1
00:00:00,000 --> 00:00:04,073
Audrow Nash: Hi everyone. Audrow
here. In this episode, I speak

2
00:00:04,073 --> 00:00:07,758
with Jenny Read, who's a program
director at the Advanced

3
00:00:07,758 --> 00:00:11,896
Research and Innovation Agency,
or ARIA for short. ARIA is a new

4
00:00:11,896 --> 00:00:15,581
initiative by the UK to fund
work and research that's not

5
00:00:15,581 --> 00:00:19,460
getting enough funding, but
could be really impactful in the

6
00:00:19,460 --> 00:00:22,951
long term. They're loosely
modeled on the US's Defense

7
00:00:22,951 --> 00:00:26,572
Advanced Research Projects
Agency, or DARPA, but without

8
00:00:26,572 --> 00:00:30,580
the defense angle. As a program
director, Jenny has carved out

9
00:00:30,580 --> 00:00:34,653
an area that she thinks is ripe
for investment. And you guessed

10
00:00:34,653 --> 00:00:38,597
it, it's in robotics, mostly
around manipulation. To me, this

11
00:00:38,597 --> 00:00:42,282
kind of initiative is very
exciting because it might bust

12
00:00:42,282 --> 00:00:46,226
some plateaus we're seeing in
robotics and let us get to some

13
00:00:46,226 --> 00:00:49,717
new applications and markets. I
think you'll like this

14
00:00:49,717 --> 00:00:53,661
interview. If you're into far
sighted technologies that could

15
00:00:53,661 --> 00:00:57,281
make a big difference, or want
to understand some of the

16
00:00:57,281 --> 00:01:01,290
challenges and opportunities in
the current state of robotics.

17
00:01:01,290 --> 00:01:05,169
Also, if you're interested in
doing some of this far sighted

18
00:01:05,169 --> 00:01:09,177
work. ARIA is in the UK, but is
not just for people in the UK.

19
00:01:09,177 --> 00:01:13,250
So you might consider applying.
All right. With that, let's get

20
00:01:13,250 --> 00:01:17,000
to the interview. Hi, Jenny.
Would you introduce yourself?

21
00:01:17,000 --> 00:01:20,593
Jenny Read: Absolutely. I'm
Jenny Read, I'm a program

22
00:01:20,593 --> 00:01:24,525
director at ARIA, the UK's
advanced research and invention

23
00:01:24,525 --> 00:01:25,000
agency.

24
00:01:25,000 --> 00:01:27,000
Audrow Nash: And tell me about
ARIA.

25
00:01:27,000 --> 00:01:31,088
Jenny Read: Yeah. ARIA is quite
a new funding organization in

26
00:01:31,088 --> 00:01:34,975
the UK. We were founded by an
Act of Parliament in January

27
00:01:34,975 --> 00:01:38,929
2023, and our mission is to
produce transformative societal

28
00:01:38,929 --> 00:01:42,481
benefit to science and
technology. So we were kind of

29
00:01:42,481 --> 00:01:46,704
new an experiment basically in a
new way of funding for the UK.

30
00:01:46,704 --> 00:01:50,725
We're loosely modeled on the
APRA DARPA model in the US that

31
00:01:50,725 --> 00:01:54,612
is widely perceived as having
been very successful, and we

32
00:01:54,612 --> 00:01:58,834
haven't really had that kind of
program driven, mission focused

33
00:01:58,834 --> 00:02:02,922
approach to funding in the UK
before. And that's what ARIA is

34
00:02:02,922 --> 00:02:05,000
trying to do. So they kick off.

35
00:02:05,000 --> 00:02:08,291
Audrow Nash: Now you say loosely
focused. What what aspects are

36
00:02:08,291 --> 00:02:11,582
you taking and what aspects are
you leaving from the APRA DARPA

37
00:02:11,582 --> 00:02:12,000
program.

38
00:02:12,000 --> 00:02:15,087
Jenny Read: Yeah. Well one
important thing that we're

39
00:02:15,087 --> 00:02:18,000
leaving from DARPA is the
defense aspect. So yeah.

40
00:02:18,000 --> 00:02:20,000
Audrow Nash: Just ARPA no D

41
00:02:20,000 --> 00:02:23,729
Jenny Read: Exactly Defense, the
military is actually the one

42
00:02:23,729 --> 00:02:27,457
area that ARIA won't fund. So
that's an important change. And

43
00:02:27,457 --> 00:02:31,369
that has various consequences as
well. So, you know, DARPA has a

44
00:02:31,369 --> 00:02:34,914
sort of built in customer in the
form of the Department of

45
00:02:34,914 --> 00:02:38,643
Defense or it doesn't. So we
have to think hard about how are

46
00:02:38,643 --> 00:02:42,555
we going to take the science and
tech that we deliver or develop

47
00:02:42,555 --> 00:02:45,000
and translate that into societal
impact.

48
00:02:45,000 --> 00:02:48,856
Audrow Nash: And so yeah, it's
about tech transfer basically.

49
00:02:48,856 --> 00:02:52,333
So you're funding research so
that it can you can build

50
00:02:52,333 --> 00:02:56,000
companies out of it and it can
improve society in general.

51
00:02:56,000 --> 00:02:59,944
Jenny Read: That's right. So
it's an interesting mix because

52
00:02:59,944 --> 00:03:04,150
we are not translational in the
sense of being super high TRL or

53
00:03:04,150 --> 00:03:08,291
close to commercialization. We
can fund very low. So technology

54
00:03:08,291 --> 00:03:10,000
hazardous level because I.

55
00:03:10,000 --> 00:03:13,000
Audrow Nash: Would never know
that acronym okay. Yeah.

56
00:03:13,000 --> 00:03:16,755
Jenny Read: Yeah. All right. We
fund work that's very basic and

57
00:03:16,755 --> 00:03:20,629
speculative. But importantly, we
always have a sense of how we're

58
00:03:20,629 --> 00:03:23,907
hoping it's going to deliver
benefit for society. So we

59
00:03:23,907 --> 00:03:27,424
wouldn't fund or we haven't
typically funded work that sort

60
00:03:27,424 --> 00:03:31,238
of blue sky in the sense we have
no idea how this might benefit.

61
00:03:31,238 --> 00:03:34,993
We're like, we have a purpose in
mind, and then we're trying to

62
00:03:34,993 --> 00:03:38,450
get the knowledge and the
technology together that we need

63
00:03:38,450 --> 00:03:41,967
to solve the problem we're
focusing on. And we have at that

64
00:03:41,967 --> 00:03:45,305
moment seven programs, and each
program is focusing on a

65
00:03:45,305 --> 00:03:48,642
different problem area or
challenge that we're aiming to

66
00:03:48,642 --> 00:03:49,000
solve.

67
00:03:49,000 --> 00:03:53,162
Audrow Nash: Okay. Very
interesting. Now, would you tell

68
00:03:53,162 --> 00:03:57,176
me a bit of your role in all of
this? And then I guess

69
00:03:57,176 --> 00:04:00,000
specifically about the robotics
focus?

70
00:04:00,000 --> 00:04:03,769
Jenny Read: Absolutely. So I'm a
program director, which is

71
00:04:03,769 --> 00:04:07,409
ARIA's version of the RPA
program managers. And that's an

72
00:04:07,409 --> 00:04:10,795
interesting part of our model
actually. So one of our

73
00:04:10,795 --> 00:04:14,436
principles is we always say we
fund people, then projects

74
00:04:14,436 --> 00:04:18,013
rather than focusing on the
project first. And I applied

75
00:04:18,013 --> 00:04:21,717
that approach by starting out
not thinking like, oh, okay,

76
00:04:21,717 --> 00:04:24,975
here's a particular area of
science or a particular

77
00:04:24,975 --> 00:04:28,871
challenge that we want to solve,
but rather let's recruit the

78
00:04:28,871 --> 00:04:32,512
program directors. So they
started out and they recruited

79
00:04:32,512 --> 00:04:36,345
eight program directors like
myself. And then each of us was

80
00:04:36,345 --> 00:04:39,794
given the challenge of try and
find what ARIA calls an

81
00:04:39,794 --> 00:04:43,754
opportunity space. So an area
which feels really important, if

82
00:04:43,754 --> 00:04:47,459
true, would be hugely important
for society, but it's also

83
00:04:47,459 --> 00:04:50,780
underserved, so it hasn't
received the attention and

84
00:04:50,780 --> 00:04:54,613
resources commensurate with its
potential importance. And it

85
00:04:54,613 --> 00:04:58,445
should also be in some sense,
right. So there's a reason why

86
00:04:58,445 --> 00:05:01,000
now is the time to dive into
that space.

87
00:05:01,000 --> 00:05:03,628
Audrow Nash: Tell me more about
that underserved part of it.

88
00:05:03,628 --> 00:05:06,036
Because that's an interesting
thing I remember from our

89
00:05:06,036 --> 00:05:07,000
previous conversation.

90
00:05:07,000 --> 00:05:10,748
Jenny Read: Yeah that's right. I
think we're trying we're trying

91
00:05:10,748 --> 00:05:14,261
to shift the dial. We're trying
to find new spaces. We don't

92
00:05:14,261 --> 00:05:17,716
want to jump on the bandwagon
where there's already lots of

93
00:05:17,716 --> 00:05:21,171
attention because, you know,
that's great. It's being dealt

94
00:05:21,171 --> 00:05:24,802
with. So we wouldn't have much
counterfactual impact there. So

95
00:05:24,802 --> 00:05:28,550
we're trying to find things that
are currently neglected because

96
00:05:28,550 --> 00:05:31,946
nobody else will do them for
whatever reason, whether it's

97
00:05:31,946 --> 00:05:35,635
because it's too risky or it's
not the sort of thing where it's

98
00:05:35,635 --> 00:05:39,032
easy to capture value. So it's
not a attractive investment

99
00:05:39,032 --> 00:05:42,545
proposition. And so for those
reasons that particular spaces

100
00:05:42,545 --> 00:05:46,000
underserved and yet potentially
hugely valuable to society.

101
00:05:46,000 --> 00:05:49,649
Audrow Nash: So you're, you're
not going into areas that are

102
00:05:49,649 --> 00:05:53,236
receiving a ton of investment
from maybe it's like American

103
00:05:53,236 --> 00:05:56,824
venture capital or something
like this. And you're focusing

104
00:05:56,824 --> 00:06:00,291
on areas that they're missing
that are also probably high

105
00:06:00,291 --> 00:06:04,000
impact. And for one reason or
another are not getting funded.

106
00:06:04,000 --> 00:06:05,000
Jenny Read: Exactly.

107
00:06:04,000 --> 00:06:07,434
Audrow Nash: And there's
probably also, you loo you leave

108
00:06:07,434 --> 00:06:10,867
a little bit of like the hype
space because certain areas

109
00:06:10,867 --> 00:06:14,482
really blow up and the investors
pay unbelievable amounts of

110
00:06:14,482 --> 00:06:17,916
money for it. And then it's like
you won't make much of a

111
00:06:17,916 --> 00:06:21,590
contribution there. So you're
focusing on really high impact,

112
00:06:21,590 --> 00:06:24,000
potentially or preferably
ignored areas.

113
00:06:24,000 --> 00:06:26,000
Jenny Read: That's right. Where
can we change the conversation.

114
00:06:26,000 --> 00:06:31,000
Audrow Nash: Okay. And so where
in robotics are you focusing.

