1
00:00:00,280 --> 00:00:02,400
Hi, and welcome to the Neil
Ashton Podcast.

2
00:00:03,080 --> 00:00:05,920
In each episode, we explained
some of the fascinating ways

3
00:00:05,920 --> 00:00:09,080
that science and engineering are
changing the world around us.

4
00:00:09,800 --> 00:00:12,760
We talked to leading engineers
from elite level sports like

5
00:00:12,840 --> 00:00:16,880
cycling in Formula One to some
of the world's top academics to

6
00:00:16,880 --> 00:00:20,960
understand how fluid dynamics,
machine learning, supercomputing

7
00:00:21,360 --> 00:00:23,000
are bringing in a new era
discovery.

8
00:00:23,960 --> 00:00:27,080
We also hear some of their life
stories, their career advice,

9
00:00:27,640 --> 00:00:30,080
the lessons they've learned on
the way that I hope will be

10
00:00:30,080 --> 00:00:33,800
helpful to you too.
So sit back and enjoy this

11
00:00:33,800 --> 00:00:40,680
episode.
Hi, and welcome back to the Neil

12
00:00:40,680 --> 00:00:43,960
Ashton Podcast.
So today's guest is Yoris Port,

13
00:00:43,960 --> 00:00:47,080
who is the CEO and founder of
Rescale.

14
00:00:47,320 --> 00:00:51,000
He's somebody who I wanted to
talk to for a while as part of

15
00:00:51,160 --> 00:00:54,280
the opportunity to speak to
people who have had that bold

16
00:00:54,280 --> 00:00:57,160
vision to, to create a start up,
to create a new company.

17
00:00:57,240 --> 00:00:59,920
And I thought it was really
interesting to try and learn

18
00:00:59,920 --> 00:01:03,000
from these people, you know,
what motivated them to do in the

19
00:01:03,000 --> 00:01:06,120
1st place, what some of the
lessons they've learnt in doing

20
00:01:06,120 --> 00:01:08,160
that.
And it's particularly relevant,

21
00:01:08,480 --> 00:01:10,840
I guess, for the themes of of
this podcast.

22
00:01:11,560 --> 00:01:15,920
Rescale is, is used by actually
a lot of companies these days,

23
00:01:15,920 --> 00:01:17,760
you know, if they want to
integrate more like high

24
00:01:17,760 --> 00:01:21,280
performance computing and and
more recently more applied AI,

25
00:01:21,280 --> 00:01:24,360
you know, they've got hundreds
of customers in this space,

26
00:01:24,360 --> 00:01:26,480
enterprise customers.
So it's kind of interesting to

27
00:01:26,480 --> 00:01:30,240
speak to somebody who looks
after that company and they have

28
00:01:30,240 --> 00:01:32,000
a good sense of what's coming
next.

29
00:01:32,120 --> 00:01:34,600
It's been a topic also this
podcast is looking at, you know,

30
00:01:34,600 --> 00:01:38,240
what's the future of AI and we
we ended up talking for quite

31
00:01:38,240 --> 00:01:42,240
some time about the potential of
agentic AI.

32
00:01:42,240 --> 00:01:45,680
This is something that, you
know, myself and yours turns out

33
00:01:45,680 --> 00:01:48,600
are quite aligned on in this
could be quite transformative.

34
00:01:48,600 --> 00:01:51,360
You know, we previously mainly
spoken about AI surrogates and

35
00:01:51,360 --> 00:01:54,640
we we also do talk about that.
But I think the yeah, the the

36
00:01:54,640 --> 00:01:57,640
agentic AI and the potential
this has for engineering was a

37
00:01:57,640 --> 00:01:59,840
really interesting discussion
that comes probably in the

38
00:01:59,880 --> 00:02:03,520
second-half of of the chat.
So it definitely TuneIn if you

39
00:02:03,520 --> 00:02:06,720
want to hear some interesting
thoughts on that, given my

40
00:02:06,720 --> 00:02:09,440
background, having worked Adbs
before and therefore being

41
00:02:09,440 --> 00:02:12,840
immersed in the cloud computing
space, I always find that

42
00:02:12,840 --> 00:02:16,000
interesting as well.
And how that has evolved.

43
00:02:16,080 --> 00:02:21,200
You know, Rescale was created
like 2011, Oh, quite a long time

44
00:02:21,360 --> 00:02:25,320
before, I guess ways today where
cloud is more mainstream and and

45
00:02:25,320 --> 00:02:27,200
accepted, you know, and
Rescale's had some pretty

46
00:02:27,200 --> 00:02:31,960
impressive founders, sorry, fund
funding from companies like

47
00:02:32,120 --> 00:02:35,440
people like Salt and Jeff Bezos,
Paul Graham.

48
00:02:35,440 --> 00:02:38,280
So talk a bit about that with
the, you know, Y Combinator,

49
00:02:38,280 --> 00:02:41,320
NVIDIA, Microsoft.
So, you know, it's, it's

50
00:02:41,360 --> 00:02:44,120
impressive defeat for someone to
create a company that's had

51
00:02:44,120 --> 00:02:45,840
hundreds of millions of dollars
in funding.

52
00:02:46,280 --> 00:02:50,280
And so I hope that you learn a
lot from some of his advice for

53
00:02:50,280 --> 00:02:52,600
for maybe one of you who is
thinking about creating a, a

54
00:02:53,200 --> 00:02:56,240
start up yourself.
I always talk about doing this

55
00:02:56,240 --> 00:02:58,600
and I'm probably just too risk
adverse to do it.

56
00:02:58,600 --> 00:03:01,280
So, but don't listen to me,
listen to yours and hopefully

57
00:03:01,280 --> 00:03:03,880
you get some, some, some in.
So this was a wide-ranging

58
00:03:03,880 --> 00:03:05,760
discussion.
You know, we probably could have

59
00:03:05,760 --> 00:03:09,600
taken it even deeper or in other
areas, but at an hour and a

60
00:03:09,600 --> 00:03:12,160
half, I thought he was already
taking a lot of his time.

61
00:03:12,200 --> 00:03:15,000
But yeah, I certainly learnt a
lot from this discussion and

62
00:03:15,040 --> 00:03:18,280
have a renewed and even more
sense of, you know, appreciation

63
00:03:18,280 --> 00:03:21,280
for what people like him do and
push the boundaries and try and

64
00:03:21,280 --> 00:03:24,240
create new companies and ideas.
So hopefully it's an inspiration

65
00:03:24,240 --> 00:03:27,680
for you listening as well.
So sit back and enjoy this

66
00:03:27,680 --> 00:03:30,840
episode with your support.
Thank you for for coming and

67
00:03:30,840 --> 00:03:33,800
doing this.
You're actually on the quite

68
00:03:33,800 --> 00:03:37,560
high up the list of people I
wanted to speak coming I guess

69
00:03:37,560 --> 00:03:41,520
from a cloud provider background
before sort of seeing cloud.

70
00:03:41,800 --> 00:03:45,920
I think I joined 80 Business in
2020 and thinking, oh, this is

71
00:03:45,920 --> 00:03:49,120
quite novel, this is quite new.
And then looking back and

72
00:03:49,120 --> 00:03:53,520
realized that Rescale was part
of the Y Combinator in 2011.

73
00:03:54,640 --> 00:03:58,080
So back then, HPC in the cloud
must have been even more

74
00:03:58,080 --> 00:04:00,480
radical, even more sort of crazy
ideas.

75
00:04:00,480 --> 00:04:04,320
So maybe it's a starting point.
I'd be interested to know like

76
00:04:04,320 --> 00:04:08,440
what was your pitch for that
start up for for risk?

77
00:04:09,840 --> 00:04:11,560
Yeah, absolutely.
So first of all, thanks for

78
00:04:11,880 --> 00:04:15,000
thanks for having me on.
I think this is, you know a nice

79
00:04:15,000 --> 00:04:18,880
opportunity to to take some time
and chat about the background.

80
00:04:18,880 --> 00:04:23,600
I think for HPC in the cloud,
certainly when we started in

81
00:04:23,600 --> 00:04:29,400
2011, it was a very new concept.
I will be honest, when we

82
00:04:29,400 --> 00:04:32,920
founded the company, we thought
we were late to market because

83
00:04:33,680 --> 00:04:35,160
cloud computing had already
started.

84
00:04:35,400 --> 00:04:39,360
You had big data.
And it seemed pretty obvious to

85
00:04:39,360 --> 00:04:42,120
me that there would be like a
sort of a big compute company.

86
00:04:42,360 --> 00:04:46,920
And I actually looked for a
company to join myself to say,

87
00:04:46,920 --> 00:04:51,000
hey, who who's doing this,
solving this problem, right?

88
00:04:51,480 --> 00:04:53,880
And I had personally sort of
experienced this problem before.

89
00:04:53,880 --> 00:04:57,600
So it seemed like we were late
to market.

90
00:04:58,000 --> 00:05:00,200
Looking back now, we're quite
early to market, right?

91
00:05:00,720 --> 00:05:03,560
So at that time there was
definitely 0% HPC happening in

92
00:05:03,560 --> 00:05:05,480
cloud.
But it's, you know, it's been a

93
00:05:05,480 --> 00:05:07,880
fun journey and things have, you
know, changed over time.

94
00:05:08,640 --> 00:05:13,720
It was a tough process to get
people excited about actually,

95
00:05:13,920 --> 00:05:16,600
you know, investing in this and,
and sort of joining the team to

96
00:05:16,600 --> 00:05:19,960
pursue this.
But you know, like most good

97
00:05:19,960 --> 00:05:24,120
start-ups, you can start with a
a great idea and and and lots of

98
00:05:24,120 --> 00:05:27,280
effort and eventually you know
you can make it.

99
00:05:28,000 --> 00:05:31,720
Yeah.
So maybe taking a step before

100
00:05:31,720 --> 00:05:35,840
that, where had you been before?
What gave you that original

101
00:05:35,920 --> 00:05:40,320
motivational idea even to
overcome these problems and and

102
00:05:40,320 --> 00:05:42,720
create a start up, right.
That's still a a leap of faith

103
00:05:43,120 --> 00:05:44,400
to leave a company to a start
up.

104
00:05:44,680 --> 00:05:47,200
Yeah, absolutely.
I think for for me, my

105
00:05:47,200 --> 00:05:51,160
background is quite technical.
So I'd studied, I grew up doing

106
00:05:51,160 --> 00:05:53,400
a lot of computer science.
I'd studied applied math,

107
00:05:53,400 --> 00:05:55,840
mechanical engineering,
aeronautics, astronautics.

108
00:05:55,880 --> 00:05:59,440
I was sort of part of did a lot
of work in the field of multi

109
00:05:59,440 --> 00:06:02,920
display optimization.
I know you come from a strong

110
00:06:02,920 --> 00:06:07,520
CFD background, right.
So by the professor I studied

111
00:06:07,520 --> 00:06:12,040
under actually as a sort of
elasticity expert who was a

112
00:06:12,040 --> 00:06:15,240
disciple of Lucian Schmidt from
the sort of structure of the

113
00:06:15,240 --> 00:06:17,560
first person to kind of do
structural optimization.

114
00:06:18,440 --> 00:06:19,640
So I had sort of a background
there.

115
00:06:19,640 --> 00:06:25,320
And then I, I spent some time
working at Boeing applying a lot

116
00:06:25,320 --> 00:06:27,920
of these different kind of
tools.

117
00:06:27,920 --> 00:06:31,480
So a lot of software, a lot of
math, a lot of different kind of

118
00:06:31,480 --> 00:06:35,560
physics calculations for the 787
Dreamliner program.

119
00:06:35,800 --> 00:06:38,520
And the, the big challenge we
had there was trying to solve,

120
00:06:39,360 --> 00:06:42,760
it's the first kind of fully
carbon fiber airplane and the

121
00:06:42,760 --> 00:06:46,360
wing optimization, wing being
kind of the most important part

122
00:06:46,360 --> 00:06:48,240
of an airplane.
A lot of big technical

123
00:06:48,240 --> 00:06:51,080
challenges are the main
difference being since it's

124
00:06:51,080 --> 00:06:52,800
carbon fiber, many more
variables.

125
00:06:52,800 --> 00:06:54,680
And how do you sort of optimize
this design?

126
00:06:55,240 --> 00:06:58,440
And so I had a background there
kind of using different

127
00:06:58,440 --> 00:07:02,040
techniques in in sort of
parameters trying to optimize

128
00:07:02,040 --> 00:07:04,760
what it would be the lightest
weight wing design for the best

129
00:07:04,760 --> 00:07:07,680
performance.
Long story short, took us many

130
00:07:07,680 --> 00:07:10,960
years, but eventually we got
there and that's the wing if

131
00:07:10,960 --> 00:07:13,680
you've ever flown on 787, that's
the that's the wing that's on

132
00:07:13,680 --> 00:07:16,720
there.
So it's a very efficient

133
00:07:16,720 --> 00:07:19,400
airplane, right.
But in order to do that and get

134
00:07:19,400 --> 00:07:22,400
to that answer, we we really
have to leverage a lot of

135
00:07:22,400 --> 00:07:24,040
different computing
capabilities.

136
00:07:24,160 --> 00:07:25,960
And this was more than 20 years
ago.

137
00:07:25,960 --> 00:07:27,440
So there was no cloud computing
yet.

138
00:07:28,160 --> 00:07:31,360
So it was really much more about
how do we scale sort of a

139
00:07:31,360 --> 00:07:35,280
distributed systems problem from
a software perspective inside of

140
00:07:35,640 --> 00:07:37,880
Boeing with, you know, different
resources from different

141
00:07:37,960 --> 00:07:40,560
business units that we would
sort of over the weekend be able

142
00:07:40,560 --> 00:07:43,920
to kind of gather a bunch of
compute capacity service

143
00:07:43,920 --> 00:07:46,680
together from different teams
and, and solve some of these

144
00:07:46,680 --> 00:07:49,640
larger scale problems.
And eventually that that really

145
00:07:49,640 --> 00:07:52,880
got me into HPC because it was
like in academia, I actually

146
00:07:52,880 --> 00:07:55,560
studied more how do you solve
these equations more

147
00:07:55,560 --> 00:07:57,680
efficiently, right.
So trying to combine an air

148
00:07:57,680 --> 00:08:00,960
elasticity, try to kind of
combine mechanical engineering

149
00:08:01,480 --> 00:08:03,920
physics with, with fluid
dynamics and, and things like

150
00:08:03,920 --> 00:08:05,640
that.
And with the ultimate goal of

151
00:08:05,640 --> 00:08:08,560
just, you know, more efficiently
calculating all these different

152
00:08:08,560 --> 00:08:11,760
complex multi physics responses.
And at Boeing, we were

153
00:08:11,760 --> 00:08:14,200
implementing with this sort of
multi physics problem, but in a

154
00:08:14,200 --> 00:08:16,560
very different way than how
academics kind of looked at the

155
00:08:16,560 --> 00:08:18,400
problem.
It was much more about the,

156
00:08:18,880 --> 00:08:21,040
let's throw some more compute at
this problem and like where are

157
00:08:21,040 --> 00:08:23,480
the real bottlenecks and like
sort of how do you scale this?

158
00:08:24,480 --> 00:08:27,320
And from my own experience, I
really enjoyed the ability to

159
00:08:27,320 --> 00:08:31,040
kind of gather a lot of these
compute resources and just solve

160
00:08:31,240 --> 00:08:34,640
interesting problems faster.
So that got me into this whole

161
00:08:34,640 --> 00:08:38,240
sort of category of HBCI.
Think one kind of insight as

162
00:08:38,240 --> 00:08:41,120
well was like, you know, if you
look at kind of the people who

163
00:08:41,120 --> 00:08:45,040
were the best better name assist
at Boeing, for example, or the

164
00:08:45,040 --> 00:08:48,360
best at some of these like large
physics computational problems,

165
00:08:48,920 --> 00:08:52,960
they became the really good at
running HPC, right, Basically in

166
00:08:52,960 --> 00:08:54,800
practice, right.
Like these are people at in

167
00:08:54,800 --> 00:08:56,200
industry sort of working on
this.

168
00:08:56,200 --> 00:08:59,720
And I, I, I think I sort of had
a background because of the

169
00:08:59,720 --> 00:09:02,280
software and the math to be able
to solve those kinds of

170
00:09:02,280 --> 00:09:04,400
problems.
But it's a really like a sort of

171
00:09:04,400 --> 00:09:08,160
untapped potential to be able to
kind of unlock, you know, the,

172
00:09:08,200 --> 00:09:11,200
the possibility of, of
leveraging really large scale

173
00:09:11,200 --> 00:09:13,000
compute for many different
problems.

174
00:09:13,120 --> 00:09:16,520
And so that sort of.
Led to hey.

175
00:09:17,280 --> 00:09:19,840
This is an area I'm like pretty
passionate about.

176
00:09:20,160 --> 00:09:23,080
I also tried some other things.
I was a management consultant

177
00:09:23,080 --> 00:09:26,520
for some period of time, short
period of time, yeah.

178
00:09:26,520 --> 00:09:28,880
And so like I, I went to
Business School.

179
00:09:29,360 --> 00:09:32,400
So as I looked at it, I saw a
pretty broad set of different

180
00:09:33,480 --> 00:09:37,160
things you could do.
But the thing that kept pulling

181
00:09:37,160 --> 00:09:39,680
me back was I, I did really
think there was like a really

182
00:09:39,680 --> 00:09:42,360
interesting problem and and
really impactful problem if we

183
00:09:42,360 --> 00:09:45,920
could solve this sort of large
scale physics calculations for

184
00:09:45,920 --> 00:09:49,320
like engineers and scientists
that really push the sort of

185
00:09:50,240 --> 00:09:53,360
boundaries forward, not only in
a place like Boeing, but also in

186
00:09:53,360 --> 00:09:57,000
many other industries, right, in
automotive and in life sciences,

187
00:09:57,000 --> 00:09:59,240
semiconductor.
And so that seemed like a a

188
00:09:59,920 --> 00:10:01,800
worthwhile sort of mission to
pursue.

189
00:10:03,400 --> 00:10:07,760
But I'm, I'm kind of intrigued
the practicals of doing it

190
00:10:07,760 --> 00:10:10,360
because I often feel that a lot
of engineers and maybe it's

191
00:10:10,360 --> 00:10:13,360
changed now, haven't got the
mindset to create a start up

192
00:10:13,600 --> 00:10:16,920
that, you know, they, they think
a little bit more linearly, you

193
00:10:16,920 --> 00:10:19,080
know, OK, I'm going to do
engineering and make something

194
00:10:19,080 --> 00:10:21,440
better.
Was it going to Business School

195
00:10:21,440 --> 00:10:24,880
or going to McKinsey that gave
you a bit more confidential

196
00:10:24,880 --> 00:10:27,240
awareness to go and do a start
up?

197
00:10:28,840 --> 00:10:31,840
Yeah, I think, well, there's,
there's many different things.

