1
00:00:00,000 --> 00:00:03,399
Fred Parietti: Try to convince a
scientist to bring a

2
00:00:03,399 --> 00:00:07,311
teleoperation rig into a clean
room, right? and then use, you

3
00:00:07,311 --> 00:00:11,415
know, not their own hands, but a
teleoperation rig to manipulate

4
00:00:11,415 --> 00:00:15,455
the cells. They're going to tell
you you're crazy, right? Yeah.

5
00:00:15,455 --> 00:00:19,174
And I can tell you how many
robots I saw that make pizzas.

6
00:00:19,174 --> 00:00:23,343
And I'm like, guys, right? Like,
you know, there are two kinds of

7
00:00:23,343 --> 00:00:27,319
pizzas. Either pizzas for which
minimum wage labor is fine and

8
00:00:27,319 --> 00:00:31,295
there are pizza for which you
really want the human experience

9
00:00:31,295 --> 00:00:32,000
of someone.

10
00:00:32,000 --> 00:00:35,000
Audrow Nash: like an incredible
chef. Yeah.

11
00:00:35,000 --> 00:00:39,364
Fred Parietti: And in both
cases, the value per pizza is

12
00:00:39,364 --> 00:00:43,649
very low. The same amount of
mass in terms of like gene

13
00:00:43,649 --> 00:00:48,636
edited cells for actually saving
people's life, has a value of a

14
00:00:48,636 --> 00:00:53,000
few million dollars. Not like
the $20 of a pizza, right?

15
00:00:53,000 --> 00:00:56,000
Audrow Nash: Alright. Hi, Fred.
Would you introduce yourself?

16
00:00:56,000 --> 00:01:00,516
Fred Parietti: Yes. Hi Audrow.
It’s great to be, you know,

17
00:01:00,516 --> 00:01:05,033
here. Yeah. My name is Fred. I'm
the co-founder and CEO at

18
00:01:05,033 --> 00:01:09,471
Multiply Labs, which is a
robotics company that develops,

19
00:01:09,471 --> 00:01:13,754
automation technology for
manufacturing next generation

20
00:01:13,754 --> 00:01:15,000
pharmaceuticals.

21
00:01:15,000 --> 00:01:17,000
Audrow Nash: Oh, go ahead.

22
00:01:17,000 --> 00:01:20,586
Fred Parietti: My background is
in mechanical engineer

23
00:01:20,586 --> 00:01:24,703
engineering, and so I've always
been a roboticist. And so, you

24
00:01:24,703 --> 00:01:28,820
know, it's very exciting to be
working in the space where, you

25
00:01:28,820 --> 00:01:32,672
know, you started and where all
the things that you always

26
00:01:32,672 --> 00:01:34,000
wanted to build are.

27
00:01:34,000 --> 00:01:37,490
Audrow Nash: Hell, yeah. Yeah.
It's so cool. And we met. And,

28
00:01:37,490 --> 00:01:40,693
like. I don't even know, it
might have been like 2014 or

29
00:01:40,693 --> 00:01:44,126
something. And I came to MIT
while. You were doing your PhD.

30
00:01:44,126 --> 00:01:47,272
You gave me a tour of your lab,
and you were working on

31
00:01:47,272 --> 00:01:50,419
supernumerary limbs, so, like,
extra arms for people in

32
00:01:50,419 --> 00:01:53,908
manufacturing. How did you get
from here? From there to here.

33
00:01:53,908 --> 00:01:56,769
Like, what was your path and how
did. You discover

34
00:01:56,769 --> 00:01:59,000
pharmaceuticals as an
interesting area?

35
00:01:59,000 --> 00:02:02,948
Fred Parietti: Yeah, that's a
very, very interesting, maybe

36
00:02:02,948 --> 00:02:07,097
unusual story. But, it's meet
me. My, my dream has always been

37
00:02:07,097 --> 00:02:11,112
to build robots. And that's all
I always wanted to do. And I

38
00:02:11,112 --> 00:02:15,059
guess I knew I was not patient
enough to be a professor. So

39
00:02:15,059 --> 00:02:17,000
that was like, off the table.

40
00:02:17,000 --> 00:02:19,000
Audrow Nash: Yeah.

41
00:02:19,000 --> 00:02:23,055
Fred Parietti: But, you know, a
robot, I want to build the,

42
00:02:23,055 --> 00:02:27,246
like, the most advanced possible
robots. And so I always knew

43
00:02:27,246 --> 00:02:31,645
that was very expensive. Right.
And so. Okay. Like, they must be

44
00:02:31,645 --> 00:02:35,218
very useful. They must do
something that's extremely

45
00:02:35,218 --> 00:02:38,998
valuable. I must be absolutely
sure that the robots I'm

46
00:02:38,998 --> 00:02:42,915
building are doing something
super valuable, because that

47
00:02:42,915 --> 00:02:46,832
would justify their cost. And
the cost of all the amazing

48
00:02:46,832 --> 00:02:50,749
engineers that need to hire to
work with me and build the

49
00:02:50,749 --> 00:02:54,803
robots. And. And so during their
PhD, I met our co-founder,

50
00:02:54,803 --> 00:02:58,514
Alice, and she was not a
robotics person. She is a now

51
00:02:58,514 --> 00:03:02,844
she's a professor. At that time,
she was also a PhD student, in

52
00:03:02,844 --> 00:03:07,173
pharmaceutical technologies. And
she told me how, manual pharma

53
00:03:07,173 --> 00:03:11,296
manufacturing is, and I could
not believe it. I just assumed

54
00:03:11,296 --> 00:03:15,145
that making drugs was like
making chips. So I assumed it

55
00:03:15,145 --> 00:03:17,000
had already been automated.

56
00:03:17,000 --> 00:03:19,211
Audrow Nash: You mean like,
silicon chips? Like computer

57
00:03:19,211 --> 00:03:20,000
chips kind of thing?

58
00:03:20,000 --> 00:03:24,419
Fred Parietti: Exactly, exactly,
exactly. Right. And so, but it

59
00:03:24,419 --> 00:03:28,908
turns out that factoring is even
more valuable than silicon, you

60
00:03:28,908 --> 00:03:32,555
know, semiconductor
manufacturing, it's probably the

61
00:03:32,555 --> 00:03:36,833
only thing that's more available
on a per month or per volume

62
00:03:36,833 --> 00:03:40,972
basis. And yet, the from the
point of view of manufacturing

63
00:03:40,972 --> 00:03:45,461
it, it's still entirely manual,
especially for the most advanced

64
00:03:45,461 --> 00:03:49,809
drugs, which are the most, you
know, difficult to make. And so

65
00:03:49,809 --> 00:03:53,667
in other words bio biological
drugs. So therapy is gene

66
00:03:53,667 --> 00:03:58,086
therapy is RNA. Those are just
very hard to make. Most of them,

67
00:03:58,086 --> 00:04:01,943
not all of them, but most of
them. And I just could not

68
00:04:01,943 --> 00:04:05,731
believe it. And so I was like,
wow, this is definitely

69
00:04:05,731 --> 00:04:10,080
justifies developing extremely
advanced robots. You know, it's

70
00:04:10,080 --> 00:04:14,358
a huge need, for the technology.
There's a huge value because

71
00:04:14,358 --> 00:04:18,286
nothing is more valuable than
saving on improving lives.

72
00:04:18,286 --> 00:04:21,442
Right. And it requires
precision. It requires

73
00:04:21,442 --> 00:04:25,019
sterility. It sounds like a
perfect application for

74
00:04:25,019 --> 00:04:29,508
robotics. It also looked at that
time like a pretty empty market

75
00:04:29,508 --> 00:04:33,857
meaning that pharma companies
didn't have and still don't have

76
00:04:33,857 --> 00:04:38,346
engineering departments. They're
really focused on approving the

77
00:04:38,346 --> 00:04:42,203
biology to the FDA and then
selling the drugs, but they

78
00:04:42,203 --> 00:04:46,552
don't have engineers. And most
robotics companies at the time,

79
00:04:46,552 --> 00:04:50,690
like we're talking about like 7
or 8 years ago, right? Most

80
00:04:50,690 --> 00:04:55,039
robotics companies then, were
focusing on lab automation. That

81
00:04:55,039 --> 00:04:58,546
was something that was
non-regulated. So they were

82
00:04:58,546 --> 00:05:02,404
focusing on non FDA stuff
because they were afraid, you

83
00:05:02,404 --> 00:05:06,402
know, of the FDA. And we
thought, wait a minute. Like FDA

84
00:05:06,402 --> 00:05:10,820
compliant manufacturing is where
the value is because those are

85
00:05:10,820 --> 00:05:15,169
the drugs that go to patients.
And those are the drugs are you

86
00:05:15,169 --> 00:05:19,377
really need to be sterile. So
they are the ones where robots

87
00:05:19,377 --> 00:05:23,726
are the most useful. So that's
how you know, you know, it came

88
00:05:23,726 --> 00:05:28,005
to be. And so we would start in
multiple labs to address that

89
00:05:28,005 --> 00:05:32,283
need through the most cutting
edge possible robotics. So very

90
00:05:32,283 --> 00:05:36,562
different from my PhD, but I
have to say, the PhD research on

91
00:05:36,562 --> 00:05:40,770
super supernumerary robotic
limbs. You know, it told me, you

92
00:05:40,770 --> 00:05:44,908
know, everything I know about
robotics. And so, it was just

93
00:05:44,908 --> 00:05:48,906
about. Okay, how do I apply
that? In a way that can help,

94
00:05:48,906 --> 00:05:52,974
the most people and quite
frankly, let us develop the most

95
00:05:52,974 --> 00:05:56,832
advanced possible robotic
technology. I like to us, you

96
00:05:56,832 --> 00:06:01,110
know, I'll be building robots
anyways. So it's about what are

97
00:06:01,110 --> 00:06:05,389
the most useful robots I can
build. Provided that I'd we also

98
00:06:05,389 --> 00:06:09,457
building them, you know, for
free on my own. Right. So, so

99
00:06:09,457 --> 00:06:11,000
that's how it started.

100
00:06:11,000 --> 00:06:14,697
Audrow Nash: Totally. And yeah,
by going for the highest value

101
00:06:14,697 --> 00:06:18,037
area, then. You have the most
justification for the most

102
00:06:18,037 --> 00:06:21,317
advanced robot and the best
people like, so you can pay

103
00:06:21,317 --> 00:06:24,000
their salaries and all the
things like. This.

104
00:06:24,000 --> 00:06:28,084
Fred Parietti: Yes I, I just
hate to be constrained by all.

105
00:06:28,084 --> 00:06:32,029
There's no budget for that
sensor. For that material or I

106
00:06:32,029 --> 00:06:36,181
hate to be constrained by oh I
somehow can not work with the

107
00:06:36,181 --> 00:06:40,057
best people and you know in
these, you know, environment

108
00:06:40,057 --> 00:06:43,933
with these application both
constraints are gone. Like I

109
00:06:43,933 --> 00:06:48,017
know I can you know, do a very
convincing pitch to the best

110
00:06:48,017 --> 00:06:52,377
engineers to be here. And I know
everybody in our team is just,

111
00:06:52,377 --> 00:06:53,000
you know.

112
00:06:53,000 --> 00:06:54,000
Audrow Nash: Super good.

113
00:06:54,000 --> 00:06:57,936
Fred Parietti: And, you know, at
that level and then at the same

114
00:06:57,936 --> 00:07:01,810
time, I never need to tell them,
oh, you know what? You know, I

115
00:07:01,810 --> 00:07:05,316
would love to develop that
technology, but we can't. It's

116
00:07:05,316 --> 00:07:09,190
too expensive. Like it's never
the problem in such a high value

117
00:07:09,190 --> 00:07:12,634
field. If the robots work,
pretty much any price will do

118
00:07:12,634 --> 00:07:16,016
because the robots are, you
know, basically making such

119
00:07:16,016 --> 00:07:17,000
valuable things.

120
00:07:17,000 --> 00:07:20,544
Audrow Nash: Yeah. Yeah. That's
so interesting. And it's

121
00:07:20,544 --> 00:07:24,152
interesting to me because I have
never talked. To another

122
00:07:24,152 --> 00:07:27,696
robotics company in the
pharmaceutical industry for this

123
00:07:27,696 --> 00:07:31,494
kind of thing. I've talked to,
like, I don't even know, like

124
00:07:31,494 --> 00:07:35,481
maybe it's 100 or 100 and 50 or
200 startups at this point. And

125
00:07:35,481 --> 00:07:39,089
it's like, I don't think I've
ever met another one that's

126
00:07:39,089 --> 00:07:43,013
working in the pharmaceutical
space. Why do you think that is?

127
00:07:43,013 --> 00:07:46,747
Just they're scared. Away by the
very large frustrations of

128
00:07:46,747 --> 00:07:50,861
dealing with the FDA, or they're
just not aware of. It, or why do

129
00:07:50,861 --> 00:07:52,000
you think that is?

130
00:07:52,000 --> 00:07:57,854
Fred Parietti: I think there is
maybe three things. One, yes.

131
00:07:57,854 --> 00:08:03,995
Regulatory. There's a regulatory
fear. Be, biology and, biology.

132
00:08:03,995 --> 00:08:10,040
It's not like engineering yet.
And so it's very scary. Scary to

133
00:08:10,040 --> 00:08:11,000
engineers.

134
00:08:11,000 --> 00:08:12,000
Audrow Nash: Why is that? What
do you mean?

135
00:08:12,000 --> 00:08:16,242
Fred Parietti: And three. Yeah.
So, and the third one was, is

136
00:08:16,242 --> 00:08:20,276
just very different from the day
to day experience of most

137
00:08:20,276 --> 00:08:24,170
people, right? So, yeah. So, but
let's go back. So first

138
00:08:24,170 --> 00:08:27,926
regulatory, right? Yes.
Traditionally, robotics people

139
00:08:27,926 --> 00:08:32,029
are in engineers in general,
mechanical software engineers.

140
00:08:32,029 --> 00:08:36,271
They're afraid of the FDA. They
feel like it's something that

141
00:08:36,271 --> 00:08:40,165
they cannot control. And
investors are also terrified by

142
00:08:40,165 --> 00:08:41,000
the FDA and.

143
00:08:41,000 --> 00:08:43,000
Audrow Nash: Song cycles and.
Everything.

144
00:08:43,000 --> 00:08:46,526
Fred Parietti: So the venture
world is divided in two, and

145
00:08:46,526 --> 00:08:49,809
there is people that invest in
FDA stuff, life science

146
00:08:49,809 --> 00:08:53,334
investors, and then every
everything else is lumped on and

147
00:08:53,334 --> 00:08:56,739
they're tech investor. So. A
tech investor can invest in

148
00:08:56,739 --> 00:09:00,325
nuclear fusion or in financial
services. And that's totally

149
00:09:00,325 --> 00:09:03,000
fine. But they will not touch
medical drugs.

150
00:09:03,000 --> 00:09:04,000
Audrow Nash: Yeah.

151
00:09:04,000 --> 00:09:08,539
Fred Parietti: Going to the FDA.
So like that's how scary that

152
00:09:08,539 --> 00:09:12,933
is. So that's was the first
obstacle. Luckily from our point

153
00:09:12,933 --> 00:09:17,033
of view we actually provide the
technology to the pharma

154
00:09:17,033 --> 00:09:21,645
companies. So it is the pharma
companies that still approve the

155
00:09:21,645 --> 00:09:25,672
drugs and around their
facilities. But at least we were

156
00:09:25,672 --> 00:09:29,919
not afraid to operate in this
era, in this market, which I

157
00:09:29,919 --> 00:09:34,385
think a lot of people are still
held back by that fear. Yeah.

158
00:09:34,385 --> 00:09:38,266
The second part was by
essentially biology. So, yeah,

159
00:09:38,266 --> 00:09:42,000
to an engineer, biology is like
black magic, right?

160
00:09:42,000 --> 00:09:44,000
Audrow Nash: I agree. Yes.

161
00:09:44,000 --> 00:09:47,145
Fred Parietti: It's in honestly,
even to a scientist is just from

162
00:09:47,145 --> 00:09:50,145
a scientist point of view. It's
all they know. So they I think

163
00:09:50,145 --> 00:09:53,000
they think it's normal that
everything is black magic. But.

164
00:09:53,000 --> 00:09:54,836
Audrow Nash: You know, it's
different than your point of

165
00:09:54,836 --> 00:09:55,000
view.

166
00:09:55,000 --> 00:09:59,270
Fred Parietti: Right? Right.
Like, I can, you know, honestly,

167
00:09:59,270 --> 00:10:03,610
like it is easier to go to mass
like we know we can go to mass

168
00:10:03,610 --> 00:10:08,020
is just how many million dollars
it would take, right? Yes. But

169
00:10:08,020 --> 00:10:11,800
we know we can do it right?
Finding a single drug that

170
00:10:11,800 --> 00:10:16,000
treats Alzheimer's. Nobody on
the planet can guarantee that.

171
00:10:16,000 --> 00:10:19,444
Audrow Nash: Right. You don't
know how. Big of a sinkhole that

172
00:10:19,444 --> 00:10:20,000
is needed.

173
00:10:20,000 --> 00:10:23,727
Fred Parietti: And I would say
more money, as many more

174
00:10:23,727 --> 00:10:27,860
billions have been invested in
to find a cure for Alzheimer's

175
00:10:27,860 --> 00:10:31,655
than going on to Mars. And I
think we're, we're going on

176
00:10:31,655 --> 00:10:35,652
Mars, probably. Wow. And so
yeah. So like, biology is hard.

177
00:10:35,652 --> 00:10:39,921
It's basically alien technology
that we don't fully understand.

178
00:10:39,921 --> 00:10:43,987
And, and so it is extremely
scary for an engineer to put the

179
00:10:43,987 --> 00:10:48,052
success of their company and
their technology in the, in the

180
00:10:48,052 --> 00:10:52,050
hands of substantially opaque,
mysterious alien technology,

181
00:10:52,050 --> 00:10:55,845
i.e. biology. Right. So people
much rather do. Hey, if I

182
00:10:55,845 --> 00:11:00,114
develop these whatever payments
you know app I know I can do it

183
00:11:00,114 --> 00:11:04,383
is just a, you know, a market.
You can build a crank that makes

184
00:11:04,383 --> 00:11:08,516
money. I feel like much more
easily in biology, you know, you

185
00:11:08,516 --> 00:11:12,717
might have the best science in
the best current knowledge, and

186
00:11:12,717 --> 00:11:17,121
you do the clinical trial and it
fails and you're screwed. Right.

187
00:11:17,121 --> 00:11:21,390
So I, I don't know if we can say
that in the, in the podcast. I

188
00:11:21,390 --> 00:11:22,000
will say.

189
00:11:22,000 --> 00:11:23,963
Audrow Nash: Whatever. It
doesn't matter, I swear. Do

190
00:11:23,963 --> 00:11:25,000
whatever you like. You know.

191
00:11:25,000 --> 00:11:28,716
Fred Parietti: You are in
trouble. You're in deep, deep

192
00:11:28,716 --> 00:11:32,838
trouble. Your company might be
worth zero. Like, not -10%. It

193
00:11:32,838 --> 00:11:36,757
might be worth zero. Even with
the best science behind it.

194
00:11:36,757 --> 00:11:40,000
Because the biology is not yet
fully understood.