115
00:06:31,000 --> 00:06:34,999
Jenny Read: Yeah. So I when I
came into ARIA and start looking

116
00:06:34,999 --> 00:06:39,126
around in robotics very much, I
was an outsider. My background's

117
00:06:39,126 --> 00:06:42,867
actually originally in physics.
And then for many years of

118
00:06:42,867 --> 00:06:46,350
neuroscience, it's quickly
dawned on me that obviously

119
00:06:46,350 --> 00:06:49,768
robotics is having a bit of a
moment. There's so much

120
00:06:49,768 --> 00:06:53,638
potential because AI and so in
lots of robotics, I feel it's

121
00:06:53,638 --> 00:06:57,379
exactly like you say, there's
not a lot of space for AI to

122
00:06:57,379 --> 00:07:01,248
make a difference, but it felt
to me like there was one area

123
00:07:01,248 --> 00:07:04,989
that was being relatively
neglected and that's hardware. I

124
00:07:04,989 --> 00:07:08,923
was very struck as a as I say,
as an outsider to the field by

125
00:07:08,923 --> 00:07:12,922
how little robot hardware had
advanced really over the past 50

126
00:07:12,922 --> 00:07:16,727
years or even longer. People are
saying to me, well, that's

127
00:07:16,727 --> 00:07:20,791
because we know the bottleneck
is the control and the autonomy,

128
00:07:20,791 --> 00:07:24,402
right? You can tell operate
existing robots to carry out

129
00:07:24,402 --> 00:07:28,401
complex tasks, but we can't yet
have them do those same tasks,

130
00:07:28,401 --> 00:07:32,335
autonomy autonomously. So that
proves that the problem is the

131
00:07:32,335 --> 00:07:36,398
control and the autonomy and not
the hardware. And I think that

132
00:07:36,398 --> 00:07:40,268
has been true, but I doubt it's
going to be true. Yeah, it's

133
00:07:40,268 --> 00:07:44,138
arguable, but I think it's been
yeah, certainly a compelling

134
00:07:44,138 --> 00:07:48,201
case can be made. Up till now I
think. But my sense is over the

135
00:07:48,201 --> 00:07:51,684
next few years, precisely
because of this explosion in

136
00:07:51,684 --> 00:07:55,425
novel AI techniques, that's
going to cease to be true. And

137
00:07:55,425 --> 00:07:58,649
there's going to be a new
bottleneck, which is the

138
00:07:58,649 --> 00:08:02,132
capability of the robot bodies.
And again, coming from

139
00:08:02,132 --> 00:08:06,002
neuroscience and spending my
time studying biological brains

140
00:08:06,002 --> 00:08:10,001
and bodies. It was very striking
just how crude current robots

141
00:08:10,001 --> 00:08:12,000
are compared to our own bodies.

142
00:08:12,000 --> 00:08:15,990
Audrow Nash: Yeah, simple
heuristics get us a long way in

143
00:08:15,990 --> 00:08:20,190
robotics. So it's like, yeah,
pretty crude. A lot of things.

144
00:08:20,190 --> 00:08:23,830
They just kind of work and
pretty robustly. Okay. So

145
00:08:23,830 --> 00:08:28,310
hardware. Where within hardware
is it hardware in general. Is it

146
00:08:28,310 --> 00:08:32,580
manufacturing. Like what are you
thinking within the hardware

147
00:08:32,580 --> 00:08:33,000
space.

148
00:08:33,000 --> 00:08:36,959
Jenny Read: There's a few areas
that I'm concentrating on. So

149
00:08:36,959 --> 00:08:40,724
one is actuation. So I feel like
at the moment robotics is

150
00:08:40,724 --> 00:08:44,034
largely using a DC electric
motors as a very mature

151
00:08:44,034 --> 00:08:48,058
technology. But it's a very old
technology at this point. It's

152
00:08:48,058 --> 00:08:52,212
nearly 200 years old. And yeah,
and it obviously wasn't designed

153
00:08:52,212 --> 00:08:56,236
for robotics. And in many ways,
do you see electric motors are

154
00:08:56,236 --> 00:09:00,000
not suitable for robotics
relativity just today. Yeah. No.

155
00:09:00,000 --> 00:09:01,000
Audrow Nash: I don't know too
much.

156
00:09:01,000 --> 00:09:04,293
Jenny Read: Yeah. Low to high
frequency motion. Whereas what

157
00:09:04,293 --> 00:09:07,421
you would typically want in
robotics it's high torque and

158
00:09:07,421 --> 00:09:10,000
much lower frequency motion. So
then oh you're.

159
00:09:10,000 --> 00:09:13,204
Audrow Nash: Saying that drone
motors aren't appropriate for

160
00:09:13,204 --> 00:09:16,354
like, I don't know, robot arms
or something like that where

161
00:09:16,354 --> 00:09:19,451
they spin that at very high
speed, low torque. Yeah. These

162
00:09:19,451 --> 00:09:21,000
kind of things totally agree.

163
00:09:21,000 --> 00:09:24,240
Jenny Read: Right. And so then
you need gears right to

164
00:09:24,240 --> 00:09:27,000
translate into a low to high
torque low speed.

165
00:09:27,000 --> 00:09:29,486
Audrow Nash: And then they
reflected inertia and all sorts

166
00:09:29,486 --> 00:09:30,000
of problems.

167
00:09:30,000 --> 00:09:32,909
Jenny Read: Exactly. You've got
all this inertia that you didn't

168
00:09:32,909 --> 00:09:35,455
want and you're just basically
fighting all the way down

169
00:09:35,455 --> 00:09:38,136
against the problem caused by
the fact that this technology

170
00:09:38,136 --> 00:09:40,000
fundamentally isn't suited for
the needs.

171
00:09:40,000 --> 00:09:43,485
Audrow Nash: Well, the the high
gear ratio ones, I mean, there

172
00:09:43,485 --> 00:09:46,970
are low gear ratio ones that I
think are better. Do you do you

173
00:09:46,970 --> 00:09:50,006
think so or like the ones
there's some companies where

174
00:09:50,006 --> 00:09:53,379
they're doing like 1 to 1 or 1
to 4 gear ratios. And so it's

175
00:09:53,379 --> 00:09:56,976
very direct. It's high torque. I
think I'm not too knowledgeable

176
00:09:56,976 --> 00:09:59,000
in this area, but what do you
think?

177
00:09:59,000 --> 00:10:02,611
Jenny Read: Yeah, I think there
are a number of different

178
00:10:02,611 --> 00:10:06,475
solutions, but none of them are
ideal. It won't have problems

179
00:10:06,475 --> 00:10:10,403
like efficiency and weight. Back
drive ability. And, you know,

180
00:10:10,403 --> 00:10:13,000
when you compare it to
biological muscle.

181
00:10:13,000 --> 00:10:14,000
Audrow Nash: That's very
different.

182
00:10:14,000 --> 00:10:17,205
Jenny Read: Yeah. Very
different. I'd say one of the

183
00:10:17,205 --> 00:10:20,966
things I'm interested in, in
trying to develop in the program

184
00:10:20,966 --> 00:10:23,000
is fund novel forms of
actuation.

185
00:10:23,000 --> 00:10:26,093
Audrow Nash: Okay. All right.
Actuation sounds really cool. I

186
00:10:26,093 --> 00:10:29,237
think we'll, we'll I'd like to
go through all of the different

187
00:10:29,237 --> 00:10:32,177
areas in hardware that you'd
like to focus, and then maybe

188
00:10:32,177 --> 00:10:34,966
circle back to them, in more
detail because I have more

189
00:10:34,966 --> 00:10:37,806
questions, definitely for
actuation, but I don't want to

190
00:10:37,806 --> 00:10:41,000
get bogged down there right now.
I want to see the whole thing.

191
00:10:41,000 --> 00:10:45,181
Jenny Read: Yeah. So another one
is sensing, again, I was struck

192
00:10:45,181 --> 00:10:48,905
by the relative crudity and
poverty of robot senses apart

193
00:10:48,905 --> 00:10:52,302
from vision. So typically
robots, you cameras. I'm a

194
00:10:52,302 --> 00:10:56,222
vision scientist, so you don't
need to convince me about the

195
00:10:56,222 --> 00:11:00,273
benefits of vision. And we are
highly visual creatures, right.

196
00:11:00,273 --> 00:11:03,997
And what the consequence of that
is that, you know, we've

197
00:11:03,997 --> 00:11:07,524
invented painting and
photography, and we love looking

198
00:11:07,524 --> 00:11:11,118
at pictures. So cameras are
ubiquitous to this wall and

199
00:11:11,118 --> 00:11:14,580
cheap and available to use in
robots. But if you were

200
00:11:14,580 --> 00:11:18,304
designing a robot a priori, I
don't think you necessarily

201
00:11:18,304 --> 00:11:21,963
start with cameras as top
priority. And the reason I say

202
00:11:21,963 --> 00:11:25,817
that is if you look at biology,
there are animals that cope

203
00:11:25,817 --> 00:11:29,868
without vision at all. I, cave
salamanders with no eyes had so

204
00:11:29,868 --> 00:11:33,853
hunt. But I don't think there's
a single animal that has that

205
00:11:33,853 --> 00:11:37,577
lacks touch sense. The kind of
sensation is clearly super

206
00:11:37,577 --> 00:11:41,497
important in biology, and I
just. It's not like I'm obsessed

207
00:11:41,497 --> 00:11:45,286
with robots having to copy
biology. But evolution has been

208
00:11:45,286 --> 00:11:48,748
exploring the design space for
billions of years, and

209
00:11:48,748 --> 00:11:52,603
apparently it hasn't found any
workable part of that design

210
00:11:52,603 --> 00:11:56,000
space that doesn't have really
rich tactile sensing.

211
00:11:56,000 --> 00:11:59,395
Audrow Nash: So sensing and
specifically touch sensing if

212
00:11:59,395 --> 00:12:02,909
you're interested in I think
that's really cool. I mean, if

213
00:12:02,909 --> 00:12:06,066
we had robots with skin, all of
the sensor and sensor

214
00:12:06,066 --> 00:12:09,759
information you could get would
be amazing. And, in having the

215
00:12:09,759 --> 00:12:13,213
robot navigate the world and
operate without like, pushing

216
00:12:13,213 --> 00:12:15,000
things over and whatever else.

217
00:12:15,000 --> 00:12:17,963
Jenny Read: Right, right. The
other thing, of course, is

218
00:12:17,963 --> 00:12:21,138
sensing about its internal state
as well, because that's the

219
00:12:21,138 --> 00:12:23,995
other thing that surprised me
when I entered this is a

220
00:12:23,995 --> 00:12:27,063
robotics field, was to learn
about how fragile robots are.

221
00:12:27,063 --> 00:12:30,450
This is remember visiting a lab
and they were showing me a robot

222
00:12:30,450 --> 00:12:33,730
that they were trying to train
to do household cleaning tasks.

223
00:12:33,730 --> 00:12:37,116
But the the lab chief was like,
but I mustn't touch it because I

224
00:12:37,116 --> 00:12:40,291
probably break it. So I was
like, how I going, could this be

225
00:12:40,291 --> 00:12:43,042
a problem? And I think
presumably one of the reasons

226
00:12:43,042 --> 00:12:46,005
robots are fragile is that they
lack sensing about their

227
00:12:46,005 --> 00:12:48,810
internal state or not
perception. So they don't sense

228
00:12:48,810 --> 00:12:52,037
when they're doing themselves
damage. So they're a little bit

229
00:12:52,037 --> 00:12:55,212
like these humans who have some,
you know, genetic disorder,

230
00:12:55,212 --> 00:12:58,280
which means they don't feel
pain. They have to live really

231
00:12:58,280 --> 00:13:01,296
careful lives because otherwise
they die incredibly young

232
00:13:01,296 --> 00:13:04,048
because of all the damage
they've done to their body

233
00:13:04,048 --> 00:13:05,000
without realizing.

