680
trying to do this
thermodynamically, talking about

1518
01:22:09,680 --> 01:22:13,360
how you move from systems that
have the kind of end, the

1519
01:22:13,360 --> 01:22:17,160
teleometric end that the second
law of thermodynamics suggests.

1520
01:22:17,360 --> 01:22:23,560
And then the kind of stable
dynamical end that say Benard

1521
01:22:23,560 --> 01:22:28,160
convection or, or various self
organizing processes have.

1522
01:22:28,440 --> 01:22:33,480
And then the kind of end in
which the, the end is not just a

1523
01:22:33,480 --> 01:22:37,880
pattern, but it's, it's a
representation in which there

1524
01:22:37,880 --> 01:22:40,920
can be a representation to that
end, for example, in a

1525
01:22:40,920 --> 01:22:44,800
compression like like DNA or, or
whatever else.

1526
01:22:45,600 --> 01:22:47,480
That's that's what evolution is
about.

1527
01:22:47,480 --> 01:22:49,080
Evolution is about generating
that.

1528
01:22:49,280 --> 01:22:53,360
But in that case, what I always
found troubling about the

1529
01:22:53,360 --> 01:22:58,160
telonomic perspective is it
always stayed at the observer

1530
01:22:58,160 --> 01:23:00,880
perspective.
I'm outside of the system and I

1531
01:23:00,880 --> 01:23:05,440
would describe its behavior, but
not actually explaining how it

1532
01:23:05,440 --> 01:23:08,520
generates its behavior.
And so I basically want to

1533
01:23:08,520 --> 01:23:10,920
approach this in the other way
around.

1534
01:23:10,920 --> 01:23:14,600
Ask what kind of a system has
these behaviors and generate

1535
01:23:14,600 --> 01:23:17,760
something that I would feel
comfortable of saying it's

1536
01:23:18,400 --> 01:23:22,400
intrinsically about something
and that about this is something

1537
01:23:22,400 --> 01:23:25,400
that it's not yet that could
happen, that might happen, that

1538
01:23:25,400 --> 01:23:28,320
might be the result of repairing
damage and so on.

1539
01:23:28,320 --> 01:23:31,160
So that's, that's, that's
basically how I've approached it

1540
01:23:32,120 --> 01:23:34,640
and and one of the reasons why I
had a problem with the term

1541
01:23:34,640 --> 01:23:38,000
teleonomy over the years, yeah.
Yeah.

1542
01:23:38,000 --> 01:23:40,880
No, I, I agree with all of that
because because the interesting

1543
01:23:40,880 --> 01:23:45,360
systems have their own inner
perspective on on the goals that

1544
01:23:45,360 --> 01:23:48,800
they have, which is not the same
as as what we see from third

1545
01:23:48,800 --> 01:23:53,120
person observation.
Yeah, so, so, so a few of us,

1546
01:23:53,200 --> 01:23:56,240
Josh Bond guard and I and and
and and Chris Fields and some

1547
01:23:56,240 --> 01:23:58,840
other people have been, have
been working on very minimal

1548
01:23:58,840 --> 01:24:02,200
systems to see what it takes to
evolve exactly what you just

1549
01:24:02,200 --> 01:24:04,040
said.
So observe, so observers that

1550
01:24:04,920 --> 01:24:08,920
where where their various
components are about something,

1551
01:24:08,920 --> 01:24:10,520
you know, that that that have
that reference.

1552
01:24:10,960 --> 01:24:14,480
And one of the things that we're
finding is that, and this, this

1553
01:24:14,480 --> 01:24:18,160
gets to, to the distinction that
you just made around, you know,

1554
01:24:18,160 --> 01:24:20,680
there are goals that you give a
system when you create the

1555
01:24:20,680 --> 01:24:22,800
thermostat and then there are
sort of intrinsic goals that

1556
01:24:22,840 --> 01:24:24,000
evolve and, and things like
that.

1557
01:24:24,720 --> 01:24:27,160
One of the things that we're
finding and, and there's not

1558
01:24:27,160 --> 01:24:30,200
time to go into all the details,
but I'll send you some of this

1559
01:24:30,200 --> 01:24:34,680
stuff after we're finding that
there are very minimal systems,

1560
01:24:34,720 --> 01:24:38,000
very minimal systems to in fact,
deterministic minimal systems

1561
01:24:38,320 --> 01:24:42,960
where there is, there is no
magic in the sense that there is

1562
01:24:42,960 --> 01:24:45,560
no, you know, in biology,
there's always some mechanisms

1563
01:24:45,560 --> 01:24:47,600
you just haven't found yet.
Or you can say, well, there must

1564
01:24:47,600 --> 01:24:48,760
be something that you just
haven't.