195
00:11:40,000 --> 00:11:43,506
Audrow Nash: Yes, but so what
you're doing. So our number two,

196
00:11:43,506 --> 00:11:47,125
what you're doing. Is something
different. You're not developing

197
00:11:47,125 --> 00:11:50,235
the drugs per se. You're
developing the technology that

198
00:11:50,235 --> 00:11:53,515
makes the drugs. So like you're
selling the shovels to the

199
00:11:53,515 --> 00:11:56,908
companies to go build. They're
trying to do the gold rush of

200
00:11:56,908 --> 00:11:59,000
drug development through. But
it's a.

201
00:11:59,000 --> 00:12:03,091
Fred Parietti: Very it's a very
safe shovel that we're selling,

202
00:12:03,091 --> 00:12:07,312
meaning that because we know and
deeply respect. Right. We are we

203
00:12:07,312 --> 00:12:11,143
are humble engineers, meaning
that we never walk into their

204
00:12:11,143 --> 00:12:14,909
homes and tell the pharma
companies, you know what, you're

205
00:12:14,909 --> 00:12:19,000
going to change your biology in
order to feed it to our robots.

206
00:12:19,000 --> 00:12:22,000
Audrow Nash: Right. We do
literally the opposite.

207
00:12:22,000 --> 00:12:25,949
Fred Parietti: We're like, hey,
guys, we know robotics is easier

208
00:12:25,949 --> 00:12:29,650
than biology. Therefore, we will
build a robotic system that

209
00:12:29,650 --> 00:12:33,290
adapts to your process. Right?
So, so that if you spent ten

210
00:12:33,290 --> 00:12:37,054
years developing this process,
and if the only thing by trial

211
00:12:37,054 --> 00:12:40,817
and error that you know works,
we are not going to modify it.

212
00:12:40,817 --> 00:12:44,272
We will teach our robots to
replicate it of course, more

213
00:12:44,272 --> 00:12:46,000
efficiently, more with more.

214
00:12:46,000 --> 00:12:47,000
Audrow Nash: And with robots.

215
00:12:47,000 --> 00:12:51,119
Fred Parietti: Too. Right,
right. They're robots. They do.

216
00:12:51,119 --> 00:12:55,452
They do that very well. But we
adapt. We adapt the lower risk

217
00:12:55,452 --> 00:12:59,713
area robotics to the higher risk
area biology. The other way

218
00:12:59,713 --> 00:13:03,619
around would not work so
effectively. Our technology is

219
00:13:03,619 --> 00:13:07,952
baked in. It. The assumption or
the knowledge that biology is

220
00:13:07,952 --> 00:13:12,000
black magic and you do not want
to mess with black magic.

221
00:13:12,000 --> 00:13:14,893
Audrow Nash: Yeah. Hey, that's
so funny. Yeah, like, we. Like,

222
00:13:14,893 --> 00:13:17,693
have these spells and we know
they work. And therefore we do

223
00:13:17,693 --> 00:13:19,000
that over every time versus.

224
00:13:19,000 --> 00:13:21,000
Fred Parietti: Finding that
spell.

225
00:13:21,000 --> 00:13:24,000
Audrow Nash: Yeah. Took years
ever of effort.

226
00:13:24,000 --> 00:13:26,865
Fred Parietti: From the
brightest minds in the space.

227
00:13:26,865 --> 00:13:30,000
And in most cases they actually
discovered it by accident.

228
00:13:30,000 --> 00:13:33,954
Audrow Nash: Yeah. It's amazing
that that seems to be a frequent

229
00:13:33,954 --> 00:13:37,537
medical thing where it's the
discoveries by accident where

230
00:13:37,537 --> 00:13:41,429
they just like, notice something
and they're like, why is that?

231
00:13:41,429 --> 00:13:45,382
Okay. And then point three, did
you want to touch on for in more

232
00:13:45,382 --> 00:13:46,000
detail or.

233
00:13:46,000 --> 00:13:48,944
Fred Parietti: Yeah. Is the
difference with the lived

234
00:13:48,944 --> 00:13:52,222
experience. So a lot of robotics
people, obviously they are

235
00:13:52,222 --> 00:13:55,389
absorbed by robotics. And so
they never think so. They're

236
00:13:55,389 --> 00:13:58,444
like well what is what are my
daily problems except for

237
00:13:58,444 --> 00:14:01,722
robotics? I need to follow my
clothes, clean my house right

238
00:14:01,722 --> 00:14:02,000
hook.

239
00:14:02,000 --> 00:14:05,000
Audrow Nash: Because I need to,
you know, you know, live. Yeah.

240
00:14:05,000 --> 00:14:08,504
Fred Parietti: You live and I
don't have time because I need

241
00:14:08,504 --> 00:14:12,007
to work on the robots. And so I
think that's very normal for

242
00:14:12,007 --> 00:14:15,453
them. And it was normal for me
as well. I had absolutely no

243
00:14:15,453 --> 00:14:19,015
clue. It was more intuitive for
me to develop wearable robots

244
00:14:19,015 --> 00:14:21,000
then to develop, you know,
robots.

245
00:14:21,000 --> 00:14:22,000
Audrow Nash: For pharmaceutical.

246
00:14:21,000 --> 00:14:24,176
Fred Parietti: Processing,
because I had to meet our

247
00:14:24,176 --> 00:14:27,779
co-founder, and she told me I
would have never come up with

248
00:14:27,779 --> 00:14:29,000
that idea on my own.

249
00:14:29,000 --> 00:14:32,428
Audrow Nash: That's awesome
though. I mean, it's like you,

250
00:14:32,428 --> 00:14:35,796
you stumbled on a massive
problem and then you went, That

251
00:14:35,796 --> 00:14:39,342
could be really valuable. And
then you like that can make it

252
00:14:39,342 --> 00:14:42,533
so you can remove those
constraints you mentioned. And

253
00:14:42,533 --> 00:14:46,079
then you went into this area and
it's really awesome. I feel

254
00:14:46,079 --> 00:14:49,271
like. The way I've been.
Thinking of it for a while is

255
00:14:49,271 --> 00:14:52,935
it's like technology or like, I
don't know, computers, robots,

256
00:14:52,935 --> 00:14:56,362
these kinds of things. And
biology are like the two things

257
00:14:56,362 --> 00:14:59,435
that the future is built on.
Yeah. And I feel pretty

258
00:14:59,435 --> 00:15:03,159
knowledgeable about robotics. I
know nothing about biology, and

259
00:15:03,159 --> 00:15:06,705
it's cool that you sit right in
the middle, of this, I don't

260
00:15:06,705 --> 00:15:07,000
know.

261
00:15:07,000 --> 00:15:10,454
Fred Parietti: It's very
exciting. We love it. We can

262
00:15:10,454 --> 00:15:14,559
learn a lot. We can. We get to
be at the edge of, you know, you

263
00:15:14,559 --> 00:15:18,338
know, organic technology,
robotics, hardware and software.

264
00:15:18,338 --> 00:15:22,118
And that, by the way, also
enables, you know, you know, AI

265
00:15:22,118 --> 00:15:26,288
and advanced, you know, learning
techniques and so on. But also,

266
00:15:26,288 --> 00:15:30,003
you know, we interface with
biology, which is a whole new

267
00:15:30,003 --> 00:15:33,782
alien technology that we don't
understand. We get to learn

268
00:15:33,782 --> 00:15:37,561
basically from the brightest
minds in biology. And totally

269
00:15:37,561 --> 00:15:41,732
it's great because in their own
there is a lot of respect. So in

270
00:15:41,732 --> 00:15:45,837
engineers are very intimidated
by biologists because, you know,

271
00:15:45,837 --> 00:15:49,682
they are magicians, right? And
they're really like they are

272
00:15:49,682 --> 00:15:53,461
like kind of like a wizards, you
know, and, and but what I

273
00:15:53,461 --> 00:15:57,566
discovered is that the biology
is, are also very intimidated by

274
00:15:57,566 --> 00:15:59,000
the robotic engineers.

275
00:15:59,000 --> 00:16:00,000
Audrow Nash: Yeah.

276
00:15:59,000 --> 00:16:01,642
Fred Parietti: Yeah. And so as
long as nobody else tries to

277
00:16:01,642 --> 00:16:04,328
drown out the other voice, you
can work together and be very

278
00:16:04,328 --> 00:16:07,149
productive and everybody feels
like they're learning a ton, and

279
00:16:07,149 --> 00:16:08,000
everybody loves it.

280
00:16:08,000 --> 00:16:11,038
Audrow Nash: Hell, yeah. Yeah, I
would imagine, because it's

281
00:16:11,038 --> 00:16:14,025
super cool. You have all the
experts going, oh you could do

282
00:16:14,025 --> 00:16:17,063
that. Like, oh. That's super
cool. Oh, we could. Do this for

283
00:16:17,063 --> 00:16:19,949
that. That'd be a lot of fun.
Tell me about it, because I

284
00:16:19,949 --> 00:16:22,886
don't know almost. Anything
about pharmaceuticals. Tell me

285
00:16:22,886 --> 00:16:26,076
about what you mean by next gen
pharmaceuticals. Like, what are

286
00:16:26,076 --> 00:16:29,215
they? You mentioned them at the
beginning. But how do you make

287
00:16:29,215 --> 00:16:32,152
them? What are the processes
involved, like, conceptually?

288
00:16:32,152 --> 00:16:35,139
How should I understand these to
understand later what your

289
00:16:35,139 --> 00:16:36,000
robots are doing.

290
00:16:36,000 --> 00:16:40,421
Fred Parietti: Yeah. No. Super
important. And you know, keep in

291
00:16:40,421 --> 00:16:44,351
mind if there's someone that is
scientists or biologists

292
00:16:44,351 --> 00:16:48,491
listening, you know, keep in
mind that this is a mechanical

293
00:16:48,491 --> 00:16:52,000
engineer interpretation and
simplification. Right.

294
00:16:52,000 --> 00:16:54,000
Audrow Nash: So having said
that.

295
00:16:54,000 --> 00:16:57,789
Fred Parietti: You know,
effectively, you know, there is

296
00:16:57,789 --> 00:17:01,850
a degree of complexity in all
medicines, you know, you know,

297
00:17:01,850 --> 00:17:05,774
you know, drugs. And so the
simplest one are the ones that

298
00:17:05,774 --> 00:17:09,429
are they're called small
molecule, you know, drugs and

299
00:17:09,429 --> 00:17:13,218
they're essentially chemicals
that are so simple you can

300
00:17:13,218 --> 00:17:17,346
actually synthesize them with
chemical reactions okay. Right.

301
00:17:17,346 --> 00:17:21,338
So small molecules, just a few,
you know, limited number, a

302
00:17:21,338 --> 00:17:25,534
series of atoms. Right. And you
can just, you know, like it is

303
00:17:25,534 --> 00:17:29,391
chemistry. You just mix it
together artificially. You can

304
00:17:29,391 --> 00:17:33,654
start making, with that metal,
some more complicated, molecules

305
00:17:33,654 --> 00:17:37,511
that are typical of life, you
know, like, you know, like,

306
00:17:37,511 --> 00:17:41,571
yeah, I guess some, you know,
some biological molecules. Not

307
00:17:41,571 --> 00:17:45,226
just simple small molecule
chemicals, but what you can

308
00:17:45,226 --> 00:17:49,421
synthesize artificially is very
limited. When you actually get

309
00:17:49,421 --> 00:17:53,617
into, like, antibodies that are,
you know, gigantic, you know,

310
00:17:53,617 --> 00:17:57,000
from the point of view of just
how, how, you know.

311
00:17:57,000 --> 00:17:59,000
Audrow Nash: Like, like
structure, right? Yeah.

312
00:17:59,000 --> 00:18:01,000
Fred Parietti: Like you actually
cannot synthesize that stuff.

313
00:18:01,000 --> 00:18:02,000
Audrow Nash: You have to.

314
00:18:02,000 --> 00:18:06,292
Fred Parietti: Use a genetically
engineered organism and convince

315
00:18:06,292 --> 00:18:09,991
it to make that thing for you.
So you're using life as a

316
00:18:09,991 --> 00:18:13,689
factory because our technology
is not advanced enough to

317
00:18:13,689 --> 00:18:16,000
synthesize those type of
molecules.

318
00:18:16,000 --> 00:18:19,448
Audrow Nash: Okay. So you said
small molecule. So that's the

319
00:18:19,448 --> 00:18:22,552
first one. The second one, this
is this one that we're

320
00:18:22,552 --> 00:18:26,000
synthesizing with life as a
factory. What's that one called.

321
00:18:26,000 --> 00:18:29,813
Fred Parietti: Large molecules.
Right. Or and in that for

322
00:18:29,813 --> 00:18:33,692
example, a typical example is
antibodies, right? You don't

323
00:18:33,692 --> 00:18:37,906
synthesize that chemically. You
ask a genetic engineer organism

324
00:18:37,906 --> 00:18:42,120
right in a, in a that to make as
much as possible. And then you

325
00:18:42,120 --> 00:18:46,000
blend them, you destroy them and
then you filter out only.

326
00:18:46,000 --> 00:18:47,000
Audrow Nash: The antibodies and
of.

327
00:18:47,000 --> 00:18:51,465
Fred Parietti: Those that, that
target, you know, crazy, entity.

328
00:18:51,465 --> 00:18:53,000
And but it gets worse.

329
00:18:53,000 --> 00:18:54,000
Audrow Nash: Okay.

330
00:18:54,000 --> 00:18:58,833
Fred Parietti: What if you want
what if your drug is not a

331
00:18:58,833 --> 00:19:03,000
molecule, however big? What if
the drug is a cell?

332
00:19:03,000 --> 00:19:06,816
Audrow Nash: Okay, so just like
leaving cell. Yeah. So my

333
00:19:06,816 --> 00:19:10,699
knowledge is super molecules
would. Make up cells. Is that

334
00:19:10,699 --> 00:19:14,381
true? Like you would have a cell
being composed of many

335
00:19:14,381 --> 00:19:18,464
molecules. Okay I don't I don't
know how they're related. You

336
00:19:18,464 --> 00:19:19,000
can see.

337
00:19:19,000 --> 00:19:22,452
Fred Parietti: That that way.
Yeah. Basically many I guess

338
00:19:22,452 --> 00:19:26,262
millions of them. And so and so
because the cell is an organism,

339
00:19:26,262 --> 00:19:29,000
it's now living it can
duplicate. Right? Yeah.

340
00:19:29,000 --> 00:19:31,000
Audrow Nash: So how do you make
those for sure.

341
00:19:31,000 --> 00:19:34,871
Fred Parietti: You know, the
outer tone of things. Right. And

342
00:19:34,871 --> 00:19:38,678
so again we cannot synthesize
cells from scratch. Right. And

343
00:19:38,678 --> 00:19:42,612
so in the case of cell therapy
you start from the cells of the

344
00:19:42,612 --> 00:19:46,355
patient. Right. Like you know a
blood sample or a sample of

345
00:19:46,355 --> 00:19:50,162
inner tissue or the sales of a
donor. So the raw material is

346
00:19:50,162 --> 00:19:53,906
itself living. And then you
genetically engineer that right

347
00:19:53,906 --> 00:19:56,000
to perform a particular
function.

348
00:19:56,000 --> 00:19:59,429
Audrow Nash: Crazy. Okay. And
when we say we are genetically

349
00:19:59,429 --> 00:20:02,971
engineered what does that mean?
Like you expose it to specific

350
00:20:02,971 --> 00:20:06,171
conditions that make it express
in a specific way that's

351
00:20:06,171 --> 00:20:09,829
desirable. Is it something like
that, like you expose it to this

352
00:20:09,829 --> 00:20:13,486
new exposure to that and somehow
it turns it into whatever it is

353
00:20:13,486 --> 00:20:14,000
you want?

354
00:20:14,000 --> 00:20:17,385
Fred Parietti: Yeah. You know.
Exactly. And so, so, for

355
00:20:17,385 --> 00:20:21,138
example, cell therapy right now,
the most, you know, the most

356
00:20:21,138 --> 00:20:24,954
common ones that are already a
few of them approved, you know,

357
00:20:24,954 --> 00:20:28,338
about a dozen. And so, you know,
you gene modified cell

358
00:20:28,338 --> 00:20:31,785
therapies, you start from a
blood sample of the patient,

359
00:20:31,785 --> 00:20:35,046
right? You separate out a
subset, typically of immune

360
00:20:35,046 --> 00:20:38,000
cells, but it can also be other
cells. And then.

361
00:20:38,000 --> 00:20:39,000
Audrow Nash: You.

362
00:20:39,000 --> 00:20:42,724
Fred Parietti: Modify them
genetically, either using a

363
00:20:42,724 --> 00:20:46,931
virus that that delivers the
payload to those cells. Wow. And

364
00:20:46,931 --> 00:20:51,138
modifies, you know, them or you
can also modify them in other

365
00:20:51,138 --> 00:20:55,000
ways, like zapping them with
electricity and then the so

366
00:20:55,000 --> 00:20:59,000
that, you know, if there is
that, you know, basically, the

367
00:20:59,000 --> 00:21:03,138
modifying agent around them is
synapse in. Yeah. And so, and

368
00:21:03,138 --> 00:21:07,552
then the cell now, you know, has
been modified. So if your cell,

369
00:21:07,552 --> 00:21:11,690
if you're the patient or the
donor, it has now been modified

370
00:21:11,690 --> 00:21:15,000
genetically. So it can maybe get
a new receptor.

371
00:21:15,000 --> 00:21:17,338
Audrow Nash: Or in your body
doesn't rejected or see it as a

372
00:21:17,338 --> 00:21:18,000
threat. And then.

373
00:21:18,000 --> 00:21:21,657
Fred Parietti: There's that. Or
you need to give it a new sensor

374
00:21:21,657 --> 00:21:25,029
to recognize that cancer or that
particular tissue that you

375
00:21:25,029 --> 00:21:26,000
wanted to target.

376
00:21:26,000 --> 00:21:29,448
Audrow Nash: Cool. So I see why.
You're saying this is black

377
00:21:29,448 --> 00:21:31,000
magic. This is crazy. Yeah.

378
00:21:31,000 --> 00:21:32,697
Fred Parietti: It is. It's
happening. Like, I don't know

379
00:21:32,697 --> 00:21:33,000
right now.

380
00:21:33,000 --> 00:21:36,066
Audrow Nash: I have no idea how
people would come to these ideas

381
00:21:36,066 --> 00:21:38,988
of how to do these things. And I
suspect you're about to list

382
00:21:38,988 --> 00:21:41,000
even more complex ones, too,
which is not.

383
00:21:41,000 --> 00:21:44,827
Fred Parietti: I mean, these is
kind of like the pinnacle. And

384
00:21:44,827 --> 00:21:48,407
the idea is, you know what? If
you could edit that cell in

385
00:21:48,407 --> 00:21:52,235
multiple ways and so giving them
multiple, you know, it stores

386
00:21:52,235 --> 00:21:55,753
or, you know, so effective you
have to in the cell like a

387
00:21:55,753 --> 00:21:59,642
little robot. Yeah. But it's a
robot we cannot build. It's only

388
00:21:59,642 --> 00:22:01,000
a robot we can modify.

389
00:22:01,000 --> 00:22:03,000
Audrow Nash: We can configure it
through exposure to specific.

390
00:22:03,000 --> 00:22:05,916
Fred Parietti: Yeah. It's almost
like, I don't know, like you

391
00:22:05,916 --> 00:22:08,594
have a horse and you give that
horse, like, night vision

392
00:22:08,594 --> 00:22:11,653
goggles or whatever, right? And
now I can see. But you, we don't

393
00:22:11,653 --> 00:22:14,713
have the technology to actually
fabricate that horse on our own,

394
00:22:14,713 --> 00:22:15,000
right.