234
00:13:05,000 --> 00:13:07,918
Audrow Nash: Oh, for sure. Or
like, if you lose nerves in your

235
00:13:07,918 --> 00:13:10,929
feet and then you have a pebble
in your shoe and that really the

236
00:13:10,929 --> 00:13:13,000
infection and wounds that you
can get there.

237
00:13:13,000 --> 00:13:14,000
Jenny Read: Exactly.

238
00:13:14,000 --> 00:13:17,948
Audrow Nash: Okay. So I like
those points quite a bit. So

239
00:13:17,948 --> 00:13:21,273
sensing so actuation
specifically around motors.

240
00:13:21,273 --> 00:13:25,152
We're not novel sensing not
novel forms of actuation and

241
00:13:25,152 --> 00:13:29,169
then sensing around touch and
proprioception. Is that what

242
00:13:29,169 --> 00:13:30,000
it's called?

243
00:13:30,000 --> 00:13:33,316
Jenny Read: Yeah. Well, there's
proprioception and nociceptors

244
00:13:33,316 --> 00:13:36,364
and so forth. Perception is
sensing your current state of

245
00:13:36,364 --> 00:13:39,358
your body and nociceptors
sensing what is being damaged,

246
00:13:39,358 --> 00:13:40,000
essentially.

247
00:13:40,000 --> 00:13:42,000
Audrow Nash: Oh, cool. I don't
know that word. Sweet.

248
00:13:42,000 --> 00:13:43,000
Jenny Read: Yeah.

249
00:13:43,000 --> 00:13:45,000
Audrow Nash: Any other focus
areas?

250
00:13:45,000 --> 00:13:48,969
Jenny Read: Yeah. So the other
area that we're getting quite a

251
00:13:48,969 --> 00:13:52,747
lot of attention to, and again,
coming back to this idea of

252
00:13:52,747 --> 00:13:56,268
evolution, one of the things
that again struck me about

253
00:13:56,268 --> 00:14:00,237
robotics is how little attention
has been given to things like

254
00:14:00,237 --> 00:14:04,143
morphological computing. The
advantages you can get by making

255
00:14:04,143 --> 00:14:06,960
control more efficient by
designing the body

256
00:14:06,960 --> 00:14:10,929
appropriately. And that's again,
something that I think we see

257
00:14:10,929 --> 00:14:14,514
very commonly in biology. And
it's because in evolution,

258
00:14:14,514 --> 00:14:18,100
brains and bodies obviously
evolve together in lockstep.

259
00:14:18,100 --> 00:14:21,941
Where is my sense to that in
robotics? Yeah. Often you get a

260
00:14:21,941 --> 00:14:25,590
mechanical engineer designing a
robot. Yeah. Choosing the

261
00:14:25,590 --> 00:14:29,495
actuators, choosing the limb
lengths and so on. And then they

262
00:14:29,495 --> 00:14:33,209
sort of hand it over to a
computer scientist or a software

263
00:14:33,209 --> 00:14:37,050
engineer or go, go figure out
how to control this thing. And

264
00:14:37,050 --> 00:14:40,827
it's always a bit like it might
be of like the Frankenstein

265
00:14:40,827 --> 00:14:44,861
movie, you know, you've got this
body that is like, bring it to

266
00:14:44,861 --> 00:14:48,702
life. Yeah. And then looking
back, even I realize I actually

267
00:14:48,702 --> 00:14:52,672
do this. I can trace this back
in sort of Western tradition is

268
00:14:52,672 --> 00:14:56,577
basically the book of Genesis,
right? God forms man from clay

269
00:14:56,577 --> 00:15:00,546
and then breathes life in. And
obviously that's not at all how

270
00:15:00,546 --> 00:15:04,324
we think of evolution having
happened. And I think that the

271
00:15:04,324 --> 00:15:08,229
tight integration between the
control and bodies in evolution

272
00:15:08,229 --> 00:15:12,006
is why we see this incredible
efficiency. So that's how you

273
00:15:12,006 --> 00:15:16,104
can get something like a B with
like a million neurons doing all

274
00:15:16,104 --> 00:15:17,000
these amazing.

275
00:15:17,000 --> 00:15:18,000
Audrow Nash: Incredible.

276
00:15:17,000 --> 00:15:20,792
Jenny Read: Things. Right? Yeah.
Flight obviously of all these

277
00:15:20,792 --> 00:15:24,462
sort of decisions and memory and
communication like in which

278
00:15:24,462 --> 00:15:28,254
direction is the best honey and
so on. And I think part of the

279
00:15:28,254 --> 00:15:31,862
reason they're able to do that
is it just needs to spend so

280
00:15:31,862 --> 00:15:35,226
many fewer computational
resources compensating for the

281
00:15:35,226 --> 00:15:37,000
inadequacies of their bodies.

282
00:15:37,000 --> 00:15:39,965
Audrow Nash: Yeah. Yeah. They're
just driven by impulses in a

283
00:15:39,965 --> 00:15:42,979
sense that are highly efficient
and very appropriate generally

284
00:15:42,979 --> 00:15:44,000
to their environment.

285
00:15:44,000 --> 00:15:47,478
Jenny Read: Right. Exactly. So
when I see people saying, oh,

286
00:15:47,478 --> 00:15:50,897
look, the hardware is fine,
because, you know, I can tell I

287
00:15:50,897 --> 00:15:54,433
operate a robot to do this. It's
kind of like, well, yes, you

288
00:15:54,433 --> 00:15:57,563
can, but you're throwing an
entire human brain at that

289
00:15:57,563 --> 00:16:01,098
problem, right? And the human
brain is a lot of computational

290
00:16:01,098 --> 00:16:04,402
resources, power. And what
you're achieving could be done

291
00:16:04,402 --> 00:16:08,053
by a mouse brain. If that robot
brain is deeply integrated into

292
00:16:08,053 --> 00:16:11,473
a mouse body, so can we find
ways of achieving that kind of

293
00:16:11,473 --> 00:16:15,066
efficiency gain in robotics? A
few people have looked at this,

294
00:16:15,066 --> 00:16:18,370
right? Because obviously you
can't actually evolve robots

295
00:16:18,370 --> 00:16:22,022
like one that it reproduce. And
two, we haven't got billions of

296
00:16:22,022 --> 00:16:25,615
years to play around with. So
you have to do it in simulation.

297
00:16:25,615 --> 00:16:29,267
And people have obviously tried
to explore similar ideas, but I

298
00:16:29,267 --> 00:16:32,512
think up till now you just
haven't been able to simulate

299
00:16:32,512 --> 00:16:36,222
things of sufficient complexity
to really make any progress. And

300
00:16:36,222 --> 00:16:39,699
that's why up till now, actually
the best deal was to have a

301
00:16:39,699 --> 00:16:43,467
smart mechanical engineer design
your robot body. But maybe we're

302
00:16:43,467 --> 00:16:47,002
reaching the point now where
that's no longer true, and maybe

303
00:16:47,002 --> 00:16:50,538
we can have we can experiment
with generative AI exploring in

304
00:16:50,538 --> 00:16:53,842
simulated worlds. You know,
suppose this control strategy

305
00:16:53,842 --> 00:16:57,087
gets integrated with this
particular body design. Can we

306
00:16:57,087 --> 00:16:59,000
discover unexpected
efficiencies?

307
00:16:59,000 --> 00:17:02,370
Audrow Nash: Yeah, that kind of
thing is really cool. Yeah. I

308
00:17:02,370 --> 00:17:05,575
remember several years ago
looking into work like this and

309
00:17:05,575 --> 00:17:08,890
it's super cool. And in theory,
it's amazing. I just haven't

310
00:17:08,890 --> 00:17:12,425
seen that compelling of examples
recently, but I haven't been in

311
00:17:12,425 --> 00:17:15,685
the academic space for quite a
while. So maybe that kind of

312
00:17:15,685 --> 00:17:18,834
thing has happened, or the
results that you achieve might

313
00:17:18,834 --> 00:17:22,204
be amazing in simulation, but
they're very hard to generalize

314
00:17:22,204 --> 00:17:25,740
to reality because they overfit
the simulator. And the simulator

315
00:17:25,740 --> 00:17:29,000
has all sorts of shorthands for
complex physics phenomenon.

316
00:17:29,000 --> 00:17:32,152
Jenny Read: Right, exactly. So
clearly, a big part of this

317
00:17:32,152 --> 00:17:35,630
effort is going to be improving
our physics based simulation and

318
00:17:35,630 --> 00:17:38,728
developing techniques for
closing this into real gap, for

319
00:17:38,728 --> 00:17:39,000
sure.

320
00:17:39,000 --> 00:17:41,093
Audrow Nash: Yeah. Very cool.
Okay. Do you have any more

321
00:17:41,093 --> 00:17:43,000
focus? Because that already
seems like a lot to me.

322
00:17:43,000 --> 00:17:46,100
Jenny Read: Yeah, that pretty
much covers it. But yeah, you're

323
00:17:46,100 --> 00:17:47,000
right. It's a lot.

324
00:17:47,000 --> 00:17:51,047
Audrow Nash: Okay, so I hear
novel actuation, sensing that's

325
00:17:51,047 --> 00:17:54,958
around like touch sensing and
then also like state sensing

326
00:17:54,958 --> 00:17:59,207
where they're like, am I in pain
or something like that so they

327
00:17:59,207 --> 00:18:03,388
can realize their state and not
the word for it. But then, and

328
00:18:03,388 --> 00:18:07,502
then morphological computing,
which you said, and that's like

329
00:18:07,502 --> 00:18:11,347
a tighter integration between
hardware design and control

330
00:18:11,347 --> 00:18:15,595
strategy. Probably with a lot of
leaning on simulation. Is that

331
00:18:15,595 --> 00:18:16,000
right?

332
00:18:16,000 --> 00:18:18,685
Jenny Read: Correct. Yeah. I
should say we even though

333
00:18:18,685 --> 00:18:21,867
nociceptors and proprioception I
think are important, we haven't

334
00:18:21,867 --> 00:18:25,000
actually we're not funding any
work in that area just based on.

335
00:18:25,000 --> 00:18:28,252
Audrow Nash: Yeah, that seems
very hard to me. Yeah. But I

336
00:18:28,252 --> 00:18:31,000
don't know the technology state
in that area too.

337
00:18:31,000 --> 00:18:35,326
Jenny Read: Like, okay. Well, so
I just I think proprioception

338
00:18:35,326 --> 00:18:39,581
does come in with some of the
novel artificial muscles, which

339
00:18:39,581 --> 00:18:43,000
can, report on their own state
effectively. Yeah.

340
00:18:43,000 --> 00:18:45,175
Audrow Nash: And if you have
encoders and, motors, you can

341
00:18:45,175 --> 00:18:46,000
tell where the arm is.

342
00:18:46,000 --> 00:18:47,000
Jenny Read: Right.