1565
01:24:48,960 --> 01:24:51,720
So these things, there's that
we, we can simulate them, but we

1566
01:24:51,720 --> 01:24:54,080
on the computer, we know exactly
what all the ingredients are.

1567
01:24:54,680 --> 01:24:59,360
And what we see is that there
are the things that it's doing

1568
01:24:59,360 --> 01:25:02,680
because you made it do that,
because the algorithm says to do

1569
01:25:02,680 --> 01:25:05,160
that.
But then there are also these

1570
01:25:05,160 --> 01:25:08,680
weird side quests that The thing
is doing that are nowhere in the

1571
01:25:08,680 --> 01:25:11,880
algorithm.
And these additional sort of

1572
01:25:11,880 --> 01:25:15,880
intrinsic these, these weird,
some of them, some of them are

1573
01:25:16,480 --> 01:25:18,760
what you would immediately
recognize as competencies,

1574
01:25:18,760 --> 01:25:23,040
things like goal delay,
gratification and things like

1575
01:25:23,040 --> 01:25:25,360
that.
But but others are just are just

1576
01:25:25,360 --> 01:25:29,360
new sort of new goals that the
thing we see obviously pursues

1577
01:25:29,600 --> 01:25:32,000
that are not specified in the
algorithm at all.

1578
01:25:32,280 --> 01:25:37,760
So, yeah, I, I, I'm starting to
think that that ability to, to

1579
01:25:37,760 --> 01:25:41,600
have its own goals that are not
nailed down by the mechanisms or

1580
01:25:41,600 --> 01:25:44,280
the algorithms or the materials
or all the stuff that we're used

1581
01:25:44,280 --> 01:25:47,800
to is really pretty baked in to
the world.

1582
01:25:47,800 --> 01:25:50,320
And I mean, I think evolution
does an amazing job at scaling

1583
01:25:50,320 --> 01:25:52,280
that up to the point where we
start, you know, where it

1584
01:25:52,280 --> 01:25:54,560
becomes obvious to us and that
everybody can see that, that

1585
01:25:54,560 --> 01:25:55,840
this is what this thing is
doing.

1586
01:25:55,840 --> 01:25:58,520
Or or even to the point where he
can self report like us.

1587
01:26:00,000 --> 01:26:04,120
But my at this point, my view is
that that's actually I and in

1588
01:26:04,120 --> 01:26:06,120
fact, Josh and I are writing a
paper called It Doesn't Take

1589
01:26:06,120 --> 01:26:08,520
Much, which is basically all
about these very basic, like

1590
01:26:08,520 --> 01:26:12,080
very minimal systems that are
already doing things that nobody

1591
01:26:12,160 --> 01:26:15,440
told it to do.
So yeah, I think, I think it it

1592
01:26:15,520 --> 01:26:18,040
it permeates the the world that
we live in.

1593
01:26:18,520 --> 01:26:22,000
Well, Jed's it's.
We've got a hard cut off soon,

1594
01:26:22,080 --> 01:26:24,000
so yeah, that's.
Right.

1595
01:26:24,640 --> 01:26:25,800
Got a 4/4.
Minutes left.

1596
01:26:25,800 --> 01:26:28,280
I just want to say this has been
amazing.

1597
01:26:28,280 --> 01:26:31,960
It's been it's been wonderful to
be a family on the wall watching

1598
01:26:31,960 --> 01:26:34,040
the two of you.
I'm surprised it's taking both

1599
01:26:34,040 --> 01:26:35,520
of you so long to have a
conversation.

1600
01:26:35,520 --> 01:26:36,920
There's any final words from the
two of you?

1601
01:26:36,920 --> 01:26:40,640
We've only got about 3 minutes
left so it's a quick one please.

1602
01:26:40,680 --> 01:26:46,360
Terry wow, there's, there's,
there's so much to deal with.