198
00:10:31,880 --> 00:10:35,360
I I think ultimately for like
the best founders, they have

199
00:10:35,360 --> 00:10:40,080
like a few traits that are
pretty common that together are

200
00:10:40,080 --> 00:10:44,040
very uncommon, right.
So like it's see if I can sort

201
00:10:44,040 --> 00:10:46,840
of recall, I think Marc
Andreessen sort of shared what

202
00:10:46,840 --> 00:10:50,240
his perspective on this is,
which are like, like, I think a

203
00:10:50,240 --> 00:10:53,280
pretty good viewpoint.
So what is you have to be very

204
00:10:53,280 --> 00:10:55,480
open and curious, right?
So you have to be kind of

205
00:10:55,480 --> 00:10:56,720
willing to learn many different
things.

206
00:10:56,720 --> 00:10:59,480
And so that's probably also
pretty common with like people

207
00:10:59,480 --> 00:11:02,480
in academia, things like that.
You also have to.

208
00:11:04,400 --> 00:11:07,600
Be be pretty.
Willing to stick with something

209
00:11:07,600 --> 00:11:10,000
and work through a lot of
challenges for a long period of

210
00:11:10,000 --> 00:11:12,760
time.
And then there's this element of

211
00:11:12,760 --> 00:11:15,040
like he calls it
disagreeableness, but it's like

212
00:11:15,040 --> 00:11:18,880
you sort of have to be sober
enough, right to be able it's

213
00:11:18,880 --> 00:11:21,000
it's you know, starting a
company is not the most like

214
00:11:21,000 --> 00:11:23,600
rational thing to do, right?
Like no matter kind of which

215
00:11:23,600 --> 00:11:29,000
field you're in, it's, it's,
it's really hard, right?

216
00:11:29,000 --> 00:11:32,360
And so it's like people do it
because they kind of have to,

217
00:11:32,480 --> 00:11:35,160
not the people who just do it
because they want to.

218
00:11:35,160 --> 00:11:40,520
It's, it's often for maybe not
the, the sort of ideal reasons,

219
00:11:40,520 --> 00:11:42,800
right?
But you'd be sure to have these

220
00:11:42,800 --> 00:11:45,440
different traits, you need a lot
of like, I think it's, it's

221
00:11:45,440 --> 00:11:47,920
pretty risky.
And so like, and a lot of

222
00:11:47,920 --> 00:11:52,080
people, you know, will tell you
to do something different or why

223
00:11:52,080 --> 00:11:54,000
it's not going to work or why
everybody's already thought of

224
00:11:54,000 --> 00:11:55,880
this idea from the founders.
I know.

225
00:11:56,160 --> 00:11:58,360
I think it's just like, yeah,
it's, it's a combination of

226
00:11:58,360 --> 00:11:59,840
these, all these different
traits, right?

227
00:12:00,360 --> 00:12:03,720
And, and everyone is sort of
unique, also uniquely flawed,

228
00:12:03,720 --> 00:12:09,000
probably in many ways.
I certainly AM, but but it is

229
00:12:09,000 --> 00:12:12,560
hard to find.
I would say like and out of all

230
00:12:12,560 --> 00:12:14,680
these capabilities, you have to
be really smart of course and

231
00:12:14,680 --> 00:12:17,640
things like that.
But like the probably the one in

232
00:12:17,640 --> 00:12:19,840
lease supply is, is I think the
courage, right?

233
00:12:19,920 --> 00:12:24,400
It's like sort of the
willingness to kind of jump in

234
00:12:24,400 --> 00:12:26,720
and do it.
There are times though, when

235
00:12:26,720 --> 00:12:29,120
entrepreneurship is much more
popular, right?

236
00:12:29,120 --> 00:12:33,000
Like so when certain companies
are taking off and there's a lot

237
00:12:33,000 --> 00:12:35,640
of funding and things look a
little like a little easier, you

238
00:12:35,640 --> 00:12:37,160
get a lot of people jumping into
the game.

239
00:12:37,600 --> 00:12:40,880
But most big companies are
really successful ones are built

240
00:12:40,880 --> 00:12:42,520
over a very long period of time,
right?

241
00:12:42,520 --> 00:12:43,760
And there's lots of ups and
downs.

242
00:12:44,480 --> 00:12:46,960
Even if you look at the absolute
most successful companies, there

243
00:12:46,960 --> 00:12:53,320
are really challenging periods.
And I think I think you have to

244
00:12:53,320 --> 00:12:56,920
be willing to enable to kind of
work through those, those

245
00:12:56,920 --> 00:12:59,880
difficult times, right.
And yeah, it's maybe something

246
00:12:59,880 --> 00:13:03,800
from Jensen, I think I shared
before where it's like, you

247
00:13:03,800 --> 00:13:06,280
know, like it, it's really the
challenges that that form the

248
00:13:06,280 --> 00:13:10,000
character that allow you to to
become like one of these kind of

249
00:13:10,000 --> 00:13:12,800
leaders of these companies.
It's probably not just, yeah,

250
00:13:12,800 --> 00:13:14,480
just being smart or something
like that.

251
00:13:14,640 --> 00:13:17,320
Yeah, what, what practically
though were the steps?

252
00:13:17,320 --> 00:13:21,160
I'm always kind of intrigued.
So the beginning, you have an

253
00:13:21,160 --> 00:13:23,200
idea.
How did that form?

254
00:13:23,200 --> 00:13:25,400
Did you have an idea?
And then you, you know, you go

255
00:13:25,400 --> 00:13:29,360
around to the various species to
get to get funding.

256
00:13:29,720 --> 00:13:32,080
You had just a tiny idea.
And then it evolved like what?

257
00:13:32,480 --> 00:13:35,880
How did this thing come to be,
essentially?

258
00:13:35,920 --> 00:13:39,240
Yeah, there's a good book
written on this from by Peter

259
00:13:39,240 --> 00:13:41,480
Thiel 0 to one, right?
It's like like how do you start

260
00:13:41,480 --> 00:13:45,520
something from nothing?
I think the how how it

261
00:13:45,520 --> 00:13:48,840
tactically works, right?
It's it's for me, it was about

262
00:13:49,840 --> 00:13:52,280
sort of came to the conclusion
you asked about this as well.

263
00:13:52,280 --> 00:13:53,800
Like what does Business School
really teach you?

264
00:13:53,960 --> 00:13:56,360
I mean, I think the biggest
lesson from Business School is

265
00:13:56,360 --> 00:13:59,960
maybe, you know, all these, at
least the one I went to, all

266
00:13:59,960 --> 00:14:01,880
these different CE OS come
through, right?

267
00:14:01,880 --> 00:14:04,040
And they give these talks.
These are all very impressive

268
00:14:04,040 --> 00:14:07,800
people.
But I think the one thing once

269
00:14:07,800 --> 00:14:10,280
you've seen enough of them,
right, you do all these case

270
00:14:10,280 --> 00:14:11,600
studies and you learn all these
things.

271
00:14:11,600 --> 00:14:16,080
It does give you the feeling
that at least for me, you know,

272
00:14:16,160 --> 00:14:18,760
sort of you can really do
anything, right?

273
00:14:18,800 --> 00:14:20,920
Like you hear these kind of
stories, the same thing if you

274
00:14:20,920 --> 00:14:23,040
listen to like sort of founder
stories.

275
00:14:23,320 --> 00:14:27,160
And I think you could kind of do
anything you set your mind to

276
00:14:27,160 --> 00:14:29,680
do, right?
And it's in some ways one of the

277
00:14:29,680 --> 00:14:32,200
best insurance policies, right?
Because like, look, you get this

278
00:14:32,200 --> 00:14:34,800
Business School degree, you
know, you can go get a job

279
00:14:34,800 --> 00:14:37,480
somewhere.
Like I had sort of, you know,

280
00:14:37,480 --> 00:14:39,480
done this internship at
McKenzie.

281
00:14:39,480 --> 00:14:44,640
I could go back there, right?
And that gives you, you know,

282
00:14:44,640 --> 00:14:47,320
maybe sort of a floor of like,
OK, now I can take a lot more

283
00:14:47,320 --> 00:14:49,920
risk right now.
This was a time I had to, I had,

284
00:14:49,920 --> 00:14:52,760
I did not have a spouse that
didn't have kids.

285
00:14:52,960 --> 00:14:57,240
I, I was actually at sort of a
point in time, I think where I

286
00:14:57,240 --> 00:14:59,000
was able to take the most risk,
right?

287
00:14:59,080 --> 00:15:01,920
And so I think that's sort of
sets a good foundation to jump

288
00:15:01,920 --> 00:15:03,920
into to do it.
I would recommend, right?

289
00:15:03,920 --> 00:15:06,000
Like you do have to kind of burn
the boats, right?

290
00:15:06,000 --> 00:15:08,600
Like, so you can't be comparing
yourself to your peers who are

291
00:15:08,600 --> 00:15:11,240
going to go work in finance or
whatever and make a bunch of

292
00:15:11,240 --> 00:15:14,960
money or go work in academia and
publish the most papers and the

293
00:15:14,960 --> 00:15:16,640
do most innovative kind of
thinking.

294
00:15:16,960 --> 00:15:19,720
I have a lot of peers like that.
And I think you have to be kind

295
00:15:19,720 --> 00:15:23,320
of get yourself to the point
where like you, you really do

296
00:15:23,320 --> 00:15:27,840
want to start this company kind
of no matter what, right?

297
00:15:28,360 --> 00:15:30,800
And then you kind of have to
like most founders describe it

298
00:15:30,800 --> 00:15:32,880
as like, I would say that you
kind of have to right, Like you

299
00:15:32,880 --> 00:15:35,720
just don't see any other way.
And that's kind of the how I saw

300
00:15:35,720 --> 00:15:37,120
it.
It was not to just for the

301
00:15:37,120 --> 00:15:38,640
purpose of starting company.
It was like.

302
00:15:39,440 --> 00:15:42,640
To make the impact.
That I thought would be possible

303
00:15:42,640 --> 00:15:45,720
like that I could make myself
right and I saw kind of was

304
00:15:45,720 --> 00:15:49,080
possible at Boeing and then it's
like OK there's many other

305
00:15:49,080 --> 00:15:51,480
industries there's many other
engineers and scientists who are

306
00:15:51,480 --> 00:15:54,920
all like bottleneck by compute
basically right and then there's

307
00:15:54,920 --> 00:15:56,920
this thing called cloud
computing and like everybody's

308
00:15:56,920 --> 00:16:02,960
access to these and it seemed
again I felt I was late right

309
00:16:02,960 --> 00:16:07,400
and so it seemed pretty obvious
to get to go pursue that the.

310
00:16:08,040 --> 00:16:11,600
The tactical steps would be I
decided to just move to Silicon

311
00:16:11,600 --> 00:16:14,560
Valley.
So like, I came out Silicon

312
00:16:14,560 --> 00:16:18,240
Valley and I knew nothing right
about.

313
00:16:18,240 --> 00:16:21,080
Silicon Valley, really like I'd,
you know, like read some books,

314
00:16:21,320 --> 00:16:24,280
listened to some podcasts,
probably weren't really, I don't

315
00:16:24,280 --> 00:16:24,880
even know.
What they were.

316
00:16:24,960 --> 00:16:27,440
Called podcasts at that time,
but the you know, Stanford had

317
00:16:27,440 --> 00:16:31,440
this like I think they still do
this entrepreneurial thought

318
00:16:31,440 --> 00:16:34,160
leaders program.
And so I remember just listening

319
00:16:34,240 --> 00:16:38,440
to these people come in and this
is a different generation.

320
00:16:38,440 --> 00:16:40,040
So you talk about like, you
know, I was listening to this

321
00:16:40,040 --> 00:16:42,280
stuff like like, like way back
in the day, right?

322
00:16:42,280 --> 00:16:47,640
Like sort of 20, probably 20
tenths maybe earlier to yeah,

323
00:16:47,760 --> 00:16:49,880
probably earlier.
But they give you some

324
00:16:50,400 --> 00:16:54,040
inspiration and sort of a road
map of of like, you know, how

325
00:16:54,040 --> 00:16:56,640
you can start a company.
There is something special about

326
00:16:56,640 --> 00:16:57,960
Silicon Valley.
So you have this very high

327
00:16:57,960 --> 00:17:02,480
concentration of founders, you
know, engineers who want to

328
00:17:02,480 --> 00:17:06,160
build companies as sort of
builders, investors.

329
00:17:06,720 --> 00:17:09,400
You have sort of a cultural
appetite for risk, right?

330
00:17:09,520 --> 00:17:12,079
Like it's a, it's a very special
place.

331
00:17:12,760 --> 00:17:16,680
And I do think if it's like my
mentality is a little bit like,

332
00:17:16,680 --> 00:17:19,480
you know, if you're going to, if
you're going to go do something

333
00:17:19,480 --> 00:17:25,000
like this, you might as well,
you know, try to sort of play in

334
00:17:25,000 --> 00:17:28,760
the NBA, so to speak, right?
Like so many places you can

335
00:17:28,760 --> 00:17:33,320
start a company, but I think to
give yourself the best chances

336
00:17:33,320 --> 00:17:36,400
of success, it seems like moving
to Silicon Valley would be a a

337
00:17:36,600 --> 00:17:39,640
good move, right?
But I literally moved out

338
00:17:39,640 --> 00:17:43,680
without really, I had no family
or you know, I had some

339
00:17:44,240 --> 00:17:46,480
classmates that also moved to
the Bay Area.

340
00:17:46,520 --> 00:17:51,760
But I did not have any like like
sort of real good reason to be

341
00:17:51,760 --> 00:17:53,720
there other than to kind of try
to start a company.

342
00:17:54,640 --> 00:17:56,320
And I just started working at
writing coat.

343
00:17:56,640 --> 00:17:59,920
So I just started building the
product in parallel.

344
00:17:59,920 --> 00:18:02,960
It was like, you know, trying to
see, hey, you know, can we raise

345
00:18:02,960 --> 00:18:05,760
some money, etcetera.
But it was just by myself and

346
00:18:05,760 --> 00:18:08,640
just writing some code.
That's really how it started,

347
00:18:08,880 --> 00:18:10,240
right?
And I think you mentioned

348
00:18:10,240 --> 00:18:13,680
earlier like sort of eventually
led to Y Combinator, which is

349
00:18:13,680 --> 00:18:15,440
one of the incubators.
So that that can really help a

350
00:18:15,440 --> 00:18:17,160
lot, right?
Because they that can give you a

351
00:18:17,160 --> 00:18:20,880
pretty quick start and a very
fast network of other founders

352
00:18:21,080 --> 00:18:24,680
and great mentors that I can get
you off the ground.

353
00:18:25,480 --> 00:18:26,560
It's a little bit different than
today.

354
00:18:26,560 --> 00:18:29,360
This is back in 2011, right?
So at that time you got to

355
00:18:29,360 --> 00:18:32,080
remember it was like social,
local, mobile or all the hot

356
00:18:32,080 --> 00:18:35,800
trends, right?
And so like sort of working on,

357
00:18:37,080 --> 00:18:39,400
you know, a, a product that
would do like.

358
00:18:39,400 --> 00:18:42,520
You.
Know complex super computing for

359
00:18:42,520 --> 00:18:45,960
multidisciplinary optimization
of physics where the aerospace

360
00:18:45,960 --> 00:18:49,680
market was definitely not the
hottest idea, right yeah, yeah,

361
00:18:50,200 --> 00:18:55,000
but I think you know that what
is nice is I do think the

362
00:18:55,000 --> 00:18:59,320
culture in Silicon Valley
embraces this sort of anybody

363
00:18:59,320 --> 00:19:01,760
with an idea can come there
right and you know ideas are

364
00:19:01,760 --> 00:19:04,240
cheap right so it's all about
the building and the execution

365
00:19:04,240 --> 00:19:05,400
so.
Yeah.

366
00:19:06,800 --> 00:19:10,320
It's in a very meritocratic
place, like, unlike like many

367
00:19:10,320 --> 00:19:13,480
other sort of games, so to
speak, in life.

368
00:19:13,560 --> 00:19:16,800
I feel like Silicon Valley is
quite meritocratic, right?

369
00:19:17,400 --> 00:19:19,920
Where sort of anybody can be the
next like Zuck.

370
00:19:20,840 --> 00:19:22,760
And so everybody kind of has to
be nice to each other and help

371
00:19:22,760 --> 00:19:27,040
each other.
And so, you know, it's yeah, it

372
00:19:27,040 --> 00:19:29,920
has a nice.
I think that dynamic is really,

373
00:19:29,920 --> 00:19:31,960
there's a lot of paying it for
it, like a lot of founders help

374
00:19:31,960 --> 00:19:34,000
each other.
People are very accessible,

375
00:19:34,040 --> 00:19:37,680
right.
So for me that was yeah, it was

376
00:19:37,680 --> 00:19:39,960
a it was a great journey.
It was, it was pretty tough

377
00:19:39,960 --> 00:19:40,960
though.
Like I like I was saying, I was

378
00:19:40,960 --> 00:19:43,240
pitching this company.
Yeah, yeah.

379
00:19:43,360 --> 00:19:47,400
That, you know, the exact
opposite of social locomobile.

380
00:19:47,400 --> 00:19:49,760
There's like literally that if
you have the Venn diagrams, we

381
00:19:49,760 --> 00:19:52,120
would be the one that like does
not overlap with anything that

382
00:19:52,200 --> 00:19:54,120
like investors were interested
in at that time.

383
00:19:54,120 --> 00:19:56,080
But you know, everything goes
through these waves, right?

384
00:19:56,080 --> 00:20:00,880
And so I do think eventually,
like, you know, if you just have

385
00:20:00,880 --> 00:20:02,200
a good mission, you're on,
right?

386
00:20:02,200 --> 00:20:03,600
Like you'll find some funding,
right?

387
00:20:03,600 --> 00:20:07,360
Like it is a there's a lot of
investors, right?

388
00:20:07,360 --> 00:20:09,760
And so ultimately you just need
one to write a check.

389
00:20:10,320 --> 00:20:13,160
Yeah, yeah.
So what was the first few years

390
00:20:13,160 --> 00:20:14,440
like?
What were the some standout

391
00:20:14,440 --> 00:20:16,520
moments?
Who were the was a returning

392
00:20:16,520 --> 00:20:21,080
point from a big investor or
like in those early days, 2011,

393
00:20:21,080 --> 00:20:25,680
what when did you really feel
that he was going to work out?

394
00:20:25,760 --> 00:20:28,080
Because I guess there were
moments when you maybe thought,

395
00:20:28,080 --> 00:20:30,960
OK, I should just stop doing
this and get a job somewhere

396
00:20:30,960 --> 00:20:33,240
else.
Yeah, I mean, I think like, you

397
00:20:33,240 --> 00:20:35,840
know, what you call working out
versus like success, these are

398
00:20:35,840 --> 00:20:39,280
all like your own definition, I
would say, right, Like so and

399
00:20:39,280 --> 00:20:41,280
that definition for most people
changes over time.

400
00:20:41,960 --> 00:20:44,720
I remember making a promise to
my significant other at the

401
00:20:44,720 --> 00:20:48,240
time, we were not married yet
that, you know, like I'll just

402
00:20:48,240 --> 00:20:50,440
do this thing.
And she was working really hard

403
00:20:51,560 --> 00:20:53,880
and and, you know, like really
grinding.