343
00:18:47,000 --> 00:18:50,450
Audrow Nash: So we have
proprioception or however you

344
00:18:50,450 --> 00:18:54,485
say it, just to a decent level.
But we don't know if like, you

345
00:18:54,485 --> 00:18:58,260
don't have skin so you can't
tell if you're bumping things

346
00:18:58,260 --> 00:19:02,101
and a lot of it's more open loop
and you can observe at the

347
00:19:02,101 --> 00:19:06,331
motors if they're bumping things
because they'll, you'll generate

348
00:19:06,331 --> 00:19:10,302
a force on them and they can
observe that, but it's still not

349
00:19:10,302 --> 00:19:14,207
as good. And especially with the
high gear ratios, you might

350
00:19:14,207 --> 00:19:18,308
clobber stuff before you realize
that you have indeed clobbered

351
00:19:18,308 --> 00:19:22,148
it. Right? Okay, so tell me
about so I get the areas you're

352
00:19:22,148 --> 00:19:26,314
focusing. That's really awesome.
Your background in neuroscience

353
00:19:26,314 --> 00:19:30,284
and biology. Tell me how you're
here and how you are. Like, I

354
00:19:30,284 --> 00:19:33,799
guess I'd love to hear your
background and how you got

355
00:19:33,799 --> 00:19:37,639
involved with all this. And you
mentioned they were funding

356
00:19:37,639 --> 00:19:41,675
people, not programs, per se.
Tell me how you started with all

357
00:19:41,675 --> 00:19:42,000
this.

358
00:19:42,000 --> 00:19:45,222
Jenny Read: Yeah. So as I
mentioned, I started out in

359
00:19:45,222 --> 00:19:48,261
physics. I actually did a
doctorate in theoretical

360
00:19:48,261 --> 00:19:51,604
astrophysics, but after that I
moved into computational

361
00:19:51,604 --> 00:19:55,494
neuroscience. I started modeling
synaptic transmission. And then

362
00:19:55,494 --> 00:19:59,020
I got into vision science. And I
spent many years studying

363
00:19:59,020 --> 00:20:02,606
stereoscopic vision. So how do
you get a sense of depth? By

364
00:20:02,606 --> 00:20:06,557
combining the images in your two
eyes, which is it sounds kind of

365
00:20:06,557 --> 00:20:09,900
niche, but it's actually a
really nice model system for

366
00:20:09,900 --> 00:20:13,000
studying perception and
understanding. Yeah, right.

367
00:20:13,000 --> 00:20:16,099
Audrow Nash: How is it I mean,
robotics, it's like we use it

368
00:20:16,099 --> 00:20:19,300
all the time. Stereo vision for
depth, cheap cameras. You just

369
00:20:19,300 --> 00:20:22,554
need processing and you can get
a lot of information, very rich

370
00:20:22,554 --> 00:20:24,000
information about the world.

371
00:20:24,000 --> 00:20:27,140
Jenny Read: Right? Right.
Exactly. And it's very nice.

372
00:20:27,140 --> 00:20:30,860
Question for perception as well.
Yeah. How is it that electrical

373
00:20:30,860 --> 00:20:34,291
activity in neurons can give you
a sense of depth? So I was

374
00:20:34,291 --> 00:20:37,895
studying stereoscopic vision in
humans, and other primates for

375
00:20:37,895 --> 00:20:41,558
many years. And then my career
took a bit of a swerve because I

376
00:20:41,558 --> 00:20:45,163
discovered that insects have got
stereotypes. Just to a chance

377
00:20:45,163 --> 00:20:48,942
conversation in the coffee room,
a colleague was mentioning to me

378
00:20:48,942 --> 00:20:52,314
that praying mantis have got
stereoscopic vision. That was

379
00:20:52,314 --> 00:20:55,744
news to me at the time. And it
really surprised me, because

380
00:20:55,744 --> 00:20:59,116
within the field, there's a
sense that stereoscopic vision

381
00:20:59,116 --> 00:21:02,605
is something really expensive.
Exactly what you told us told

382
00:21:02,605 --> 00:21:04,000
the correspondents from.

383
00:21:04,000 --> 00:21:07,000
Audrow Nash: Their little head
in the praying mantis has the

384
00:21:07,000 --> 00:21:08,000
stereoscopic vision.

385
00:21:08,000 --> 00:21:11,711
Jenny Read: Exactly. I'm like,
wait, hang on. Insects got a

386
00:21:11,711 --> 00:21:15,107
stereoscopic vision. So I
applied for funding to study

387
00:21:15,107 --> 00:21:19,006
insect stereo is to understand,
like, how are they doing this?

388
00:21:19,006 --> 00:21:22,214
Is it similar to human
stereoscopic vision or is it

389
00:21:22,214 --> 00:21:25,547
different? Long story short,
it's kind of similar and

390
00:21:25,547 --> 00:21:29,509
different. But that was a really
interesting, research program.

391
00:21:29,509 --> 00:21:33,031
But it also got me thinking
about robotics for the first

392
00:21:33,031 --> 00:21:37,119
time. And I think that's because
insects are so simple. It's kind

393
00:21:37,119 --> 00:21:40,830
of natural to think of them as
little robots. So I thought,

394
00:21:40,830 --> 00:21:44,415
yeah, right. So I was trying to
reproduce whatever stereo

395
00:21:44,415 --> 00:21:48,000
algorithm was going on in the
head of the praying mantis.

396
00:21:48,000 --> 00:21:50,759
Audrow Nash: Then that that's
really cool. Yeah, yeah. I see

397
00:21:50,759 --> 00:21:52,000
how that lends to robotics.

398
00:21:52,000 --> 00:21:55,219
Jenny Read: That's right. You
start to have to think about the

399
00:21:55,219 --> 00:21:58,489
motor behavior as well, because,
I mean, one of the things that

400
00:21:58,489 --> 00:22:01,656
it's clear is that they that's
the story vision I think works

401
00:22:01,656 --> 00:22:04,719
best in the center of their
vision. So, you know, they have

402
00:22:04,719 --> 00:22:07,782
to detect something and then
move their head towards it. So

403
00:22:07,782 --> 00:22:10,949
now you really have to focus.
Yeah. You have to put the close

404
00:22:10,949 --> 00:22:14,220
the loop. Right. You have to put
the motor behavior in when you

405
00:22:14,220 --> 00:22:17,283
think about the sensory input.
So yeah I was thinking about

406
00:22:17,283 --> 00:22:20,398
robotics from that perspective
other than an idea came along

407
00:22:20,398 --> 00:22:23,357
and we have recruited for a
program directors and made it

408
00:22:23,357 --> 00:22:26,265
clear that you could apply to
work on something that was

409
00:22:26,265 --> 00:22:29,483
outside your existing area of
expertise. I just thought, well,

410
00:22:29,483 --> 00:22:32,494
it's probably crazy, but why
don't I pitch to them to come

411
00:22:32,494 --> 00:22:34,000
and do something on robotics?

412
00:22:34,000 --> 00:22:38,034
Audrow Nash: Very cool. Okay.
And then so how does this all so

413
00:22:38,034 --> 00:22:42,003
you have your focus areas. How
do you go about selecting like

414
00:22:42,003 --> 00:22:45,842
what were you looking for. How
much funding and support are

415
00:22:45,842 --> 00:22:49,616
these companies or groups
getting? How does all this work?

416
00:22:49,616 --> 00:22:53,000
Kind of from a logistics of the
program perspective.

417
00:22:53,000 --> 00:22:56,086
Jenny Read: Yeah. So the first
thing to say is it varies a lot

418
00:22:56,086 --> 00:22:59,222
from program to program. Because
program directors have quite a

419
00:22:59,222 --> 00:23:02,357
lot of freedom to set things up
how they think is best going to

420
00:23:02,357 --> 00:23:04,000
advance their particular
program.

421
00:23:04,000 --> 00:23:07,000
Audrow Nash: Did how are you
doing it? Yeah.

422
00:23:07,000 --> 00:23:11,276
Jenny Read: So in my case, I had
a budget of around 50 million

423
00:23:11,276 --> 00:23:15,552
pounds. I in my original call
for proposals, I said, you know,

424
00:23:15,552 --> 00:23:19,552
the maximum 20 individual
project. 20 million. Actually, I

425
00:23:19,552 --> 00:23:23,690
haven't I'm not planning to make
any such large awards. I've

426
00:23:23,690 --> 00:23:27,552
probably made more smaller
awards, but, you know, fewer,

427
00:23:27,552 --> 00:23:31,414
larger ones, do. The thing about
my program is extremely

428
00:23:31,414 --> 00:23:35,690
collaborative. So a lot of the
creators, as we call them, the,

429
00:23:35,690 --> 00:23:40,034
people who are actually going to
be doing the research, even if

430
00:23:40,034 --> 00:23:44,103
they applied separately, going
to work with each other. And

431
00:23:44,103 --> 00:23:48,034
that's. Yeah, that's something
that I certainly feel very

432
00:23:48,034 --> 00:23:52,379
positive about. But I was struck
by what an appetite there was.

433
00:23:52,379 --> 00:23:56,862
From the community for that. And
I think that's one of the things

434
00:23:56,862 --> 00:24:00,931
that attracts me about the idea
of sort of synergy and, you

435
00:24:00,931 --> 00:24:05,069
know, the whole be more the sum
of the parts because you can

436
00:24:05,069 --> 00:24:08,793
work together collectively on
this common goal and the

437
00:24:08,793 --> 00:24:12,724
co-design area that I mentioned,
whether you're trying to

438
00:24:12,724 --> 00:24:16,862
optimize the design of a body
and also the control, that's a

439
00:24:16,862 --> 00:24:21,000
great area where I have a whole
bunch of different creators,

440
00:24:21,000 --> 00:24:24,586
taking particular skills in
areas like probabilistic

441
00:24:24,586 --> 00:24:29,000
programing or coziness into real
gap or Bayesian optimization or

442
00:24:29,000 --> 00:24:32,862
the nitty gritty of actually
building robot hardware. So

443
00:24:32,862 --> 00:24:36,931
they're all coming together with
these different lenses and

444
00:24:36,931 --> 00:24:41,000
perspectives. And I think that's
going to be super helpful.

445
00:24:41,000 --> 00:24:44,705
Audrow Nash: Yeah, I think so
too. Well, that's really cool

446
00:24:44,705 --> 00:24:48,473
when you're saying that they
work together, are they working

447
00:24:48,473 --> 00:24:51,990
together inside of your program
or across the whole area

448
00:24:51,990 --> 00:24:54,000
program? Is it like oh go ahead.

449
00:24:54,000 --> 00:24:57,419
Jenny Read: It's within a
program. So they're working

450
00:24:57,419 --> 00:24:58,000
together.

451
00:24:58,000 --> 00:25:01,856
Audrow Nash: So maybe like
sharing code, sharing resources,

452
00:25:01,856 --> 00:25:05,843
sharing know how all towards the
advancement of all the goals

453
00:25:05,843 --> 00:25:08,000
that are okay. I like that a
lot.

454
00:25:08,000 --> 00:25:11,459
Jenny Read: I like that I should
say it's not mandatory. So there

455
00:25:11,459 --> 00:25:14,493
are some creators within the
program who, you know, maybe

456
00:25:14,493 --> 00:25:17,739
they're a company. They don't
want to share anything, because

457
00:25:17,739 --> 00:25:20,454
of their, you know,
commercialization goals. That's

458
00:25:20,454 --> 00:25:23,647
fine too. But I'm happy that
there are many creators who are

459
00:25:23,647 --> 00:25:26,415
very open about sharing code,
sharing ideas, sharing

460
00:25:26,415 --> 00:25:27,000
approaches.