395
00:22:15,000 --> 00:22:16,000
Audrow Nash: Yeah.

396
00:22:16,000 --> 00:22:20,067
Fred Parietti: We but we can
modify, I guess, genetic code to

397
00:22:20,067 --> 00:22:22,000
develop night vision. And so.

398
00:22:22,000 --> 00:22:24,000
Audrow Nash: And I love the
metaphor and.

399
00:22:24,000 --> 00:22:28,019
Fred Parietti: Basically after,
you know, after you've done

400
00:22:28,019 --> 00:22:31,833
that, then you need, a few
millions to a few billions of

401
00:22:31,833 --> 00:22:35,783
those cells. So you let them
duplicate and then you infuse

402
00:22:35,783 --> 00:22:40,074
them back into patients. So now
the patient has a better immune

403
00:22:40,074 --> 00:22:44,229
system if their immune system.
But now it sees, you know, the

404
00:22:44,229 --> 00:22:46,000
illness and can target it.

405
00:22:46,000 --> 00:22:49,469
Audrow Nash: Wild. Okay. So
those. Are kind of the three

406
00:22:49,469 --> 00:22:53,000
types small molecules, large
molecules and cell. Therapy.

407
00:22:53,000 --> 00:22:57,759
Fred Parietti: They said that B
two is gene therapy. So in some

408
00:22:57,759 --> 00:23:01,386
cases you actually give the
virus containing the

409
00:23:01,386 --> 00:23:05,767
modification directly to the
patient. And that's then when

410
00:23:05,767 --> 00:23:10,300
you want to do editing, you
know, in, in, in, you know, more

411
00:23:10,300 --> 00:23:15,060
broadly. Right. Or if you think
the virus can target that part,

412
00:23:15,060 --> 00:23:19,517
of, you know, the body of the
patient, honestly. And so and

413
00:23:19,517 --> 00:23:24,050
then, of course, there is RNA.
RNA does not have a permanent

414
00:23:24,050 --> 00:23:28,734
modification of, you know, the
genetic code. Okay. But it does

415
00:23:28,734 --> 00:23:31,000
tell the sales of the patient,

416
00:23:31,000 --> 00:23:32,000
Audrow Nash: To do some.

417
00:23:32,000 --> 00:23:36,343
Fred Parietti: Instructions to.
Yeah, to make different proteins

418
00:23:36,343 --> 00:23:40,687
on the Z. So it's basically kind
of like temporarily tricks them

419
00:23:40,687 --> 00:23:44,963
into behaving as if they had a
different genetic code, but they

420
00:23:44,963 --> 00:23:48,899
don't. And so, so those are all
like, you know, that's it.

421
00:23:48,899 --> 00:23:52,632
There's no chemical reactions.
And so if you don't have

422
00:23:52,632 --> 00:23:57,043
chemical reactions, but you need
to convince or modify, you know,

423
00:23:57,043 --> 00:24:01,386
living beings, cells, right, to
synthesize this stuff, then your

424
00:24:01,386 --> 00:24:05,391
process becomes much less like
chemistry and much more like

425
00:24:05,391 --> 00:24:09,395
cooking or growing plants.
Right. Excel that you don't need

426
00:24:09,395 --> 00:24:13,467
to be sterile when you're cook,
right? In this case, if your

427
00:24:13,467 --> 00:24:17,471
batch of cells has even a single
wrong cell, imaging like a

428
00:24:17,471 --> 00:24:21,000
bacteria should not be there.
Yeah, well, it can be.

429
00:24:21,000 --> 00:24:22,000
Audrow Nash: In the old.

430
00:24:21,000 --> 00:24:24,524
Fred Parietti: Fashioned right
now. It becomes dangerous as

431
00:24:24,524 --> 00:24:28,168
opposed to, you know, healthful
and curative. Right. And it's

432
00:24:28,168 --> 00:24:31,215
not like a normal drug. If
you're making a bunch of

433
00:24:31,215 --> 00:24:35,097
chemicals, you can sterilize it.
All right. Or in fact, if you're

434
00:24:35,097 --> 00:24:38,622
taken by mouth, you don't even
need to sterilize it because

435
00:24:38,622 --> 00:24:42,086
your stomach is sterilized,
right? Yeah. It's amazing. But

436
00:24:42,086 --> 00:24:45,730
you can't sterilize a batch of
cells because you are going to

437
00:24:45,730 --> 00:24:48,000
kill the bad as well as the good
ones.

438
00:24:48,000 --> 00:24:50,000
Audrow Nash: Yeah. So the only
way.

439
00:24:50,000 --> 00:24:54,213
Fred Parietti: To do it is to do
this very manual kind of culture

440
00:24:54,213 --> 00:24:58,296
process, in a completely sterile
way, from day one to the final

441
00:24:58,296 --> 00:25:02,315
day. And without making a single
mistake. Single mistake means

442
00:25:02,315 --> 00:25:04,000
you need to throw it away.

443
00:25:04,000 --> 00:25:08,063
Audrow Nash: Wow. And so give me
an idea of. The, like, timescale

444
00:25:08,063 --> 00:25:11,625
and number of steps for these
kinds of things. Like is it

445
00:25:11,625 --> 00:25:15,188
like, because if I'm cooking,
like, I could be done in an

446
00:25:15,188 --> 00:25:19,188
hour, which I assume that these
processes do not take an hour, I

447
00:25:19,188 --> 00:25:23,000
assume they take a while. How
long. Roughly, or. For some of.

448
00:25:23,000 --> 00:25:27,179
Fred Parietti: The shortest
we've seen, if, you know, take a

449
00:25:27,179 --> 00:25:31,498
few days. And so maybe hundreds
of individual tasks and steps.

450
00:25:31,498 --> 00:25:35,816
Wow. Typical length right now is
1 to 2 weeks and thousands of

451
00:25:35,816 --> 00:25:37,000
individual steps.

452
00:25:37,000 --> 00:25:40,845
Audrow Nash: Wow. And these are
currently being. Performed by

453
00:25:40,845 --> 00:25:44,689
people, like, in lab suits that
are it's kind of like factory

454
00:25:44,689 --> 00:25:48,660
work, but for pharmaceuticals in
the sense. Like, you can think

455
00:25:48,660 --> 00:25:52,000
of it like a factory or it's,
it's, it's is to is to.

456
00:25:52,000 --> 00:25:55,671
Fred Parietti: Qualified, you
know, the type of workforce you

457
00:25:55,671 --> 00:25:59,222
need to, to consider, like a
factory. Yeah. You're in clean

458
00:25:59,222 --> 00:26:02,894
rooms, so it's not a lab. It's a
clean room. Yeah. And so you

459
00:26:02,894 --> 00:26:05,000
need people that are super
trained.

460
00:26:05,000 --> 00:26:07,897
Audrow Nash: And probably, like,
all have at least a bachelor's

461
00:26:07,897 --> 00:26:09,000
in some sort of biology.

462
00:26:09,000 --> 00:26:12,449
Fred Parietti: Or minimum. And
in many cases, like master

463
00:26:12,449 --> 00:26:16,140
masters, PhD PhDs. Wow. I mean,
we've seen like MD, PhD who's

464
00:26:16,140 --> 00:26:19,771
working on this stuff. And so
and obviously it's never their

465
00:26:19,771 --> 00:26:23,341
dream to be a manufacturing
operator in a clean room. Yeah.

466
00:26:23,341 --> 00:26:27,213
I mean, it's a mess. You need to
gown up, right? Because you are

467
00:26:27,213 --> 00:26:28,000
a risk to do.

468
00:26:28,000 --> 00:26:31,190
Audrow Nash: You have a lot of
bacteria and all sorts of

469
00:26:31,190 --> 00:26:34,892
contaminants coming from you all
the time? Yeah. For sure. That's

470
00:26:34,892 --> 00:26:37,000
wild. So it's an expensive
workforce?

471
00:26:37,000 --> 00:26:38,000
Fred Parietti: That doesn't want
to do the job.

472
00:26:38,000 --> 00:26:41,830
Audrow Nash: That does it? Yeah.
That hates it. But they are

473
00:26:41,830 --> 00:26:45,979
needed and they're probably paid
pretty well. Is it true? Because

474
00:26:45,979 --> 00:26:49,362
the it's like undesirable. It's
hard work. It's is it

475
00:26:49,362 --> 00:26:53,000
repetitive. Because they end up
doing the same processes.

476
00:26:53,000 --> 00:26:55,588
Fred Parietti: Super, super
repetitive, but also repetitive

477
00:26:55,588 --> 00:26:58,000
and stressful because it's
repetitive. But if you mess.

478
00:26:58,000 --> 00:26:59,000
Audrow Nash: Up at all.

479
00:26:59,000 --> 00:27:01,959
Fred Parietti: You throw
anything, anything like you

480
00:27:01,959 --> 00:27:05,203
could touch the wrong thing
there on the surface for like

481
00:27:05,203 --> 00:27:06,000
half a second.

482
00:27:06,000 --> 00:27:09,000
Audrow Nash: It's that wow,
that's crazy.

483
00:27:09,000 --> 00:27:12,412
Fred Parietti: And or breathe on
it or sneeze or whatever.

484
00:27:12,412 --> 00:27:16,059
Scratch your forehead and you
didn't realize now your glove is

485
00:27:16,059 --> 00:27:19,706
contaminated. And so, so it's
expensive as operators, but from

486
00:27:19,706 --> 00:27:22,000
the point of view of these
people that.

487
00:27:22,000 --> 00:27:24,000
Audrow Nash: They don't even
like doing it. Yeah.

488
00:27:24,000 --> 00:27:27,106
Fred Parietti: They would like
to be in drug discovery as

489
00:27:27,106 --> 00:27:30,647
opposed to drug manufacturing or
maybe process optimization as a,

490
00:27:30,647 --> 00:27:34,189
as opposed to process execution.
So from the point of view of the

491
00:27:34,189 --> 00:27:37,567
employer, they're expensive. But
from the point of view of the

492
00:27:37,567 --> 00:27:40,455
employees, undesirable,
underpaid with respect to the

493
00:27:40,455 --> 00:27:41,000
ideal job.

494
00:27:41,000 --> 00:27:45,138
Audrow Nash: Okay. Yeah. They
don't love it. And I wonder, so

495
00:27:45,138 --> 00:27:48,937
that sounds interesting. You
have all these people doing

496
00:27:48,937 --> 00:27:53,279
this. I was thinking factory is
in, like a series of steps where

497
00:27:53,279 --> 00:27:57,349
people are doing repetitive
work. Not necessarily the skill,

498
00:27:57,349 --> 00:28:01,487
because there's some skilled
manufacturing also. But so it's,

499
00:28:01,487 --> 00:28:04,812
it's kind of like a
pharmaceutical factory in the

500
00:28:04,812 --> 00:28:08,000
way I kind of defined it right
now. So for the.

501
00:28:08,000 --> 00:28:11,561
Fred Parietti: Simpler drugs can
be scaled easily. So you can

502
00:28:11,561 --> 00:28:15,239
make like, you know, 5 million
tablets of aspirin. But for this

503
00:28:15,239 --> 00:28:18,508
new biological next generation
therapeutics. Oh, it's so

504
00:28:18,508 --> 00:28:22,069
manual. It is so manual. And by
the way, if you try to change

505
00:28:22,069 --> 00:28:25,571
the process and simplify it,
most likely the cells are going

506
00:28:25,571 --> 00:28:28,957
to die in or the results are
going to be different so that

507
00:28:28,957 --> 00:28:31,000
the FDA now needs to look again
at.

508
00:28:31,000 --> 00:28:32,000
Audrow Nash: 3 to.

509
00:28:32,000 --> 00:28:35,000
Fred Parietti: 5 approve it.
Right. So the best time.

510
00:28:35,000 --> 00:28:38,806
Audrow Nash: Consuming and
expensive everything. So imagine

511
00:28:38,806 --> 00:28:39,000
it.

512
00:28:39,000 --> 00:28:42,751
Fred Parietti: You go to one
that your favorite restaurant in

513
00:28:42,751 --> 00:28:46,748
the world. And there's only that
chef in that location with local

514
00:28:46,748 --> 00:28:50,438
ingredients that makes that
dish. And now imagine, oh, okay,

515
00:28:50,438 --> 00:28:53,943
you want to replicate that
restaurant in your backyard in

516
00:28:53,943 --> 00:28:57,756
a, in a, in another, in another
continent. Yeah. Right. And so

517
00:28:57,756 --> 00:29:01,446
it's not just about following
the same recipe. You know, you

518
00:29:01,446 --> 00:29:05,258
also need first of all, you need
to have the same skill of the

519
00:29:05,258 --> 00:29:09,255
chef, you know, and you know, in
the same implicit knowledge that

520
00:29:09,255 --> 00:29:12,945
they had about their process.
But the thing is you have that

521
00:29:12,945 --> 00:29:16,696
you need the same ingredients.
Yeah. And even the same tomato

522
00:29:16,696 --> 00:29:20,570
seed. If you primed it outside
of your house, as opposed to the

523
00:29:20,570 --> 00:29:24,445
other continent where different
sun, different soil, the tomato

524
00:29:24,445 --> 00:29:28,380
is going to taste different. So
any change the process when life

525
00:29:28,380 --> 00:29:32,254
is involved, right? When biology
is involved results inevitably

526
00:29:32,254 --> 00:29:36,006
in something that is different
taste right and that it's good

527
00:29:36,006 --> 00:29:39,757
if you're talking about wine or
food, that you can appreciate

528
00:29:39,757 --> 00:29:43,385
the variation. But if you're
talking about a drug that goes

529
00:29:43,385 --> 00:29:44,000
inside of.

530
00:29:44,000 --> 00:29:45,000
Audrow Nash: You, you.

531
00:29:45,000 --> 00:29:46,758
Fred Parietti: Don't want to
see, oh, what's the flavor of

532
00:29:46,758 --> 00:29:47,000
the day.

533
00:29:47,000 --> 00:29:49,000
Audrow Nash: Right? Yes. Yeah.

534
00:29:49,000 --> 00:29:53,909
Fred Parietti: And so that's why
it's so hard to scale biology.

535
00:29:53,909 --> 00:29:55,000
Bioprocessing.

536
00:29:55,000 --> 00:29:58,817
Audrow Nash: Yeah. Because it's
like what I'm imagining. Is like

537
00:29:58,817 --> 00:30:02,575
a really difficult optimization
problem to go from beginning to

538
00:30:02,575 --> 00:30:06,273
end process or to the medication
being. Developed. And you can

539
00:30:06,273 --> 00:30:09,971
introduce uncertainty at every
level when you change anything.

540
00:30:09,971 --> 00:30:13,848
So it's like yes, you have drugs
made in one spot because they've

541
00:30:13,848 --> 00:30:17,546
learned how to make it there and
they can quality control it a

542
00:30:17,546 --> 00:30:21,065
lot more easily. And they've
nailed down their supply lines

543
00:30:21,065 --> 00:30:24,346
and all sorts of things like
this. So you don't make it

544
00:30:24,346 --> 00:30:28,163
somewhere different. So it just
sounds very. Very it sounds like

545
00:30:28,163 --> 00:30:31,563
a very. Thorny set of problems
with a. Lot of implicitly.

546
00:30:31,563 --> 00:30:34,963
Defined relationships between
all the steps that we don't

547
00:30:34,963 --> 00:30:38,780
understand if they're important
or not. So it makes it very hard

548
00:30:38,780 --> 00:30:42,478
to automate, from an engineering
perspective where we just go,

549
00:30:42,478 --> 00:30:45,938
you kind of like, give this
wiggly path that gets from one

550
00:30:45,938 --> 00:30:49,337
spot to another, and we just
want to make a straight line

551
00:30:49,337 --> 00:30:53,035
from there, from start to end.
But they need to go through all

552
00:30:53,035 --> 00:30:56,555
those wiggly steps because
they're not exactly. Sure where.

553
00:30:56,555 --> 00:30:59,000
There's the errors that they're
avoiding.

554
00:30:59,000 --> 00:31:02,560
Fred Parietti: Exactly.
Actually, that's exactly it.

555
00:31:02,560 --> 00:31:06,530
Like it would be simple for an
engineer to say, oh, it's a

556
00:31:06,530 --> 00:31:10,226
wiggly path. Let me draw a
straight line. And but that

557
00:31:10,226 --> 00:31:14,265
would assume you know what
you're talking about in terms of

558
00:31:14,265 --> 00:31:18,167
biology. And because this
biology, we need to assume that

559
00:31:18,167 --> 00:31:22,000
every wiggles is motivated even
when you don't know why.

560
00:31:22,000 --> 00:31:25,495
Audrow Nash: Yeah, right. And so
the best performance fence. Have

561
00:31:25,495 --> 00:31:27,000
you ever heard of that idea.

562
00:31:27,000 --> 00:31:29,000
Fred Parietti: Oh no. No. What's
that?

563
00:31:29,000 --> 00:31:32,494
Audrow Nash: You cannot take
down a fence unless you know.

564
00:31:32,494 --> 00:31:36,048
Why it's there. This kind of
thing. I don't know. I've just

565
00:31:36,048 --> 00:31:39,663
one. One colleague mentioned
that quite a bit. So, like, you

566
00:31:39,663 --> 00:31:43,337
don't take down the thing unless
you know why it's there and.

567
00:31:43,337 --> 00:31:44,000
That is it.

568
00:31:44,000 --> 00:31:48,716
Fred Parietti: That is it. And
in biology you know. Yeah.

569
00:31:48,716 --> 00:31:54,011
Anybody on the fence looks very
irrational. But you just need to

570
00:31:54,011 --> 00:31:58,644
have you know a lot of respect
honestly for OSI how many

571
00:31:58,644 --> 00:32:03,608
scientists year, year went into
finding out that things that

572
00:32:03,608 --> 00:32:08,324
miraculously work. Yeah it
happens to works and so and so

573
00:32:08,324 --> 00:32:13,453
our job is okay. How can I learn
how to make that fence with a

574
00:32:13,453 --> 00:32:18,086
robotic technology? And. Yeah.
And so it's, it's, it's a

575
00:32:18,086 --> 00:32:22,802
honestly like to me, it's
actually no different from, you

576
00:32:22,802 --> 00:32:27,766
know, the good there's two other
good examples, like, like a

577
00:32:27,766 --> 00:32:30,000
couture dress or, you know.

578
00:32:30,000 --> 00:32:32,000
Audrow Nash: I don't know what
it is.

579
00:32:32,000 --> 00:32:35,937
Fred Parietti: Like, you know,
like a, like a high fashion. Oh

580
00:32:35,937 --> 00:32:39,937
yeah. Right. You know, there's,
there's dozens of artisans that

581
00:32:39,937 --> 00:32:43,746
do that. Right. And then the
stitching and so on. Right. And

582
00:32:43,746 --> 00:32:47,683
so you couldn't simplify it a
lot. Right. But you not look the

583
00:32:47,683 --> 00:32:48,000
same.

584
00:32:48,000 --> 00:32:49,000
Audrow Nash: Right.

585
00:32:49,000 --> 00:32:53,451
Fred Parietti: Because you know,
like they did all of that to

586
00:32:53,451 --> 00:32:57,828
appeal to the taste of the time
for that season or whatever.