404
00:20:53,960 --> 00:20:56,960
And I was just kind of sitting
in our, I was working really

405
00:20:56,960 --> 00:20:58,960
hard, but I was just writing
software and not making any

406
00:20:58,960 --> 00:21:00,640
money and, and she was paying
the rent.

407
00:21:01,360 --> 00:21:04,560
And so we sort of made this deal
where like, OK, well, if you

408
00:21:04,560 --> 00:21:06,440
know, let's, you know, at what
point are you going to say, how

409
00:21:06,440 --> 00:21:07,880
can I get a get a real job,
right?

410
00:21:08,480 --> 00:21:09,800
And my parents were wondering
the same thing.

411
00:21:09,920 --> 00:21:13,680
And we basically said, well,
like, you know, let's give it

412
00:21:13,680 --> 00:21:15,680
two years.
And if you can pay yourself a

413
00:21:15,680 --> 00:21:19,560
salary of something, right?
Like enough to clear like the

414
00:21:19,560 --> 00:21:23,600
minimum medical benefits and
things like that, then that's a

415
00:21:23,600 --> 00:21:25,040
win, right?
Like that, then we can keep

416
00:21:25,040 --> 00:21:26,720
going basically.
So that was kind of the.

417
00:21:26,720 --> 00:21:33,800
Deal I made with her and.
I think that's you know, like,

418
00:21:33,880 --> 00:21:38,320
like, yeah, I would say really
made it it it changes all the

419
00:21:38,320 --> 00:21:40,800
time.
If I think if you have the

420
00:21:40,800 --> 00:21:43,160
ambition and sort of the mission
that we're on, right?

421
00:21:43,160 --> 00:21:44,720
It's a it's a very big mission,
right?

422
00:21:45,520 --> 00:21:49,480
And so, you know, I think I
think we still have a lot of our

423
00:21:49,480 --> 00:21:51,120
work cut out for us even today,
right?

424
00:21:51,520 --> 00:21:55,240
But but some big milestones,
important lesson I think for

425
00:21:55,240 --> 00:21:58,160
founders is also like, you know,
I think Paul Graham says this

426
00:21:58,160 --> 00:22:04,280
like, you know, companies don't
die, founders give up, right?

427
00:22:04,280 --> 00:22:06,360
And you can't just keep going
right now.

428
00:22:06,360 --> 00:22:09,200
There's there's a limit probably
like you take market feedback,

429
00:22:09,200 --> 00:22:11,240
you're like, Hey, is this still,
is this investable?

430
00:22:11,240 --> 00:22:13,360
Is this like smart?
Like did you learn some new

431
00:22:13,360 --> 00:22:15,240
things?
But if you're kind of iterating

432
00:22:15,240 --> 00:22:17,040
quickly, you're learning a lot
of things.

433
00:22:17,040 --> 00:22:20,320
You're adapting your company to
like meet the market, so to

434
00:22:20,320 --> 00:22:22,080
speak, right?
Kind of get this product market

435
00:22:22,080 --> 00:22:26,680
fit and then any set of like
really smart people that, that

436
00:22:26,680 --> 00:22:28,920
work really hard together and
sort of have these attributes

437
00:22:30,080 --> 00:22:31,680
they, they will build success,
right?

438
00:22:31,760 --> 00:22:35,000
Like, and so I think it is in
that way an amazing place where

439
00:22:35,000 --> 00:22:36,880
you can kind of pursue, pursue
what you want to do.

440
00:22:37,280 --> 00:22:40,720
For us, a big milestone was
first check.

441
00:22:40,720 --> 00:22:44,280
So the first investor is always
very special for us.

442
00:22:44,280 --> 00:22:47,880
That was, you know, we'd gone
around like pitched all these

443
00:22:47,880 --> 00:22:49,880
like VCs kept getting turned
down.

444
00:22:49,880 --> 00:22:51,760
And at this point we were doing
Y Combinator.

445
00:22:52,240 --> 00:22:55,520
And you know, Paul Graham teed
up like one of his buddies.

446
00:22:56,320 --> 00:22:57,760
It's like, OK, just give him the
pitch.

447
00:22:58,480 --> 00:23:00,920
Here's the thing, I followed
exactly the playbook that he

448
00:23:00,920 --> 00:23:05,280
gave me and you know, we didn't,
we didn't know what price, what,

449
00:23:06,400 --> 00:23:08,680
you know, amount of the company
we should give up for, for

450
00:23:08,680 --> 00:23:10,520
whatever investment around.
We're just trying to get some

451
00:23:10,520 --> 00:23:14,320
investor right.
And it was like, OK, well, like,

452
00:23:14,520 --> 00:23:17,680
like, how about these terms or
whatever based on guidance from

453
00:23:17,680 --> 00:23:19,600
Paul Graham.
And, you know, eventually he

454
00:23:19,600 --> 00:23:22,880
goes through a bunch of things
and still don't invest.

455
00:23:23,280 --> 00:23:25,920
So I come back to Paul and I'm
just like, man, you know, we did

456
00:23:25,920 --> 00:23:28,680
everything you said.
This is like the, you know, 50th

457
00:23:28,680 --> 00:23:34,080
time and, you know, we didn't
get the check and but but you

458
00:23:34,080 --> 00:23:36,480
know, I did follow like all the
things that you said, right?

459
00:23:36,480 --> 00:23:37,880
And so like, why don't you just
invest?

460
00:23:37,880 --> 00:23:41,640
And he's like, sure, he whips
out his checkbook and writes a

461
00:23:41,640 --> 00:23:44,440
check.
And, and I think that was very

462
00:23:44,440 --> 00:23:46,600
special.
Not really about the actual

463
00:23:46,600 --> 00:23:50,440
dollar amount invested.
It was much more about his, you

464
00:23:50,440 --> 00:23:53,000
know, when somebody kind of
makes a bet on you like that,

465
00:23:53,080 --> 00:23:54,920
right?
And it's like real money out of

466
00:23:54,920 --> 00:23:58,160
his pocket.
It, you know, means so much to

467
00:23:58,160 --> 00:24:01,080
you that I'm sure you've seen
this with maybe like a thesis

468
00:24:01,080 --> 00:24:03,640
advisor or other people who've
been in your career in the past.

469
00:24:04,160 --> 00:24:07,000
That that is a a really special
moment, right?

470
00:24:07,560 --> 00:24:10,520
And I think that certainly gave
me a lot of confidence to kind

471
00:24:10,520 --> 00:24:14,080
of just keep, keep going.
And eventually we, we got lots

472
00:24:14,080 --> 00:24:16,520
of great investors, but that was
that was a special one.

473
00:24:16,520 --> 00:24:19,240
Another one was, you know, we,
we'd started working with some

474
00:24:19,240 --> 00:24:20,920
early customers like we couldn't
raise money.

475
00:24:20,920 --> 00:24:24,520
So we just like built the
product and started to get who

476
00:24:24,520 --> 00:24:27,840
wants to go run like large scale
CFD in the cloud, right?

477
00:24:29,200 --> 00:24:33,480
And it's like, you know, there's
like Boeing, Airbus, etcetera.

478
00:24:33,480 --> 00:24:35,800
At that time there were these
really small space companies,

479
00:24:35,800 --> 00:24:37,880
we're talking about like 50
person companies, right?

480
00:24:39,440 --> 00:24:43,680
And we went to, there were
basically only two private space

481
00:24:43,680 --> 00:24:46,800
companies at the time.
We went to both of them and

482
00:24:47,000 --> 00:24:48,680
because of my aerospace
background, I've known some

483
00:24:48,680 --> 00:24:51,920
people that working there.
So we're able to kind of chat

484
00:24:51,920 --> 00:24:55,480
with them and eventually we sort
of got them on the hook that

485
00:24:55,480 --> 00:25:00,760
they they would do this right.
And that that was also very

486
00:25:00,760 --> 00:25:03,000
special because I think that's
when you kind of.

487
00:25:03,520 --> 00:25:05,840
Could see hey.
Like, and they're making a bet

488
00:25:05,840 --> 00:25:07,640
on you, right?
And, and, you know, when the

489
00:25:07,640 --> 00:25:09,920
company is like one or two
people, it's very personal,

490
00:25:10,080 --> 00:25:12,880
right?
So when these with these first

491
00:25:12,880 --> 00:25:15,440
few customers, like at this
point, I had my Co founder,

492
00:25:15,440 --> 00:25:17,840
who's my old boss from Boeing,
Adam McKenzie join.

493
00:25:17,840 --> 00:25:21,480
And, you know, we sit there in
the room with some of these like

494
00:25:21,480 --> 00:25:24,800
aerodynamicists and they're just
like this concept of putting it

495
00:25:24,800 --> 00:25:27,720
in the cloud is, is pretty
foreign.

496
00:25:27,760 --> 00:25:32,040
And remember, like web services,
So EWS had just started and web

497
00:25:32,040 --> 00:25:34,800
services are built for like
almost the exact opposite of

498
00:25:34,800 --> 00:25:36,800
HPC, right?
Like it, it's like very loosely

499
00:25:36,800 --> 00:25:40,480
coupled, very, you know, you
kind of have to write your

500
00:25:40,480 --> 00:25:42,640
software very tolerant to
failure, etcetera.

501
00:25:43,920 --> 00:25:46,280
The actual performance is like
not great.

502
00:25:47,080 --> 00:25:52,160
And so it, you know, but you
could kind of see where this was

503
00:25:52,160 --> 00:25:55,400
going.
And so we pitched these

504
00:25:55,400 --> 00:25:58,640
companies and I think that was a
big breakthrough when when they

505
00:25:58,640 --> 00:26:00,480
were like, yes, we'll buy the
software, We'll pay something

506
00:26:00,480 --> 00:26:03,480
for this basically right.
And we'll buy the software and

507
00:26:03,480 --> 00:26:06,040
again, that was many iterations
to get there.

508
00:26:06,120 --> 00:26:08,240
But once you get that
breakthrough, I mean, that was

509
00:26:08,240 --> 00:26:12,400
very exciting, right?
And and those little winds along

510
00:26:12,400 --> 00:26:15,320
the way, I think it's like at no
point, I think the entire

511
00:26:15,320 --> 00:26:17,080
journey I felt like I've really
made it.

512
00:26:17,600 --> 00:26:20,720
Yeah.
But you know, like I, I think

513
00:26:20,720 --> 00:26:23,560
those are very material in those
early days, right?

514
00:26:23,560 --> 00:26:25,640
Like like just just getting off
the ground.

515
00:26:25,680 --> 00:26:29,600
I think many great ideas just
don't have enough energy on

516
00:26:29,600 --> 00:26:32,240
execution put behind them,
right?

517
00:26:32,280 --> 00:26:34,760
And, and even to this day,
right, like I still have people

518
00:26:34,760 --> 00:26:37,800
come to me and be like, well, I
mean, we're doing much more than

519
00:26:37,800 --> 00:26:40,240
just HPC these days.
But you get some, you know, you

520
00:26:40,240 --> 00:26:42,720
still have this sort of old
network of HPC people, you know,

521
00:26:42,720 --> 00:26:45,520
I both know who we're talking
about and.

522
00:26:45,640 --> 00:26:48,200
You know there'll be.
People who say you get both

523
00:26:48,200 --> 00:26:50,240
messages, right?
There were people at that time

524
00:26:50,240 --> 00:26:52,880
who were like, you get these
Silicon Valley investors who

525
00:26:52,880 --> 00:26:55,880
were like, what do you mean like
HPC, Like it already exists,

526
00:26:55,880 --> 00:26:57,800
It's cloud computing and it's
AWS, right?

527
00:26:57,800 --> 00:27:01,520
So you have that class of people
which are basically sort of the,

528
00:27:02,120 --> 00:27:03,320
yeah, it's like it's already
been done.

529
00:27:03,320 --> 00:27:04,720
Like why are you wasting your
time on this?

530
00:27:04,720 --> 00:27:07,120
And then you have the other
class, which are the HPC people,

531
00:27:07,120 --> 00:27:08,160
right?
Like it's super computing.

532
00:27:08,240 --> 00:27:11,640
And these these folks are like,
this is absolutely impossible,

533
00:27:11,760 --> 00:27:13,840
like never possible.
All that hardware is not good

534
00:27:13,840 --> 00:27:17,200
enough, like etcetera.
And you have an entire ecosystem

535
00:27:17,200 --> 00:27:19,760
that these days it's a little
bit different, right?

536
00:27:19,760 --> 00:27:22,840
But remember then it was like
nobody was really doing this and

537
00:27:22,840 --> 00:27:24,800
nobody believed it was really
possible.

538
00:27:26,720 --> 00:27:29,960
And then you have people who
say, well, I had that idea.

539
00:27:31,040 --> 00:27:33,840
Right, as if having the idea is
an important thing.

540
00:27:33,840 --> 00:27:37,720
Yeah, I think the execution is
the the big thing that really

541
00:27:37,720 --> 00:27:38,960
matters.
And that's where like those

542
00:27:38,960 --> 00:27:41,440
traits of a good founder, I
think you, you, you know,

543
00:27:42,280 --> 00:27:44,240
willing to persevere.
If you look at most companies,

544
00:27:44,240 --> 00:27:46,840
if you're just willing to kind
of go for more than a decade,

545
00:27:46,840 --> 00:27:50,160
most of the companies are very
successful, right?

546
00:27:51,960 --> 00:27:54,560
But it's not easy, right?
So like I, I think that's also

547
00:27:54,560 --> 00:27:58,320
like, it's not for everybody.
Well, and, and I was thinking

548
00:27:58,320 --> 00:28:02,160
that you were probably really
facing an uphill battle because,

549
00:28:03,240 --> 00:28:06,040
you know, if I look at from a
technology point of view, you

550
00:28:06,040 --> 00:28:10,400
know, the hyperscale has took
quite a long time to really

551
00:28:11,760 --> 00:28:13,640
have, for example, the
interconnect.

552
00:28:14,120 --> 00:28:18,680
You, you, you know, like when I
joined Adbs in what 2020, things

553
00:28:18,680 --> 00:28:21,920
like EFA or they, they were
quite new.

554
00:28:21,920 --> 00:28:26,440
So you were, did you feel that
was part of the challenge before

555
00:28:26,440 --> 00:28:29,880
that, the hyperscalers that I
guess you know, you were

556
00:28:29,880 --> 00:28:36,480
simplifying the use of when
moving as quick as you wanted in

557
00:28:36,480 --> 00:28:41,600
terms of like technology?
Yeah, very good question because

558
00:28:41,600 --> 00:28:45,240
like that was the remember we
started in 2011, right.

559
00:28:45,240 --> 00:28:49,640
So there were some some days in
the wilderness there, but.

560
00:28:52,080 --> 00:28:54,720
With cloud computing was kind of
really gaining traction, right?

561
00:28:54,840 --> 00:28:56,960
And even back then, you got to
remember people were way

562
00:28:56,960 --> 00:28:59,880
underestimating these markets by
like X, right?

563
00:28:59,960 --> 00:29:03,240
Like from where it is today?
It shows how difficult it is.

564
00:29:03,240 --> 00:29:04,880
Even when you see something
happening, it's like very

565
00:29:04,880 --> 00:29:07,080
difficult to forecast, right?
We could talk about AI later,

566
00:29:07,360 --> 00:29:12,120
yeah.
Early on, a big one was always

567
00:29:12,120 --> 00:29:14,640
like, well, like, like why do I
need Rescale, right?

568
00:29:14,640 --> 00:29:18,640
Like I could just go to a cloud
company and get this myself,

569
00:29:18,640 --> 00:29:20,760
right?
And so and.

570
00:29:21,080 --> 00:29:23,760
A lot of people perceived even
to this day, probably people

571
00:29:23,760 --> 00:29:28,120
perceived the cloud companies
that are very close partners as

572
00:29:28,120 --> 00:29:31,120
like competitors in this market
and that's sort of the wrong

573
00:29:31,120 --> 00:29:34,960
framing.
The reality was I would be going

574
00:29:34,960 --> 00:29:37,560
in at that time.
It was going to be Andy, Jesse

575
00:29:38,160 --> 00:29:42,320
and begging him to build
InfiniBand networking and

576
00:29:42,320 --> 00:29:45,000
begging him to build like Specs
that.

577
00:29:45,880 --> 00:29:48,680
Would do really well for this
market, right?

578
00:29:48,720 --> 00:29:52,400
And the answer often wasn't
necessarily no, but it was like,

579
00:29:52,400 --> 00:29:56,600
well, an incremental dollar
spent today, you know, I'm

580
00:29:56,600 --> 00:29:59,760
better off just building more,
let's call it for simplification

581
00:29:59,760 --> 00:30:02,680
commodity compute like because
there's more market share to

582
00:30:02,680 --> 00:30:06,560
grab there versus this like
specialized compute, everything

583
00:30:06,560 --> 00:30:11,040
costs X more, workloads are much
more volatile, etcetera,

584
00:30:11,040 --> 00:30:12,880
etcetera, etcetera, right.
So there were sort of many

585
00:30:12,880 --> 00:30:15,120
reasons to not do it.
There's also like some real

586
00:30:15,120 --> 00:30:18,440
difficult networking challenges
like security challenges.

587
00:30:18,440 --> 00:30:20,280
So the the way you build like a
cloud.

588
00:30:20,280 --> 00:30:23,880
Service as, as you know very
well is a little bit different

589
00:30:23,880 --> 00:30:27,320
than the HPC system, right.
And so at that time, the

590
00:30:27,320 --> 00:30:32,800
maturity wasn't there to easily
sort of pursue this market, but.

591
00:30:33,120 --> 00:30:34,680
But from first.
Principles, it seems like a

592
00:30:34,680 --> 00:30:36,760
really a lost opportunity,
right, because if you kind of

593
00:30:36,760 --> 00:30:39,720
look at it, it's like, well, you
know, commodity confused great.

594
00:30:39,720 --> 00:30:41,880
Like the whole thing about AWS
was, hey, it was built for

595
00:30:41,880 --> 00:30:45,480
scaling e-commerce peak loads
during the like, Black Friday or

596
00:30:45,480 --> 00:30:48,320
something, right?
And but you know, like the rest

597
00:30:48,320 --> 00:30:51,760
of the time and even today, like
if you just look at it wasn't

598
00:30:51,760 --> 00:30:54,840
the money makers at AWS, it's
like the simple, it's just like

599
00:30:54,840 --> 00:30:59,160
EC2S3 like basic stuff, right?
And most people consume just

600
00:30:59,160 --> 00:31:03,160
that what and super computing
and high performance computing

601
00:31:03,160 --> 00:31:05,400
was always the case is that, you
know, these systems are super

602
00:31:05,400 --> 00:31:07,560
expensive.
So time sharing them only the

603
00:31:07,560 --> 00:31:10,080
biggest companies in the world
or the biggest labs could afford

604
00:31:10,080 --> 00:31:12,000
to kind of build these systems,
right?

605
00:31:12,080 --> 00:31:15,040
And it makes a lot of sense to
share those systems, right,

606
00:31:15,040 --> 00:31:16,200
because they're so expensive,
right?