461
00:25:27,000 --> 00:25:30,976
Audrow Nash: Oh, yeah. I mean,
especially I think when you go

462
00:25:30,976 --> 00:25:34,821
deal with problems that are as
hard as these, there ends up

463
00:25:34,821 --> 00:25:38,796
being a lot of boilerplate, like
there ends up being a lot of

464
00:25:38,796 --> 00:25:41,664
things that are
undifferentiated, and don't,

465
00:25:41,664 --> 00:25:45,248
don't allow you to test your
idea or like, they're just

466
00:25:45,248 --> 00:25:49,289
necessary steps on the way there
and like, might as well share

467
00:25:49,289 --> 00:25:53,395
all those. And so then, so you
have companies and research labs

468
00:25:53,395 --> 00:25:57,501
or like, who are the people that
are applying for some chunk of

469
00:25:57,501 --> 00:25:59,000
your 50 million pounds?

470
00:25:59,000 --> 00:26:02,776
Jenny Read: Yeah, all of the
above. So we're very open as to

471
00:26:02,776 --> 00:26:06,112
who we fund. Big companies,
small companies, academic

472
00:26:06,112 --> 00:26:10,203
research labs, even individuals,
everyone's eligible, we can also

473
00:26:10,203 --> 00:26:13,979
fund outside the UK everything
within a set of programs. So.

474
00:26:13,979 --> 00:26:18,007
Yeah. And in my case as well, I
don't think I've actually, ended

475
00:26:18,007 --> 00:26:21,720
up selecting any individuals who
have had it as part of the

476
00:26:21,720 --> 00:26:25,559
program. But apart from that,
pretty much every other type of

477
00:26:25,559 --> 00:26:26,000
entity.

478
00:26:26,000 --> 00:26:30,026
Audrow Nash: That's cool. So you
mentioned you can fund people

479
00:26:30,026 --> 00:26:34,247
outside of the UK. Oh, actually,
before we getting to that, where

480
00:26:34,247 --> 00:26:38,468
are you guys now in the process?
Like do we have I think you were

481
00:26:38,468 --> 00:26:42,364
saying before that you have a
few people or you have groups,

482
00:26:42,364 --> 00:26:45,740
but you're finishing up
negotiating or where are you

483
00:26:45,740 --> 00:26:46,000
now?

484
00:26:46,000 --> 00:26:49,832
Jenny Read: That's right. We
haven't yet signed any contracts

485
00:26:49,832 --> 00:26:53,162
for the main program funding.
We're in the process of

486
00:26:53,162 --> 00:26:56,932
negotiating the details and hope
to sign our first contracts

487
00:26:56,932 --> 00:26:58,000
before Christmas.

488
00:26:58,000 --> 00:27:02,262
Audrow Nash: That's awesome. How
many? Roughly how many entities?

489
00:27:02,262 --> 00:27:06,000
Like companies, research labs.
Are you? Possibly funding.

490
00:27:06,000 --> 00:27:09,000
Jenny Read: About 26 for the
program.

491
00:27:09,000 --> 00:27:13,608
Audrow Nash: That's awesome. And
they run the gamut of the areas

492
00:27:13,608 --> 00:27:18,000
that you said actuation, sensing
and morphological computing.

493
00:27:18,000 --> 00:27:21,449
Jenny Read: That's right. And
some of them are also what I

494
00:27:21,449 --> 00:27:25,077
called challenge specification
creators. So I was keen to get

495
00:27:25,077 --> 00:27:28,645
people involved who are not in
robotics themselves, but have

496
00:27:28,645 --> 00:27:32,034
some problem that they would
like robots to solve so that

497
00:27:32,034 --> 00:27:35,186
they could work with the
creators. Yeah, and I should

498
00:27:35,186 --> 00:27:38,992
say, I think I haven't mentioned
yet that within robot hardware,

499
00:27:38,992 --> 00:27:42,382
the focus of the program is
specifically on dexterity. So

500
00:27:42,382 --> 00:27:46,069
we're interested in people who
have dexterous tasks, like, for

501
00:27:46,069 --> 00:27:49,815
example, you know, assembly in
manufacturing, the sort of tasks

502
00:27:49,815 --> 00:27:53,502
that still have to be done by
human beings. What would it take

503
00:27:53,502 --> 00:27:56,000
to have a robot be able to do
those tasks?

504
00:27:56,000 --> 00:27:59,698
Audrow Nash: I like that a lot.
You want to tell me a bit more

505
00:27:59,698 --> 00:28:03,277
about the dexterity challenge
and like what skills I get the

506
00:28:03,277 --> 00:28:06,558
importance of it. And I also
think that it's an area in

507
00:28:06,558 --> 00:28:10,435
robotics where there needs to be
a lot of work. And you see a lot

508
00:28:10,435 --> 00:28:13,477
of companies working on this.
This actually from my

509
00:28:13,477 --> 00:28:17,056
perspective, this seems like a
much hotter area or more more

510
00:28:17,056 --> 00:28:20,456
money and more investment and
more attention all over, is

511
00:28:20,456 --> 00:28:24,035
going into this area, but I
suppose maybe not as much on the

512
00:28:24,035 --> 00:28:27,674
hardware. Like you still see two
finger grippers. But tell me

513
00:28:27,674 --> 00:28:30,000
about this area a little bit
more like.

514
00:28:30,000 --> 00:28:33,901
Jenny Read: Well, I think, you
know, just as you say, dexterity

515
00:28:33,901 --> 00:28:37,678
is just so important because
basically anything we would want

516
00:28:37,678 --> 00:28:41,578
robots to do for us is going to
involve the precise application

517
00:28:41,578 --> 00:28:44,984
of forces to objects in the
world, which is essentially

518
00:28:44,984 --> 00:28:48,637
dexterous manipulation. You
know, we're seeing robots start

519
00:28:48,637 --> 00:28:52,661
to be used, you know, to inspect
factories and so on. But I think

520
00:28:52,661 --> 00:28:56,500
that their use is going to be so
limited until they can really

521
00:28:56,500 --> 00:29:00,029
manipulate objects in a way
comparable to human beings. I

522
00:29:00,029 --> 00:29:03,249
mean, not a new point,
obviously, as you say, that's

523
00:29:03,249 --> 00:29:06,840
why there's so much tension
going into it. But my sense is

524
00:29:06,840 --> 00:29:10,121
this is an area where the
hardware really is going to

525
00:29:10,121 --> 00:29:13,960
prove limiting. And the tactile
sensing, for example, is going

526
00:29:13,960 --> 00:29:17,799
to be really key here, but also
the actuation. Right. And it's

527
00:29:17,799 --> 00:29:21,266
just hard to have dexterous
fingers if each one is being

528
00:29:21,266 --> 00:29:23,000
driven by an electric motor.

529
00:29:23,000 --> 00:29:27,462
Audrow Nash: Interesting. Yeah.
So this is and so what it sounds

530
00:29:27,462 --> 00:29:31,924
like to me is this is kind of a
everything, combines around this

531
00:29:31,924 --> 00:29:36,386
area of the population because I
can see how hardware might be a

532
00:29:36,386 --> 00:29:40,708
limit and your sensors might be
a limit, and then your ability

533
00:29:40,708 --> 00:29:44,821
to design hardware in tandem
with how it's being controlled

534
00:29:44,821 --> 00:29:49,353
would be a limit. And simulation
would be a limit, right? This is

535
00:29:49,353 --> 00:29:53,118
kind of like all of those
points. They all focus under

536
00:29:53,118 --> 00:29:55,000
manipulation in some sense.

537
00:29:55,000 --> 00:29:58,000
Jenny Read: Yeah. Yeah, exactly.

538
00:29:58,000 --> 00:30:01,168
Audrow Nash: Okay. Yeah, that
seems really cool. One thing

539
00:30:01,168 --> 00:30:04,008
that's interesting to me is,
I've talked to a lot of

540
00:30:04,008 --> 00:30:07,503
companies that use, like the two
finger grippers or the solution

541
00:30:07,503 --> 00:30:10,998
grippers and they're like, yeah,
we tried a bunch of things, but

542
00:30:10,998 --> 00:30:14,493
these are just the best working
ones kind of thing. Where do you

543
00:30:14,493 --> 00:30:17,825
imagine there's a lot of room
for improvement? I'm sure there

544
00:30:17,825 --> 00:30:20,938
is. But, what are kind of what
are the things that you're

545
00:30:20,938 --> 00:30:24,215
excited about trying or I don't
know, where do you see there

546
00:30:24,215 --> 00:30:27,000
being like, low hanging fruit
perhaps in this area?

547
00:30:27,000 --> 00:30:30,000
Jenny Read: Yeah. I hesitate to
describe any fruit as low

548
00:30:30,000 --> 00:30:33,158
hanging in this area, but I
think I already mentioned middle

549
00:30:33,158 --> 00:30:35,000
hanging. Yeah, I already
mentioned.

550
00:30:35,000 --> 00:30:36,000
Audrow Nash: Lowest possible
hanging.

551
00:30:36,000 --> 00:30:40,023
Jenny Read: Yeah. Assembly
tasks. So, the sort of things

552
00:30:40,023 --> 00:30:44,118
where you're trying to clip
something on or maybe stretch

553
00:30:44,118 --> 00:30:48,428
and pull over. So. Right. You
may have to handle, deformable

554
00:30:48,428 --> 00:30:53,026
materials. You may want to, you
know, it's plugging in cables in

555
00:30:53,026 --> 00:30:57,264
the cable harness. I think all
of those things, it's really

556
00:30:57,264 --> 00:31:01,000
going to be challenging to do
with a pincer gripper.

557
00:31:01,000 --> 00:31:02,000
Audrow Nash: Yes.

558
00:31:02,000 --> 00:31:04,000
Jenny Read: Okay. Disassembly
using.

559
00:31:04,000 --> 00:31:07,710
Audrow Nash: Mostly on
deformable materials within

560
00:31:07,710 --> 00:31:12,237
manipulation space because that
that is something I don't see

561
00:31:12,237 --> 00:31:16,912
any good. I haven't I mean, I'm
not terribly in the space, but,

562
00:31:16,912 --> 00:31:21,438
I haven't seen any compelling,
like, solutions for deformable

563
00:31:21,438 --> 00:31:25,000
objects. Like, that's a very
hard area for sure.

564
00:31:25,000 --> 00:31:28,256
Jenny Read: What's interesting
to hear? I wouldn't say we're

565
00:31:28,256 --> 00:31:31,729
focusing on it specifically, but
it's definitely something we're

566
00:31:31,729 --> 00:31:34,930
aiming to address as part of
this. For example, taking care

567
00:31:34,930 --> 00:31:38,240
that the simulation techniques
we develop can hand deformable

568
00:31:38,240 --> 00:31:39,000
objects. Yeah.

569
00:31:39,000 --> 00:31:42,574
Audrow Nash: Yeah, that's really
hard. Like it seems like people

570
00:31:42,574 --> 00:31:45,870
work really hard to develop very
good simulation models for

571
00:31:45,870 --> 00:31:49,388
deformable things. And if they
work well, they're very slow and

572
00:31:49,388 --> 00:31:52,907
then that doesn't. That's kind
of the opposite of what you need

573
00:31:52,907 --> 00:31:56,090
if you want to train things on
it. So say you want like a

574
00:31:56,090 --> 00:31:59,609
reinforcement learning circuit
in there. Then that really limit

575
00:31:59,609 --> 00:32:00,000
to you.

576
00:32:00,000 --> 00:32:01,000
Jenny Read: Yeah.

577
00:32:01,000 --> 00:32:04,735
Audrow Nash: Okay. Yeah. Well
that's really cool. Okay. So we

578
00:32:04,735 --> 00:32:08,286
have those big areas, and we
have 26 companies that you're

579
00:32:08,286 --> 00:32:10,000
funding and they're getting.