587
00:32:57,828 --> 00:33:02,498
Right. And so and so how can you
do that at scale, how that, you

588
00:33:02,498 --> 00:33:06,730
know, now it's clearly not
possible in high fashion to get

589
00:33:06,730 --> 00:33:11,253
to every single person the same
quality and scale. You need to

590
00:33:11,253 --> 00:33:15,850
teach a robot to be as skill as
the artisan. Right? Because any

591
00:33:15,850 --> 00:33:20,373
simplification would come at a
cost. In terms of, you know, in

592
00:33:20,373 --> 00:33:23,000
that case, their, esthetic, I
guess.

593
00:33:23,000 --> 00:33:27,013
Audrow Nash: Yeah. It's not the
same thing basically, as that

594
00:33:27,013 --> 00:33:28,000
was being made.

595
00:33:28,000 --> 00:33:32,084
Fred Parietti: Correct. And so
we need to do that in biology.

596
00:33:32,084 --> 00:33:36,167
And in fact, the price point and
the value and the cost of an

597
00:33:36,167 --> 00:33:40,519
advanced next generation drug is
similar to the top high fashion,

598
00:33:40,519 --> 00:33:44,669
you know, dresses. Yeah. And yet
our challenge is to make that

599
00:33:44,669 --> 00:33:48,218
high fashion available to
everybody and affordable to

600
00:33:48,218 --> 00:33:52,368
everybody. And so the level of
refinement you need in robotics

601
00:33:52,368 --> 00:33:56,318
to take that product and not
change it, not to ruin it, but

602
00:33:56,318 --> 00:34:00,000
still scale it. That's a big
robotics challenge, I bet.

603
00:34:00,000 --> 00:34:04,442
Audrow Nash: Yeah, it sounds
very, very hard. Okay. So I feel

604
00:34:04,442 --> 00:34:08,811
like I get. That making drugs is
really, really complex. And

605
00:34:08,811 --> 00:34:12,816
there's a lot of steps and it's
very difficult and time

606
00:34:12,816 --> 00:34:16,383
consuming. And you need
sanitation. The whole way

607
00:34:16,383 --> 00:34:20,607
through. What are your robots
do? How are they involved in

608
00:34:20,607 --> 00:34:25,267
this whole thing? What kinds of
tasks are they doing and how are

609
00:34:25,267 --> 00:34:28,835
you working with the
pharmaceutical companies for

610
00:34:28,835 --> 00:34:30,000
different drugs?

611
00:34:30,000 --> 00:34:33,590
Fred Parietti: Yeah. So,
Essentially the most important

612
00:34:33,590 --> 00:34:37,702
thing for us is not messing with
the biology. And so instead of

613
00:34:37,702 --> 00:34:41,749
designing new instruments or new
ways to manipulate, you know,

614
00:34:41,749 --> 00:34:45,599
the cells or the biology, we
teach the robots to do exactly

615
00:34:45,599 --> 00:34:49,254
what the scientists are doing
right. And so we teach the

616
00:34:49,254 --> 00:34:52,779
robots to use the same
instruments, the same reagents.

617
00:34:52,779 --> 00:34:56,499
Reagents are typically, you
know, liquids containing, you

618
00:34:56,499 --> 00:35:00,285
know, you know, reagents or
molecules that the cells need,

619
00:35:00,285 --> 00:35:04,201
right? And consumables, those
consumables are typically sets

620
00:35:04,201 --> 00:35:08,378
of boxes, bags, tubes, the, the
sales things you go through with

621
00:35:08,378 --> 00:35:12,621
right. And you use them once and
you throw them away because they

622
00:35:12,621 --> 00:35:16,668
need to be sterile. Right. And
you can never have the sales of

623
00:35:16,668 --> 00:35:20,519
two patients being in contact.
Right. So the most important

624
00:35:20,519 --> 00:35:24,565
part for us is, robots that use
the same instruments, the same

625
00:35:24,565 --> 00:35:27,959
consumables and the same
reagents. So we have a very

626
00:35:27,959 --> 00:35:31,875
modular system because every
pharma company uses a different

627
00:35:31,875 --> 00:35:35,335
set of instruments or
consumables or reagents. And so

628
00:35:35,335 --> 00:35:39,251
we basically surround imaging
like a robot, a set of robotic

629
00:35:39,251 --> 00:35:43,167
arms surrounded by a set of
modules or cubes, kind of like a

630
00:35:43,167 --> 00:35:47,018
shed to shelves on the sides.
And these shelves are made by

631
00:35:47,018 --> 00:35:50,738
modules, and each modules
contains an instrument that was

632
00:35:50,738 --> 00:35:54,393
used by people originally and
now is being operated by a

633
00:35:54,393 --> 00:35:58,440
robot. So to teach the robots to
do that, and it depends if an

634
00:35:58,440 --> 00:36:02,552
instrument is the bottleneck, we
can duplicate or triplicate or

635
00:36:02,552 --> 00:36:06,077
more that particular module. If
the robotic arm is the

636
00:36:06,077 --> 00:36:10,254
bottleneck, we can duplicate or
triplicate the robotic arms. And

637
00:36:10,254 --> 00:36:14,301
in fact, right now in our latest
robots, we actually have four

638
00:36:14,301 --> 00:36:18,151
robotic arms at the same. Oh.
Cool. So I'm back at multiple

639
00:36:18,151 --> 00:36:19,000
robotic arms.

640
00:36:19,000 --> 00:36:21,000
Audrow Nash: You're back? Yeah.

641
00:36:21,000 --> 00:36:25,000
Fred Parietti: Just in a very,
very different context.

642
00:36:25,000 --> 00:36:26,000
Audrow Nash: Yeah.

643
00:36:25,000 --> 00:36:29,031
Fred Parietti: So yeah. So yeah,
my mom was like, oh, you're back

644
00:36:29,031 --> 00:36:32,442
at doing robotics. I'm like,
mom, I never stopped doing

645
00:36:32,442 --> 00:36:33,000
robotics.

646
00:36:33,000 --> 00:36:34,000
Audrow Nash: Yeah.

647
00:36:34,000 --> 00:36:36,529
Fred Parietti: It's just
happened. So that now it looks

648
00:36:36,529 --> 00:36:39,149
kind of, you know, like before,
but obviously they're not

649
00:36:39,149 --> 00:36:42,000
wearable anymore. And they are
doing something very different.

650
00:36:42,000 --> 00:36:45,925
Audrow Nash: Yes. Okay. So you
have all these robotic arms. And

651
00:36:45,925 --> 00:36:49,913
they have all these bins and in
these bins they might have tools

652
00:36:49,913 --> 00:36:53,651
or things that they'll interact
with. And then you have them

653
00:36:53,651 --> 00:36:57,514
doing I suppose they have like a
little work station or like a

654
00:36:57,514 --> 00:37:00,941
work site where they're
manipulating something. And how

655
00:37:00,941 --> 00:37:02,000
does all that go?

656
00:37:02,000 --> 00:37:06,910
Fred Parietti: Yeah. So you know
what? I also have a quick video

657
00:37:06,910 --> 00:37:11,436
here that I wanted to show you
and will describe it for the

658
00:37:11,436 --> 00:37:16,192
people listening. So here it is.
Okay. Okay. And so, so, yeah.

659
00:37:16,192 --> 00:37:20,948
So, on the left you can see, the
manual process. So scientists

660
00:37:20,948 --> 00:37:25,551
doing stuff by hand and on the
right you can see the robotic

661
00:37:25,551 --> 00:37:30,000
process. And in particular, a
robot robotic arm on a rail.

662
00:37:30,000 --> 00:37:31,000
Audrow Nash: Mat.

663
00:37:31,000 --> 00:37:34,429
Fred Parietti: Surround is
surrounded by these bays or

664
00:37:34,429 --> 00:37:38,111
modules, right. And every
module, every kind of like space

665
00:37:38,111 --> 00:37:41,349
opening on the sides of the
robotic arm, contains a

666
00:37:41,349 --> 00:37:44,968
particular instrument. And so
the robotic arm is handling

667
00:37:44,968 --> 00:37:48,905
syringes and vials and so on.
And the same way that the person

668
00:37:48,905 --> 00:37:52,905
is doing it. But now in a much
more repeatable and sterile way,

669
00:37:52,905 --> 00:37:56,714
and so, yeah, so, you know, you
kind of see like human touch

670
00:37:56,714 --> 00:38:00,524
points, people are touching
everything, of course, when they

671
00:38:00,524 --> 00:38:04,397
need to, you know, handle stuff
manually. And, but the robots

672
00:38:04,397 --> 00:38:08,079
has zero human touch points and
human touch points are the

673
00:38:08,079 --> 00:38:11,000
number one source of
contamination and errors.

674
00:38:11,000 --> 00:38:12,000
Audrow Nash: Right.

675
00:38:12,000 --> 00:38:16,606
Fred Parietti: And so obviously
the robot doesn't breathe,

676
00:38:16,606 --> 00:38:21,053
right. It's much easier to
filter the air, you know, and

677
00:38:21,053 --> 00:38:25,659
control the air quality inside
the robotic system. So it's

678
00:38:25,659 --> 00:38:30,026
effectively a much, you know,
more repeatable, and more

679
00:38:30,026 --> 00:38:34,791
sterile process. And so these
example only contains a single

680
00:38:34,791 --> 00:38:39,000
robotic arm. But I have another
example with oh four.

681
00:38:39,000 --> 00:38:41,000
Audrow Nash: So I can see that.

682
00:38:41,000 --> 00:38:42,000
Fred Parietti: Okay. Here it is.

683
00:38:42,000 --> 00:38:44,000
Audrow Nash: Oh, they're on the
ceiling. I love it.

684
00:38:44,000 --> 00:38:45,000
Fred Parietti: Yes.

685
00:38:44,000 --> 00:38:46,000
Audrow Nash: So ceiling is such
a good spot for robots.

686
00:38:46,000 --> 00:38:49,535
Fred Parietti: Yes. We need to
cram so many robotic arms that

687
00:38:49,535 --> 00:38:52,721
we need to put a few on the
ceiling than the few on the

688
00:38:52,721 --> 00:38:56,430
floor. The ones on the floor are
also on a rail. Sometimes it is

689
00:38:56,430 --> 00:38:59,617
also a vertical axis. Yeah, the
ones on the ceiling are

690
00:38:59,617 --> 00:39:03,267
stationary, but they are still
quite helpful. And yeah. So here

691
00:39:03,267 --> 00:39:06,454
you can see train moving. Oh, by
the way, you see these

692
00:39:06,454 --> 00:39:09,989
imperfections that we need to
inject to convince people that,

693
00:39:09,989 --> 00:39:13,639
you know, this is not a render.
So we always add at least that,

694
00:39:13,639 --> 00:39:16,710
you know, if a frame or a
sequence with people around

695
00:39:16,710 --> 00:39:17,000
them.

696
00:39:17,000 --> 00:39:22,000
Audrow Nash: So people see that
this is true. Super cool.

697
00:39:22,000 --> 00:39:25,609
Fred Parietti: And so yeah. So
these are the multiple robotic

698
00:39:25,609 --> 00:39:29,278
arms and in this case, you know,
a single robotic arm could do

699
00:39:29,278 --> 00:39:32,000
the same, right? Yeah. But it
would be slower.

700
00:39:32,000 --> 00:39:34,787
Audrow Nash: Yes for sure. Yeah.
You might as. Well make it

701
00:39:34,787 --> 00:39:37,433
faster for sure. And then it
becomes. A more interesting

702
00:39:37,433 --> 00:39:40,362
problem too because you have to
parallelize them. And then you

703
00:39:40,362 --> 00:39:43,102
get to work on some of that
awesome tech that you like for

704
00:39:43,102 --> 00:39:44,000
all of this exotic.

705
00:39:44,000 --> 00:39:48,414
Fred Parietti: So what about,
you know, you know, for example,

706
00:39:48,414 --> 00:39:52,900
avoidance of collisions, right?
What about optimizing the plans

707
00:39:52,900 --> 00:39:57,385
for multiple robotic arms that
they're facing with, you know, a

708
00:39:57,385 --> 00:40:01,586
dozen or two different modules
so that if the planned space

709
00:40:01,586 --> 00:40:05,644
becomes huge, exponentially
complicated. And so there's a

710
00:40:05,644 --> 00:40:10,201
lot of optimization there that,
you know, our software team, you

711
00:40:10,201 --> 00:40:13,974
know, works on. And yeah, so
this is this is all very

712
00:40:13,974 --> 00:40:18,388
exciting. But it's also very,
very valuable and very impactful

713
00:40:18,388 --> 00:40:22,589
application. And so we just
enjoy that intersection between

714
00:40:22,589 --> 00:40:26,718
top robotics and also top, you
know, good impact, positive

715
00:40:26,718 --> 00:40:28,000
impact in society.

716
00:40:28,000 --> 00:40:31,875
Audrow Nash: Oh yeah. Love it.
That is so cool. Okay. So I see

717
00:40:31,875 --> 00:40:35,688
that these. Are doing a bunch of
similar tasks to what people

718
00:40:35,688 --> 00:40:38,000
would do. The. Thing. Yeah. Go
ahead.

719
00:40:38,000 --> 00:40:42,155
Fred Parietti: Yeah. It's a it's
even them it's more than just

720
00:40:42,155 --> 00:40:46,378
similar. Sorry I stopped no. Go
ahead now. It is like they need

721
00:40:46,378 --> 00:40:50,131
to be the same equivalent. They
need to be statistically

722
00:40:50,131 --> 00:40:53,684
equivalent by virtue of
measuring the biology and the

723
00:40:53,684 --> 00:40:57,504
effects on the cells. If the
robots are not statistically

724
00:40:57,504 --> 00:41:01,794
equivalent to the scientist, we
are messing with the biology and

725
00:41:01,794 --> 00:41:03,000
we have a problem.

726
00:41:03,000 --> 00:41:05,813
Audrow Nash: Cool. So how do
you. Evaluate if they're

727
00:41:05,813 --> 00:41:08,998
statistically equivalent? I'm
imagining like you look at the

728
00:41:08,998 --> 00:41:11,917
results and you make a
distribution of it in some form,

729
00:41:11,917 --> 00:41:14,783
and then you see if the
distributions look similar. So

730
00:41:14,783 --> 00:41:17,968
it might be like we generate
this many cells for this batch.

731
00:41:17,968 --> 00:41:20,887
And this many of them are of
high quality. And then you

732
00:41:20,887 --> 00:41:24,231
compare across many runs. And if
the distributions look similar

733
00:41:24,231 --> 00:41:27,469
then you say that's the same or
good. Is that is it something

734
00:41:27,469 --> 00:41:28,000
like that.

735
00:41:28,000 --> 00:41:32,067
Fred Parietti: It is exactly
that. But it is another layer of

736
00:41:32,067 --> 00:41:36,200
complexity, which is biology is
very random, predictable, even

737
00:41:36,200 --> 00:41:40,400
with these, you know, checks in
places. And so what you need to

738
00:41:40,400 --> 00:41:44,600
do is you need to start from the
same cells, donor cells, human

739
00:41:44,600 --> 00:41:45,000
cells.

740
00:41:45,000 --> 00:41:48,276
Audrow Nash: And so you have
them do. It and you have you

741
00:41:48,276 --> 00:41:50,000
compare a and a kind of thing.

742
00:41:50,000 --> 00:41:53,295
Fred Parietti: So and we do at
the same time from the same

743
00:41:53,295 --> 00:41:55,000
donor, from the same cell bag.

744
00:41:55,000 --> 00:41:57,657
Audrow Nash: And you do the same
room and you see that you get

745
00:41:57,657 --> 00:41:58,000
similar.

746
00:41:58,000 --> 00:42:01,680
Fred Parietti: The main process
in the robotic process, because

747
00:42:01,680 --> 00:42:05,186
if you did it with different
starting cells, the variability

748
00:42:05,186 --> 00:42:08,515
might be caused by the cells.
Right? It because maybe one

749
00:42:08,515 --> 00:42:12,196
patient was different from the
other. We know not probably. And

750
00:42:12,196 --> 00:42:15,000
so you need to always add a
control right there.

751
00:42:15,000 --> 00:42:19,429
Audrow Nash: So you just split
in two. Yeah. Yes. That is super

752
00:42:19,429 --> 00:42:23,506
cool. Very clever. I imagine
with enough data you wouldn't

753
00:42:23,506 --> 00:42:27,723
need to do that approach. But
the thing is, if you are. Just

754
00:42:27,723 --> 00:42:32,011
starting or if it's expensive or
time consuming, say you have

755
00:42:32,011 --> 00:42:36,370
like. A. 300 step in process,
where you don't want to do it to

756
00:42:36,370 --> 00:42:40,517
get there every time. This
approach probably makes a lot of

757
00:42:40,517 --> 00:42:44,875
sense to kind of converge on the
results quickly and tune your

758
00:42:44,875 --> 00:42:46,000
systems quickly.

759
00:42:46,000 --> 00:42:49,773
Fred Parietti: Yeah. And you
know, from the engineering point

760
00:42:49,773 --> 00:42:53,237
of view, it might look like a
lot of work, but from the,

761
00:42:53,237 --> 00:42:56,701
biology point of view, like is
like, oh my God, we can't

762
00:42:56,701 --> 00:43:00,351
believe we just need to prove
equivalence. This way. And it

763
00:43:00,351 --> 00:43:03,753
only takes two months of
testing. So that's really like

764
00:43:03,753 --> 00:43:04,000
the.

765
00:43:04,000 --> 00:43:05,000
Audrow Nash: Huge wind.

766
00:43:05,000 --> 00:43:07,000
Fred Parietti: Speed of light
for pharma. Yes.

767
00:43:07,000 --> 00:43:10,000
Audrow Nash: And wow, by the
way, it's so funny.

768
00:43:10,000 --> 00:43:14,010
Fred Parietti: Yeah, yeah, yeah.
It's crazy. And so we're the

769
00:43:14,010 --> 00:43:17,757
only company that has ever
published, peer reviewed data,

770
00:43:17,757 --> 00:43:21,438
showing equivalence between a
manual process and robotic

771
00:43:21,438 --> 00:43:25,645
process. Right. So, like, nobody
else has a, you know, published

772
00:43:25,645 --> 00:43:29,721
is people have, you know, white
papers and all this stuff, but

773
00:43:29,721 --> 00:43:33,205
like, we went on a peer
reviewed, you know, you know,

774
00:43:33,205 --> 00:43:37,281
journal publications. And so, so
it's very solid data, because

775
00:43:37,281 --> 00:43:41,357
that's the kind of credibility
you need, you know, I work with

776
00:43:41,357 --> 00:43:43,000
companies. Yeah, exactly.

777
00:43:43,000 --> 00:43:47,071
Audrow Nash: Yeah. They just
wouldn't use it. It would be

778
00:43:47,071 --> 00:43:51,000
like an R&D thing, versus
something used in production.

779
00:43:51,000 --> 00:43:53,241
Fred Parietti: Correct? Correct.
And the whole point for us is

780
00:43:53,241 --> 00:43:54,000
the using production.

781
00:43:54,000 --> 00:43:59,382
Audrow Nash: Yes. Very cool. So
I'm imagining this isn't so I

782
00:43:59,382 --> 00:44:05,029
guess what's the scale of your.
Operation or deployments or. How

783
00:44:05,029 --> 00:44:10,412
many robots or. Cells are there
like work cells like this are

784
00:44:10,412 --> 00:44:16,059
deployed or. How, I don't know,
how large would you say? I don't

785
00:44:16,059 --> 00:44:21,176
know exactly what the metric
would be. But how big and how

786
00:44:21,176 --> 00:44:24,000
much deployment do you have? So.