607
00:31:16,200 --> 00:31:18,840
So share the CapEx investment,
the sort of timeshare of these

608
00:31:18,840 --> 00:31:20,600
systems, so?
Used to be like mainframe time

609
00:31:20,600 --> 00:31:22,920
sharing, but like with cloud
this becomes so much easier.

610
00:31:23,440 --> 00:31:28,400
And so like it really did seem
like that was A at that time

611
00:31:28,800 --> 00:31:31,960
like a a lost opportunity and as
Rescale we did make a conscious

612
00:31:31,960 --> 00:31:33,800
choice.
It's like we are not at least at

613
00:31:33,800 --> 00:31:35,520
that time, right.
We're not going to go get into

614
00:31:35,520 --> 00:31:39,360
this CapEx game like the venture
capital dollars are not made.

615
00:31:39,400 --> 00:31:40,640
You can see that today with
these.

616
00:31:40,840 --> 00:31:43,720
AI companies in the
infrastructure side are not

617
00:31:43,720 --> 00:31:46,920
really made for, you know,
making big CapEx compute

618
00:31:46,920 --> 00:31:50,760
investments, right.
And so, yeah, we would be trying

619
00:31:50,760 --> 00:31:53,520
to work really closely with the.
Cloud providers to actually

620
00:31:53,520 --> 00:31:55,520
build the right infrastructure
for our customers.

621
00:31:55,520 --> 00:31:57,080
But you could kind of see where
it's going, right?

622
00:31:57,080 --> 00:31:59,360
Like at that time it was to me
again, I thought it was late.

623
00:32:00,040 --> 00:32:03,720
It was pretty obvious that like
this is a great market.

624
00:32:03,720 --> 00:32:07,000
Opportunity, right, But like
yeah, yeah, I think of the eyes

625
00:32:07,000 --> 00:32:10,080
of say somebody like Andy Jassy
or you know, Sacho was running

626
00:32:10,080 --> 00:32:11,640
advanced computing at that time
in Azure.

627
00:32:11,640 --> 00:32:15,320
I think they made some smart
bets too, I think.

628
00:32:15,320 --> 00:32:20,640
I think they, yeah, HPC is still
this like small sub segment that

629
00:32:20,640 --> 00:32:22,840
seems very difficult relative to
all of computing.

630
00:32:23,560 --> 00:32:25,320
So all these things are all
about your frame, right?

631
00:32:25,320 --> 00:32:27,320
Like if you come from the HPC
frame, this seems like pretty

632
00:32:27,320 --> 00:32:29,920
obvious I think.
If you come from the, you're

633
00:32:29,920 --> 00:32:33,200
running like the fastest
growing, most successful cloud

634
00:32:33,200 --> 00:32:35,760
company in the world, like, you
know, there's many bets you

635
00:32:35,760 --> 00:32:38,080
could be making.
And yeah, it took a while to get

636
00:32:38,080 --> 00:32:39,360
this one off the ground even
today.

637
00:32:39,360 --> 00:32:41,600
And as far as I understand,
there's no infinite band

638
00:32:41,600 --> 00:32:45,640
networking at at AWS.
No, that's, that's true.

639
00:32:45,760 --> 00:32:48,560
And we should say as well, I
guess that when we use the term

640
00:32:48,560 --> 00:32:52,120
HPC and I guess the relevant
first discussing this is, I mean

641
00:32:52,120 --> 00:32:54,520
you can argue, but I would say
from a hardware point of view,

642
00:32:54,520 --> 00:32:59,640
what AI training needs today has
many, many, many of the same

643
00:32:59,640 --> 00:33:02,120
components as HPC.
There's some subtleties, but in

644
00:33:02,120 --> 00:33:07,320
terms of a low latency network,
that's basically what AI

645
00:33:07,320 --> 00:33:10,480
training needs as well.
So interestingly, like the

646
00:33:11,080 --> 00:33:14,680
reason that we're talking about
prior in those early days, it

647
00:33:14,680 --> 00:33:17,400
was only really for loosely
coupled, wasn't it, from like a

648
00:33:17,400 --> 00:33:19,360
hardcore performance point of
view.

649
00:33:19,440 --> 00:33:24,400
So my question is, do you think
those early days almost made it

650
00:33:24,400 --> 00:33:27,920
harder to convince people when
there was actually the hardware?

651
00:33:27,920 --> 00:33:32,960
Are you still getting people who
are like it's it's slow running

652
00:33:32,960 --> 00:33:35,560
on the cloud or there's a
because they they're still just

653
00:33:35,560 --> 00:33:40,280
thinking back to the mid twenty
10s when maybe the hardware

654
00:33:40,920 --> 00:33:42,360
wasn't there.
Do you know what I mean?

655
00:33:42,360 --> 00:33:45,120
Are people lazy to be thinking
of what's current?

656
00:33:45,560 --> 00:33:48,000
Yeah, and and like you and I are
probably on the same side of the

657
00:33:48,000 --> 00:33:54,000
table on this, but, but I think
if I had to steal a man, like

658
00:33:54,000 --> 00:33:57,320
what's the sort of reason to not
not go to cloud, right?

659
00:33:57,320 --> 00:33:59,640
Like a lot of people bring up
costs and we can talk about

660
00:33:59,640 --> 00:34:03,600
that.
But I think if I'm sort of put

661
00:34:03,600 --> 00:34:06,400
my hat on as one of our
customers, right like like a

662
00:34:06,400 --> 00:34:09,880
manufacturing company it seems
if you're.

663
00:34:10,280 --> 00:34:14,120
Certainly starting a new company
like like really obvious.

664
00:34:14,440 --> 00:34:16,880
To to sort.
Of buy this compute as a service

665
00:34:17,239 --> 00:34:19,400
as opposed to start building
your own data centers etcetera,

666
00:34:19,639 --> 00:34:21,159
right?
Like unless there is some sort

667
00:34:21,159 --> 00:34:23,480
of reason that's your
competitive advantage, right?

668
00:34:24,320 --> 00:34:26,440
I can tell you Boeing thought it
was their competitive advantage.

669
00:34:26,440 --> 00:34:27,920
They probably still take that to
this day.

670
00:34:27,920 --> 00:34:31,560
And yeah, there's, there are
some I would say like edge cases

671
00:34:31,560 --> 00:34:34,400
where it can make sense.
But in general, right, you want

672
00:34:34,400 --> 00:34:38,440
to kind of a provider.
Like any major cloud provider

673
00:34:38,480 --> 00:34:41,960
has way better economics, has
way more efficient sort of

674
00:34:41,960 --> 00:34:44,840
scale.
The better supply chain they

675
00:34:44,840 --> 00:34:47,880
have like better hardware,
faster, they are able to refresh

676
00:34:47,880 --> 00:34:50,880
and resell your old hardware if
you want to rotate to new skews

677
00:34:50,880 --> 00:34:53,440
and things like that.
So there's so many benefits if

678
00:34:53,440 --> 00:34:55,400
you just.
Look at solving the real problem

679
00:34:55,400 --> 00:34:57,360
you're trying to solve as a
manufacturer, say running a

680
00:34:57,360 --> 00:35:00,920
simulation faster or designing a
vehicle faster.

681
00:35:01,400 --> 00:35:03,120
That's the real problem you're
trying to solve, right?

682
00:35:03,720 --> 00:35:07,600
The problem I think in HPC.
Is there's an entire industry

683
00:35:07,600 --> 00:35:10,080
that's been set up with a
different framing, which is like

684
00:35:10,080 --> 00:35:14,800
I'm, I'm actually here to most
efficiently invest in CapEx, run

685
00:35:14,800 --> 00:35:18,360
the infrastructure into the
ground, try to efficiently use

686
00:35:18,360 --> 00:35:21,360
this infrastructure, you know,
depreciate it in a really smart

687
00:35:21,360 --> 00:35:22,840
way.
But it's all about this sort of

688
00:35:22,840 --> 00:35:25,720
ITTCO sort of frame, right?
And.

689
00:35:26,240 --> 00:35:30,160
Purely on that frame, maybe
sometimes on Prem is competitive

690
00:35:30,160 --> 00:35:32,760
with cloud, right?
But but if you even value a

691
00:35:32,760 --> 00:35:35,480
little bit like actual
performance, which is in the

692
00:35:35,480 --> 00:35:38,200
name of the industry, right,
like high performance computing,

693
00:35:38,320 --> 00:35:41,280
the simple sort of to me.
The concept is, is like, well,

694
00:35:41,280 --> 00:35:42,800
who do you want to bear the
CapEx?

695
00:35:42,800 --> 00:35:46,400
Do you want it to be some really
low cost of capital, very big

696
00:35:46,400 --> 00:35:48,400
cloud infrastructure player,
right?

697
00:35:48,520 --> 00:35:52,160
Should that be you, right, as as
a manufacturer and especially

698
00:35:52,160 --> 00:35:56,240
with the sort of rotating to new
infrastructure, there's a lot of

699
00:35:56,240 --> 00:35:59,280
complexity, right.
So typical organization might be

700
00:35:59,280 --> 00:36:03,280
running, you say 50 different
physics codes.

701
00:36:03,760 --> 00:36:06,080
Right.
Each of these have different

702
00:36:06,080 --> 00:36:08,360
algorithms, some of them have
different sub algorithms within

703
00:36:08,360 --> 00:36:10,320
those algorithms, right,
Different ways of running them.

704
00:36:10,360 --> 00:36:13,000
And so there's like 10s of
thousands of combinations and

705
00:36:13,000 --> 00:36:15,040
way to run this.
Then they can all be compiled

706
00:36:15,040 --> 00:36:16,640
differently.
They have different optimal

707
00:36:16,640 --> 00:36:17,640
architectures.
Right.

708
00:36:17,720 --> 00:36:21,120
And it's sort of like, and then
you run them sporadically at

709
00:36:21,120 --> 00:36:23,560
different types like so if
you're a automotive company,

710
00:36:23,560 --> 00:36:25,920
there's a big part of the design
cycle in detailed design.

711
00:36:25,920 --> 00:36:27,280
You know, this is happening a
lot.

712
00:36:27,720 --> 00:36:31,400
There's other times where you
may not be running so much and

713
00:36:31,400 --> 00:36:36,600
so it is pretty tough.
I think for the average

714
00:36:36,720 --> 00:36:42,880
enterprise like you know,
consumer of HPC to sort of say,

715
00:36:43,160 --> 00:36:45,440
you know, it makes sense to kind
of run your own HPC.

716
00:36:45,480 --> 00:36:49,680
Now if you look at the market
today, right, it's, it's about

717
00:36:49,680 --> 00:36:54,960
20% cloud and say 80% on Prem,
which which you know, I don't

718
00:36:54,960 --> 00:36:56,880
know, blows my mind still today,
right.

719
00:36:58,520 --> 00:37:00,920
I think there is a, you know,
it's a big.

720
00:37:01,280 --> 00:37:03,080
Switching costs takes a long
time.

721
00:37:03,080 --> 00:37:04,800
If you're running and operating
data centers, you're maybe

722
00:37:04,800 --> 00:37:07,280
depreciating that you know if.
It's ACFO making a decision

723
00:37:07,280 --> 00:37:11,440
maybe six years, right?
And so even if you said 100% to

724
00:37:11,440 --> 00:37:14,040
cloud tomorrow, it's it still
takes a long time, right?

725
00:37:15,640 --> 00:37:18,120
And there's a lot of
complexities that I think are

726
00:37:18,600 --> 00:37:20,520
glossed over.
So I think a big reason you

727
00:37:20,520 --> 00:37:24,880
didn't see it take off earlier
was a lot of this complexity.

728
00:37:24,880 --> 00:37:29,160
It is actually very hard to run,
say, a fluid dynamics code.

729
00:37:29,880 --> 00:37:33,880
On an HPC system, very scalably,
very reliably and continuously

730
00:37:33,880 --> 00:37:36,600
keep the codes updated.
And like when you run different

731
00:37:36,600 --> 00:37:38,480
algorithms, they scale
differently and sort of like

732
00:37:38,480 --> 00:37:39,880
what's the right cluster size
so.

733
00:37:40,400 --> 00:37:43,520
There's so much that goes into
that.

734
00:37:43,520 --> 00:37:48,120
That is, it's still, you know,
it's a specialized field right

735
00:37:48,280 --> 00:37:51,080
now, every scale we write
software to simplify all that.

736
00:37:51,080 --> 00:37:52,800
Right.
And try to sort of abstract away

737
00:37:52,800 --> 00:37:55,880
all this complexity, but it's
quite complex and if you really

738
00:37:55,880 --> 00:37:58,760
dig into it.
If you ask the people who are on

739
00:37:58,760 --> 00:38:01,120
the sort of still running on
prime systems, they would often

740
00:38:01,120 --> 00:38:05,320
tell you the biggest reason why
is, is sort of this it is hard

741
00:38:05,320 --> 00:38:07,360
and you have to, you have to be
willing to kind of reinvent

742
00:38:07,360 --> 00:38:11,040
yourself, right?
So like, just imagine if you're

743
00:38:11,080 --> 00:38:15,720
a high performance computing
administrator for like 1 of the

744
00:38:15,720 --> 00:38:17,960
top three aerospace companies in
the world, right?

745
00:38:19,120 --> 00:38:21,800
All you've done last 20 years is
basically figure out how to

746
00:38:21,800 --> 00:38:25,280
procure like the right
infrastructure, you know, 18

747
00:38:25,280 --> 00:38:28,160
months of like.
Procurement processes get these

748
00:38:28,160 --> 00:38:31,200
systems tested up and running
service, you know your different

749
00:38:31,200 --> 00:38:33,320
engineering customers, internals
to your organization.

750
00:38:33,320 --> 00:38:35,200
With cloud, it's a completely
different paradigm.

751
00:38:35,200 --> 00:38:37,680
Like that entire paradigm is
like sort of hardware up

752
00:38:37,680 --> 00:38:39,640
thinking, right?
Which is like it sort of starts

753
00:38:39,640 --> 00:38:42,320
with the processor and then you
build out the systems.

754
00:38:43,360 --> 00:38:44,880
I think the right way to think
about.

755
00:38:44,880 --> 00:38:48,720
What is the purpose of HPC is
ultimately to serve the

756
00:38:48,720 --> 00:38:52,480
workloads with good performance.
And that's more like workload

757
00:38:52,480 --> 00:38:54,840
down thinking, right?
So like it's really just about

758
00:38:54,840 --> 00:38:57,000
OK, like for this workload, what
is actually the right

759
00:38:57,040 --> 00:38:59,480
infrastructure, what's the right
price, what's the right

760
00:38:59,480 --> 00:39:00,920
performance, etcetera.
Like how are you going to

761
00:39:00,920 --> 00:39:03,960
optimize that problem?
And with an elastic system,

762
00:39:04,080 --> 00:39:06,160
you're going to solve that
problem much more elegantly,

763
00:39:06,480 --> 00:39:12,200
right?
So I would say that's a very

764
00:39:12,240 --> 00:39:18,520
long answer, but I think it's
still to this day, you know,

765
00:39:18,520 --> 00:39:21,000
there's people who believe.
Like HBC should be done on Prem

766
00:39:22,320 --> 00:39:25,480
Yeah I.
I see that my big observation is

767
00:39:25,480 --> 00:39:27,920
always just timelines that
everything, as you say, just

768
00:39:27,920 --> 00:39:31,440
takes so much longer than people
realize.

769
00:39:31,440 --> 00:39:37,640
So, you know, when I started, I
remember and this is This is why

770
00:39:37,640 --> 00:39:41,600
I'm still amazed by you having
these ideas nine years earlier

771
00:39:41,640 --> 00:39:45,120
that people were like completely
anti cloud.

772
00:39:45,240 --> 00:39:48,920
A lot of people were anti cloud.
And then you know, just as the

773
00:39:48,920 --> 00:39:53,800
time as I was leaving, I
remembered that it wasn't, are

774
00:39:53,800 --> 00:39:56,320
we going to do that?
It was like how, how we could,

775
00:39:56,440 --> 00:40:01,600
oh, no, that's not true.
It was probably hybrid, hybrid

776
00:40:01,600 --> 00:40:04,280
and we are going to do some of
it on, but we'll still keep

777
00:40:04,800 --> 00:40:06,000
some.
And that depended whether it's

778
00:40:06,000 --> 00:40:08,920
to start upon enterprise.
But I definitely saw a shift.

779
00:40:08,920 --> 00:40:15,400
But then I thought that means
the time then to do Apoc to then

780
00:40:15,400 --> 00:40:20,920
roll it out to change like so.
It's probably, I guess you'd

781
00:40:20,920 --> 00:40:24,360
agree there's probably even more
upside in cloud to come because

782
00:40:24,400 --> 00:40:27,040
it takes so long to convince
people and then for them to

783
00:40:27,040 --> 00:40:30,800
change that we may not see that
consequence for another few

784
00:40:30,800 --> 00:40:32,240
years.
Yeah, I think that's right.

785
00:40:32,240 --> 00:40:35,400
I do think that, you know,
Rescale, we're really focused on

786
00:40:35,400 --> 00:40:39,160
solving kind of like what is the
real problem the customer's

787
00:40:39,160 --> 00:40:42,120
trying to solve, right?
And so often it's again, using

788
00:40:42,120 --> 00:40:46,320
this example of a manufacturer
might be an engineer doing sort

789
00:40:46,320 --> 00:40:50,560
of fluid dynamic simulation.
They just want to run that

790
00:40:50,560 --> 00:40:53,520
simulation as efficiently as
possible and like all this other

791
00:40:53,520 --> 00:40:54,960
stuff.
And so they're like all the IT

792
00:40:54,960 --> 00:40:58,240
stakeholders trying to kind of
serve that user ideal, right?

793
00:40:58,600 --> 00:41:00,480
You kind of work from that
customer backwards.

794
00:41:01,240 --> 00:41:04,480
I think you really want to solve
this.

795
00:41:04,480 --> 00:41:07,240
Problem more from first
principles as opposed to.

796
00:41:08,080 --> 00:41:10,160
Sort of you have all this
baggage of these like on Prem

797
00:41:10,160 --> 00:41:11,760
systems, hybrid, all this stuff,
right?

798
00:41:11,760 --> 00:41:15,160
And our view is sort of like,
well, if we can give you a

799
00:41:15,160 --> 00:41:17,520
service that is the optimal
router, right?

800
00:41:17,520 --> 00:41:20,000
So they can real time say, OK,
you're going to run this fluid

801
00:41:20,000 --> 00:41:22,200
dynamics code.
You know, we understand some of

802
00:41:22,200 --> 00:41:24,880
the metadata, right?
So we say, hey, it has like, you

803
00:41:24,880 --> 00:41:27,520
know, 500,000 cells or whatever,
right?

804
00:41:27,640 --> 00:41:31,680
And say you're running Star CCM.
We know that version of Star

805
00:41:31,680 --> 00:41:35,520
CCM, how well it scales.
We we know like with that kind

806
00:41:35,520 --> 00:41:37,840
of cell model, like what's the
size cluster that should go on,

807
00:41:37,880 --> 00:41:38,800
right.
And then you go through this

808
00:41:38,800 --> 00:41:42,160
sort of what we've built as a
compute recommendation engine.