580
00:32:10,000 --> 00:32:15,000
Jenny Read: Some part of what
the North.

581
00:32:15,000 --> 00:32:18,432
Audrow Nash: Companies and labs,
just organizations in general,

582
00:32:18,432 --> 00:32:21,755
that you're funding. What's the
timeline for this, or are you

583
00:32:21,755 --> 00:32:24,914
giving away all of the 50
million right at once, or is it,

584
00:32:24,914 --> 00:32:28,237
over several years, you start to
give more of it, or how does

585
00:32:28,237 --> 00:32:29,000
this all work?

586
00:32:29,000 --> 00:32:32,779
Jenny Read: Yeah, we're
basically committing it, all at

587
00:32:32,779 --> 00:32:37,107
once. These, each creator has a
particular timeline, typically,

588
00:32:37,107 --> 00:32:40,885
but ranging from about two to
about four years for this

589
00:32:40,885 --> 00:32:45,076
program. So we'll be spending
the money over that period. But

590
00:32:45,076 --> 00:32:47,000
we're allocate doing it now.

591
00:32:47,000 --> 00:32:50,477
Audrow Nash: That's awesome.
Okay. And so if a it with your

592
00:32:50,477 --> 00:32:53,836
allocation how do you like one
of the things that venture

593
00:32:53,836 --> 00:32:57,372
capitalists do that I think is
kind of clever is they fund a

594
00:32:57,372 --> 00:33:00,849
bunch of startups a little bit,
or sometimes they'll fund a

595
00:33:00,849 --> 00:33:04,149
bunch of startups a little bit
later rounds that make it

596
00:33:04,149 --> 00:33:07,449
different, especially early
ones. They'll fund a bunch a

597
00:33:07,449 --> 00:33:10,867
little bit and then they'll
reinvest potentially later as,

598
00:33:10,867 --> 00:33:14,403
traction is being shown with the
startups. Will you be doing

599
00:33:14,403 --> 00:33:17,939
anything similar or is it all
allocated right away or how do

600
00:33:17,939 --> 00:33:19,000
you think of that?

601
00:33:19,000 --> 00:33:22,847
Jenny Read: We have a few where
we are. We've got a couple we

602
00:33:22,847 --> 00:33:26,568
were explicitly allocating.
Exactly. So either we have a go

603
00:33:26,568 --> 00:33:30,542
no go point where we'll decide
whether to run dog in most cases

604
00:33:30,542 --> 00:33:33,947
we have a we're making the
allocation. Yeah. Upfront I

605
00:33:33,947 --> 00:33:37,605
should say all our funding. We
remain very involved in the

606
00:33:37,605 --> 00:33:40,948
delivery of the projects and
we'll be discussing with

607
00:33:40,948 --> 00:33:44,606
creators, you know, whether
things are going well, whether

608
00:33:44,606 --> 00:33:48,453
we need to pivot, worst case,
potentially terminating if it's

609
00:33:48,453 --> 00:33:52,048
turning out things or not
working quite well. Yeah. Which

610
00:33:52,048 --> 00:33:55,832
is obviously to be expected.
Right. It's high risk. Oh yeah.

611
00:33:55,832 --> 00:33:59,238
So you know you build in well
let's try this and put a

612
00:33:59,238 --> 00:34:03,085
milestone here and see if we
reached that point. So depending

613
00:34:03,085 --> 00:34:06,743
how things pan out then if we've
decided not to pursue one

614
00:34:06,743 --> 00:34:10,149
particular project that
potentially frees up resources

615
00:34:10,149 --> 00:34:14,122
to then double down on another
project with. But we're somewhat

616
00:34:14,122 --> 00:34:18,033
limited in our ability to, you
know, if we have a large amount

617
00:34:18,033 --> 00:34:22,132
of follow on funding within this
program because each program has

618
00:34:22,132 --> 00:34:25,916
a particular budget, what it
will say is, I think looking at

619
00:34:25,916 --> 00:34:29,637
offer to go back to what we can
learn from that. One of the

620
00:34:29,637 --> 00:34:33,484
things that's impressed me about
the offer is how they have a

621
00:34:33,484 --> 00:34:37,205
great record of pursuing ideas
over a very long time scale,

622
00:34:37,205 --> 00:34:40,927
over decades. I would love to
see how you're doing that. So

623
00:34:40,927 --> 00:34:44,395
even though program directors
tenure is limited, so was

624
00:34:44,395 --> 00:34:48,179
appointed for initially three
years, it will be a maximum of

625
00:34:48,179 --> 00:34:52,090
five years and then I'll move
on. But maybe another program to

626
00:34:52,090 --> 00:34:56,000
will come into this space and
then they can double down. Yeah.

627
00:34:56,000 --> 00:34:59,795
Audrow Nash: Interesting, Is
that a common thing with RPA? I

628
00:34:59,795 --> 00:35:03,527
mean, I've talked with like I
talked with Tim Chung a while

629
00:35:03,527 --> 00:35:06,880
ago who was, DARPA program
director. He did the sub t

630
00:35:06,880 --> 00:35:10,295
subterranean challenge, a bunch
of robots in caves and

631
00:35:10,295 --> 00:35:14,343
underground environments. And he
is no longer there, so, I don't

632
00:35:14,343 --> 00:35:17,000
know, maybe their tenure is not
that long.

633
00:35:17,000 --> 00:35:20,053
Jenny Read: For this kind
funding. Yeah, that that's a key

634
00:35:20,053 --> 00:35:21,000
part of the model.

635
00:35:21,000 --> 00:35:23,269
Audrow Nash: The churn. So you
keep getting fresh ideas and

636
00:35:23,269 --> 00:35:24,000
that's pretty cool.

637
00:35:24,000 --> 00:35:26,000
Jenny Read: Yeah, I think it's
cool.

638
00:35:26,000 --> 00:35:29,588
Audrow Nash: So you have these
companies they're funded for 2

639
00:35:29,588 --> 00:35:33,059
to 4. I keep saying companies
you have these organizations,

640
00:35:33,059 --> 00:35:36,706
they're funded for 2 to 4 years.
They're doing these high risk

641
00:35:36,706 --> 00:35:40,294
things. Now, you said they're
not all in the UK or it's not a

642
00:35:40,294 --> 00:35:43,294
requirement that you fund
companies that go against

643
00:35:43,294 --> 00:35:46,588
organizations in the UK. How are
you working with global

644
00:35:46,588 --> 00:35:49,941
communities or like global
organizations that may want to

645
00:35:49,941 --> 00:35:53,353
apply for this kind of thing?
And what requirements do you

646
00:35:53,353 --> 00:35:57,000
have? Like do they have to move
to the UK or how does it work.

647
00:35:57,000 --> 00:36:00,660
Jenny Read: That they
demonstrate UK benefit? Which

648
00:36:00,660 --> 00:36:04,608
could be through offering
something really essential to

649
00:36:04,608 --> 00:36:08,340
the program? We certainly
encourage them to consider

650
00:36:08,340 --> 00:36:12,000
moving to the UK or setting up
an office in the UK.

651
00:36:12,000 --> 00:36:15,298
Audrow Nash: Or is it anywhere
in the UK or is it like, do you

652
00:36:15,298 --> 00:36:17,000
guys have a hub in one location?

653
00:36:17,000 --> 00:36:20,645
Jenny Read: No, I mean, we have
our headquarters in London, but

654
00:36:20,645 --> 00:36:23,711
we, in many ways I would
actually say is because, you

655
00:36:23,711 --> 00:36:24,000
know.

656
00:36:24,000 --> 00:36:26,000
Audrow Nash: Space constraints
and things.

657
00:36:26,000 --> 00:36:29,624
Jenny Read: Just like, you know,
geographic diversity, we value

658
00:36:29,624 --> 00:36:33,248
diversity of any axes. So if you
want to come to the UK but not

659
00:36:33,248 --> 00:36:36,642
London, then that's great. You
know. So yeah, it's about UK

660
00:36:36,642 --> 00:36:39,000
benefit not being in any
particular place

661
00:36:39,000 --> 00:36:42,073
Audrow Nash: Do you have any
like what. Do you have any

662
00:36:42,073 --> 00:36:44,642
examples of company,
organizations that you're

663
00:36:44,642 --> 00:36:47,994
considering that are not in the
UK? And what kind of problem

664
00:36:47,994 --> 00:36:49,000
space are they in?

665
00:36:49,000 --> 00:36:53,989
Jenny Read: Yeah. So one company
was a nonprofit is basis. It's.

666
00:36:53,989 --> 00:36:58,745
Yeah, it's a nonprofit applied
research organization studying

667
00:36:58,745 --> 00:37:03,578
intelligence. It's currently, in
New York, but, it's following

668
00:37:03,578 --> 00:37:08,255
the, are the funding, which, as
I say, when you negotiate in

669
00:37:08,255 --> 00:37:13,323
that funding to set up an office
in the UK and they'll be working

670
00:37:13,323 --> 00:37:17,220
with other creators on this
co-design optimization

671
00:37:17,220 --> 00:37:18,000
challenge.

672
00:37:18,000 --> 00:37:19,000
Audrow Nash: Okay.

673
00:37:19,000 --> 00:37:24,200
Jenny Read: That really cool.
Yeah, yeah. Go ahead. Because

674
00:37:24,200 --> 00:37:29,400
another example is, a company in
Denmark, plastics, startup

675
00:37:29,400 --> 00:37:34,071
that's developing, electro
elastic actuators for soft

676
00:37:34,071 --> 00:37:39,007
motion. Cool. Yeah. So it'll
also be working with the UK

677
00:37:39,007 --> 00:37:44,031
creators and enabling UK
creators to have early access to

678
00:37:44,031 --> 00:37:45,000
prototypes.

679
00:37:45,000 --> 00:37:48,730
Audrow Nash: Very cool. Let's
see. So what are some other

680
00:37:48,730 --> 00:37:52,330
examples of organizations that
you have that are in the

681
00:37:52,330 --> 00:37:56,126
negotiations I suppose. Like
what are some promising ones.

682
00:37:56,126 --> 00:38:00,183
What are you. I mean I'm sure
they're all promising, you know.

683
00:38:00,183 --> 00:38:04,306
What are some I don't know, it's
it's kind of spanned the space

684
00:38:04,306 --> 00:38:07,971
of different organizations
you're planning to work with.

685
00:38:07,971 --> 00:38:10,000
I'd love to hear some examples.

686
00:38:10,000 --> 00:38:14,413
Jenny Read: Yeah. So one is I
could name a small company d z p

687
00:38:14,413 --> 00:38:18,470
technologies, founded by
materials chemist called Slacker

688
00:38:18,470 --> 00:38:22,527
Stover. And they're innovating
in printed electronics. So

689
00:38:22,527 --> 00:38:26,655
they're looking to create new
components, robotics such as

690
00:38:26,655 --> 00:38:30,000
stretchable and flexible and
tronic components.

691
00:38:30,000 --> 00:38:31,000
Audrow Nash: Oh, that'd be
really nice.