787
00:44:24,000 --> 00:44:28,456
Fred Parietti: So I can't be too
specific because we only

788
00:44:28,456 --> 00:44:33,303
announced two big partners, and
I don't want people to reverse

789
00:44:33,303 --> 00:44:37,290
engineer, you know? Yeah. The,
the details of those

790
00:44:37,290 --> 00:44:42,293
collaborations I see. But what I
can tell you is that this year,

791
00:44:42,293 --> 00:44:47,140
our number of deployments will
be, you know, the robot you saw

792
00:44:47,140 --> 00:44:51,752
before. That is a pretty big
robot, by the way, with a four

793
00:44:51,752 --> 00:44:56,443
robotic arms we are talking
about, you know, 30 by ten feet,

794
00:44:56,443 --> 00:45:01,290
in area. So hopefully so it is a
big multi-million dollar, you

795
00:45:01,290 --> 00:45:05,980
know, a robotic system. We are
going to deploy these here in

796
00:45:05,980 --> 00:45:10,749
the 5 to 10 year range. Okay.
And we already deployed several

797
00:45:10,749 --> 00:45:12,000
of them. And so.

798
00:45:12,000 --> 00:45:15,519
Audrow Nash: Okay. So you have
like. Tens of systems deployed

799
00:45:15,519 --> 00:45:18,000
that are of similar size. So
that probably.

800
00:45:18,000 --> 00:45:21,333
Fred Parietti: The idea, yes, is
that we are effectively already

801
00:45:21,333 --> 00:45:23,000
sold out for the next two years.

802
00:45:23,000 --> 00:45:24,000
Audrow Nash: So congrats.

803
00:45:24,000 --> 00:45:27,010
Fred Parietti: We know it's
because there is so much demand

804
00:45:27,010 --> 00:45:30,224
for these therapies. You can't
make them by hand. It's just not

805
00:45:30,224 --> 00:45:33,235
possible. So now it's on our
problem. Yes. Okay. We need to

806
00:45:33,235 --> 00:45:34,000
address how do.

807
00:45:34,000 --> 00:45:35,000
Audrow Nash: You scale.

808
00:45:35,000 --> 00:45:38,791
Fred Parietti: Chain right to
double in next year and then

809
00:45:38,791 --> 00:45:42,712
double again the year after.
Right. Wow. Just to get through

810
00:45:42,712 --> 00:45:45,000
the backlog. Well, that's and
soul.

811
00:45:45,000 --> 00:45:48,933
Audrow Nash: Gazing, though. I
mean, I'm sure that that's a

812
00:45:48,933 --> 00:45:52,600
very strong position for
investors that want to come in

813
00:45:52,600 --> 00:45:56,733
and like, you guys are in great
terms, for seeking investment,

814
00:45:56,733 --> 00:46:00,000
for finding more, customers, all
sorts of things.

815
00:46:00,000 --> 00:46:02,986
Fred Parietti: Yeah. I mean,
right now, the kind of the

816
00:46:02,986 --> 00:46:06,243
paradox is that when people
reach out and they're like, hey,

817
00:46:06,243 --> 00:46:09,717
guys, we would really like some
robots right now because they're

818
00:46:09,717 --> 00:46:12,866
always like, when they reach out
to you, there are oh, no,

819
00:46:12,866 --> 00:46:16,231
yesterday, the robots yesterday.
Totally. Like, okay, there is

820
00:46:16,231 --> 00:46:19,000
about a, you know, two year
waiting list right now.

821
00:46:19,000 --> 00:46:23,000
Audrow Nash: Okay. Wow. Oh,
that's so exciting.

822
00:46:23,000 --> 00:46:27,306
Fred Parietti: So we basically
need to, to work very closely

823
00:46:27,306 --> 00:46:31,468
with them to extend the supply
chain, maintaining the same

824
00:46:31,468 --> 00:46:35,630
quality. Right. And then find
ways to make it happen. And,

825
00:46:35,630 --> 00:46:40,079
yeah. Yeah. And so, it's I think
it's very exciting. Is it, it

826
00:46:40,079 --> 00:46:44,313
reflects the thing where if you
go and look for a available

827
00:46:44,313 --> 00:46:48,762
application of robotics, like
the market is there, right? And,

828
00:46:48,762 --> 00:46:53,140
you know, I really sometimes
don't like them. So I don't like

829
00:46:53,140 --> 00:46:57,374
wishy washy, argument that
sometimes people have. Well, you

830
00:46:57,374 --> 00:47:01,177
know, oh, we are robotics
company and the robots will

831
00:47:01,177 --> 00:47:04,837
eventually be used for
everything. So our market is

832
00:47:04,837 --> 00:47:09,358
infinite. So you don't need to
worry investor about our revenue

833
00:47:09,358 --> 00:47:13,520
today because we're doing
general purpose robotics. And so

834
00:47:13,520 --> 00:47:17,826
our market is infinite, right.
And that I think, is too many

835
00:47:17,826 --> 00:47:22,490
times using that as an excuse to
justify like zero revenue. Yeah.

836
00:47:22,490 --> 00:47:26,365
Right. To justify a poor
business. You know, it's like

837
00:47:26,365 --> 00:47:30,671
the Silicon Valley show, when
there was that joke, you know,

838
00:47:30,671 --> 00:47:35,048
and it's HBO, there was a joke
that, you know, they said, you

839
00:47:35,048 --> 00:47:39,354
know, the you must try to stay a
zero revenue for as much as

840
00:47:39,354 --> 00:47:40,000
possible.

841
00:47:40,000 --> 00:47:41,000
Audrow Nash: Because zero.

842
00:47:41,000 --> 00:47:44,171
Fred Parietti: Revenue is better
than revenue because it lets you

843
00:47:44,171 --> 00:47:45,000
paint a gigantic.

844
00:47:45,000 --> 00:47:46,000
Audrow Nash: Infinite potential.
Yeah.

845
00:47:46,000 --> 00:47:49,934
Fred Parietti: Right. When you
start having revenue, well, you

846
00:47:49,934 --> 00:47:54,058
actually measured on how much is
it growing. Right. And so we are

847
00:47:54,058 --> 00:47:57,421
definitely on the camp that you
should have available

848
00:47:57,421 --> 00:48:00,975
application measurable. You
know, growing real you know,

849
00:48:00,975 --> 00:48:04,909
revenue. Right. But I'm seeing
you dating too many times right

850
00:48:04,909 --> 00:48:08,716
now in robotics, the sort of
like infinite potential kind of

851
00:48:08,716 --> 00:48:11,000
like hype approach. And I hope
that.

852
00:48:11,000 --> 00:48:13,000
Audrow Nash: Yeah, yeah.

853
00:48:13,000 --> 00:48:16,136
Fred Parietti: I just hope that
that doesn't create a bad name,

854
00:48:16,136 --> 00:48:19,172
you know, for robotics, because
when the bubble bursts and it

855
00:48:19,172 --> 00:48:22,009
will burst, right? Only the
people with real revenue, you

856
00:48:22,009 --> 00:48:24,000
know, will have a business will
survive.

857
00:48:24,000 --> 00:48:25,000
Audrow Nash: Yeah.

858
00:48:24,000 --> 00:48:26,631
Fred Parietti: But I hope that
people don't conclude, oh,

859
00:48:26,631 --> 00:48:29,492
robotics is like that because,
you know, the work we are doing

860
00:48:29,492 --> 00:48:30,000
is so real.

861
00:48:30,000 --> 00:48:34,041
Audrow Nash: Hell, yeah. Yeah. I
think my impression, I mean, if

862
00:48:34,041 --> 00:48:37,703
you. Look like the.com bust or
something, it's like, yeah,

863
00:48:37,703 --> 00:48:41,617
there was a bunch of bubbles.
But clearly the internet is huge

864
00:48:41,617 --> 00:48:45,216
now. It's like, I think that
robotics will be similar for

865
00:48:45,216 --> 00:48:49,320
this, where there are very clear
applications for robotics, where

866
00:48:49,320 --> 00:48:53,424
you're making a ton of money for
companies. And those will be the

867
00:48:53,424 --> 00:48:57,339
ones that carry everything. The
hype. I'm a lot more skeptical

868
00:48:57,339 --> 00:49:01,190
about. I feel similar to you
about the general purpose robots

869
00:49:01,190 --> 00:49:04,789
for this kind of thing. Yeah, it
feels like when I was in

870
00:49:04,789 --> 00:49:08,641
academia, One thing I thought
was really funny was that a lot

871
00:49:08,641 --> 00:49:12,176
of the research that was very
cool and like, you'd see a

872
00:49:12,176 --> 00:49:16,028
quadrotor doing awesome things.
It's like, well, why do we do

873
00:49:16,028 --> 00:49:20,006
this? And it's like search and
rescue. Why do we do quadrupeds?

874
00:49:20,006 --> 00:49:23,857
It's like search and rescue.
Every. Everything that couldn't.

875
00:49:23,857 --> 00:49:27,645
Even, like. Pipe robots or, blow
up robots. It was like, oh,

876
00:49:27,645 --> 00:49:31,497
search and rescue for sure. It
was like the catchall. And now

877
00:49:31,497 --> 00:49:34,906
it's similar as a general
purpose robots when we don't

878
00:49:34,906 --> 00:49:38,000
know exactly what the market is
we're going into.

879
00:49:38,000 --> 00:49:41,847
Fred Parietti: Yeah. And I can
tell you how many robots I saw

880
00:49:41,847 --> 00:49:45,695
that, like, make pizzas. And I'm
like, guys, right? Like, you

881
00:49:45,695 --> 00:49:49,479
know, there are two kinds of
pizzas, either pizzas for which

882
00:49:49,479 --> 00:49:53,264
minimum wage labor is fine and
there are pizza for which you

883
00:49:53,264 --> 00:49:57,048
really want the human experience
of someone making, like, an

884
00:49:57,048 --> 00:50:01,022
incredible. Right? Yeah. And in
both cases, the value per pizza

885
00:50:01,022 --> 00:50:04,680
is very low. The same amount of
math in terms of like gene

886
00:50:04,680 --> 00:50:08,717
edited cells for actually saving
people's life, has a value of a

887
00:50:08,717 --> 00:50:12,501
few million dollars. Not like
the $20 on a pizza. Right. And

888
00:50:12,501 --> 00:50:16,601
so like, like it doesn't make it
in the technology we're using is

889
00:50:16,601 --> 00:50:20,259
substantially the same, right?
Yeah, we're using imitation

890
00:50:20,259 --> 00:50:24,107
learning, machine learning and
reinforcement learning. We are

891
00:50:24,107 --> 00:50:27,828
using only collaborative robots.
So it's literally the same

892
00:50:27,828 --> 00:50:31,549
technology. In one case it is a
business. In the other case

893
00:50:31,549 --> 00:50:33,000
there's not a business.

894
00:50:33,000 --> 00:50:36,708
Audrow Nash: Yes. Yeah. We were
talking. Before we started

895
00:50:36,708 --> 00:50:40,161
recording like for house
cleaning and stuff where it's

896
00:50:40,161 --> 00:50:43,870
like the humanoid will clean
your house. It's like. But. A

897
00:50:43,870 --> 00:50:48,026
person that you pay $25 an hour,
we'll also clean your house, and

898
00:50:48,026 --> 00:50:51,734
they'll probably do a better
job, and they're not going to

899
00:50:51,734 --> 00:50:55,762
fall over. These kinds of things
I really like your perspective

900
00:50:55,762 --> 00:50:58,000
on go right to the high value
task.

901
00:50:58,000 --> 00:51:00,000
Fred Parietti: Versus mean.

902
00:51:00,000 --> 00:51:02,000
Audrow Nash: Go low end. Yeah.

903
00:51:02,000 --> 00:51:06,446
Fred Parietti: It's exactly it.
And so these you know, our goals

904
00:51:06,446 --> 00:51:10,475
is you know, our goal is to
develop a very cool robot. All

905
00:51:10,475 --> 00:51:14,782
it is mine is right. And so I'm
like okay so I need you know I

906
00:51:14,782 --> 00:51:19,089
want to the value of the best
possible robotic systems. Right.

907
00:51:19,089 --> 00:51:23,326
And so I'm like which is smarter
going for $20 an hour all is

908
00:51:23,326 --> 00:51:27,633
equal. Or going for $20,000 an
hour or maybe $200,000 an hour,

909
00:51:27,633 --> 00:51:32,148
which is honestly like the order
of magnitude we're talking about

910
00:51:32,148 --> 00:51:35,760
for next gen therapy
manufacturing. Right. And so if

911
00:51:35,760 --> 00:51:40,067
you can have a business that is
for order of magnitude better,

912
00:51:40,067 --> 00:51:44,444
literally, right, 3 or 4, right?
Of course I can develop better

913
00:51:44,444 --> 00:51:48,681
robots higher, better engineers.
Right. And it makes sense to

914
00:51:48,681 --> 00:51:53,058
start from there. Now, I'm not
saying that we're never going to

915
00:51:53,058 --> 00:51:57,364
get to, you know, humanize it.
Clean the house. Of course, I'm

916
00:51:57,364 --> 00:52:01,671
looking forward to that I can't
wait. Very happy to buy for 30

917
00:52:01,671 --> 00:52:05,770
K, a humanoid that cleans my
house. I would love it. Right.

918
00:52:05,770 --> 00:52:09,660
But, you know, at the same time,
you know, it's way more

919
00:52:09,660 --> 00:52:13,758
reasonable to start from high
value. And then 20 years from

920
00:52:13,758 --> 00:52:17,718
now, get 2 or 10 or whatever.
Five, certainly, though not

921
00:52:17,718 --> 00:52:21,608
today. Get to those mass
manufacturing or mass or sorry,

922
00:52:21,608 --> 00:52:25,707
low added value tasks. Right.
And we need to also avoid the

923
00:52:25,707 --> 00:52:29,875
risk that we promise that the
human is for the House already

924
00:52:29,875 --> 00:52:34,043
today and two years from now,
people get bored of it, right?

925
00:52:34,043 --> 00:52:38,349
Oh, of course is not happening.
And so they then conclude that

926
00:52:38,349 --> 00:52:42,448
all robotics is BS and hype.
Right? And I don't want to get

927
00:52:42,448 --> 00:52:46,199
into a post .com bubble
situation for robotics. Right.

928
00:52:46,199 --> 00:52:50,159
And I think with AI, with
digital, I like, you know, like

929
00:52:50,159 --> 00:52:54,466
large language models, they are
useful enough that I am pretty

930
00:52:54,466 --> 00:52:58,008
sure there will be some
disillusionment, but is now

931
00:52:58,008 --> 00:53:02,037
going to go back to zero. Yeah,
but the ChatGPT moment for

932
00:53:02,037 --> 00:53:06,205
robotics has not arrived yet,
and so I just don't want to go

933
00:53:06,205 --> 00:53:10,095
back into a robotic winter
because people were promising

934
00:53:10,095 --> 00:53:13,916
cheap humanoid to sweep the
floor and could not deliver

935
00:53:13,916 --> 00:53:16,000
businesses that could do that.

936
00:53:16,000 --> 00:53:20,289
Audrow Nash: Yeah. I wonder, my
perspective on this is thinking.

937
00:53:20,289 --> 00:53:24,376
About the autonomous cars and
all the hype around that. So in

938
00:53:24,376 --> 00:53:28,598
like 2015, it was like next year
we're going to have autonomous

939
00:53:28,598 --> 00:53:32,485
cars. It's going to change
everything. There's going to be

940
00:53:32,485 --> 00:53:36,706
no more truck drivers, no more
Uber drivers. All transportation

941
00:53:36,706 --> 00:53:40,928
will be revolutionized. And then
it's like, okay, in 2018, it's

942
00:53:40,928 --> 00:53:45,015
probably going to happen. And
then in. 2020, whatever. And so

943
00:53:45,015 --> 00:53:48,768
now, but then. What you see
actually. Is the public just

944
00:53:48,768 --> 00:53:52,722
totally. Loses interest. Yeah.
It was so happy. But I don't

945
00:53:52,722 --> 00:53:56,809
think it's maybe the investors
because actually, if you look.

946
00:53:56,809 --> 00:54:00,830
At the funding, over the last
five years in robotics, it was

947
00:54:00,830 --> 00:54:04,784
majority autonomous vehicles.
And then that went to zero in

948
00:54:04,784 --> 00:54:08,067
the last little bit. But,
because talking to some

949
00:54:08,067 --> 00:54:12,021
investors who were on reports,
they've shown me some of the

950
00:54:12,021 --> 00:54:16,309
reports, I had an interview with
F Prime Sanjay from it where we

951
00:54:16,309 --> 00:54:20,397
talked about this and, but the,
the slice of more specialized

952
00:54:20,397 --> 00:54:24,284
robotics applications, kind of
like what you and I seem to

953
00:54:24,284 --> 00:54:27,835
value is just gradually
increasing. And you have this

954
00:54:27,835 --> 00:54:31,856
big overhead. On top of it for
the autonomous cars that then

955
00:54:31,856 --> 00:54:34,000
goes to zero when the hype dies.

956
00:54:34,000 --> 00:54:39,167
Fred Parietti: Yeah. Yeah. No.
Absolutely, absolutely. So your

957
00:54:39,167 --> 00:54:40,000
finishing.

958
00:54:40,000 --> 00:54:43,324
Audrow Nash: Yeah. But the
thing. That I think that's

959
00:54:43,324 --> 00:54:46,962
interesting about that is it
feels like so we have a Tesla

960
00:54:46,962 --> 00:54:50,161
and it just drives itself
everywhere. It makes some

961
00:54:50,161 --> 00:54:53,799
mistakes. But so there was so
much hype ten. Years ago for

962
00:54:53,799 --> 00:54:57,625
this. And then in. My opinion,
it's like Tesla quietly solved

963
00:54:57,625 --> 00:55:01,514
it. And there's no attention on
it. Now that it's solved. It's

964
00:55:01,514 --> 00:55:05,341
not totally solved, but it's
really, really good. Where I can

965
00:55:05,341 --> 00:55:09,167
drive anywhere in town and I
just ding it on to autopilot and

966
00:55:09,167 --> 00:55:12,867
it does it almost free. But I
wonder if it'll be similar to

967
00:55:12,867 --> 00:55:15,000
this. But. Yes. What do you
think?

968
00:55:15,000 --> 00:55:20,451
Fred Parietti: It is? True. But
keep in mind. So who survived?

969
00:55:20,451 --> 00:55:23,000
All right, I Waymo and Tesla.

970
00:55:23,000 --> 00:55:24,000
Audrow Nash: Yeah.

971
00:55:23,000 --> 00:55:25,000
Fred Parietti: Right. Pretty
much right.

972
00:55:25,000 --> 00:55:30,333
Audrow Nash: And Zoox I don't
know much Zoox and a cruise. I

973
00:55:30,333 --> 00:55:33,000
think is basically dead. Yeah.

974
00:55:33,000 --> 00:55:36,949
Fred Parietti: And so I mean I
had several friends working

975
00:55:36,949 --> 00:55:40,898
there and nobody works there
anymore. And so, nobody of my

976
00:55:40,898 --> 00:55:45,119
friends. Right. And so, and so,
so the idea here is, when, you

977
00:55:45,119 --> 00:55:49,272
know, when you hype and overhype
all these companies were in.

978
00:55:49,272 --> 00:55:53,494
Right, you know, in kind of like
hyping this so much, then the

979
00:55:53,494 --> 00:55:57,374
public then is already very
jaded, you know, because, you

980
00:55:57,374 --> 00:56:01,664
know, when you get promises, a
lot of things that don't happen,

981
00:56:01,664 --> 00:56:05,000
I think it's natural to be
skeptical, right? Yes.