809
00:41:42,160 --> 00:41:45,320
So you actually so say, OK, you
can plug in your own on Prem

810
00:41:45,320 --> 00:41:46,760
resource.
So that's one of the fixed

811
00:41:46,760 --> 00:41:49,840
resources you could choose from.
It comes at some price layer

812
00:41:49,840 --> 00:41:51,160
internal sort of cost to
operate.

813
00:41:51,160 --> 00:41:53,320
Maybe they have your cloud
resources, right.

814
00:41:53,320 --> 00:41:55,560
And then, you know, just have
one, you have all the major

815
00:41:55,560 --> 00:41:58,480
cloud providers, you have NEO
clouds, maybe have some special

816
00:41:58,480 --> 00:42:00,640
supercomputing relationships
with universities, etcetera.

817
00:42:01,000 --> 00:42:04,520
In this whole network, hundreds
of architecture choices, right?

818
00:42:04,640 --> 00:42:06,320
All of these had a little
nuances, right?

819
00:42:06,320 --> 00:42:07,720
Oh, but that's an Intel
processor.

820
00:42:07,720 --> 00:42:11,520
So you need to use Intel MPI or
like hey here this like version

821
00:42:11,520 --> 00:42:13,160
of this code.
Like needs to be compiled

822
00:42:13,160 --> 00:42:17,480
differently whatever.
So it's a, it's a we want an

823
00:42:17,480 --> 00:42:20,040
analysis in this little data,
but how many combinations are

824
00:42:20,040 --> 00:42:22,040
there is like more than 50
million combinations.

825
00:42:23,160 --> 00:42:24,840
So it's like, how do you solve
that problem?

826
00:42:24,840 --> 00:42:28,960
Well, you basically want to go
through some sort of algorithm

827
00:42:28,960 --> 00:42:31,520
and we.
Recommendation system was kind

828
00:42:31,520 --> 00:42:34,080
of our solution to this problem
to automatically kind of figure

829
00:42:34,080 --> 00:42:36,520
figure out and route that right.
And that's now you have the

830
00:42:36,520 --> 00:42:39,480
service for this user for any
workload.

831
00:42:39,880 --> 00:42:44,240
That you know, you can kind of
use both the all the metadata

832
00:42:44,240 --> 00:42:46,480
and all the knowledge Rescale
has from running many of these

833
00:42:46,480 --> 00:42:50,840
similar workloads before, but
you can also leverage any of the

834
00:42:50,840 --> 00:42:53,240
resources you want.
In the way you as an

835
00:42:53,240 --> 00:42:55,840
administrator can say, hey,
like, I first want to fill up my

836
00:42:55,840 --> 00:42:58,480
own system or whatever, right?
And that is, I think that's the

837
00:42:58,480 --> 00:42:59,920
right way to solve this problem,
right?

838
00:43:00,000 --> 00:43:05,400
And it is, it is a pretty
technical way, right?

839
00:43:05,440 --> 00:43:08,160
I think there's also like
usability of this is, is super

840
00:43:08,160 --> 00:43:10,320
important.
So for us, how this shows up is

841
00:43:10,320 --> 00:43:13,280
literally like you open up
rescale, you drag in your like

842
00:43:13,280 --> 00:43:15,920
input file, right?
Or if you upload the input file

843
00:43:16,440 --> 00:43:18,480
and you and you select the
software, you press run, right?

844
00:43:18,480 --> 00:43:20,960
It is that simple.
But in that is this highly

845
00:43:20,960 --> 00:43:24,000
complex Configurator which can
kind of optimize and solve these

846
00:43:24,000 --> 00:43:26,040
problems for you.
And, you know, I, I think that's

847
00:43:26,040 --> 00:43:28,160
a.
Great way to solve this problem.

848
00:43:29,600 --> 00:43:31,720
But like you said, like there's
a, there's a lot of practical

849
00:43:31,720 --> 00:43:34,360
challenges, right?
Like, so that's the technical

850
00:43:34,360 --> 00:43:37,000
person and he says, OK, like
this is a great solution, right?

851
00:43:37,120 --> 00:43:39,880
There's a lot of practical
things in enterprise software.

852
00:43:39,880 --> 00:43:42,400
And like go to market and like
how do you use customers like

853
00:43:42,400 --> 00:43:46,640
transition and etcetera.
So that are still challenges for

854
00:43:46,640 --> 00:43:50,760
customers today.
But I think increasingly, like

855
00:43:50,760 --> 00:43:51,760
you said, it's easier and
easier.

856
00:43:51,760 --> 00:43:54,000
Like this POC process you
described, you got to remember,

857
00:43:54,000 --> 00:43:56,360
like if you're buying on Prem,
you're probably going through

858
00:43:56,360 --> 00:44:00,720
like a sort of 9 to 12 month
like procurement process at a

859
00:44:00,720 --> 00:44:03,440
minimum, right?
For many companies, it's closer

860
00:44:03,440 --> 00:44:04,840
to 18 months.
You go through all this

861
00:44:04,840 --> 00:44:07,840
evaluation, then your system is
finally live 18 months later.

862
00:44:07,840 --> 00:44:11,320
And that's just not like, you
know, if you're an engineer

863
00:44:11,320 --> 00:44:15,160
trying to run this.
Fluid dynamic simulation, you

864
00:44:15,160 --> 00:44:16,840
can't do that right?
Like you got to solve the

865
00:44:16,840 --> 00:44:18,120
problem in a different way,
right?

866
00:44:18,280 --> 00:44:21,800
And I do think today if you go
to rescale, right, and go

867
00:44:21,800 --> 00:44:24,960
through this process, you can do
this all in like 5 minutes,

868
00:44:25,120 --> 00:44:28,080
right?
But there are practical

869
00:44:28,080 --> 00:44:29,360
challenges.
So if you think again, think of

870
00:44:29,360 --> 00:44:32,080
a like a large enterprise
organization, there's a lot of

871
00:44:32,480 --> 00:44:34,440
new ways of thinking that's sort
of required.

872
00:44:35,560 --> 00:44:37,600
And it comes back I think to
like the courage, right?

873
00:44:37,640 --> 00:44:38,880
Like are you, are you willing
to?

874
00:44:40,080 --> 00:44:42,360
Say, hey, like we got to do it a
different way and I know I'm

875
00:44:42,360 --> 00:44:44,600
going to go through lots of
challenges, etcetera.

876
00:44:44,600 --> 00:44:49,120
But and I think increasingly a
lot of companies are seeing so

877
00:44:49,120 --> 00:44:52,360
much success that it is hard to
argue the other way, right?

878
00:44:52,360 --> 00:44:54,880
Like if I had to argue where if
you're using cloud, like if I go

879
00:44:54,880 --> 00:44:56,480
to on Prem.
Yeah.

880
00:44:57,080 --> 00:45:00,080
There's a few reasons, but for
most organizations this would

881
00:45:00,080 --> 00:45:03,440
make very little sense.
And I guess now it seems, and I

882
00:45:03,440 --> 00:45:06,640
didn't predict it, but it seems
to be happening that there's

883
00:45:06,640 --> 00:45:10,680
even more because of UCB sort of
neo clouds, which I guess has

884
00:45:10,680 --> 00:45:14,320
been really because of AI.
There seems to be even more

885
00:45:14,920 --> 00:45:17,560
competition to the big cloud
providers.

886
00:45:17,720 --> 00:45:20,960
I guess there's more options,
but even more reason to have a

887
00:45:20,960 --> 00:45:23,760
single place to root that
through.

888
00:45:23,800 --> 00:45:26,000
Because if you have an if you
have an account on cloud

889
00:45:26,000 --> 00:45:30,960
provider ABCDE like to you for
that enterprise to build out all

890
00:45:30,960 --> 00:45:33,160
themselves.
I mean it can be done.

891
00:45:33,160 --> 00:45:37,840
But it's a lot of work.
Does it make it even more

892
00:45:37,840 --> 00:45:40,960
valuable to have like a go
between it than it was when

893
00:45:40,960 --> 00:45:43,960
there was mainly just AWS as the
biggest one?

894
00:45:45,560 --> 00:45:47,320
Yeah, I think.
Now of course we can, we can

895
00:45:47,320 --> 00:45:48,880
optimize within a single cloud,
right?

896
00:45:48,880 --> 00:45:51,800
So like if a customer really
loves Microsoft or Amazon or

897
00:45:51,800 --> 00:45:54,920
Google like we can, we can
certainly do that and and you

898
00:45:54,920 --> 00:45:56,720
sort of solve the problem at a
smaller scale.

899
00:45:56,760 --> 00:46:00,120
I think there's this margins
piece, there's the speed, right?

900
00:46:00,120 --> 00:46:01,640
That's why I think.
You see the newer clouds as

901
00:46:01,640 --> 00:46:03,960
well, like they're just able to
stand up a cluster much faster,

902
00:46:03,960 --> 00:46:06,160
like it is the bigger
organization is it is hard,

903
00:46:06,160 --> 00:46:06,800
right?
And it's pretty.

904
00:46:06,880 --> 00:46:09,680
Interesting to see Microsoft
that they outsource a lot of

905
00:46:09,680 --> 00:46:13,560
their GPU compute to core Weave,
right?

906
00:46:13,560 --> 00:46:17,080
So public information, you know,
there's, there's a few dynamics

907
00:46:17,080 --> 00:46:20,920
I think why like 1 is.
Sort of just pure GPU scarcity.

908
00:46:20,960 --> 00:46:24,320
So like they just have them, not
only GPUs, they also have the

909
00:46:24,320 --> 00:46:26,400
power infrastructure, right?
Like I'm sitting in a data

910
00:46:26,400 --> 00:46:28,560
center working to power them and
cool them.

911
00:46:28,640 --> 00:46:31,560
So there's sort of that timing
effect, but there is a setting

912
00:46:31,560 --> 00:46:33,120
up.
Those systems are like HPC

913
00:46:33,120 --> 00:46:36,360
systems, right?
And that is very different than

914
00:46:36,360 --> 00:46:38,480
building public cloud resources
historically.

915
00:46:38,880 --> 00:46:42,000
Now, if you only have like one
or two customers, you can build

916
00:46:42,000 --> 00:46:44,360
these bespoke systems, right?
Like, like what cloud prevention

917
00:46:44,360 --> 00:46:47,560
to get at is serving thousands
or 10s of thousands, millions of

918
00:46:47,560 --> 00:46:50,760
customers, right?
And so there's a challenge for a

919
00:46:50,760 --> 00:46:52,360
Neocloud is how do you scale
that?

920
00:46:52,360 --> 00:46:55,120
There's also, I think an element
of, but the speed really

921
00:46:55,120 --> 00:46:56,960
matters.
Like I, I think right now in AI,

922
00:46:56,960 --> 00:46:58,280
it's just all about speed,
right?

923
00:46:58,280 --> 00:47:01,640
And so people are willing to do
absolutely crazy.

924
00:47:01,640 --> 00:47:06,600
Thanks for speed, right?
And it's because of the sort of

925
00:47:06,600 --> 00:47:09,160
ultimate end use skate like this
sort of AI war between all the

926
00:47:09,160 --> 00:47:12,600
big tech companies is fueling
all of that, right.

927
00:47:12,600 --> 00:47:16,320
And and if you're just a few
months faster, I'm training a

928
00:47:16,320 --> 00:47:17,920
large language model that can
make a big.

929
00:47:17,920 --> 00:47:20,000
Difference right in in the long
term game here.

930
00:47:20,000 --> 00:47:23,600
And so it's, it's, it's
interesting the bottlenecks,

931
00:47:23,600 --> 00:47:24,560
right?
Like if you, I don't know if

932
00:47:24,560 --> 00:47:26,520
you've been following this, but
there's like, it's all these neo

933
00:47:26,520 --> 00:47:29,680
clouds, of course.
But there's also these practical

934
00:47:29,680 --> 00:47:31,560
problems people are running
into, like there's not enough

935
00:47:31,560 --> 00:47:34,680
electricians to like, you know,
data centers, right?

936
00:47:36,440 --> 00:47:40,240
And there's like, you know,
like, like a meta doesn't have

937
00:47:40,240 --> 00:47:43,400
enough like buildings.
So they'd like operating data

938
00:47:43,400 --> 00:47:45,800
centers in tents, right?
Like, and then there's like this

939
00:47:45,800 --> 00:47:51,800
really efficient cooling systems
built by NVIDIA that like run

940
00:47:51,840 --> 00:47:55,360
super efficiently.
But then like XAI is, is is like

941
00:47:55,360 --> 00:48:00,120
literally like, like using
generators to do sort of cool

942
00:48:00,120 --> 00:48:03,680
water and pump it into the
cooling system in the Super

943
00:48:03,680 --> 00:48:04,960
inefficient way.
But they, they got their

944
00:48:04,960 --> 00:48:07,000
training model up much faster
than anybody else, right?

945
00:48:07,000 --> 00:48:10,120
And so like, yeah, like running
fast matters here.

946
00:48:10,600 --> 00:48:13,880
And that's probably where, to be
fair, the, there is a difference

947
00:48:13,880 --> 00:48:19,520
between the HPC for CAE and the,
and the HPC for AI because it's

948
00:48:19,520 --> 00:48:22,280
probably true.
If if you're a very large AI

949
00:48:22,280 --> 00:48:26,360
company where you need so much
compute, but it's basically just

950
00:48:26,360 --> 00:48:29,080
for you, you could probably
argue that it might be actually

951
00:48:29,080 --> 00:48:30,280
better for you to build it
yourself.

952
00:48:30,280 --> 00:48:33,480
If that's your core
differentiator is to get that

953
00:48:33,480 --> 00:48:35,600
training model out.
Two months later, you can kind

954
00:48:35,600 --> 00:48:38,280
of see like, why would we go and
wait on a cloud?

955
00:48:38,280 --> 00:48:40,560
We can just do it ourselves.
We, we have the same type of

956
00:48:40,560 --> 00:48:42,480
GPU.
We have 100,000 of them.

957
00:48:42,840 --> 00:48:45,240
That's it.
Then it's I can see the logic

958
00:48:45,720 --> 00:48:48,400
but for most manufacturing
companies or aerospace

959
00:48:48,400 --> 00:48:53,120
companies, they don't have
100,000 GPUs to train 1 bottle

960
00:48:53,120 --> 00:48:54,840
right?
It's a different world.

961
00:48:54,840 --> 00:48:59,960
So I guess it's different needs
as you say.

962
00:49:01,040 --> 00:49:04,320
And probably they don't have 150
different applications, legacy

963
00:49:04,320 --> 00:49:07,000
applications like CAE does.
I guess the AI world is

964
00:49:07,000 --> 00:49:09,120
something newer.
Yeah, I do think like you know,

965
00:49:09,240 --> 00:49:13,640
AI is is changing a lot and you
know the CAE engineering world

966
00:49:13,680 --> 00:49:15,480
as well, right.
And, and we could talk more

967
00:49:15,480 --> 00:49:17,720
about that, but like you spend a
lot of time talking about the

968
00:49:17,720 --> 00:49:20,680
HPC layer.
I think the way, if you sort of

969
00:49:20,680 --> 00:49:22,040
Fast forward to rescale today,
right?

970
00:49:22,040 --> 00:49:24,720
Like the way the way we look at
it today is there's the HPC

971
00:49:24,720 --> 00:49:29,040
layer or the compute layer, but
you the the the data layer,

972
00:49:29,040 --> 00:49:33,920
super important, right?
Like, so, so one of the big

973
00:49:34,280 --> 00:49:36,600
argue like, OK, why do these big
training models?

974
00:49:36,600 --> 00:49:39,440
Why, why are these then sort of
on Prem is like a natural

975
00:49:39,440 --> 00:49:41,080
question, right?
Like, but like if you're sort of

976
00:49:41,080 --> 00:49:44,760
the downloading entire web and
then like training on that,

977
00:49:44,840 --> 00:49:47,880
that's a very heavy data gravity
kind of situation, right?

978
00:49:48,400 --> 00:49:52,520
And of course you want that data
as close as possible to the

979
00:49:52,520 --> 00:49:53,960
compute.
So there's always like memory

980
00:49:53,960 --> 00:49:56,400
implications.
There's there's how these, you

981
00:49:56,400 --> 00:49:58,160
know, all the detailed
architecture really matters, but

982
00:49:58,160 --> 00:49:59,600
also it's like, how do you serve
all these things?

983
00:50:00,320 --> 00:50:03,640
So that's one, I would say class
of HPC problems.

984
00:50:04,600 --> 00:50:08,840
But data starts, you know, it's
true that some of this is only

985
00:50:08,840 --> 00:50:13,320
for these large LLM providers.
But you know, I think everybody

986
00:50:13,320 --> 00:50:15,560
like if I speak with our
customers, they're all building

987
00:50:15,560 --> 00:50:18,280
AI models themselves as well.
And these are much smaller

988
00:50:18,280 --> 00:50:20,720
scale.
But ultimately they need

989
00:50:20,720 --> 00:50:23,880
efficient infrastructure to
train their own data as well.

990
00:50:24,480 --> 00:50:26,240
Yeah.
And then obviously inference is

991
00:50:26,240 --> 00:50:28,320
a big one too, right?
And like running inference

992
00:50:28,320 --> 00:50:31,320
efficiently is going to get
really important right today,

993
00:50:32,600 --> 00:50:34,640
you know, if you if you sort of
breakdown the economics.

994
00:50:34,640 --> 00:50:37,400
So like Silicon Valley is
funding a lot of very cheap

995
00:50:37,400 --> 00:50:38,360
services, right?
Why?

996
00:50:38,360 --> 00:50:41,920
Because everybody wants to win
the AI platform war, so to

997
00:50:41,920 --> 00:50:43,840
speak, right?
And so people are willing to

998
00:50:43,840 --> 00:50:46,280
lose a lot of money today on
inference, right?

999
00:50:46,320 --> 00:50:50,760
Just to win the customer, right?
But over over time, all these

1000
00:50:50,760 --> 00:50:54,360
economics need to be figured out
and things do need to be run

1001
00:50:54,360 --> 00:50:57,920
actually efficiently, right?
And so a manufacturer, because

1002
00:50:58,480 --> 00:51:01,280
here say a Tier 1 manufacturer
and automotive, you're kind of

1003
00:51:01,280 --> 00:51:04,680
running your it's like a 510%
margin business often, right?

1004
00:51:04,760 --> 00:51:08,080
And so it's like you already
have to run really efficiently.

1005
00:51:08,080 --> 00:51:10,040
Why that's why IT focus so much
on TCO.

1006
00:51:10,040 --> 00:51:12,760
That's that's why it's like,
hey, I can run like the system

1007
00:51:12,760 --> 00:51:15,600
for like 10 years and divide it,
you know, like it all makes

1008
00:51:15,600 --> 00:51:18,280
sense to me.
But I think if you think about

1009
00:51:18,280 --> 00:51:20,400
the engineering.
Problems you want to solve if

1010
00:51:20,400 --> 00:51:24,320
you're a little bit, you know
for thinking of like OK, you

1011
00:51:24,320 --> 00:51:26,720
know how to get out of being a
10% margin business as a.