692
00:38:31,000 --> 00:38:35,151
Jenny Read: Right. Exactly. No
need to try to build skin on

693
00:38:35,151 --> 00:38:39,303
robots then. Those kind of
components be handy. So that's a

694
00:38:39,303 --> 00:38:43,735
nice example. I think of. One of
the things I wanted to do with

695
00:38:43,735 --> 00:38:48,098
this program was bring people
from outside robotics to work on

696
00:38:48,098 --> 00:38:52,319
the problem, and that's a good
example. I think. Another one

697
00:38:52,319 --> 00:38:56,048
we've got. So it's some
academics also working on the

698
00:38:56,048 --> 00:39:00,622
co-design, problem. So we have a
johns at Imperial College London

699
00:39:00,622 --> 00:39:04,562
and they the Bristol are
collaborating closely together,

700
00:39:04,562 --> 00:39:08,502
on that challenge, looking to
develop new techniques for

701
00:39:08,502 --> 00:39:12,302
optimizing, robot bodies and
control policies and then

702
00:39:12,302 --> 00:39:16,242
bridging the SIM to VR gap by
covertly prototyping robot

703
00:39:16,242 --> 00:39:20,252
manipulators, you know, seeing
how well, how closely they

704
00:39:20,252 --> 00:39:24,052
performed what was expected.
Cool. Yeah. So they'll be

705
00:39:24,052 --> 00:39:27,640
working with basis and with
Quaker and Nova. It's a

706
00:39:27,640 --> 00:39:31,862
researcher at Cambridge. Again,
all bringing their different

707
00:39:31,862 --> 00:39:35,872
skills and unique talents to
bear on different aspects of

708
00:39:35,872 --> 00:39:40,094
this challenge. So there's the
company I could mention, it's

709
00:39:40,094 --> 00:39:44,105
wave drives. So that's a, small
company near Bristol, and

710
00:39:44,105 --> 00:39:48,326
they're developing a really
interesting non-contact magnetic

711
00:39:48,326 --> 00:39:51,000
magnetic transmission linear
actuator.

712
00:39:51,000 --> 00:39:53,333
Audrow Nash: That's going to
take me a minute to unpack.

713
00:39:53,333 --> 00:39:55,000
Non-contact. What is that? Say
it again.

714
00:39:55,000 --> 00:39:58,346
Jenny Read: So it's a linear
actuator. Think like a piston

715
00:39:58,346 --> 00:40:01,865
sliding in and out, but it's not
actually in contact with the

716
00:40:01,865 --> 00:40:04,000
cylinder around it. Oh they tell
the.

717
00:40:04,000 --> 00:40:05,000
Audrow Nash: Magnets to do that.

718
00:40:05,000 --> 00:40:07,871
Jenny Read: And Yeah. That's
right. So there are the

719
00:40:07,871 --> 00:40:10,908
davantage of that. There is
there isn't friction. There

720
00:40:10,908 --> 00:40:14,000
isn't where there isn't a need
for lubrication. So it's.

721
00:40:14,000 --> 00:40:18,388
Audrow Nash: Yeah, interesting
sounds finicky to me, but it

722
00:40:18,388 --> 00:40:23,000
could be very, very useful to.
Yeah. No friction and all that.

723
00:40:23,000 --> 00:40:26,608
Jenny Read: It's likely they'd
be developing their technology

724
00:40:26,608 --> 00:40:30,038
in areas like aerospace and
prosthetics. And obviously for

725
00:40:30,038 --> 00:40:33,231
this program they're developing
it for robot dexterity

726
00:40:33,231 --> 00:40:34,000
specifically.

727
00:40:34,000 --> 00:40:37,017
Audrow Nash: What a cool thing I
can imagine. Like, in like sci

728
00:40:37,017 --> 00:40:39,890
fi movies and things, when you
have like a floating arm that

729
00:40:39,890 --> 00:40:42,476
you can control for all this
thing because magnets are

730
00:40:42,476 --> 00:40:45,349
holding it up, like that kind of
thing would be super, super

731
00:40:45,349 --> 00:40:48,318
cool. I wonder if we'll ever get
to something like that. Okay,

732
00:40:48,318 --> 00:40:51,000
these are great. And any other
ones you want to mention?

733
00:40:51,000 --> 00:40:54,532
Jenny Read: Well, you were
talking about magnets and flow

734
00:40:54,532 --> 00:40:58,497
too. It reminded me of a project
that's actually not part of the

735
00:40:58,497 --> 00:41:02,091
dexterity program itself, but
part of our seed funding. So

736
00:41:02,091 --> 00:41:05,437
another thing we do, oh yeah, is
that within the wider

737
00:41:05,437 --> 00:41:09,030
opportunity space, in my case,
robot hardware, we can make

738
00:41:09,030 --> 00:41:12,686
small seed awards, to explore
really interesting ideas. And

739
00:41:12,686 --> 00:41:16,651
one of the seed awards we made,
is it again interesting to bring

740
00:41:16,651 --> 00:41:19,997
people from outside robotics? So
this is a chap called

741
00:41:19,997 --> 00:41:23,839
Christopher Wool. He's actually
a cardiologist. Yeah. And he's

742
00:41:23,839 --> 00:41:27,618
founded a company called Colored
Cardiology Devices. And he's

743
00:41:27,618 --> 00:41:31,150
looking at, developing a
ventricular assist device to aid

744
00:41:31,150 --> 00:41:34,000
patients with heart failure. But
it's talk to.

745
00:41:34,000 --> 00:41:37,000
Audrow Nash: A catheter or what
kind of things?

746
00:41:37,000 --> 00:41:41,128
Jenny Read: No, I don't believe
so. I think just to aid the

747
00:41:41,128 --> 00:41:45,185
blood flow. Although. Don't kiss
me too closely. Oh, yeah.

748
00:41:45,185 --> 00:41:49,453
Exactly. Cool. So based on that,
he got interested in motors,

749
00:41:49,453 --> 00:41:53,580
right? Because he needs a motor
for that purpose. And so he

750
00:41:53,580 --> 00:41:57,778
applied for seed funding for
ARIA to develop, a bearing less

751
00:41:57,778 --> 00:42:01,975
motor. So, yeah, rather than
resting or bearing C devotes is

752
00:42:01,975 --> 00:42:06,033
essentially suspended in a
magnetic field. So again the is

753
00:42:06,033 --> 00:42:10,230
a rotary motor this time. But
the same idea that you don't a

754
00:42:10,230 --> 00:42:11,000
cool. Yeah.

755
00:42:11,000 --> 00:42:14,139
Audrow Nash: Yeah I can't wait
till that stuff works. Yeah. No

756
00:42:14,139 --> 00:42:15,000
friction nowhere.

757
00:42:15,000 --> 00:42:16,000
Jenny Read: Right.

758
00:42:15,000 --> 00:42:17,940
Audrow Nash: It's bearings not
seizing up at some point. That

759
00:42:17,940 --> 00:42:19,000
would be so wonderful.

760
00:42:19,000 --> 00:42:20,000
Jenny Read: Exactly.

761
00:42:20,000 --> 00:42:22,000
Audrow Nash: Gosh, we're getting
into the future quickly.

762
00:42:22,000 --> 00:42:25,726
Jenny Read: Is it? And so you
can see the strategies that

763
00:42:25,726 --> 00:42:29,779
we've got a kind of portfolio of
actuation techniques, some of

764
00:42:29,779 --> 00:42:33,505
which are different forms of
motors and some of which are

765
00:42:33,505 --> 00:42:36,316
completely different
technologies. Like the

766
00:42:36,316 --> 00:42:40,173
electrode, active isolators or
the Hazel actuators rules of

767
00:42:40,173 --> 00:42:44,161
funding. A US company called
Artemus that's developing, hazel

768
00:42:44,161 --> 00:42:47,822
hydraulically amplified soft
electrostatic actuators. So

769
00:42:47,822 --> 00:42:51,613
we're kind of spreading out
batsmen. So a lot of different

770
00:42:51,613 --> 00:42:54,947
actuation technologies. Cool.
Very cool. And ensure

771
00:42:54,947 --> 00:42:59,000
roboticists in the future have
more components to choose from.

772
00:42:59,000 --> 00:43:02,911
Audrow Nash: That would be
awesome. How do you. So one of

773
00:43:02,911 --> 00:43:06,342
the things that's a big
bottleneck because so this

774
00:43:06,342 --> 00:43:10,596
sounds like it's very wonderful.
And future focused for this I

775
00:43:10,596 --> 00:43:14,987
view a massive bottleneck being
like you come up with an amazing

776
00:43:14,987 --> 00:43:18,693
way to do motors that are
frictionless. So, it's still

777
00:43:18,693 --> 00:43:22,947
like manufacturing capacity is a
tremendous difficulty. Like a

778
00:43:22,947 --> 00:43:27,269
lot of the research stuff, like
maybe it'll work once, but it's

779
00:43:27,269 --> 00:43:31,249
like way too expensive to
manufacture or it's just hard to

780
00:43:31,249 --> 00:43:35,503
manufacture, especially with,
like, aging demographics and all

781
00:43:35,503 --> 00:43:39,071
sorts of, like, everything
getting more expensive in

782
00:43:39,071 --> 00:43:43,325
general. What, do you have any.
Are you guys maybe it's out of

783
00:43:43,325 --> 00:43:46,893
focus, but what are your
thoughts around potentially

784
00:43:46,893 --> 00:43:51,147
trying to help the manufacturing
bottleneck that some of these

785
00:43:51,147 --> 00:43:53,000
technologies may encounter?

786
00:43:53,000 --> 00:43:55,923
Jenny Read: Yeah. You right.
It's definitely something we

787
00:43:55,923 --> 00:43:59,051
need to think about. I think to
be honest, that probably when

788
00:43:59,051 --> 00:44:01,718
you really get close to
manufacturing at scale, that

789
00:44:01,718 --> 00:44:04,846
probably tends to require more
resources that we're likely to

790
00:44:04,846 --> 00:44:07,974
be able to make it very true.
Right. So we would be aiming to

791
00:44:07,974 --> 00:44:10,795
get to the point where it's
attractive enough for other

792
00:44:10,795 --> 00:44:13,872
people to push it. Yeah, but
we're definitely thinking about

793
00:44:13,872 --> 00:44:16,744
how can we demonstrate the
portability of that. Yeah. So

794
00:44:16,744 --> 00:44:19,872
we're working, for example, with
the Manufacturing Technology

795
00:44:19,872 --> 00:44:22,846
Center in the UK. Oh, which
basically exists for this kind

796
00:44:22,846 --> 00:44:25,615
of reason, to try and help
companies figure out how to

797
00:44:25,615 --> 00:44:27,000
manufacture what they need.

798
00:44:27,000 --> 00:44:31,113
Audrow Nash: Okay. I'm pretty
ignorant how how significant

799
00:44:31,113 --> 00:44:35,156
like, does the UK do a lot of
manufacturing or does it? I

800
00:44:35,156 --> 00:44:37,000
don't know much I suppose.

801
00:44:37,000 --> 00:44:39,353
Jenny Read: Yeah, no we do. We
have quite a strong

802
00:44:39,353 --> 00:44:41,000
manufacturing sector actually.
Yep.

803
00:44:41,000 --> 00:44:44,955
Audrow Nash: Yeah. Awesome.
Okay. So those are all great

804
00:44:44,955 --> 00:44:49,263
examples of different companies
and initiatives. It's cool to

805
00:44:49,263 --> 00:44:53,359
hear that you guys. I mean,
you're leveraging and thinking

806
00:44:53,359 --> 00:44:57,737
about how it goes once it gets
out of kind of the research and

807
00:44:57,737 --> 00:45:01,268
early stages. And then are
actually are any of the

808
00:45:01,268 --> 00:45:05,576
technologies more mature or are
they all very high risk, very

809
00:45:05,576 --> 00:45:06,000
early?