982
00:56:05,000 --> 00:56:06,000
Audrow Nash: For sure.

983
00:56:05,000 --> 00:56:08,576
Fred Parietti: And so public
loses interest, investors, even

984
00:56:08,576 --> 00:56:12,093
though they love to say that
they are decoupled from public

985
00:56:12,093 --> 00:56:14,000
markets and but sentiment. Yeah.

986
00:56:14,000 --> 00:56:15,000
Audrow Nash: Right.

987
00:56:15,000 --> 00:56:18,667
Fred Parietti: Like carbon
copies. In fact, I would say

988
00:56:18,667 --> 00:56:20,000
they are amplifiers.

989
00:56:20,000 --> 00:56:22,000
Audrow Nash: Yeah, I think so
too. Right.

990
00:56:22,000 --> 00:56:24,652
Fred Parietti: Like they are
they try to be as sensitive as

991
00:56:24,652 --> 00:56:26,000
possible to the public trends.

992
00:56:26,000 --> 00:56:28,000
Audrow Nash: Yep.

993
00:56:28,000 --> 00:56:32,421
Fred Parietti: And so basically
the investment dries up and 99%

994
00:56:32,421 --> 00:56:36,772
of companies die. The only ones
who survived are Tesla. Tesla.

995
00:56:36,772 --> 00:56:41,333
Because basically Elon can raise
infinitely cheap capital. Right.

996
00:56:41,333 --> 00:56:45,825
But that's not a industry thing.
That's an know like, oh, he has

997
00:56:45,825 --> 00:56:49,614
infinite access to capital.
Therefore he can fund says

998
00:56:49,614 --> 00:56:52,000
anything forever. Right. And
that.

999
00:56:52,000 --> 00:56:55,084
Audrow Nash: And they just make
good. Cars anyways too. So like

1000
00:56:55,084 --> 00:56:58,168
there's even if they didn't do
the self-driving, it's like it's

1001
00:56:58,168 --> 00:56:59,000
still a good car.

1002
00:56:59,000 --> 00:57:02,983
Fred Parietti: Cool, right? But
the self-driving on its own

1003
00:57:02,983 --> 00:57:06,966
would have not been a yes even
in business. Right? It was a

1004
00:57:06,966 --> 00:57:11,017
sense to say it was saved by
infinite access to capital. And

1005
00:57:11,017 --> 00:57:15,000
the same for Waymo, right.
Because it's Google with Google.

1006
00:57:15,000 --> 00:57:16,000
Audrow Nash: Yeah.

1007
00:57:16,000 --> 00:57:20,586
Fred Parietti: Right. In the
same for a Zoox that was bought

1008
00:57:20,586 --> 00:57:25,019
by Amazon. So the only right
like and so like so basically

1009
00:57:25,019 --> 00:57:28,000
when disillusionment sets in
post hype.

1010
00:57:28,000 --> 00:57:31,598
Audrow Nash: Everyone dies if
they don't attach to a larger

1011
00:57:31,598 --> 00:57:33,000
entity that funds them.

1012
00:57:33,000 --> 00:57:35,697
Fred Parietti: Unless you have
inflated money. And obviously

1013
00:57:35,697 --> 00:57:37,000
that's not a reasonable plan.

1014
00:57:37,000 --> 00:57:38,000
Audrow Nash: For a startup.

1015
00:57:38,000 --> 00:57:40,988
Fred Parietti: What I meant to
do robotics and hype it up, and

1016
00:57:40,988 --> 00:57:43,976
then try to attach myself to an
infinite source of money, like

1017
00:57:43,976 --> 00:57:46,000
that's not a business plan. And
so I have.

1018
00:57:46,000 --> 00:57:49,214
Audrow Nash: Some businesses.
Are bought, made to be bought,

1019
00:57:49,214 --> 00:57:52,000
which I don't really like, but
that it is happening.

1020
00:57:52,000 --> 00:57:56,405
Fred Parietti: But then say that
you just want an exit as rapidly

1021
00:57:56,405 --> 00:58:00,268
as possible because you know
you're going to die. And the

1022
00:58:00,268 --> 00:58:04,199
only way to not die is to be
bought. Right? And so and so.

1023
00:58:04,199 --> 00:58:08,266
But to me, you're much stronger
if you're actually trying to

1024
00:58:08,266 --> 00:58:12,535
build something that matters. So
apply that technology in a way

1025
00:58:12,535 --> 00:58:16,195
that, has a, a sustainable
business. That's what we're

1026
00:58:16,195 --> 00:58:20,329
trying to do. And that's what I
worry a lot of people are not

1027
00:58:20,329 --> 00:58:24,734
trying to do with these infinite
market pitches. And, and so when

1028
00:58:24,734 --> 00:58:28,530
the winter comes. Right, you
know, then I think a lot of

1029
00:58:28,530 --> 00:58:32,732
companies would be caught like
that. And if it's too many, the

1030
00:58:32,732 --> 00:58:36,663
public will conclude that
robotics is useless or not ready

1031
00:58:36,663 --> 00:58:41,000
yet. And I don't want to repeat
another self-driving car winter.

1032
00:58:41,000 --> 00:58:44,825
Audrow Nash: Was it that bad?
Because from my perspective, a

1033
00:58:44,825 --> 00:58:48,586
lot of the smart. People from
the self-driving car industry

1034
00:58:48,586 --> 00:58:52,219
went into other robotics
startups. They started companies

1035
00:58:52,219 --> 00:58:55,916
in agriculture and mining and
like a bunch of really other

1036
00:58:55,916 --> 00:58:59,932
good robotics companies. I don't
know that it was like. Because

1037
00:58:59,932 --> 00:59:04,012
as. I mentioned, the investment
across. Other like more specific

1038
00:59:04,012 --> 00:59:07,709
robotics applications. Has just
been growing over the last

1039
00:59:07,709 --> 00:59:11,661
years. It's a fraction of what
was poured into autonomous cars

1040
00:59:11,661 --> 00:59:13,000
but is still growing.

1041
00:59:13,000 --> 00:59:15,000
Fred Parietti: It is true. It is
true. But I don't.

1042
00:59:15,000 --> 00:59:16,000
Audrow Nash: Want to winter too.

1043
00:59:16,000 --> 00:59:19,978
Fred Parietti: But I also like,
if I want a self-driving car

1044
00:59:19,978 --> 00:59:24,088
right now, I either need to have
a Tesla. Right? But then, you

1045
00:59:24,088 --> 00:59:28,000
know, with the caveat that you
still need behind the wheel.

1046
00:59:28,000 --> 00:59:29,000
Audrow Nash: Yes. And you have
to pay attention.

1047
00:59:29,000 --> 00:59:32,848
Fred Parietti: Right. Or I need
to jump on a Waymo, which I do

1048
00:59:32,848 --> 00:59:36,324
all the time here in San
Francisco. Right. And, but it's

1049
00:59:36,324 --> 00:59:38,000
it there's not there's not.

1050
00:59:38,000 --> 00:59:40,000
Audrow Nash: Yeah. You got the
big. Players and that's it.

1051
00:59:40,000 --> 00:59:43,847
Fred Parietti: So it's kind of
like a little disappointing that

1052
00:59:43,847 --> 00:59:47,084
I can't get a reliable
self-driving taxi everywhere I

1053
00:59:47,084 --> 00:59:50,931
go. And that's basically I think
that was totally within reach.

1054
00:59:50,931 --> 00:59:54,595
Because if Waymo did it, if
Tesla, you know, is almost doing

1055
00:59:54,595 --> 00:59:57,771
it right, then many other
players could have done it

1056
00:59:57,771 --> 01:00:01,740
technically. And so I think they
just messed up AI from the point

1057
01:00:01,740 --> 01:00:04,000
of view of the hype and the
business.

1058
01:00:04,000 --> 01:00:08,241
Audrow Nash: Yeah, that's a good
point. I wonder my thought with.

1059
01:00:08,241 --> 01:00:12,287
Humanoids and I wonder what you
think is that it's going to go

1060
01:00:12,287 --> 01:00:16,267
kind of like autonomous cars.
Where. It will take longer than

1061
01:00:16,267 --> 01:00:20,182
everyone thinks. And the public
will lose interest. And then

1062
01:00:20,182 --> 01:00:24,096
Tesla will quietly make a 30 K
humanoid that can actually do

1063
01:00:24,096 --> 01:00:28,077
useful stuff. But it might be 15
years from now or something,

1064
01:00:28,077 --> 01:00:32,252
maybe faster. Who knows? Do you
think it'll go similar for this?

1065
01:00:32,252 --> 01:00:36,167
And it'll be kind of like the
companies either get bought on

1066
01:00:36,167 --> 01:00:39,430
maybe favorable, maybe
unfavorable terms by larger

1067
01:00:39,430 --> 01:00:43,149
companies. And then a lot of
them go out of business, but

1068
01:00:43,149 --> 01:00:47,194
Tesla kind of survives. Or maybe
who knows what the other ones

1069
01:00:47,194 --> 01:00:50,000
will do. But what is. Your
thought on this?

1070
01:00:50,000 --> 01:00:53,391
Fred Parietti: I think it's more
likely than not that it would go

1071
01:00:53,391 --> 01:00:56,000
that way. However, I don't want
it to go that way.

1072
01:00:56,000 --> 01:00:58,000
Audrow Nash: And because it's
too centralized. Yeah.

1073
01:00:58,000 --> 01:01:01,720
Fred Parietti: We're doing. Yeah
yeah yeah. But also like, you

1074
01:01:01,720 --> 01:01:05,380
know, new winter and so on and
you know, and then the winter.

1075
01:01:05,380 --> 01:01:09,280
Yeah. Ten years later. And so no
like we don't want it to go that

1076
01:01:09,280 --> 01:01:13,180
way. And so we're actually doing
everything we can to not go that

1077
01:01:13,180 --> 01:01:16,720
way. So for example we got a
humanoid robot for 30 K. Yeah.

1078
01:01:16,720 --> 01:01:20,200
Not from Tesla. They're not
selling it yet. We got it from

1079
01:01:20,200 --> 01:01:23,980
unitary. All right. And so I can
tell you that the hardware you

1080
01:01:23,980 --> 01:01:27,400
can already find the hardware
for reasonably, you know, a

1081
01:01:27,400 --> 01:01:30,640
reasonable cost like 30 K.
Right. And so I assume that

1082
01:01:30,640 --> 01:01:34,000
Tesla and figure hopefully will
get there at some point.

1083
01:01:34,000 --> 01:01:37,068
Audrow Nash: That they're
selling. You're saying or that

1084
01:01:37,068 --> 01:01:38,000
they get to 30 K.

1085
01:01:38,000 --> 01:01:42,000
Fred Parietti: That they, that
they hopefully get to sell them

1086
01:01:42,000 --> 01:01:46,065
for 30 K. Reasonable. Right now
though, the only company can do

1087
01:01:46,065 --> 01:01:49,548
that is unitary. And so, so
that's what we got. So the

1088
01:01:49,548 --> 01:01:53,613
constrained AI now is actually
not the hardware. The constraint

1089
01:01:53,613 --> 01:01:57,419
is the software that nobody,
like. The humanoid is not very

1090
01:01:57,419 --> 01:02:01,097
useful even though it's fully
functional Uber wise. Yeah.

1091
01:02:01,097 --> 01:02:05,226
Because you need to, teach it to
you useful stuff. And so we are

1092
01:02:05,226 --> 01:02:09,097
basically experimenting to use
humanoids in conjunction with

1093
01:02:09,097 --> 01:02:12,710
our robotic systems. And the
idea is, inside our robotic

1094
01:02:12,710 --> 01:02:16,581
system is completely autonomous
with all those robotic arms.

1095
01:02:16,581 --> 01:02:20,774
Yeah. And the human outside good
enough. Exactly. The outside. So

1096
01:02:20,774 --> 01:02:24,516
the biggest bottleneck right now
is loading and unloading.

1097
01:02:24,516 --> 01:02:26,000
That's. Yeah, manually.

1098
01:02:26,000 --> 01:02:28,000
Audrow Nash: I guess that that's
pretty funny.

1099
01:02:28,000 --> 01:02:29,000
Fred Parietti: Yeah. Right.

1100
01:02:29,000 --> 01:02:30,000
Audrow Nash: And so if.

1101
01:02:30,000 --> 01:02:33,300
Fred Parietti: You now have
people kind of bored to death

1102
01:02:33,300 --> 01:02:37,006
loading and unloading a bunch of
carts, you know, and cartridges

1103
01:02:37,006 --> 01:02:40,480
into a robotic system, well, can
I use humanoid? I think the

1104
01:02:40,480 --> 01:02:43,954
hardware is good enough. And so
now I'm not pitching to you.

1105
01:02:43,954 --> 01:02:47,544
Potential customer potential for
my company. I'm not pitching.

1106
01:02:47,544 --> 01:02:51,134
You know what? I'm going to the
humanoid, and I'm going to use

1107
01:02:51,134 --> 01:02:54,666
it to do a bunch of, like, far
fetched stuff that we all know

1108
01:02:54,666 --> 01:02:58,082
doesn't work yet. Now I'm going
to tell you we will use the

1109
01:02:58,082 --> 01:03:01,382
general purpose humanoid that
only costs 30 K to pick and

1110
01:03:01,382 --> 01:03:04,799
place boxes. So instead of
having, you know, 2 or 3 people,

1111
01:03:04,799 --> 01:03:08,041
you have like ten cheap
humanoids doing that slowly. But

1112
01:03:08,041 --> 01:03:11,283
I can do it with current
technology, and that's actually

1113
01:03:11,283 --> 01:03:14,815
very valuable to you, because
now your cleanroom doesn't have

1114
01:03:14,815 --> 01:03:18,000
people, it only has robots
loading other robots. Right.

1115
01:03:18,000 --> 01:03:19,000
Audrow Nash: Yeah.

1116
01:03:19,000 --> 01:03:23,200
Fred Parietti: And so I think we
are the current technology can

1117
01:03:23,200 --> 01:03:27,200
be deployed effectively with a
solid business model in those

1118
01:03:27,200 --> 01:03:30,933
applications. And I believe if
there are enough of these

1119
01:03:30,933 --> 01:03:34,200
applications, we can perhaps
avoid the winter and

1120
01:03:34,200 --> 01:03:38,000
disillusionment where maybe
people will be, oh, okay, you

1121
01:03:38,000 --> 01:03:42,133
mind for sweeping? The floor is
still far, but I can see human

1122
01:03:42,133 --> 01:03:45,867
as being very useful for these
high value add, you know,

1123
01:03:45,867 --> 01:03:49,600
applications. And so and so.
Maybe it's not all to throw

1124
01:03:49,600 --> 01:03:53,267
away. I really want to avoid
that, because that sort of

1125
01:03:53,267 --> 01:03:57,333
public disillusionment slows
down the technology a lot. And I

1126
01:03:57,333 --> 01:04:00,000
don't want the development to
slow down.

1127
01:04:00,000 --> 01:04:03,854
Audrow Nash: I wonder though, I
mean I, I love the. Perspective

1128
01:04:03,854 --> 01:04:07,647
and I think it's so cool that
you can have a, you. Said do you

1129
01:04:07,647 --> 01:04:11,256
think, do you think there were.
So the humanoid loading and

1130
01:04:11,256 --> 01:04:14,682
unloading the other robotic
system. How much per hour of

1131
01:04:14,682 --> 01:04:18,169
value do you think they're
creating? Just because I think

1132
01:04:18,169 --> 01:04:21,779
you said a very high number, and
I want to I want to see if

1133
01:04:21,779 --> 01:04:25,449
that's right, because that is
crazy and inspiring. If that's

1134
01:04:25,449 --> 01:04:26,000
the case.

1135
01:04:26,000 --> 01:04:29,676
Fred Parietti: Oh, the loading
unloading is not that. So like

1136
01:04:29,676 --> 01:04:33,353
if we're talking about tens of
thousands of dollars per hour,

1137
01:04:33,353 --> 01:04:36,306
you are considering the
throughput of the robotic

1138
01:04:36,306 --> 01:04:39,982
system. Keep in mind of this
single cell therapy right now is

1139
01:04:39,982 --> 01:04:43,658
like around 400 K all the way to
$2 million per dose. Wow. So

1140
01:04:43,658 --> 01:04:47,214
you're talking about a box of
cells like, you know, imagine

1141
01:04:47,214 --> 01:04:50,770
like a leader or, you know,
roughly a third to a quarter of

1142
01:04:50,770 --> 01:04:53,000
a gallon, right? Yeah. A little
milk.

1143
01:04:53,000 --> 01:04:55,000
Audrow Nash: One of those little
boxes.

1144
01:04:55,000 --> 01:04:57,682
Fred Parietti: Is half $1
million, and we're making a lot

1145
01:04:57,682 --> 01:04:59,000
of those things in parallel.

1146
01:04:59,000 --> 01:05:01,000
Audrow Nash: Holy cow. So, so.

1147
01:05:01,000 --> 01:05:04,044
Fred Parietti: Imagine, like, we
make, you know, dozens or

1148
01:05:04,044 --> 01:05:07,140
hundreds of them in parallel,
and, you know, they take, you

1149
01:05:07,140 --> 01:05:10,394
know, a few days to a couple of
week, it clicks. Right? So the

1150
01:05:10,394 --> 01:05:13,542
average throughput per hour is
actually tens of thousands of

1151
01:05:13,542 --> 01:05:16,638
dollars. Wow. Loading is just
you know, you could load them

1152
01:05:16,638 --> 01:05:19,000
manually. So I would say 2 or 2.
But you can.

1153
01:05:19,000 --> 01:05:22,107
Audrow Nash: Extend the whole
system in some sense. You're

1154
01:05:22,107 --> 01:05:25,429
making your system have softer
boundaries with the rest of the

1155
01:05:25,429 --> 01:05:28,589
other world. Pharmaceutical
company. Yeah, the rest. Of the

1156
01:05:28,589 --> 01:05:31,696
world. So that's really cool.
And that adds a lot of value

1157
01:05:31,696 --> 01:05:35,018
because your system could just
run continuously. Yeah. And you

1158
01:05:35,018 --> 01:05:38,446
don't have to rely on people and
hiring issues. And all sorts of

1159
01:05:38,446 --> 01:05:41,661
other things that might occur
and sanitation risks that come

1160
01:05:41,661 --> 01:05:43,000
with people for all this.

1161
01:05:43,000 --> 01:05:46,606
Fred Parietti: Exactly, exactly,
exactly, exactly. And so and so

1162
01:05:46,606 --> 01:05:50,268
in and so there's, so there is a
place for humanoids. And now you

1163
01:05:50,268 --> 01:05:53,592
might argue, what right do you
really need the legs? Not we

1164
01:05:53,592 --> 01:05:55,000
don't need legs. We just.

1165
01:05:55,000 --> 01:05:56,000
Audrow Nash: I was going to ask.

1166
01:05:55,000 --> 01:05:58,168
Fred Parietti: But two arms.
Right. Yeah. You don't need to

1167
01:05:58,168 --> 01:06:01,389
face. Those are just added for
esthetic purposes. Right? You

1168
01:06:01,389 --> 01:06:03,000
can put those. They're pretty.

1169
01:06:03,000 --> 01:06:05,520
Audrow Nash: Cheap. So, like,
why not? You don't have to worry

1170
01:06:05,520 --> 01:06:08,000
about manufacturing it. That's
actually a big valuable thing.