1012
00:51:27,480 --> 00:51:31,000
Tier one supplier, right?
It's I think the way out is

1013
00:51:31,000 --> 00:51:35,800
actually using the latest
technologies, which is, you

1014
00:51:35,800 --> 00:51:38,600
know, there's, there's a compute
layer, but especially your data,

1015
00:51:38,600 --> 00:51:42,040
say you're a seat manufacturer.
I know how to like do seat

1016
00:51:42,040 --> 00:51:43,880
design really well.
And I might be really good at

1017
00:51:43,880 --> 00:51:46,800
crash simulation, right?
And so and so have all this

1018
00:51:46,800 --> 00:51:49,240
crash simulation data.
If I'm like one of the top three

1019
00:51:49,240 --> 00:51:51,880
companies in seat manufacturing,
I probably have some of the best

1020
00:51:51,880 --> 00:51:54,680
crash simulation data for seats
out of anybody, right?

1021
00:51:54,840 --> 00:51:57,360
If you can build what we call
some of these like AI physics

1022
00:51:57,360 --> 00:51:59,560
models around that, you can
build a lot of the intelligence

1023
00:51:59,560 --> 00:52:01,440
into a much more compressed
timeline, right?

1024
00:52:01,440 --> 00:52:03,600
So now instead of like trying to
speed up the algorithm, you're

1025
00:52:03,600 --> 00:52:06,000
actually like changing the
algorithm and going to from

1026
00:52:06,000 --> 00:52:09,600
deterministic to probabilistic.
But then maybe I can serve my

1027
00:52:09,600 --> 00:52:12,560
OEM like way faster, like not a
little bit faster.

1028
00:52:12,680 --> 00:52:14,640
Like I can actually say, hey,
like I, I think I could do that

1029
00:52:14,640 --> 00:52:19,000
seat with like 98% confidence
and I could give you that answer

1030
00:52:19,080 --> 00:52:23,720
in maybe an hour instead of like
2 months of analysis, right?

1031
00:52:24,680 --> 00:52:28,760
And so this is, that's, I think
where the future's headed.

1032
00:52:28,760 --> 00:52:32,440
And then like you have, that's
on the simulation side.

1033
00:52:32,920 --> 00:52:36,000
Then there's also this, you
know, we're working with

1034
00:52:36,000 --> 00:52:38,240
customers to build sort of
engineering agents.

1035
00:52:38,240 --> 00:52:39,160
Right.
It's like.

1036
00:52:39,240 --> 00:52:42,520
You, you kind of get this
capability to this, you're

1037
00:52:42,520 --> 00:52:44,840
actually starting for the
simulation engineers.

1038
00:52:44,840 --> 00:52:48,680
They have this this automotive
company empowering them to just

1039
00:52:48,680 --> 00:52:52,320
do their job like way faster.
So all the mundane tasks you're

1040
00:52:52,320 --> 00:52:54,000
starting to sort of automate,
right?

1041
00:52:54,680 --> 00:52:57,440
And that's I think every
function and every industry

1042
00:52:57,440 --> 00:53:00,160
frankly is going to go through
that sort of process, right.

1043
00:53:01,160 --> 00:53:04,480
The nice thing in engineering is
there is a lot of data and

1044
00:53:04,480 --> 00:53:07,080
there's a lot of like
intelligence already built into

1045
00:53:07,080 --> 00:53:11,440
all these processes, etcetera.
And so it's rescale with our

1046
00:53:11,440 --> 00:53:12,920
customers.
We're in a unique position where

1047
00:53:12,920 --> 00:53:16,000
we, I think, have a really good
understanding of like how the

1048
00:53:16,000 --> 00:53:18,960
actual end user engineer or
scientist runs all these work

1049
00:53:18,960 --> 00:53:20,560
flows.
What are the problems they're

1050
00:53:20,560 --> 00:53:22,360
trying to solve?
It can help them do that, you

1051
00:53:22,360 --> 00:53:24,560
know, 10 times better.
Everybody wants, right?

1052
00:53:24,840 --> 00:53:28,280
And this entire compute
conversation, right, becomes

1053
00:53:28,280 --> 00:53:31,920
pretty secondary, right, Because
it's sort of like, you know,

1054
00:53:31,920 --> 00:53:34,280
like that's just it's just
electricity, right?

1055
00:53:34,280 --> 00:53:37,480
It's it's like you're just
powering the ability for an

1056
00:53:37,480 --> 00:53:39,560
engineer or scientist to come up
with a cool innovation.

1057
00:53:40,640 --> 00:53:44,520
Well, that's kind of yeah, a
good segue to talk about that a

1058
00:53:44,520 --> 00:53:46,560
little bit more.
The you're right, we basically

1059
00:53:46,560 --> 00:53:52,720
focused on HPCHPC was pre AI the
big changer, you know, you have

1060
00:53:52,720 --> 00:53:55,600
more HPC run your simulation
faster and it still is, don't

1061
00:53:55,600 --> 00:53:57,600
get me wrong.
And of course you know, GP us

1062
00:53:57,600 --> 00:54:04,160
help that, but the AI feels very
transformative for engineering.

1063
00:54:04,160 --> 00:54:09,200
So I mean, how have you seen
that convergence on the Rescale

1064
00:54:09,200 --> 00:54:11,600
platform and from customers that
you're speaking to, You know,

1065
00:54:11,600 --> 00:54:16,120
are they still just dabbling and
with their sort of more in the

1066
00:54:16,120 --> 00:54:21,400
R&D phase, how much more sort of
AI training or AI surrogates are

1067
00:54:21,400 --> 00:54:24,320
they doing next to their
traditional CAE?

1068
00:54:24,320 --> 00:54:27,080
How are you seeing that you're,
you're at the front with

1069
00:54:27,080 --> 00:54:28,920
customers?
How is that transition?

1070
00:54:29,640 --> 00:54:34,680
The yeah, it's happening really
fast, you know, are let's say

1071
00:54:34,680 --> 00:54:37,680
our most innovative and fast
adopting customers.

1072
00:54:37,680 --> 00:54:41,960
They have AI based automation.
You called agentic kind of

1073
00:54:42,440 --> 00:54:44,840
things built in.
They have they're definitely

1074
00:54:44,840 --> 00:54:47,760
doing AI surrogates.
They are really thinking ahead

1075
00:54:47,760 --> 00:54:51,280
of like what's the sort of let's
just assume we already have AI

1076
00:54:51,280 --> 00:54:52,920
surrogates, right?
Like what are the next level of

1077
00:54:52,920 --> 00:54:55,120
problems we can solve?
Because once you have AI

1078
00:54:55,120 --> 00:54:57,880
surrogates and sort of, you
know, domain I'm I'm pretty

1079
00:54:57,880 --> 00:55:00,800
familiar with is things like in
the, in the multi physics space,

1080
00:55:00,800 --> 00:55:02,280
right?
Like you're always trying to

1081
00:55:02,280 --> 00:55:04,320
kind of to reduce your
remodeling or some way to

1082
00:55:04,320 --> 00:55:08,120
simplify these highly complex
kind of large scale models?

1083
00:55:08,120 --> 00:55:10,320
And sort of.
If you're say building an

1084
00:55:10,320 --> 00:55:12,960
airplane, you have this
conceptual design chase where

1085
00:55:12,960 --> 00:55:16,440
you're like, it's smart to use
like carbon fiber as a material

1086
00:55:16,440 --> 00:55:19,320
system for this airplane where
all the implications.

1087
00:55:19,960 --> 00:55:22,800
The challenge with that is
always like we get to the

1088
00:55:22,800 --> 00:55:26,960
detailed design that that's when
all the details like it really

1089
00:55:26,960 --> 00:55:30,160
matter.
And it'll be like, you know, at

1090
00:55:30,160 --> 00:55:32,360
the at the base of like a
commercial aircraft, you might

1091
00:55:32,360 --> 00:55:34,560
have 100 to 200 plies of carbon
fiber.

1092
00:55:34,560 --> 00:55:36,560
And then you have this like bolt
that comes through there into

1093
00:55:36,560 --> 00:55:39,040
the titanium fitting.
And there's like a bunch of

1094
00:55:39,040 --> 00:55:41,360
rules on like how that bolt
supposed to interact with the

1095
00:55:41,360 --> 00:55:43,000
system and all this kind of
stuff, right?

1096
00:55:43,960 --> 00:55:48,320
And it'll be like, OK, like we
really need 200 carbon fiber.

1097
00:55:48,320 --> 00:55:51,880
Plies at that root, right?
But then it'll be like, well, we

1098
00:55:51,880 --> 00:55:55,320
can't just go from 200 carbon
fiber plies to like 10 in the

1099
00:55:55,320 --> 00:55:57,920
next panel over.
Like this needs to go very sort

1100
00:55:57,920 --> 00:56:03,560
of slowly, like reduce it by 10%
sort of every panel because you

1101
00:56:03,560 --> 00:56:06,240
can't have this disproportionate
kind of stress situation, right?

1102
00:56:06,240 --> 00:56:08,280
And then it's like, oh, well,
this panel, we need actually be

1103
00:56:08,280 --> 00:56:11,160
able to make it.
And you know, like our tool

1104
00:56:11,160 --> 00:56:13,880
forming process can only handle
XYZ.

1105
00:56:13,920 --> 00:56:20,160
So there's a lot more that
detail drives an enormous amount

1106
00:56:20,160 --> 00:56:24,040
of the actual design.
And the problem has always been

1107
00:56:24,040 --> 00:56:26,800
is like in that conceptual
design phase, how the historic

1108
00:56:26,840 --> 00:56:29,160
has been done, you get a lot of
really smart people in the room.

1109
00:56:29,400 --> 00:56:31,920
And you say, yeah, this seems
like the right decision because

1110
00:56:31,920 --> 00:56:34,160
I've seen some test data on
carbon fiber over here and I've

1111
00:56:34,160 --> 00:56:35,680
like run some experiments over
there.

1112
00:56:35,720 --> 00:56:39,280
But what you really want to do
is kind of take all that really

1113
00:56:39,280 --> 00:56:43,560
detailed complex information and
synthesize it up to help make

1114
00:56:43,600 --> 00:56:46,440
one of these kind of higher
level decisions, right?

1115
00:56:46,480 --> 00:56:49,800
And what AI circuits allow you
to do is to simplify that

1116
00:56:49,800 --> 00:56:52,880
compression process.
I do agree that it is, you know,

1117
00:56:52,880 --> 00:56:54,680
the criticism is like, hey, this
is misleading.

1118
00:56:54,680 --> 00:56:57,040
You can, you can run the
simulation 10,000 times faster,

1119
00:56:57,040 --> 00:56:58,920
etcetera.
I mean, it is, it depends on

1120
00:56:58,920 --> 00:57:00,920
your frame of reference whether
that's true or not.

1121
00:57:00,920 --> 00:57:06,080
But like what what is absolutely
true is that using neural Nets

1122
00:57:06,200 --> 00:57:10,040
to compress this highly complex
information, right and solve

1123
00:57:10,040 --> 00:57:13,080
next sort of generation problems
that you would never even

1124
00:57:13,080 --> 00:57:15,320
attempt before, right?
Are now possible, right?

1125
00:57:15,320 --> 00:57:19,400
And and you have this much more
elegant process to sort of

1126
00:57:19,400 --> 00:57:21,760
capture the intelligence of your
engineering organization.

1127
00:57:22,680 --> 00:57:25,240
So and you can look do that at
the, you know, individual

1128
00:57:25,240 --> 00:57:28,000
discipline solver level, right?
But you can also do that at

1129
00:57:28,000 --> 00:57:29,480
higher levels.
And then you can start

1130
00:57:29,480 --> 00:57:31,440
incorporating, you know, these
sort of things like

1131
00:57:31,440 --> 00:57:33,520
manufacturing constraints and
things like that.

1132
00:57:33,520 --> 00:57:36,880
And if you can, you know, the,
the faster you can do that loop

1133
00:57:36,880 --> 00:57:40,000
basically, right, like the
better design you will get.

1134
00:57:40,120 --> 00:57:42,040
Like I think what people are
usually surprised by is that

1135
00:57:42,040 --> 00:57:46,120
that loop, there's exceptions to
this, but that loop takes a long

1136
00:57:46,120 --> 00:57:48,960
time.
Like in in typical aerospace

1137
00:57:48,960 --> 00:57:53,120
company, both Airbus and Bali of
this sort of like loop of just

1138
00:57:53,120 --> 00:57:55,720
like all the different
engineering disciplines to sort

1139
00:57:55,720 --> 00:57:58,920
of say, OK, to have this shape
wing based on this ship, we're

1140
00:57:58,920 --> 00:58:01,240
going to have these like loads
in the system based on these

1141
00:58:01,240 --> 00:58:03,480
loads, like this sort of
mechanical team is going to like

1142
00:58:03,480 --> 00:58:06,440
decide what kind of like
structure we can do because of

1143
00:58:06,440 --> 00:58:08,320
the structure, it's going to
bend a certain way, which then

1144
00:58:08,320 --> 00:58:10,680
comes back to like, OK, that's
the shape of the wing at cruise,

1145
00:58:10,680 --> 00:58:12,200
right?
And so you have this big loop

1146
00:58:12,800 --> 00:58:15,400
and that loop takes like three
to four months.

1147
00:58:15,400 --> 00:58:17,240
It could take like an hour,
right?

1148
00:58:17,240 --> 00:58:19,920
Like if you sort of said, hey,
I'm not constrained by compute.

1149
00:58:21,120 --> 00:58:23,000
You're just running a whole
bunch of software, right?

1150
00:58:23,680 --> 00:58:25,560
You still need some really smart
people to make some smart

1151
00:58:25,560 --> 00:58:27,040
decisions, right?
Like it doesn't totally

1152
00:58:27,040 --> 00:58:28,880
eliminate the engineer.
No.

1153
00:58:28,880 --> 00:58:32,920
But, but I do think that that's
where the industry is going.

1154
00:58:33,200 --> 00:58:35,920
Our customers like the the
bleeding edge customers, they

1155
00:58:35,920 --> 00:58:39,840
are implementing all this stuff
right now and it's all possible

1156
00:58:39,840 --> 00:58:41,480
today.
Like the exciting thing now is

1157
00:58:41,480 --> 00:58:43,760
like, hey, it's right there.
You just have to do it.

1158
00:58:44,400 --> 00:58:47,440
That's exactly that's the way I
normally describe it to people

1159
00:58:47,440 --> 00:58:50,400
is that, you know, 10 years ago
there used to be this dream of

1160
00:58:50,600 --> 00:58:55,240
real time CAE that, you know,
HPC would become so fast that

1161
00:58:55,240 --> 00:58:57,120
you could do the simulation in
real time.

1162
00:58:57,200 --> 00:59:00,160
But it actually is impossible.
You know, there's various

1163
00:59:00,160 --> 00:59:03,440
reasons why a traditional
structures or fluid solver will

1164
00:59:03,440 --> 00:59:08,200
never really become real time.
And that's where it's, I feel

1165
00:59:08,200 --> 00:59:11,880
the AI surrogates, because by
definition they can be real

1166
00:59:11,880 --> 00:59:15,400
time, does allow for the more
agentic AI.

1167
00:59:15,400 --> 00:59:18,720
Because how can you really have
an agentic AI system when one of

1168
00:59:18,720 --> 00:59:21,320
the agents is a solver that
takes 12 hours to run?

1169
00:59:21,560 --> 00:59:26,480
Like, yeah, you can do that, but
it, it doesn't feel like the end

1170
00:59:26,480 --> 00:59:29,440
goal.
The end goal should be like

1171
00:59:29,440 --> 00:59:31,800
Google Gemini just because I
happen to use that one.

1172
00:59:32,040 --> 00:59:37,080
There's others where like I want
to ask it to do something for

1173
00:59:37,080 --> 00:59:39,200
me.
I don't want to wait 12 hours to

1174
00:59:39,200 --> 00:59:42,120
get the answer back.
It feels like you want it to

1175
00:59:42,120 --> 00:59:45,720
come in a short time and I feel
like the AI surrogates plug into

1176
00:59:45,720 --> 00:59:49,600
that AI agent theme more easily
if you know what I mean.

1177
00:59:49,600 --> 00:59:53,000
I don't know if you agree.
Yeah, it's a good.

1178
00:59:53,000 --> 00:59:57,440
It's an interesting way for any
and I guess the way I look at it

1179
00:59:57,440 --> 01:00:00,560
is you have kind of, you know,
just like you have deep

1180
01:00:00,560 --> 01:00:02,960
research, it takes a lot longer
and then you have also like

1181
01:00:02,960 --> 01:00:04,840
quick responses.
It's just like that, right?

1182
01:00:04,840 --> 01:00:09,240
Like where it's like, hey, if I
am a designer and I just want to

1183
01:00:09,240 --> 01:00:14,600
change the shape of a mirror,
right, I can get an instant

1184
01:00:14,600 --> 01:00:17,760
arrow response that's like 99%
accurate.

1185
01:00:18,400 --> 01:00:21,400
That is awesome, right?
Like before, my alternative was

1186
01:00:21,480 --> 01:00:24,720
I changed the design, sent the
CAD model over to the Arrow

1187
01:00:24,720 --> 01:00:27,120
team.
Three days later they send me

1188
01:00:27,120 --> 01:00:29,720
back.
Hey, that, that create a lot of

1189
01:00:29,720 --> 01:00:34,480
drag, right?
So now you can get like a like a

1190
01:00:34,480 --> 01:00:37,760
pretty good answer and and
pretty good.

1191
01:00:37,760 --> 01:00:41,520
Like I've seen cases where this
is like over 99.9% accurate and

1192
01:00:41,520 --> 01:00:43,800
like, of course that depends on
like the data you train on and

1193
01:00:43,800 --> 01:00:46,600
all these kind of things.
But ultimately you can get these

1194
01:00:46,600 --> 01:00:50,600
responses back that dramatically
reduce the cycle.

1195
01:00:50,600 --> 01:00:53,600
So you just eliminated 3 days by
providing like a rough estimate

1196
01:00:53,600 --> 01:00:56,480
that's almost real time.
I still think that the detailed

1197
01:00:56,480 --> 01:00:59,400
simulations do really matter.
Like that's your training data,

1198
01:00:59,680 --> 01:01:01,080
right?
Like so like, and there's a

1199
01:01:01,160 --> 01:01:04,000
couple of parts there, right?
Like I think there's the, I call

1200
01:01:04,000 --> 01:01:06,760
it the SIM to real gap, which is
a separate discussion.