810
00:45:06,000 --> 00:45:10,364
Jenny Read: Oh they vary. So
yeah, they're not all very high

811
00:45:10,364 --> 00:45:14,291
risk. If you say Artemus, the
company making haze like

812
00:45:14,291 --> 00:45:18,582
traitors, they already have some
products out. And the what

813
00:45:18,582 --> 00:45:23,018
they're doing is, I would say,
you know, towards the close to

814
00:45:23,018 --> 00:45:26,945
market and, then, you know, we
have Tawfik Hassan that

815
00:45:26,945 --> 00:45:31,164
Cambridge is developing an
electronic skin and that's very

816
00:45:31,164 --> 00:45:35,527
early. So this is a Cambridge
University and that's very lab

817
00:45:35,527 --> 00:45:38,000
based. So yeah, we fund the
range.

818
00:45:38,000 --> 00:45:42,487
Audrow Nash: Oh, yeah. Okay. So
for all of this, how do you

819
00:45:42,487 --> 00:45:46,594
measure success? Like how do you
know if the, batch of

820
00:45:46,594 --> 00:45:51,006
organizations has gone well or
how can we see if it's, the

821
00:45:51,006 --> 00:45:55,721
whole initiative was a success?
And also, I'm assuming then we

822
00:45:55,721 --> 00:46:00,056
use that idea of success to
justify future, programs. And

823
00:46:00,056 --> 00:46:04,392
continuing in this vein, maybe.
Or how do you think about

824
00:46:04,392 --> 00:46:05,000
success?

825
00:46:05,000 --> 00:46:08,927
Jenny Read: Yeah, that's a
really tough one. Right? It's

826
00:46:08,927 --> 00:46:13,415
famously hard to assess success.
I think it becomes clearer with

827
00:46:13,415 --> 00:46:17,903
hindsight. So I think, yeah, the
thing we say about, yeah, is we

828
00:46:17,903 --> 00:46:21,971
should be able in a generation
to look back and it just be

829
00:46:21,971 --> 00:46:26,529
completely obvious. How are we a
funded research has, transformed

830
00:46:26,529 --> 00:46:30,736
society and benefited the UK.
And yeah, I get back to DARPA.

831
00:46:30,736 --> 00:46:35,014
But you look at that, I sort of
point to the internet or many

832
00:46:35,014 --> 00:46:39,432
vaccines and so on. In the short
term, obviously, we're looking

833
00:46:39,432 --> 00:46:43,780
for signals that we're heading
in that kind of direction. So I

834
00:46:43,780 --> 00:46:47,987
guess it's have demonstrating
new capabilities. New products

835
00:46:47,987 --> 00:46:52,546
that we can say. Yeah, you could
see that they are we are funding

836
00:46:52,546 --> 00:46:55,000
was critical in getting them
there.

837
00:46:55,000 --> 00:46:58,397
Audrow Nash: Awesome. Yeah. It
is a hard thing to measure with

838
00:46:58,397 --> 00:47:01,575
all these especially in the
really short term. But I think

839
00:47:01,575 --> 00:47:05,082
it'd be really cool to see some
major things you can point to in

840
00:47:05,082 --> 00:47:07,000
a generation in this kind of
thing.

841
00:47:07,000 --> 00:47:10,374
Jenny Read: Absolutely. And in
the shorter term I think things

842
00:47:10,374 --> 00:47:13,422
just like new communities for
Asians, potentially no new

843
00:47:13,422 --> 00:47:15,000
interest in robotic hardware.

844
00:47:15,000 --> 00:47:18,444
Audrow Nash: Definitely. Yeah. I
think one of the things that's

845
00:47:18,444 --> 00:47:21,342
really exciting about what
you're doing to me is that

846
00:47:21,342 --> 00:47:24,349
you're getting people that are
not involved in robotics

847
00:47:24,349 --> 00:47:27,793
traditionally into this. And so
what it's doing, in my opinion,

848
00:47:27,793 --> 00:47:31,128
or from my perspective, would be
getting them exposed to this

849
00:47:31,128 --> 00:47:34,572
world that I think is going to
be very impactful in the future.

850
00:47:34,572 --> 00:47:37,633
And so we can have more hands
helping out in robotics in

851
00:47:37,633 --> 00:47:39,000
general from all of this.

852
00:47:39,000 --> 00:47:41,000
Jenny Read: But yeah. Exactly.

853
00:47:41,000 --> 00:47:45,382
Audrow Nash: How does so you
mentioned at the beginning that

854
00:47:45,382 --> 00:47:49,326
the this is already the full
program or is it like you

855
00:47:49,326 --> 00:47:53,124
mentioned it started in 2023,
this initiative, is it

856
00:47:53,124 --> 00:47:57,213
guaranteed to continue. Like is
it a commitment from the

857
00:47:57,213 --> 00:48:01,230
government to continue for
decades, like you're saying,

858
00:48:01,230 --> 00:48:05,466
measure in a generation? Or how
are we, thinking about the

859
00:48:05,466 --> 00:48:07,000
future with all this?

860
00:48:07,000 --> 00:48:10,781
Jenny Read: So are we was set up
for ten years. Pilot initial to

861
00:48:10,781 --> 00:48:14,148
parliament. And then it'll be
reviewed. Obviously, we are

862
00:48:14,148 --> 00:48:17,633
hoping that by that ten year
point, there will be enough to

863
00:48:17,633 --> 00:48:21,000
point to the government. Realize
this is really the case.

864
00:48:21,000 --> 00:48:22,000
Audrow Nash: Make it bigger.

865
00:48:22,000 --> 00:48:25,649
Jenny Read: Make it bigger, make
them carry on. But at this

866
00:48:25,649 --> 00:48:28,000
point, it's not a ten year
commitment.

867
00:48:28,000 --> 00:48:32,623
Audrow Nash: Okay. Let's see.
And then what are you so

868
00:48:32,623 --> 00:48:37,759
thinking about the listener.
What do you hope they take away

869
00:48:37,759 --> 00:48:43,066
from this. Or is there any like
like should they join the next

870
00:48:43,066 --> 00:48:48,459
batch or like what are kind of
calls to action for anyone who's

871
00:48:48,459 --> 00:48:50,000
listening to this?

872
00:48:50,000 --> 00:48:53,311
Jenny Read: That's a great
point. I think part of what I

873
00:48:53,311 --> 00:48:56,918
would like to take away is that
they've now heard about ARIA,

874
00:48:56,918 --> 00:49:00,230
which maybe they hadn't in the
past, and maybe go on our

875
00:49:00,230 --> 00:49:03,955
website and sign up for updates.
We have, as I mentioned, seven

876
00:49:03,955 --> 00:49:07,562
programs running at the moment.
Some closed calls are closed,

877
00:49:07,562 --> 00:49:10,932
some are open. We've also
recruiting our second cohort of

878
00:49:10,932 --> 00:49:14,480
program directors, who'll be
starting early in the new year.

879
00:49:14,480 --> 00:49:18,323
So there'll be a second round of
programs coming out. So yeah, if

880
00:49:18,323 --> 00:49:22,108
you sign up for updates, you'll
find out what those are and, you

881
00:49:22,108 --> 00:49:24,000
know, potentially want to apply.

882
00:49:24,000 --> 00:49:26,993
Audrow Nash: Yeah, I'll put a
link if it's out. When we

883
00:49:26,993 --> 00:49:30,313
published this in the not so
distant future, I'll put a link,

884
00:49:30,313 --> 00:49:33,524
so that people can find it more
easily to sign up, but then

885
00:49:33,524 --> 00:49:36,517
also, maybe to the submission,
because that'd be really

886
00:49:36,517 --> 00:49:40,000
exciting to have more people in
the innovation space in general.

887
00:49:40,000 --> 00:49:43,386
Jenny Read: I think that'd be
great. And I think for my

888
00:49:43,386 --> 00:49:47,080
program specifically, I would
say, you know, please get back

889
00:49:47,080 --> 00:49:50,774
to me with any feedback about
what you've heard, if you have

890
00:49:50,774 --> 00:49:54,529
any, like, hardware needs, if
you're excited about any of the

891
00:49:54,529 --> 00:49:58,408
work that I'm funding that I've
mentioned, please get in touch.

892
00:49:58,408 --> 00:50:02,163
We aim to have some workshops
and dissemination activities to

893
00:50:02,163 --> 00:50:05,857
talk to people about what we're
doing, and obviously will be

894
00:50:05,857 --> 00:50:09,058
super keen to get any new
components or actuators or

895
00:50:09,058 --> 00:50:12,998
sensors or whatever out and try
it out and test for free flow of

896
00:50:12,998 --> 00:50:16,754
interest in them. So, you know,
please get in touch if that's

897
00:50:16,754 --> 00:50:17,000
you.

898
00:50:17,000 --> 00:50:22,083
Audrow Nash: Yeah, that would be
awesome. Do you have like so

899
00:50:22,083 --> 00:50:26,917
from your perspective, three
years or five years when your

900
00:50:26,917 --> 00:50:31,917
tenure is over for the ARIA
program? What do you, what would

901
00:50:31,917 --> 00:50:36,583
be like success, in your
opinion? Like, what do you hope

902
00:50:36,583 --> 00:50:39,000
you've gotten to in the time?

903
00:50:39,000 --> 00:50:43,248
Jenny Read: Great question. I
think it's going to I'm really

904
00:50:43,248 --> 00:50:47,637
hoping that the creators that
I'm funding are going to achieve

905
00:50:47,637 --> 00:50:52,027
something amazing, and I'll just
be so excited if even half of

906
00:50:52,027 --> 00:50:55,000
what they've proposed comes off.
Oh, yeah.

907
00:50:55,000 --> 00:50:58,000
Audrow Nash: Yeah, yeah, that
would be awesome.

908
00:50:58,000 --> 00:51:01,919
Jenny Read: I think if we could
have them. Yeah. I think, if we

909
00:51:01,919 --> 00:51:05,278
could demonstrate the value of
the co-design approach,

910
00:51:05,278 --> 00:51:09,196
demonstrate that, yes, we have
actually reached the point where

911
00:51:09,196 --> 00:51:13,053
simulation is powerful enough
that this is an entirely new way

912
00:51:13,053 --> 00:51:16,411
of designing robots and
controlling them. I think that

913
00:51:16,411 --> 00:51:20,330
would be amazing. If robots are
going around with tactile skins

914
00:51:20,330 --> 00:51:24,000
that were developed, funded by
ARIA, that would be awesome.

915
00:51:24,000 --> 00:51:27,033
Audrow Nash: Yes. Skins would be
a big value add, I think,

916
00:51:27,033 --> 00:51:30,275
especially putting robots in the
home and things like this and

917
00:51:30,275 --> 00:51:33,464
simulation. I feel like if you
made significant contributions

918
00:51:33,464 --> 00:51:36,706
to the simulation space, that
would just be that would open up

919
00:51:36,706 --> 00:51:40,000
so much. I think simulation is a
real bottleneck in my opinion.

920
00:51:40,000 --> 00:51:41,000
Jenny Read: Interested.

921
00:51:41,000 --> 00:51:43,586
Audrow Nash: So anyways, Jenny
it was wonderful talking with

922
00:51:43,586 --> 00:51:46,000
you and thank you for telling me
about the ARIA program.

923
00:51:46,000 --> 00:51:49,000
Jenny Read: Great. Thank you.
It's been an absolute pleasure.

924
00:51:49,000 --> 00:51:53,461
Audrow Nash: Bye everyone.
That's all for my conversation

925
00:51:53,461 --> 00:51:58,000
with Jenny Read. I hope you
enjoyed and see you next time.