1171
01:06:08,000 --> 01:06:12,598
Fred Parietti: Honestly. We are
also evaluating different form

1172
01:06:12,598 --> 01:06:16,752
factors, for, you know, general
purpose, the loading and

1173
01:06:16,752 --> 01:06:21,202
loading, general purpose robotic
part, and provided value is

1174
01:06:21,202 --> 01:06:25,355
cheap. And we have two arms
because you need two arms to

1175
01:06:25,355 --> 01:06:29,361
pick and play stuff. We're fine.
Right. It's just that

1176
01:06:29,361 --> 01:06:33,069
technology. We believe we've
found a very valuable

1177
01:06:33,069 --> 01:06:37,000
application that I can sell
today. You know, with it.

1178
01:06:37,000 --> 01:06:39,000
Audrow Nash: And in the future.
Yeah, yeah, yeah.

1179
01:06:39,000 --> 01:06:42,308
Fred Parietti: And that I know
customers will pay for it

1180
01:06:42,308 --> 01:06:45,734
because it's very available.
Right. If I was pitching you.

1181
01:06:45,734 --> 01:06:49,397
Oh, you know what? I'm gonna ask
a humanoid to clean your car.

1182
01:06:49,397 --> 01:06:52,527
And you would have so many
questions, and it would be

1183
01:06:52,527 --> 01:06:53,000
unclear.

1184
01:06:53,000 --> 01:06:55,000
Audrow Nash: If that would be
how much you'd be willing.

1185
01:06:55,000 --> 01:06:56,000
Fred Parietti: To pay me. Right?

1186
01:06:56,000 --> 01:07:00,602
Audrow Nash: Yeah. Yeah, that
would be funny. Cleans your car,

1187
01:07:00,602 --> 01:07:05,279
does an okay job, and it costs
five K for this kind of thing or

1188
01:07:05,279 --> 01:07:09,955
something, I don't know. Yeah,
that'd be funny. Okay, so that's

1189
01:07:09,955 --> 01:07:14,261
really cool. I like that.
Perspective a lot. I hope we can

1190
01:07:14,261 --> 01:07:19,086
avoid any winters from that kind
of thing. I think getting robots

1191
01:07:19,086 --> 01:07:23,540
out there is probably the way to
do. That with this. It does

1192
01:07:23,540 --> 01:07:28,216
seem. Like humanoids in a pharma
lab toting things around while

1193
01:07:28,216 --> 01:07:32,522
it is high value and it does
involve humanoids. Maybe your

1194
01:07:32,522 --> 01:07:36,976
videos and things like this will
expose it more, but like, I

1195
01:07:36,976 --> 01:07:41,207
didn't know anything about this
really before. How do you

1196
01:07:41,207 --> 01:07:45,586
expose? Because there's so much,
like. Humanoids doing side

1197
01:07:45,586 --> 01:07:49,892
flips, videos and things like
this. Versus humanoids doing

1198
01:07:49,892 --> 01:07:54,420
useful work in. It's like, how
do you how do you. Attach this

1199
01:07:54,420 --> 01:07:58,577
to the public perception of
humanoids so that it doesn't

1200
01:07:58,577 --> 01:08:02,882
burn when humanoids are not
useful and then tomorrow or in

1201
01:08:02,882 --> 01:08:06,000
101, one year or whatever, we're
promised.

1202
01:08:06,000 --> 01:08:10,318
Fred Parietti: Honestly, to tell
the truth, I don't care about

1203
01:08:10,318 --> 01:08:14,636
the public's perception. Yeah, I
don't care about the public's

1204
01:08:14,636 --> 01:08:19,093
perception. The idea is I don't
do these because I want to shape

1205
01:08:19,093 --> 01:08:23,272
the perception. I will do this
because I want to develop the

1206
01:08:23,272 --> 01:08:27,729
best possible robotics and right
now, that means using humanoids

1207
01:08:27,729 --> 01:08:31,838
for simple stuff that you can
simulate, like picking place.

1208
01:08:31,838 --> 01:08:35,669
And that means using their
non-human eyes. The multiple

1209
01:08:35,669 --> 01:08:39,848
robotic arms inside our system
to, carry out the complicated

1210
01:08:39,848 --> 01:08:43,539
tasks the same way the
scientists will carry them out

1211
01:08:43,539 --> 01:08:47,091
and so, super important for us
to learn from expert

1212
01:08:47,091 --> 01:08:51,200
demonstration. And my point of
view is if I build something

1213
01:08:51,200 --> 01:08:55,378
that works, produces, you know,
life saving therapies and is

1214
01:08:55,378 --> 01:08:59,000
obviously commercially
successful, that will itself.

1215
01:08:59,000 --> 01:09:01,000
Audrow Nash: That's a proof of
perception.

1216
01:09:01,000 --> 01:09:04,577
Fred Parietti: I am not trying
to chase on a line to some, you

1217
01:09:04,577 --> 01:09:07,000
know, ideas that the public got
from, you.

1218
01:09:07,000 --> 01:09:09,000
Audrow Nash: Know, something
else and everything.

1219
01:09:09,000 --> 01:09:11,000
Fred Parietti: Yeah, right. I
guess.

1220
01:09:11,000 --> 01:09:12,000
Audrow Nash: To fickle
innocence.

1221
01:09:12,000 --> 01:09:16,030
Fred Parietti: Or I guess I am
the type of chase that idea. But

1222
01:09:16,030 --> 01:09:19,932
it's in my mind and, you know,
and I'm the robotics guy, and,

1223
01:09:19,932 --> 01:09:23,642
you know, this is I'm busy
robots, right? I get to see how

1224
01:09:23,642 --> 01:09:27,544
they look like and what they do.
And, Yeah. And if the public

1225
01:09:27,544 --> 01:09:31,255
wants to be different robots,
you know, they're welcome. I

1226
01:09:31,255 --> 01:09:35,157
love more robotics companies,
but, but like, but it must work

1227
01:09:35,157 --> 01:09:38,739
in the end. And sometimes
chasing too much. You know, I,

1228
01:09:38,739 --> 01:09:42,641
for example, I hate when people
edit their robotic videos and

1229
01:09:42,641 --> 01:09:46,479
they have a ton of cuts. Yeah.
You know, and I'm like, sure,

1230
01:09:46,479 --> 01:09:50,573
you know, you're just chasing an
idea in the minds of people and

1231
01:09:50,573 --> 01:09:54,091
it will turn against you the
moment they realize it. It

1232
01:09:54,091 --> 01:09:58,185
doesn't work. Yep. And so, yeah,
for me, number one is I need to

1233
01:09:58,185 --> 01:10:01,000
build something. I'm confident
it will work.

1234
01:10:01,000 --> 01:10:05,057
Audrow Nash: Oh yeah. So one
thing that I wanted to make sure

1235
01:10:05,057 --> 01:10:08,982
we talk about. So how do you
train your. Robot to do things

1236
01:10:08,982 --> 01:10:12,507
like the human does. So you
mentioned that you can do

1237
01:10:12,507 --> 01:10:16,498
statistical analysis and you
take the you split the material

1238
01:10:16,498 --> 01:10:20,755
and then you look bad before and
after or at the human's version

1239
01:10:20,755 --> 01:10:24,546
in the robots version. And
that's a way of kind of seeing

1240
01:10:24,546 --> 01:10:28,537
if you've matched it. But how do
you, teach your robot to do

1241
01:10:28,537 --> 01:10:30,000
these different steps?

1242
01:10:30,000 --> 01:10:34,208
Fred Parietti: So thank you so
much for asking that question,

1243
01:10:34,208 --> 01:10:38,624
because it's probably our number
one priority right now from the

1244
01:10:38,624 --> 01:10:42,970
robotics R&D point of view. So
the idea is, okay, you know, the

1245
01:10:42,970 --> 01:10:47,247
robots work, they manufacture
the drugs. You know, we're going

1246
01:10:47,247 --> 01:10:51,525
to, expand the supply chain. So
we can make, you know, as many

1247
01:10:51,525 --> 01:10:55,457
as the market wants. But then
those robots, when they are

1248
01:10:55,457 --> 01:10:59,389
deployed in pharma companies,
they need to learn a lot of

1249
01:10:59,389 --> 01:11:03,046
processes, you know, tens,
hundreds of processes. And

1250
01:11:03,046 --> 01:11:07,323
because biology is black magic,
right? You never know what the

1251
01:11:07,323 --> 01:11:11,463
Weezer's the biologists are
doing in their process. And they

1252
01:11:11,463 --> 01:11:15,878
there tends to be a long tail of
ten, 15%, 10 to 10%, let's say,

1253
01:11:15,878 --> 01:11:20,224
of tasks that we've never seen
before. Right? Like they shake a

1254
01:11:20,224 --> 01:11:24,571
ball in a particular way, right?
Or they squish a bag or remove

1255
01:11:24,571 --> 01:11:28,296
bubbles in a particular way.
Right. And, and those are

1256
01:11:28,296 --> 01:11:32,159
basically artisans. They're
almost like artists at work.

1257
01:11:32,159 --> 01:11:36,506
Right? There's that particular
skill scientist that somehow has

1258
01:11:36,506 --> 01:11:40,645
figured out that if you do that
to the cells, the cells will

1259
01:11:40,645 --> 01:11:44,853
have that reaction, which is
sometimes key to the performance

1260
01:11:44,853 --> 01:11:49,200
of the drug. And so our problem
after we deploy enough hardware

1261
01:11:49,200 --> 01:11:53,408
is how do we make that hardware
maximally useful to all these

1262
01:11:53,408 --> 01:11:57,547
processes and scientists, right
of companies. Right, without

1263
01:11:57,547 --> 01:12:01,825
requiring an infinite number of
engineers? Because if we go to

1264
01:12:01,825 --> 01:12:05,688
ten customers, to a hundred
customers, I cannot ten next

1265
01:12:05,688 --> 01:12:09,552
engineers, right? Yeah, for
sure. And send them to learn

1266
01:12:09,552 --> 01:12:13,691
from the scientists. Right. So
this is actually a very Tesla

1267
01:12:13,691 --> 01:12:17,899
like situation where once you
deploy a ton of hardware, there

1268
01:12:17,899 --> 01:12:21,625
must be a way to use that
hardware to collect data and

1269
01:12:21,625 --> 01:12:26,040
then auto add the new features,
right. So that effectively every

1270
01:12:26,040 --> 01:12:29,972
new deployment increases your
database and increases your

1271
01:12:29,972 --> 01:12:34,112
functionality so that you get
better and better by virtue of

1272
01:12:34,112 --> 01:12:38,251
your scale and the answer to
that for us is called imitation

1273
01:12:38,251 --> 01:12:42,459
learning. Right? So effectively
what we do is we collect from

1274
01:12:42,459 --> 01:12:46,806
the pharma companies, they send
it to us, right? We hundreds or

1275
01:12:46,806 --> 01:12:50,807
thousands of hours of, videos of
scientists performing the

1276
01:12:50,807 --> 01:12:54,947
process, which is very easy to
collect in the current manual

1277
01:12:54,947 --> 01:12:59,431
manufacturing floor. Right. It's
just imagine you have a bunch of

1278
01:12:59,431 --> 01:13:03,501
people doing their job. You just
get a lot of video of that

1279
01:13:03,501 --> 01:13:07,917
plentiful easy chip to collect.
And then via imitation learning,

1280
01:13:07,917 --> 01:13:11,918
we extract from those, you know,
long form videos, dozens,

1281
01:13:11,918 --> 01:13:16,057
hundreds of examples for every
single task. And once we have

1282
01:13:16,057 --> 01:13:20,335
the examples, wow, we can then
train a robot to either control

1283
01:13:20,335 --> 01:13:24,336
policy by imitation learning or
other techniques. Yeah. To

1284
01:13:24,336 --> 01:13:28,820
effectively, replicate that task
within the, you know, the bounds

1285
01:13:28,820 --> 01:13:32,960
of their manufacturing facility.
Right? Or, you know, or the

1286
01:13:32,960 --> 01:13:37,444
bounds of that specific process.
And, and effectively, the robots

1287
01:13:37,444 --> 01:13:41,445
are able to auto learn those
tasks. And, and this is super

1288
01:13:41,445 --> 01:13:45,723
important because you must, like
a lot of people ask us, well,

1289
01:13:45,723 --> 01:13:50,000
can you just simulate, you know,
that where you can't because.

1290
01:13:50,000 --> 01:13:51,000
Audrow Nash: It's a black.

1291
01:13:51,000 --> 01:13:54,636
Fred Parietti: Magic thing.
Yeah. The biology is absent from

1292
01:13:54,636 --> 01:13:58,091
the physics based simulations.
Yeah. So what? We actually

1293
01:13:58,091 --> 01:14:01,121
reverse engineering is the
implicit model that the

1294
01:14:01,121 --> 01:14:04,879
scientists have in their mind of
the biology, and that informs

1295
01:14:04,879 --> 01:14:08,333
their action. This is not
written sometimes it's not even

1296
01:14:08,333 --> 01:14:12,091
done at an aware level. Yeah,
but the best scientists kind of,

1297
01:14:12,091 --> 01:14:15,000
like, have figured out what
works for the cells.

1298
01:14:15,000 --> 01:14:16,000
Audrow Nash: Yeah.

1299
01:14:15,000 --> 01:14:18,856
Fred Parietti: And we need the
robots to figure that out as

1300
01:14:18,856 --> 01:14:22,582
well. And modern, you know,
imitation learning or machine

1301
01:14:22,582 --> 01:14:26,176
learning or, you know, AI
techniques, work very well to

1302
01:14:26,176 --> 01:14:29,967
discover intrinsic patterns,
provided that you have enough

1303
01:14:29,967 --> 01:14:33,824
data. So that's why the video
data, video data is plentiful

1304
01:14:33,824 --> 01:14:37,484
enough. We can get enough to be
able to extract implicit

1305
01:14:37,484 --> 01:14:41,144
knowledge with any less
plentiful data. You probably are

1306
01:14:41,144 --> 01:14:45,000
not going to have enough to
extract the implicit knowledge.

1307
01:14:45,000 --> 01:14:48,905
Audrow Nash: Yeah. So that's
such an interesting idea. So

1308
01:14:48,905 --> 01:14:53,015
video the reason. For video is
because it's accessible. It's

1309
01:14:53,015 --> 01:14:57,126
because it's easy to set up a
camera and to record it and to

1310
01:14:57,126 --> 01:15:01,236
have them do it, or they might
already be doing it. And then

1311
01:15:01,236 --> 01:15:05,210
you can take that, chop it up,
label it, do things. I it's

1312
01:15:05,210 --> 01:15:08,909
really. A very cool thing. And
it. Seems like so many.

1313
01:15:08,909 --> 01:15:12,951
Companies are working on a
similar problem, which is how do

1314
01:15:12,951 --> 01:15:17,062
you teach robots to do things in
a general way. So you think

1315
01:15:17,062 --> 01:15:21,104
video is the good way to do
that? How do you how do you map

1316
01:15:21,104 --> 01:15:25,283
what the human. Is doing to what
the robot should be doing in

1317
01:15:25,283 --> 01:15:29,530
video? Like I'm imagining you
run a pose detector that assumes

1318
01:15:29,530 --> 01:15:33,504
what the person is doing in 3D,
and you kind of infer from

1319
01:15:33,504 --> 01:15:37,683
there, and you have to maybe
have human annotators. Slice the

1320
01:15:37,683 --> 01:15:41,999
video. So you say this. Is when
they're doing this step and you

1321
01:15:41,999 --> 01:15:46,452
kind of file it somewhere so you
can group those steps. Similarly

1322
01:15:46,452 --> 01:15:47,000
or yeah.

1323
01:15:47,000 --> 01:15:51,107
Fred Parietti: Honestly, like,
very, very good question. So

1324
01:15:51,107 --> 01:15:55,145
yes, we do need to slice the
videos. Yeah. We believe that

1325
01:15:55,145 --> 01:15:59,182
can be done at least in this
context automatically. And we

1326
01:15:59,182 --> 01:16:03,220
already have very promising
results there. Awesome. So the

1327
01:16:03,220 --> 01:16:06,979
idea is that, luckily a
fundamental fat training floor

1328
01:16:06,979 --> 01:16:11,364
is much more controlled than the
general environment. And so it

1329
01:16:11,364 --> 01:16:15,610
actually possible to train, you
know, latest generation video

1330
01:16:15,610 --> 01:16:19,996
algorithms to auto segment, and
in fact, it's even better, like

1331
01:16:19,996 --> 01:16:23,825
you can actually provide,
examples of non tasks so that

1332
01:16:23,825 --> 01:16:28,001
those are like auto labeled on
their own. Yeah. But then the

1333
01:16:28,001 --> 01:16:32,178
algorithm can also say, oh,
actually this thing that was not

1334
01:16:32,178 --> 01:16:36,703
similar to any of your examples,
this thing has been repeated 100

1335
01:16:36,703 --> 01:16:39,000
times. So it is perhaps you
know.

1336
01:16:39,000 --> 01:16:42,222
Audrow Nash: And so there's
similar there's similar. Paths

1337
01:16:42,222 --> 01:16:45,389
through state space in every
demonstration. And so that's

1338
01:16:45,389 --> 01:16:48,000
you're learning something more
general from it.

1339
01:16:48,000 --> 01:16:51,000
Fred Parietti: Exactly. And so
that's super important.

1340
01:16:51,000 --> 01:16:52,000
Audrow Nash: The super cool.

1341
01:16:52,000 --> 01:16:54,929
Fred Parietti: The auto
segmentation. And then there is

1342
01:16:54,929 --> 01:16:58,231
yes trajectory extraction. And
then there is the training of a

1343
01:16:58,231 --> 01:17:01,000
control policy. And we have very
specific, you know.

1344
01:17:01,000 --> 01:17:02,000
Audrow Nash: Pipeline for this
tools.

1345
01:17:02,000 --> 01:17:06,301
Fred Parietti: And pipelines for
all of them. When we can, we use

1346
01:17:06,301 --> 01:17:10,535
off the shelf, when we can't, we
also develop our own stuff. The

1347
01:17:10,535 --> 01:17:14,108
most important part here is
plentiful data. And so the

1348
01:17:14,108 --> 01:17:18,011
problem is that, so video is
good because it is very simple

1349
01:17:18,011 --> 01:17:22,180
to manipulate. And there are a
lot of libraries to do it. But I

1350
01:17:22,180 --> 01:17:26,083
can't go on YouTube and download
1000 hours of cell therapy

1351
01:17:26,083 --> 01:17:29,987
manufacturing right here. So
that's the problem for which I

1352
01:17:29,987 --> 01:17:34,023
think robotics is being slower
than, you know, other is that,

1353
01:17:34,023 --> 01:17:38,059
you know, if I was training, you
know, whatever I am for like

1354
01:17:38,059 --> 01:17:42,161
legal language where there's so
much legal material that I can

1355
01:17:42,161 --> 01:17:46,197
either download or scrape or, I
guess, steal that, judging by

1356
01:17:46,197 --> 01:17:50,100
how a lot of those companies got
the data. Yeah. And so but

1357
01:17:50,100 --> 01:17:54,070
nobody can stop me. Right. And
that's very good data. Right.