1201
01:01:06,760 --> 01:01:09,840
But that's, it's another, I
think, interesting element,

1202
01:01:09,840 --> 01:01:12,920
which is like, OK, are these
simulations that we're running,

1203
01:01:13,400 --> 01:01:15,200
how accurate are they?
Because the real ground tooth

1204
01:01:15,200 --> 01:01:17,480
data, right?
It's not even wind tunnel

1205
01:01:17,480 --> 01:01:19,360
testing.
It's actually like the real

1206
01:01:19,360 --> 01:01:20,960
data, right?
And the wind tunnel is a proxy

1207
01:01:20,960 --> 01:01:22,680
for that.
And then your simulation tool is

1208
01:01:22,680 --> 01:01:25,520
a proxy for the wind tunnel.
And yes, all the physics

1209
01:01:25,520 --> 01:01:27,600
equations are correct, but
there's like a lot of little

1210
01:01:27,600 --> 01:01:30,240
things that still can, can
change, like real world

1211
01:01:30,240 --> 01:01:32,920
outcomes.
And so, but from the pure

1212
01:01:32,920 --> 01:01:37,000
simulation, you now have the
ability to generate and ground

1213
01:01:37,000 --> 01:01:38,920
all this data like very
efficiently with these AI

1214
01:01:38,920 --> 01:01:42,200
models, right?
And so like, yes, that designer

1215
01:01:42,200 --> 01:01:44,720
doesn't want to wait for like an
like, yeah, you can give them an

1216
01:01:44,720 --> 01:01:47,560
agentic capability spin off a
simulation, right?

1217
01:01:47,920 --> 01:01:49,720
That's not that useful to them,
right?

1218
01:01:50,960 --> 01:01:55,680
Where agentic to me means, and
you know, it's a kind of a

1219
01:01:55,680 --> 01:01:57,120
buzzword.
So everything's agentic these

1220
01:01:57,120 --> 01:01:59,960
days.
But to me it's more of this kind

1221
01:01:59,960 --> 01:02:01,600
of like like proactive thing,
right?

1222
01:02:01,600 --> 01:02:04,040
So where it's like, OK, here,
here's what happened.

1223
01:02:04,880 --> 01:02:08,520
And it'll sort of describe a
customer scenario, automotive

1224
01:02:08,600 --> 01:02:11,400
OEM, right?
And they have a supplier,

1225
01:02:11,760 --> 01:02:16,240
supplier say it's the seat
manufacturer, they changed the

1226
01:02:16,240 --> 01:02:19,320
design of the seat because some
reason for same

1227
01:02:19,320 --> 01:02:20,760
manufacturability on their end,
right?

1228
01:02:21,160 --> 01:02:25,160
That design change is is sort of
automatically synced this OEM,

1229
01:02:26,200 --> 01:02:29,600
this is happening because this
seat manufacturer is in Europe

1230
01:02:29,600 --> 01:02:32,560
and the OEM is in the US.
You know, this is happening at

1231
01:02:32,560 --> 01:02:35,320
like 2:00 AM right now because
that change happened.

1232
01:02:36,280 --> 01:02:38,120
You need to rerun your crash
analysis.

1233
01:02:38,120 --> 01:02:40,440
So you could do an instant
response of like sort of an AI

1234
01:02:40,440 --> 01:02:43,760
surrogate of like, hey, is this
going to be a is this look great

1235
01:02:43,800 --> 01:02:45,720
or is this bad?
And it's like this might be a

1236
01:02:45,720 --> 01:02:48,000
problem.
OK, now because of that and the

1237
01:02:48,000 --> 01:02:51,720
simulation engineer sleeping
because of that, a spin up like

1238
01:02:52,240 --> 01:02:55,520
ALS dyna job to go run the crash
analysis and I'm just going to

1239
01:02:55,520 --> 01:02:59,200
run it for just this component,
the subcomponents, you know, in

1240
01:02:59,200 --> 01:03:00,960
a way that's like pretty
efficient based on a bunch of

1241
01:03:00,960 --> 01:03:02,400
rules.
I gave that agent right, which

1242
01:03:02,400 --> 01:03:05,640
said, hey, you can't spend more
than $1000 and if it design

1243
01:03:05,640 --> 01:03:07,720
change like this comes in,
always use the AI surrogate

1244
01:03:07,720 --> 01:03:11,120
first.
And now I, I want to, you know,

1245
01:03:11,120 --> 01:03:13,800
like failed that test, though,
right then then run this

1246
01:03:13,800 --> 01:03:15,240
simulation.
So when I come in in the

1247
01:03:15,240 --> 01:03:19,200
morning, right open my laptop at
7 AMI got right there.

1248
01:03:19,280 --> 01:03:21,640
I have the simulation results
right to go review right?

1249
01:03:21,680 --> 01:03:24,080
And it's like it's.
Pre done the proce processing,

1250
01:03:24,120 --> 01:03:26,240
this whole like trace of all
these things that happened is

1251
01:03:26,240 --> 01:03:27,800
presented to me in like a simple
way.

1252
01:03:27,800 --> 01:03:32,920
And you know, my job just got so
much faster and more efficient,

1253
01:03:32,920 --> 01:03:36,560
right, Like, like the real way
this happened before and and

1254
01:03:36,560 --> 01:03:39,320
most organizations work this way
is like, oh, this this CAD file

1255
01:03:39,320 --> 01:03:41,840
got zipped up, right, and then
it's like sat in somebody's

1256
01:03:41,840 --> 01:03:43,600
outbox, but they were in Europe,
so they were actually on

1257
01:03:43,600 --> 01:03:45,040
vacation for another week,
right.

1258
01:03:45,400 --> 01:03:48,200
And so like it didn't actually
get make it over to the OEM

1259
01:03:48,200 --> 01:03:53,080
because like whatever random
human delay, right, then the zip

1260
01:03:53,080 --> 01:03:54,840
file comes over and then
somebody has to do this

1261
01:03:54,840 --> 01:03:57,760
analysis, right?
And then, and then all these

1262
01:03:57,760 --> 01:04:00,880
sort of steps happen, right?
And I think if you can sort of

1263
01:04:00,880 --> 01:04:04,040
shorten that entire process now,
it's not perfect, right?

1264
01:04:04,040 --> 01:04:07,080
Like if you look at the tools we
have today, there's a lot of

1265
01:04:07,080 --> 01:04:09,720
expertise to make all these
judgements and do all these

1266
01:04:09,720 --> 01:04:11,880
things.
But you can already see today

1267
01:04:11,880 --> 01:04:14,080
that like our customers are
already doing this.

1268
01:04:14,720 --> 01:04:17,040
They're automating more and more
of this process, right?

1269
01:04:18,160 --> 01:04:19,560
And you're just going to get
comfortable with it.

1270
01:04:19,560 --> 01:04:22,200
It's just like you get
comfortable with using a Gemini

1271
01:04:22,200 --> 01:04:25,400
or ChatGPT or whatever, right?
And that's the sort of agentic

1272
01:04:25,400 --> 01:04:28,080
thing is sort of like.
Proactively doing all the kind

1273
01:04:28,080 --> 01:04:29,760
of like.
Grunt work that you don't want

1274
01:04:29,760 --> 01:04:34,040
to do right?
Yeah, but I, I seriously, I

1275
01:04:34,040 --> 01:04:36,280
don't know if I'm just, you
know, drinking the Kool-aid or

1276
01:04:36,280 --> 01:04:41,080
whatever the phrase is, but I, I
really can't see how this will

1277
01:04:41,080 --> 01:04:45,920
just totally change because.
Well, the first thing to say

1278
01:04:46,480 --> 01:04:50,200
maybe what, what I was
mentioning before is even if I

1279
01:04:50,200 --> 01:04:53,760
believe that you, the AI
surrogates, make it good enough,

1280
01:04:54,240 --> 01:04:58,920
you're still going to have a
whole load of huge HPC resources

1281
01:04:58,920 --> 01:05:01,720
to create training data.
You know, so even if the

1282
01:05:01,720 --> 01:05:04,240
surrogate models become good
enough, and unless there's some

1283
01:05:04,240 --> 01:05:06,720
breakthrough that I'm not aware
of, you still need all the

1284
01:05:06,720 --> 01:05:08,680
traditional stuff to, to
generate the data.

1285
01:05:08,680 --> 01:05:13,440
So actually it's, it's not a,
it's a replacement in that the

1286
01:05:13,440 --> 01:05:17,320
end user may primarily use it,
but in the back end, somebody's

1287
01:05:17,320 --> 01:05:19,520
still having to generate this
data to go and train the models

1288
01:05:19,520 --> 01:05:21,760
so that you're not truly
replacing it.

1289
01:05:21,760 --> 01:05:23,520
It may just be in the
background, if you know what I

1290
01:05:23,520 --> 01:05:28,760
mean, but I really feel that
the, the AI engineer thing seems

1291
01:05:28,760 --> 01:05:33,000
more and more real because what
you just described to me is

1292
01:05:33,000 --> 01:05:36,760
eventually through enough
training or awareness of

1293
01:05:36,760 --> 01:05:40,280
different agents, isn't it just
that I become like a manager?

1294
01:05:41,000 --> 01:05:45,240
So I have I and I have a team of
people who are not real people

1295
01:05:45,640 --> 01:05:49,440
who are doing this analysis for
me and I'm just looking at it at

1296
01:05:49,440 --> 01:05:50,960
the end.
Yes, and I think it's awesome,

1297
01:05:50,960 --> 01:05:53,480
right?
Like I think it's so so the

1298
01:05:53,480 --> 01:05:55,640
other part's like this is
already happening today.

1299
01:05:56,680 --> 01:05:59,360
So hopefully shortly we'll be
able to share some like maybe

1300
01:05:59,360 --> 01:06:01,880
public customer case studies and
exactly how this works and have

1301
01:06:01,880 --> 01:06:05,520
the customers talk about it.
But I think, you know, there's

1302
01:06:05,520 --> 01:06:07,600
no part what I just described,
right?

1303
01:06:07,600 --> 01:06:10,800
Like all all the components that
we have for any company to go do

1304
01:06:10,800 --> 01:06:13,760
that, you don't need a magic AI
surrogate model, right?

1305
01:06:13,800 --> 01:06:20,000
Like it's just like, but I think
the right framing is more AI

1306
01:06:20,000 --> 01:06:23,760
surrogates is 1 module of AI.
There's many AI modules in the

1307
01:06:23,760 --> 01:06:26,120
entire product development
engineering process, right?

1308
01:06:26,960 --> 01:06:29,600
Some of these are more
leveraging LLMS, right?

1309
01:06:29,600 --> 01:06:32,000
Like some of them are more like
this compute recommendation

1310
01:06:32,000 --> 01:06:33,440
engine.
That's like a totally different

1311
01:06:33,440 --> 01:06:35,200
category of AI.
But it helps solve this like

1312
01:06:35,200 --> 01:06:36,720
search problem much more
efficiently.

1313
01:06:37,760 --> 01:06:40,640
You have AI surrogates who do
these sort of probabilistic

1314
01:06:40,640 --> 01:06:42,440
physics analysis much more
efficiently.

1315
01:06:42,440 --> 01:06:47,400
The, the really nice part is the
productization of these because

1316
01:06:47,400 --> 01:06:50,160
like if I tried to steal man,
sort of the cynics and I've I've

1317
01:06:50,320 --> 01:06:53,400
heard many of them in sort of
the CAE space, right where it's

1318
01:06:53,400 --> 01:06:55,840
like, but those are not real
like engineering calculations.

1319
01:06:55,840 --> 01:06:58,160
These are just like
approximations, etcetera.

1320
01:06:58,160 --> 01:06:59,480
Approximations been around
forever.

1321
01:06:59,480 --> 01:07:01,560
We've always done surrogate.
Modeling right like the.

1322
01:07:03,240 --> 01:07:05,480
I think what's really important
is like, solve the real business

1323
01:07:05,480 --> 01:07:07,440
problem, right?
Like like solve actually what,

1324
01:07:07,560 --> 01:07:09,840
why are people even doing this
kind of work, right?

1325
01:07:09,840 --> 01:07:11,480
Like, well, they're just trying
to figure out the sort of

1326
01:07:11,480 --> 01:07:14,160
physics answer for something.
And almost all simulations that

1327
01:07:14,160 --> 01:07:18,760
are run are essentially a waste
because in the end, there's only

1328
01:07:18,760 --> 01:07:20,600
like one or two that really
matter that go into

1329
01:07:20,600 --> 01:07:25,360
certification and like the Arrow
model for say a Boeing airplane,

1330
01:07:25,400 --> 01:07:28,000
right?
Like they're sort of if you want

1331
01:07:28,000 --> 01:07:30,080
the entire way of how did you
get there, right?

1332
01:07:30,080 --> 01:07:31,960
Like, yes, then all those other
data matters, right?

1333
01:07:31,960 --> 01:07:35,520
Like all the other experiments
that were run, but more than

1334
01:07:35,760 --> 01:07:38,880
like 99% of all those
computations that don't impact

1335
01:07:38,880 --> 01:07:40,480
this, like final design is
great, right?

1336
01:07:41,440 --> 01:07:42,880
And so the question is just
like, how can you go through

1337
01:07:42,880 --> 01:07:45,280
that search process as quickly
and as efficiently as possible?

1338
01:07:45,440 --> 01:07:47,960
And what changes everything is,
I think it's if you can solve

1339
01:07:47,960 --> 01:07:51,960
that search problem so much
faster, it changes how you do

1340
01:07:51,960 --> 01:07:55,000
engineering completely, right?
And not in a negative way,

1341
01:07:55,000 --> 01:07:57,160
right?
Like to me because, because

1342
01:07:57,160 --> 01:07:59,600
there's also like the other
criticism on this is like, OK,

1343
01:07:59,600 --> 01:08:01,200
well then we just replace all
these engineers.

1344
01:08:01,240 --> 01:08:02,960
Like I'm an engineer, like why
are you replacing me?

1345
01:08:02,960 --> 01:08:07,640
Right.
But this is all like, what is

1346
01:08:07,640 --> 01:08:11,480
your frame on technology and
like how the world should kind

1347
01:08:11,480 --> 01:08:16,680
of work, etcetera, right?
But like it's I think right it

1348
01:08:16,680 --> 01:08:21,560
it's super important to kind of
like jump on these waves as like

1349
01:08:21,560 --> 01:08:23,640
this is awesome.
Why is it awesome?

1350
01:08:23,640 --> 01:08:27,640
Well, what we just have to grind
through a whole bunch of CAT and

1351
01:08:27,640 --> 01:08:31,160
CAA modeling with like a team of
12 people for like 6 months to

1352
01:08:31,160 --> 01:08:34,760
get to this answer, not to get
the exact same answer with two

1353
01:08:34,760 --> 01:08:38,920
people in a week, right?
That is a win for everybody.

1354
01:08:38,960 --> 01:08:42,279
Those ten other people, right,
that you may not need any more

1355
01:08:42,279 --> 01:08:45,680
can go do five other projects.
Those ten other people can do

1356
01:08:46,240 --> 01:08:48,600
more meaningful things, learn
other things, right?

1357
01:08:49,160 --> 01:08:52,479
If you look at the history of
like CAD modelling, right?

1358
01:08:52,640 --> 01:08:55,000
Like it used to be, we have
these drafters I'm sure you've

1359
01:08:55,000 --> 01:08:59,000
seen these pictures of like, you
know, yeah, 100 people just

1360
01:08:59,319 --> 01:09:02,560
drafting the physical documents,
right, To be able to build

1361
01:09:02,560 --> 01:09:06,040
components.
I don't think it's a negative

1362
01:09:06,040 --> 01:09:07,960
thing that people don't need to
do this anymore.

1363
01:09:08,000 --> 01:09:11,520
Like you press a button in a 3D
CAD product and spits out like

1364
01:09:11,520 --> 01:09:13,920
so.
It's a productivity argument,

1365
01:09:13,920 --> 01:09:16,520
isn't it?
You know, yeah, you can do more

1366
01:09:16,520 --> 01:09:18,359
efficient.
I mean, there's a debate on

1367
01:09:18,359 --> 01:09:20,359
jobs.
I guess it's maybe the correct

1368
01:09:20,359 --> 01:09:22,200
thing to say is it's not the
job.

1369
01:09:22,439 --> 01:09:25,279
There will be people's jobs that
they don't need to do anymore,

1370
01:09:25,279 --> 01:09:28,160
but you'd hope they could trans
transition to a different job.

1371
01:09:28,279 --> 01:09:30,399
I guess is.
Yes, and like it's.

1372
01:09:30,800 --> 01:09:35,319
You know, the USI think the job
retraining stat is something

1373
01:09:35,319 --> 01:09:38,640
like 20 to 30% do rescaling
every year.

1374
01:09:38,640 --> 01:09:41,520
Yeah, yeah.
So yes, unemployment stays at

1375
01:09:41,520 --> 01:09:44,680
like X level, right.
But like in the US statistic.

1376
01:09:44,680 --> 01:09:46,080
But I imagine it's similar
globally.

1377
01:09:48,439 --> 01:09:53,359
And I don't personally find it
very satisfying if you're like

1378
01:09:53,359 --> 01:09:57,440
if your job was like, say, just
drafting documents, right?

1379
01:09:57,440 --> 01:09:59,760
But actually, like, you can just
automate it with software.

1380
01:10:01,040 --> 01:10:02,240
It's right.
Not a very fun job.

1381
01:10:02,440 --> 01:10:04,600
Right.
Like no, no, no, hadn't.

1382
01:10:04,720 --> 01:10:08,240
There's just certain stuff.
I think that machines or AI now

1383
01:10:08,240 --> 01:10:10,480
can do better.
You know, if I have to manually

1384
01:10:10,880 --> 01:10:15,920
go through 50 simulations and
try and write a report, why this

1385
01:10:15,920 --> 01:10:20,400
one moved the vortex here and AI
could do exactly the same task

1386
01:10:20,760 --> 01:10:23,920
in less than a minute.
And then you can still write the

1387
01:10:23,920 --> 01:10:25,400
report and analyze and think
about it.

1388
01:10:25,400 --> 01:10:27,920
But you didn't have to do some
of that manual stuff that

1389
01:10:27,920 --> 01:10:31,280
probably took you half a day.
To do exactly so we've we've

1390
01:10:31,280 --> 01:10:32,360
actually done this with our
customers.

1391
01:10:32,360 --> 01:10:34,280
We survey them, right.
Let's say what part about your

1392
01:10:34,280 --> 01:10:37,040
job do you dislike do or where
you don't feel like you're

1393
01:10:37,040 --> 01:10:38,560
adding value?
One of the top things that come

1394
01:10:38,560 --> 01:10:40,560
up is exactly what you just
described is write all these

1395
01:10:40,560 --> 01:10:43,760
reports.
Guess what, it's really easy to

1396
01:10:43,760 --> 01:10:46,240
take a bunch of like simulation
files, right?

1397
01:10:46,240 --> 01:10:48,560
You've decided, hey, this is the
right sort of answer and here's

1398
01:10:48,560 --> 01:10:49,960
sort of the high level reasons
why.