1603
01:26:46,360 --> 01:26:51,080
And I would say that one of the
challenges, and maybe it's a

1604
01:26:51,080 --> 01:26:53,800
difference in the way you
approach this and, and partly

1605
01:26:53,800 --> 01:26:58,200
it's maybe the result of of
Michael sort of coming from the

1606
01:26:58,200 --> 01:27:01,400
computational world and, and me
coming from this sort of Percy

1607
01:27:01,400 --> 01:27:06,160
and semiotic realm.
But I'm really interested in

1608
01:27:06,160 --> 01:27:11,320
this notion of self and, and
this notion that I tried what I

1609
01:27:11,320 --> 01:27:15,480
call teleo dynamic, a dynamic
that actually has a represented

1610
01:27:15,480 --> 01:27:20,360
end by, by represented, I mean
literally there's a, in a sense

1611
01:27:20,360 --> 01:27:24,080
of compression that that will
tell you whether you've received

1612
01:27:24,080 --> 01:27:26,560
it or not.
And that is not externally

1613
01:27:26,560 --> 01:27:29,640
generated, but that that I think
that requires self.

1614
01:27:30,120 --> 01:27:32,960
And so I would say that the
distinction between teleonomy

1615
01:27:33,200 --> 01:27:38,760
and, and, and teleology in that
respect is that that teleology

1616
01:27:38,760 --> 01:27:41,880
is semiotic.
It has, there's a notion of

1617
01:27:41,880 --> 01:27:44,240
representation.
And the problem for philosophy

1618
01:27:44,520 --> 01:27:48,520
is that these are the most
troubling concepts as, as

1619
01:27:48,520 --> 01:27:51,520
positive that we can think of,
you know, purpose and

1620
01:27:51,520 --> 01:27:59,760
directedness, self sentience,
benefit, cost, these are things

1621
01:27:59,760 --> 01:28:02,800
that are just terrible for
philosophical discussion.

1622
01:28:02,800 --> 01:28:07,000
But in effect, we have now sort
of crossed into domains where

1623
01:28:07,000 --> 01:28:12,000
we're beginning to have tools to
ask those questions, something

1624
01:28:12,000 --> 01:28:15,360
we haven't talked about.
I think we've got close.

1625
01:28:15,360 --> 01:28:17,840
To the hard cut of if OK, but
the.

1626
01:28:17,840 --> 01:28:20,200
Last thing I want to say is that
we've never talked about it at

1627
01:28:20,200 --> 01:28:22,480
all, but I think one of things
that's also bringing us these

1628
01:28:22,480 --> 01:28:25,320
questions is these changes in
so-called artificial

1629
01:28:25,320 --> 01:28:28,680
intelligence, they're forcing us
to re ask these questions.

1630
01:28:28,680 --> 01:28:30,600
So we'll touch on all of that.
We'll have we'll try and have

1631
01:28:30,600 --> 01:28:32,440
around too soon.
Mike, any final words from your

1632
01:28:32,520 --> 01:28:33,440
side?
Sorry.

1633
01:28:34,480 --> 01:28:37,960
Yeah, just to say thank you.
To me, there's nothing more

1634
01:28:37,960 --> 01:28:40,640
fascinating than this symmetry
between between the self

1635
01:28:40,640 --> 01:28:43,080
construction of minds and the
self construction of bodies.

1636
01:28:43,080 --> 01:28:45,480
I think Turing already sort of
saw that when he wrote that

1637
01:28:45,480 --> 01:28:48,200
paper on, you know, on chemical
organization, that this was,

1638
01:28:48,400 --> 01:28:50,240
this was two sides of the same
coin.

1639
01:28:50,240 --> 01:28:52,960
And yeah, I'm, I'm delighted
that that we're working on it

1640
01:28:53,120 --> 01:28:55,280
from different angles.
And yeah, happy to discuss

1641
01:28:55,280 --> 01:28:56,960
again.
Thanks so much and, and thanks.

1642
01:28:56,960 --> 01:28:58,440
Thanks.
Thanks, Michael.

1643
01:28:59,080 --> 01:29:00,200
And thanks, it was.
Great.

1644
01:29:00,560 --> 01:29:02,200
Really appreciate having you go.
Up both on.

1645
01:29:02,200 --> 01:29:05,040
But yeah, Michael, we know you
have a lecture to give, so I

1646
01:29:05,080 --> 01:29:07,240
think I got to.
Go teach.

1647
01:29:07,280 --> 01:29:08,680
Yeah, OK.
Thanks so much.