1358
01:17:54,070 --> 01:17:58,040
There's no way to get that data.
So I need a way to get that

1359
01:17:58,040 --> 01:18:01,943
level of internal size data
cheaply and scale in a scalable

1360
01:18:01,943 --> 01:18:05,781
way. So ideally I need my
partners to send that data to me

1361
01:18:05,781 --> 01:18:09,353
without me being involved
because it would be too time

1362
01:18:09,353 --> 01:18:12,000
consuming for me to personally
generate.

1363
01:18:12,000 --> 01:18:14,696
Audrow Nash: It's like text.
Data, as you were saying, like a

1364
01:18:14,696 --> 01:18:17,260
test, like on a moment where
it's like you make a pipeline

1365
01:18:17,260 --> 01:18:20,000
that just aggregates everything
and then learns from it. Yeah.

1366
01:18:20,000 --> 01:18:23,354
Fred Parietti: And you find a
way for your partners to

1367
01:18:23,354 --> 01:18:27,143
basically pay you to get the
hardware that then gives you the

1368
01:18:27,143 --> 01:18:30,870
data. And of course, they need
to agree to give you the data

1369
01:18:30,870 --> 01:18:34,534
very confidentially. Right?
Much. The difference is that we

1370
01:18:34,534 --> 01:18:38,075
are regulated and Tesla is not.
But still like this is no

1371
01:18:38,075 --> 01:18:41,739
different. And that's basically
my one, you know, the main,

1372
01:18:41,739 --> 01:18:45,217
main, you know, problem with
Teleoperation. I just don't

1373
01:18:45,217 --> 01:18:49,130
believe their operation scales
the same way. It's not so cheap,

1374
01:18:49,130 --> 01:18:53,106
is not so plentiful. And by the
way, try to convince a scientist

1375
01:18:53,106 --> 01:18:56,894
to bring a teleoperation into a
clean room right in order and

1376
01:18:56,894 --> 01:19:00,435
then use, you know, not their
own hands, but the robot to

1377
01:19:00,435 --> 01:19:04,224
manipulate the cells. They're
going to tell you you're crazy,

1378
01:19:04,224 --> 01:19:07,640
right? So like in operation, I
would have zero data. So

1379
01:19:07,640 --> 01:19:10,000
obviously, you know, there's a
ton of.

1380
01:19:10,000 --> 01:19:12,000
Audrow Nash: One data bias.

1381
01:19:12,000 --> 01:19:15,432
Fred Parietti: Right? If you're
lucky. And then there's a ton of

1382
01:19:15,432 --> 01:19:18,756
problems using videos. So what
you are saying is the change of

1383
01:19:18,756 --> 01:19:21,920
context, you know, from, you
know, a human demonstrating it

1384
01:19:21,920 --> 01:19:25,137
to what the robot should do to
replicate that. But that is a

1385
01:19:25,137 --> 01:19:28,139
technical challenge. And if I
solve it, I have something

1386
01:19:28,139 --> 01:19:31,410
scalable and that's what I like
to have in terms of technical

1387
01:19:31,410 --> 01:19:32,000
challenges.

1388
01:19:32,000 --> 01:19:35,406
Audrow Nash: Yeah, yeah. Because
I guess you can just say. Well,

1389
01:19:35,406 --> 01:19:38,386
what's the hand of the human
doing? And you can probably

1390
01:19:38,386 --> 01:19:41,738
label it like, oh, they grabbed.
This. And then they shook that

1391
01:19:41,738 --> 01:19:44,772
when it's in their hand. And
like these kinds of labeling

1392
01:19:44,772 --> 01:19:48,124
things, you could have the robot
do, you can break it down into

1393
01:19:48,124 --> 01:19:51,370
actions the robot can do. And
then that would let you do this

1394
01:19:51,370 --> 01:19:54,510
mostly autonomously. Like you
can onboard new skills fairly

1395
01:19:54,510 --> 01:19:56,000
easily, which is quite cool.

1396
01:19:56,000 --> 01:19:59,978
Fred Parietti: Precisely it. In
fact, do you really care what

1397
01:19:59,978 --> 01:20:03,826
the human hand is doing provided
you know its effect on the

1398
01:20:03,826 --> 01:20:05,000
bottle on the bio?

1399
01:20:05,000 --> 01:20:07,000
Audrow Nash: Yeah, not at all.
So you can.

1400
01:20:07,000 --> 01:20:11,026
Fred Parietti: Even just track
the objects that are in contact

1401
01:20:11,026 --> 01:20:12,000
with the cells.

1402
01:20:12,000 --> 01:20:15,100
Audrow Nash: Oh yeah, that's way
easier. Yeah. For sure. Okay.

1403
01:20:15,100 --> 01:20:18,250
Yeah. It's way. Simpler problem
planning too, because you don't

1404
01:20:18,250 --> 01:20:21,400
need to emulate the human arm.
You can just emulate what the is

1405
01:20:21,400 --> 01:20:23,000
doing or whatever it is. I mean.

1406
01:20:23,000 --> 01:20:25,827
Fred Parietti: Effectively
they're related because if

1407
01:20:25,827 --> 01:20:29,133
you're trying to emulate what
the human arm is doing, probably

1408
01:20:29,133 --> 01:20:32,227
having a structure similar to
the human arm lets you do it

1409
01:20:32,227 --> 01:20:35,640
correctly. So that's why we only
use collaborative robotic arms.

1410
01:20:35,640 --> 01:20:39,000
We're cool. Yeah. Six degrees of
freedom because they are close

1411
01:20:39,000 --> 01:20:41,987
enough. The kinematics were
never limited. Yeah, you can

1412
01:20:41,987 --> 01:20:43,000
always copy people.

1413
01:20:43,000 --> 01:20:46,997
Audrow Nash: Yeah, yeah, I would
imagine. If you like. Move it

1414
01:20:46,997 --> 01:20:50,478
similarly, whatever the object
is. But I mean, I don't

1415
01:20:50,478 --> 01:20:54,410
understand the problem that well
then you would get a similar

1416
01:20:54,410 --> 01:20:58,278
effect. But I don't know. So
that's really cool. It's a very

1417
01:20:58,278 --> 01:21:02,146
clever way of approaching this.
I also feel like if you guys

1418
01:21:02,146 --> 01:21:05,885
nail this, there's a lot of
other wonderful areas that you

1419
01:21:05,885 --> 01:21:09,753
can go and solve to. Yeah. Even
though pharmaceutical. Yeah,

1420
01:21:09,753 --> 01:21:13,879
yeah. Even though pharmaceutical
is such a big industry and it's

1421
01:21:13,879 --> 01:21:17,811
so like you they, they don't
have engineering teams generally

1422
01:21:17,811 --> 01:21:21,550
as you were saying. But yeah,
tons and tons of other areas

1423
01:21:21,550 --> 01:21:24,000
would benefit from this same
approach.

1424
01:21:24,000 --> 01:21:27,577
Fred Parietti: I agree with you.
And this is a cutting edge, I

1425
01:21:27,577 --> 01:21:30,750
think, of, robotics research
and, we just found what we

1426
01:21:30,750 --> 01:21:34,269
believe is the highest possible
impact and value application.

1427
01:21:34,269 --> 01:21:37,846
But I agree with you that the
ripple effect of this technology

1428
01:21:37,846 --> 01:21:39,000
industries are huge.

1429
01:21:39,000 --> 01:21:43,050
Audrow Nash: Yeah. For sure.
Okay. So let's see, getting.

1430
01:21:43,050 --> 01:21:47,383
Towards wrapping up. One thing
that you've brought up a bunch

1431
01:21:47,383 --> 01:21:51,220
is high value problems. So as
opposed to having like a

1432
01:21:51,220 --> 01:21:55,696
humanoid robot sweep the floors
or something like this, have it

1433
01:21:55,696 --> 01:21:59,177
do incredibly important
pharmaceutical work, in a

1434
01:21:59,177 --> 01:22:03,511
complex system and. Just high
value work. How do you find how

1435
01:22:03,511 --> 01:22:07,702
high. Value problems for this
kind of thing? Because that I

1436
01:22:07,702 --> 01:22:12,107
think a lot of. People who are
looking to start companies want

1437
01:22:12,107 --> 01:22:16,370
to find high value problems and
then it's you can go talk to

1438
01:22:16,370 --> 01:22:20,988
people and things, but it's hard
to understand what are the true.

1439
01:22:20,988 --> 01:22:25,464
High. Value things, what are the
real bottlenecks of the people

1440
01:22:25,464 --> 01:22:30,011
that you're trying to serve? Do
you have any thoughts on. How to

1441
01:22:30,011 --> 01:22:32,000
find these high value areas?

1442
01:22:32,000 --> 01:22:37,013
Fred Parietti: I mean, I guess
from the manufacturing point of

1443
01:22:37,013 --> 01:22:41,621
view from a very like first
principle, you know, point of

1444
01:22:41,621 --> 01:22:46,230
view, the most high value
problems are the ones that have

1445
01:22:46,230 --> 01:22:51,000
the largest market value per
like cubic centimeter, or the.

1446
01:22:51,000 --> 01:22:53,000
Audrow Nash: Ram or whatever.
Yeah.

1447
01:22:53,000 --> 01:22:56,060
Fred Parietti: Of, product.
Right. And so by that metric,

1448
01:22:56,060 --> 01:22:59,121
the two highest value products
in the east of mankind are

1449
01:22:59,121 --> 01:23:01,000
semiconductors and
pharmaceuticals.

1450
01:23:01,000 --> 01:23:02,000
Audrow Nash: All right.

1451
01:23:02,000 --> 01:23:06,828
Fred Parietti: Way more valuable
in gold. So that's basically how

1452
01:23:06,828 --> 01:23:11,359
we found it out. And another way
for a high value problem is,

1453
01:23:11,359 --> 01:23:15,815
would be to kind of evaluate,
you know, what is, I guess the

1454
01:23:15,815 --> 01:23:20,495
size of, of deals you can get,
right? But then again, like, you

1455
01:23:20,495 --> 01:23:25,249
need to divide it by the number
or volume or weight of products.

1456
01:23:25,249 --> 01:23:29,631
Yes. Another area that's, that's
very high value is, but it

1457
01:23:29,631 --> 01:23:34,459
metric is maybe aerospace. But I
do believe that even a plane per

1458
01:23:34,459 --> 01:23:38,841
gram or per unit volume is way
lower. Yeah. So probably the

1459
01:23:38,841 --> 01:23:43,521
only other one would basically
be military. But from our point,

1460
01:23:43,521 --> 01:23:47,754
from my point of view, I was
like, you know what? I'm not

1461
01:23:47,754 --> 01:23:52,211
going to make robots for that
application. And I much rather

1462
01:23:52,211 --> 01:23:56,519
basically, make a robot that
save lives than the other way

1463
01:23:56,519 --> 01:24:00,753
around. Yeah. And in, in the end
even obviously it is all

1464
01:24:00,753 --> 01:24:04,912
arbitrary. This is just my
personal preference. But I do

1465
01:24:04,912 --> 01:24:09,369
believe that these, this stems
from the fact that, saving or

1466
01:24:09,369 --> 01:24:13,900
extending life is more valuable
than taking it away. There is

1467
01:24:13,900 --> 01:24:18,431
fundamentally, you know, we are
alive and we want to be alive

1468
01:24:18,431 --> 01:24:23,036
and be healthy and be happy, and
we're willing to pay way more

1469
01:24:23,036 --> 01:24:27,270
for that. Yeah. Then to somehow
kill someone else. And so

1470
01:24:27,270 --> 01:24:31,652
fundamentally, I think it's a
better business. But this was

1471
01:24:31,652 --> 01:24:36,109
more, you know, basic decision
for me. I want to robots that

1472
01:24:36,109 --> 01:24:37,000
help people.

1473
01:24:37,000 --> 01:24:41,302
Audrow Nash: Yes. Totally. Yeah.
Yeah. It's nice to work on

1474
01:24:41,302 --> 01:24:45,604
something that adds value to
people. And I it's like if you

1475
01:24:45,604 --> 01:24:49,979
don't have health, you don't
have anything. Really. Yeah. So

1476
01:24:49,979 --> 01:24:54,063
yeah, I think adding making it
so that we can extend the

1477
01:24:54,063 --> 01:24:58,438
frontier of our medical system
would be phenomenal. So yeah.

1478
01:24:58,438 --> 01:25:02,813
What you guys are doing. Hell
yeah. Okay, so wrapping up, do

1479
01:25:02,813 --> 01:25:07,479
you have anything that you like
for our. Watchers and listeners.

1480
01:25:07,479 --> 01:25:12,000
What do you hope they take away.
From this whole conversation?

1481
01:25:12,000 --> 01:25:15,452
Fred Parietti: Well, I mean, I
think the, the most exciting,

1482
01:25:15,452 --> 01:25:19,134
thing for me is that, you know,
robotics right now is one of the

1483
01:25:19,134 --> 01:25:22,816
hardest, you know, areas of, you
know, technology in general. It

1484
01:25:22,816 --> 01:25:26,096
is considered like the next
frontier of AI because it the

1485
01:25:26,096 --> 01:25:29,088
application of AI to the
physical world. So I'm very

1486
01:25:29,088 --> 01:25:32,770
happy because it used to not be
like that, even just a few years

1487
01:25:32,770 --> 01:25:33,000
ago.

1488
01:25:33,000 --> 01:25:34,000
Audrow Nash: I know it's true.

1489
01:25:34,000 --> 01:25:37,669
Fred Parietti: Yeah. Right. And
so I'm like, yeah, so what is

1490
01:25:37,669 --> 01:25:41,277
cool? And let's build the best
possible robots and just, you

1491
01:25:41,277 --> 01:25:44,404
know, a so I would just
encourage people to build in

1492
01:25:44,404 --> 01:25:47,953
robotics and be you know, my
caveats are be careful because

1493
01:25:47,953 --> 01:25:51,682
you need to find, you know, an
application that justifies that

1494
01:25:51,682 --> 01:25:54,869
level of technological
advancement. Because the worst

1495
01:25:54,869 --> 01:25:58,598
thing we can do right now is to
squander all this attention by

1496
01:25:58,598 --> 01:26:02,206
making people think robotics is
useless or not ready yet. So

1497
01:26:02,206 --> 01:26:05,755
let's avoid that. The other
thing I want to tell people is,

1498
01:26:05,755 --> 01:26:09,363
yeah, like, you know, there are
so many cool applications of

1499
01:26:09,363 --> 01:26:13,152
robotics that help people. And
so I would like, you know, yeah.

1500
01:26:13,152 --> 01:26:17,061
Like, you know, that I feel like
there's so many applications and

1501
01:26:17,061 --> 01:26:20,670
it's impossible for a single
company to address all of them.

1502
01:26:20,670 --> 01:26:24,459
And so I just encourage people
to, you know, just you know, get

1503
01:26:24,459 --> 01:26:25,000
building.

1504
01:26:25,000 --> 01:26:29,447
Audrow Nash: Oh yeah. Love it.
And is there any, any places

1505
01:26:29,447 --> 01:26:33,216
you'd like to point our
listeners to. Any calls to

1506
01:26:33,216 --> 01:26:37,437
action. For people maybe
interested in joining you guys.

1507
01:26:37,437 --> 01:26:40,000
Or anything you can then go
there.

1508
01:26:40,000 --> 01:26:43,178
Fred Parietti: I mean, for me,
you know, the you know, the

1509
01:26:43,178 --> 01:26:46,356
areas where you have the most to
learn are always the most

1510
01:26:46,356 --> 01:26:49,425
interesting. And so for my
point, from my point of view,

1511
01:26:49,425 --> 01:26:52,000
like the intersection between
robotics and bio.

1512
01:26:52,000 --> 01:26:55,000
Audrow Nash: Yes, is just so
cool. Totally.

1513
01:26:55,000 --> 01:26:58,287
Fred Parietti: And, and there is
a lot of need of people that

1514
01:26:58,287 --> 01:27:01,681
have a kind of an understanding
of both areas. Right now, those

1515
01:27:01,681 --> 01:27:05,022
people are exceedingly rare. In
fact, they almost don't exist.

1516
01:27:05,022 --> 01:27:08,093
It's always a robotics person
learns about biology or the

1517
01:27:08,093 --> 01:27:11,325
other way around. But I would
definitely point people. There

1518
01:27:11,325 --> 01:27:14,504
is so much value and so many new
things to discover in that

1519
01:27:14,504 --> 01:27:17,791
intersection that they feel like
people, you know, it's just,

1520
01:27:17,791 --> 01:27:20,000
very worthwhile, you know? And
they were.

1521
01:27:20,000 --> 01:27:23,692
Audrow Nash: Yeah. How do you,
as a roboticist, like. How do

1522
01:27:23,692 --> 01:27:27,508
you learn about biology? Like,
do you go back and read a book?

1523
01:27:27,508 --> 01:27:31,138
Do you? I mean, I'm sure the
position you're in now working

1524
01:27:31,138 --> 01:27:34,892
with pharmaceutical companies is
an incredible education. But

1525
01:27:34,892 --> 01:27:38,646
like. How would you learn. If
you're a roboticist or software

1526
01:27:38,646 --> 01:27:42,400
engineer and you want to learn
more about biology but have no

1527
01:27:42,400 --> 01:27:44,000
idea how to approach that?

1528
01:27:44,000 --> 01:27:47,671
Fred Parietti: Yeah, you need to
not be, you know, afraid of

1529
01:27:47,671 --> 01:27:51,342
asking them questions because
obviously you're going to look

1530
01:27:51,342 --> 01:27:54,953
extremely dumb to an expert. But
keep in mind that they are

1531
01:27:54,953 --> 01:27:58,379
probably thinking that are
looking extremely dumb to you

1532
01:27:58,379 --> 01:28:02,295
ask questions about your field.
Right. So everybody is held back

1533
01:28:02,295 --> 01:28:06,027
by the same fear of like looking
like, you know, they have no

1534
01:28:06,027 --> 01:28:09,943
clue because they have no clue.
And so after you recognize that,

1535
01:28:09,943 --> 01:28:13,553
you know, you can just have
those conversations. And now we

1536
01:28:13,553 --> 01:28:17,286
also helped by the fact that,
you know, chat GPT doesn't care

1537
01:28:17,286 --> 01:28:21,018
if you look dumb and they're
pretty good. You know, these new

1538
01:28:21,018 --> 01:28:24,812
models to do like research and
ask basic questions. And it's a

1539
01:28:24,812 --> 01:28:28,605
very interactive way of learning
that, you know, pretty much a

1540
01:28:28,605 --> 01:28:32,277
lot of, but I would say there is
a limit to how much you can

1541
01:28:32,277 --> 01:28:36,070
learn from an LLM because it's
too tailored to please you. And

1542
01:28:36,070 --> 01:28:39,803
so, I much rather, as soon as I
have a question that might be

1543
01:28:39,803 --> 01:28:43,352
dumb, but I think is very
interesting to me. I much rather

1544
01:28:43,352 --> 01:28:46,900
ask it to a person that I know
knows in that space, and in

1545
01:28:46,900 --> 01:28:50,449
return I can answer their
questions about my space, and we

1546
01:28:50,449 --> 01:28:51,000
both win.

1547
01:28:51,000 --> 01:28:53,950
Audrow Nash: Oh, yeah. That's
awesome. All right. Thank you

1548
01:28:53,950 --> 01:28:56,000
Fred. It's been awesome talking
with you.

1549
01:28:56,000 --> 01:28:58,880
Fred Parietti: Thank you very
much. I really, really,

1550
01:28:58,880 --> 01:29:01,000
appreciate it was a great
conversation.

1551
01:29:01,000 --> 01:29:03,000
Audrow Nash: Hell yeah. Alright.
Bye everyone.