1399
01:10:51,040 --> 01:10:54,560
And you can just generate a
report using AI tool that works

1400
01:10:54,560 --> 01:10:56,680
today.
There's no reason not to do

1401
01:10:56,680 --> 01:10:57,640
that.
Now, do you need to review the

1402
01:10:57,640 --> 01:11:00,360
report?
Yes, right, but like you can

1403
01:11:00,360 --> 01:11:05,160
shortcut a lot of the process,
which actually allows the person

1404
01:11:05,240 --> 01:11:08,040
or the human to add, you know,
spend much more time on those

1405
01:11:08,040 --> 01:11:10,680
like value added tasks, right.
And and I kind of think of this

1406
01:11:10,680 --> 01:11:15,000
always as like moving up a layer
of abstraction where now you can

1407
01:11:15,000 --> 01:11:16,200
do so much more.
I think what people

1408
01:11:16,200 --> 01:11:19,120
underestimate is like when you
change something, right, like

1409
01:11:19,120 --> 01:11:23,400
say the six month process down
to a week, that changes

1410
01:11:23,440 --> 01:11:25,360
everything, right?
It changes like industry

1411
01:11:25,360 --> 01:11:27,640
dynamics because like you can
now do product development so

1412
01:11:27,640 --> 01:11:30,280
much faster.
It changes like what kind of

1413
01:11:30,320 --> 01:11:32,120
innovations you can actually
build, right?

1414
01:11:32,840 --> 01:11:34,720
It changes like the org
structures of companies,

1415
01:11:34,720 --> 01:11:37,800
obviously, but I think that's
all, you know, like value added

1416
01:11:37,800 --> 01:11:41,000
to society, right?
Like I think it's great, yeah.

1417
01:11:41,000 --> 01:11:43,560
And.
I guess full circle.

1418
01:11:43,560 --> 01:11:47,840
The underpinning of all of this
is access to compute, basically.

1419
01:11:47,840 --> 01:11:50,920
That's right, isn't it?
I do think, I mean, I, I will

1420
01:11:50,920 --> 01:11:53,960
say like, you know, pre AI wave,
right?

1421
01:11:53,960 --> 01:11:56,960
And, and, and I think, yeah,
that's when you were probably at

1422
01:11:57,080 --> 01:12:01,400
AWS, you know, compute was
really considered like very much

1423
01:12:01,400 --> 01:12:03,680
a commodity that's like, yeah,
like there's all these like

1424
01:12:03,680 --> 01:12:06,560
complexities, even HPC, but like
in the end, it's just like sort

1425
01:12:06,560 --> 01:12:09,760
of a means to the end, etcetera.
I think AI has shown, which is

1426
01:12:09,760 --> 01:12:12,400
quite exciting how important
compute still is, right?

1427
01:12:13,360 --> 01:12:15,400
And they're sort of the, the
high level thing of like, hey,

1428
01:12:15,400 --> 01:12:17,760
well, you need this massive
scale compute to solve this,

1429
01:12:17,760 --> 01:12:19,720
specifically this sort of LLM
trading model.

1430
01:12:19,720 --> 01:12:21,680
But I think that analogy applies
to many other things.

1431
01:12:21,680 --> 01:12:24,120
So if you take a typical
engineering problem at a large

1432
01:12:24,120 --> 01:12:27,280
company, a complex problem like
like like high end aerospace

1433
01:12:27,280 --> 01:12:31,000
product, these people doing this
work are highly compute bound,

1434
01:12:32,400 --> 01:12:34,640
right?
And they're compute bound at

1435
01:12:34,640 --> 01:12:38,080
the, you know, I can only run so
many simulations, but they're

1436
01:12:38,080 --> 01:12:40,160
also compute bound at like as
you move up these layers of

1437
01:12:40,160 --> 01:12:42,280
abstraction, right?
As I sort of like this agentic

1438
01:12:42,280 --> 01:12:46,520
engineering, it becomes easier
and easier to actually like sort

1439
01:12:46,520 --> 01:12:49,120
of kick off a simulation, right?
Add more training data, all

1440
01:12:49,120 --> 01:12:52,320
these kind of things, right?
That's where you need to be

1441
01:12:52,320 --> 01:12:54,600
really smart about like, well,
what are all the, like design

1442
01:12:54,600 --> 01:12:56,680
points you want to run?
Again, AI can help you do all

1443
01:12:56,680 --> 01:12:59,800
these things, right?
But the compute bill sort of

1444
01:12:59,800 --> 01:13:04,120
going up and up and up.
However, in most organizations,

1445
01:13:04,120 --> 01:13:07,080
if you look at like, what is
your R&D spent, the biggest cost

1446
01:13:07,080 --> 01:13:08,600
is the people.
Yeah.

1447
01:13:08,840 --> 01:13:11,800
So this opportunity to get much
more leverage out of an

1448
01:13:11,800 --> 01:13:14,160
individual, right?
Like they can just add much more

1449
01:13:14,360 --> 01:13:17,080
value in this cycle is is
massive.

1450
01:13:17,320 --> 01:13:19,040
I think it is underpinned by
compute.

1451
01:13:20,280 --> 01:13:22,320
And if you don't have that
foundational building block,

1452
01:13:22,320 --> 01:13:24,400
right, like you're, you're
highly constrained.

1453
01:13:24,400 --> 01:13:26,160
The good thing is engineers can
always work around those

1454
01:13:26,160 --> 01:13:27,560
constraints.
You know, they just make do with

1455
01:13:27,560 --> 01:13:30,800
whatever you're given.
But you know, like those

1456
01:13:30,800 --> 01:13:33,560
constraints, like really
constrained also innovation,

1457
01:13:33,560 --> 01:13:35,200
right?
They constrain the ability, like

1458
01:13:35,200 --> 01:13:38,320
if you can make it, SpaceX is a
good example, right?

1459
01:13:38,320 --> 01:13:40,480
Like it like sort of back when I
was working at Boeing, it was

1460
01:13:40,480 --> 01:13:43,360
like, hey, there's a SpaceX
company and, you know, Boeing

1461
01:13:43,360 --> 01:13:45,440
thought they were pretty good at
like launching rockets and

1462
01:13:45,440 --> 01:13:49,560
putting satellites in orbit.
And you know, at SpaceX there

1463
01:13:49,560 --> 01:13:52,480
will take the cost.
I think of like putting a kilo

1464
01:13:52,480 --> 01:13:57,240
in space, like down by like 20
to 50 X depending on like kind

1465
01:13:57,240 --> 01:14:00,160
of what metrics you use, but a
minimum 20X, right?

1466
01:14:01,040 --> 01:14:03,680
Once you do that, right?
Like it opens up this entire

1467
01:14:03,680 --> 01:14:05,680
market, right?
Like it's like, OK, now you can

1468
01:14:05,680 --> 01:14:08,720
do Starlink, now you can do all
kinds of other stuff in space,

1469
01:14:08,880 --> 01:14:11,680
right?
And you want those expansion

1470
01:14:11,680 --> 01:14:13,280
opportunities.
I do feel like a lot of

1471
01:14:13,280 --> 01:14:17,840
industries are pretty stuck,
like the way Boeing builds an

1472
01:14:17,840 --> 01:14:20,000
airplane or Airbus for that
matter, right?

1473
01:14:20,000 --> 01:14:21,920
Is pretty stuck in sort of these
old ways, right?

1474
01:14:21,920 --> 01:14:24,560
And I think of the way you can
get out of that is using all

1475
01:14:24,560 --> 01:14:27,000
these new tools where it's sort
of like something is at least

1476
01:14:27,000 --> 01:14:30,400
10X better.
It totally changes not only like

1477
01:14:30,400 --> 01:14:32,360
how fast you can do the product
development, but also like the

1478
01:14:32,360 --> 01:14:34,960
markets you can serve.
It changes the way you work with

1479
01:14:34,960 --> 01:14:36,400
your customers.
If you're seeing an aircraft

1480
01:14:36,400 --> 01:14:39,360
manufacturer right, and that
pushes society for it, I think

1481
01:14:39,360 --> 01:14:41,720
it's great.
I, I don't know about you, but

1482
01:14:41,720 --> 01:14:44,400
the older I get and the more I
work with enterprise and the

1483
01:14:44,400 --> 01:14:47,720
more I work with startups I
have, and it's probably not

1484
01:14:47,920 --> 01:14:52,760
practically possible, but I
often think, OK, if company A is

1485
01:14:52,760 --> 01:14:56,840
a legacy enterprise company and
they're trying to innovate and

1486
01:14:56,840 --> 01:15:00,080
come up with some new product.
I really do think their best

1487
01:15:00,080 --> 01:15:04,360
suggestion is they should go and
spin off a startup, right?

1488
01:15:04,480 --> 01:15:07,680
Because I sometimes feel as if
they cannot change into they

1489
01:15:07,680 --> 01:15:11,360
should literally go and take 50
super smart people, fund it and

1490
01:15:11,360 --> 01:15:14,680
let them do whatever the hell
they want and then integrate it

1491
01:15:14,680 --> 01:15:16,280
back.
Do you know what I mean?

1492
01:15:16,280 --> 01:15:19,920
Some companies are just too
stiff to really innovate.

1493
01:15:20,640 --> 01:15:21,200
Yes.
I don't.

1494
01:15:21,200 --> 01:15:22,680
Disagree.
I mean, I think it is much

1495
01:15:22,680 --> 01:15:26,800
harder to be running like one of
those companies and like

1496
01:15:26,800 --> 01:15:29,720
transform as an organization and
it's proven through all the

1497
01:15:29,720 --> 01:15:31,560
numbers, right?
Like like sort of the Fortune

1498
01:15:31,560 --> 01:15:35,720
500 companies, like, you know,
the stock market top companies

1499
01:15:35,720 --> 01:15:38,400
rotate like very quickly, right?
Like most companies don't last

1500
01:15:38,400 --> 01:15:40,920
like 1520 years as a public
market company.

1501
01:15:41,280 --> 01:15:46,680
So I will say though, like, you
know, like we as Boom Supersonic

1502
01:15:46,680 --> 01:15:49,600
is one of our customers, right,
man, starting a supersonic jet

1503
01:15:49,600 --> 01:15:51,960
company, right?
I always think Rescale is like a

1504
01:15:51,960 --> 01:15:55,080
pretty tough company to run, but
that is hard, right?

1505
01:15:55,080 --> 01:15:57,360
But that really takes a lot of
courage, right, to do something

1506
01:15:57,360 --> 01:15:59,200
like that.
So it's very admirable for folks

1507
01:15:59,200 --> 01:16:01,600
who have the mission and willing
to kind of take that on.

1508
01:16:01,600 --> 01:16:05,400
But what is also true is that
like, so they're super

1509
01:16:05,400 --> 01:16:06,680
innovative.
They're adopting all these

1510
01:16:06,680 --> 01:16:08,240
methods.
They're going to be way faster

1511
01:16:08,240 --> 01:16:10,800
than any other kind of large
aerospace manufacturer trying to

1512
01:16:10,800 --> 01:16:13,960
do a supersonic jet, of course,
but it's still really hard,

1513
01:16:14,120 --> 01:16:15,360
right?
Like, like startups are just

1514
01:16:15,360 --> 01:16:20,320
really hard, right?
And, you know, I'm, I would say

1515
01:16:20,320 --> 01:16:22,960
I'm glad I'm, I work in the like
field of software, right?

1516
01:16:22,960 --> 01:16:25,200
But we, our customers are all
building hardware pretty much.

1517
01:16:25,200 --> 01:16:30,320
So it's, yeah, it's awesome to
be able to serve them.

1518
01:16:30,400 --> 01:16:32,720
But I think it's, well, that's
the right answer.

1519
01:16:32,720 --> 01:16:36,080
It is still like, you know, the,
the the odds are sort of stacked

1520
01:16:36,080 --> 01:16:37,720
against you when you start a
company, right?

1521
01:16:37,720 --> 01:16:41,320
And so even today, like boom
supersonic, right?

1522
01:16:41,320 --> 01:16:42,880
Like they, they, they flew that
airplane.

1523
01:16:43,480 --> 01:16:44,920
It's it's like built with
rescales.

1524
01:16:44,920 --> 01:16:48,960
Awesome, right to test airplane,
they got to, they build the real

1525
01:16:48,960 --> 01:16:49,800
one.
They got to get all these

1526
01:16:49,800 --> 01:16:56,040
customers, you know, and it's,
it's, it's, it's a tough, yeah,

1527
01:16:56,040 --> 01:16:58,800
it's a tough problem to solve as
a as a small company, right?

1528
01:17:00,160 --> 01:17:03,120
So, but, but that is the way
like I do think cultural change,

1529
01:17:03,120 --> 01:17:06,160
like culture really matters.
So one thing we focus on at

1530
01:17:06,160 --> 01:17:10,640
Rescale is like not only making
sure our customers are adopting

1531
01:17:10,640 --> 01:17:13,120
all these capabilities, right,
but like if you, if you work at

1532
01:17:13,120 --> 01:17:17,520
Rescale, we have this.
So being an AI first company is

1533
01:17:17,520 --> 01:17:20,040
completely changes how you run
the company yourself, right?

1534
01:17:20,200 --> 01:17:22,520
And so we're very focused on
that.

1535
01:17:22,520 --> 01:17:24,200
Like this whole thing you were
talking about, hey, now I'm

1536
01:17:24,200 --> 01:17:28,000
managing a bunch of agents.
So every employee at Rescale is

1537
01:17:28,000 --> 01:17:31,280
empowered to kind of like, you
know, run their agents, right?

1538
01:17:31,640 --> 01:17:35,360
And, and we are using every
single AI framework, right?

1539
01:17:35,560 --> 01:17:37,960
Like I'm testing all of them
usually in parallel at the same

1540
01:17:37,960 --> 01:17:42,480
time, very rapidly changing
landscape.

1541
01:17:42,480 --> 01:17:45,240
But I think the way you build a
company and the way you scale a

1542
01:17:45,240 --> 01:17:49,240
company is quite different in
this like world of when you have

1543
01:17:49,240 --> 01:17:52,560
this AI tooling.
And I do think as a as if you

1544
01:17:52,560 --> 01:17:54,920
consider rescale a big or small
company, but as a 200 person

1545
01:17:54,920 --> 01:17:58,480
company versus say a 2000 person
company, we have a distinct

1546
01:17:58,480 --> 01:18:01,200
advantage.
Our ability to adopt new tooling

1547
01:18:01,280 --> 01:18:03,960
for this and sort of like dog
food and use all these tools

1548
01:18:03,960 --> 01:18:09,400
ourselves is, you know, if we
sort of execute well, is it

1549
01:18:09,880 --> 01:18:11,680
really fast, right?
Like we can, we can adopt new

1550
01:18:11,680 --> 01:18:14,840
tools really fast.
That allows us to understand how

1551
01:18:14,840 --> 01:18:18,840
our customers will also need to
like sort of change and adopt

1552
01:18:18,840 --> 01:18:20,960
tooling like this, right?
And, and it's a very different

1553
01:18:20,960 --> 01:18:21,960
business.
Our business is building

1554
01:18:21,960 --> 01:18:23,280
software.
Their business might be building

1555
01:18:23,280 --> 01:18:25,680
a vehicle, but a lot of the
principles are the same, right?

1556
01:18:25,720 --> 01:18:28,120
Like it, it's sort of like you
have to really rethink how you

1557
01:18:28,120 --> 01:18:30,640
do business and how you do
engineering and all these

1558
01:18:30,640 --> 01:18:33,080
things.
But I think that's also why it's

1559
01:18:33,080 --> 01:18:35,240
exciting, right?
Because it's like it is a new

1560
01:18:35,240 --> 01:18:38,560
paradigm.
These shifts don't happen very

1561
01:18:38,560 --> 01:18:40,760
often, right?
Like I don't know, if you asked

1562
01:18:40,760 --> 01:18:43,800
me a decade ago, are we going to
pass the Turing test, I'd have

1563
01:18:43,800 --> 01:18:45,200
been like, I don't think so,
right?

1564
01:18:45,560 --> 01:18:49,120
I would have been very wrong.
But like I think just like the

1565
01:18:49,200 --> 01:18:52,720
cloud shift like opened up these
like massive, massive markets

1566
01:18:52,720 --> 01:18:54,800
and and massive new
opportunities, right And many

1567
01:18:54,800 --> 01:18:58,160
businesses could not exist
without sort of concept of cloud

1568
01:18:58,160 --> 01:19:01,880
etcetera.
AII do think is even bigger,

1569
01:19:02,040 --> 01:19:04,240
right?
It is sort of like the probably

1570
01:19:04,240 --> 01:19:05,840
the biggest one of our
generation.

1571
01:19:05,920 --> 01:19:09,400
And that's awesome because like,
you know, I don't know if you

1572
01:19:09,400 --> 01:19:11,640
think about this, but I'm like,
man, what was it like when like

1573
01:19:11,640 --> 01:19:14,640
the I mean, I was a kid, like
the Internet sort of first came

1574
01:19:14,640 --> 01:19:17,760
out and then you could all of a
sudden like do this, like, you

1575
01:19:17,760 --> 01:19:20,040
know, you could sort of message
people across the world.

1576
01:19:20,040 --> 01:19:21,760
It's like it seemed crazy,
right?

1577
01:19:23,160 --> 01:19:25,000
And it was really a special
time.

1578
01:19:26,120 --> 01:19:29,040
And I think those, you know, you
go through the high of these

1579
01:19:29,040 --> 01:19:30,520
waves.
I think right now is a really

1580
01:19:30,520 --> 01:19:33,400
special time, right where we're
like the bleeding edge of

1581
01:19:33,520 --> 01:19:36,440
applying AI to engineering and
like how you build companies,

1582
01:19:36,520 --> 01:19:39,240
everything's changing and
that's, I think that's really

1583
01:19:39,240 --> 01:19:42,000
fun.
Yeah, I think we're living in

1584
01:19:42,000 --> 01:19:47,560
interesting times where it's
hard to predict what things will

1585
01:19:47,560 --> 01:19:49,960
be like in 10 years time.
It's hard to.

1586
01:19:50,400 --> 01:19:52,720
Yes.
Is it going to really be

1587
01:19:52,720 --> 01:19:54,480
different or is it just going to
stay the same?

1588
01:19:54,480 --> 01:19:56,600
I have a feeling it will
genuinely be different.

1589
01:19:56,600 --> 01:20:00,680
I, I just feel like I I use
these tools enough and I'm sure

1590
01:20:00,680 --> 01:20:02,240
you're the same, that I think
it's more.

1591
01:20:02,680 --> 01:20:04,760
It's more than hype and anybody
just says that.

1592
01:20:05,320 --> 01:20:07,840
I think they need to use these
tools themselves to see the

1593
01:20:08,400 --> 01:20:10,080
potential.
I mean, one of the key things,

1594
01:20:10,080 --> 01:20:12,520
so we work with customers like
we really encourage all the

1595
01:20:12,520 --> 01:20:15,520
executives CEO down, right.
I mean with a lot of like

1596
01:20:16,640 --> 01:20:21,680
CTOSCIOS, but also CEO right,
very important that they lead by

1597
01:20:21,680 --> 01:20:23,560
example like embrace these
tools, right?

1598
01:20:23,920 --> 01:20:25,480
If you want your word, if you
believe in this, right.

1599
01:20:26,000 --> 01:20:28,640
But yeah, like Silicon Valley
hype cycle is at its all time.

1600
01:20:29,000 --> 01:20:31,400
That's right.
Like it is just crazy times.

1601
01:20:31,400 --> 01:20:33,880
But you know, I I've seen these
waves before.

1602
01:20:33,880 --> 01:20:35,600
I you've seen them as well,
right?

1603
01:20:35,600 --> 01:20:40,080
Like I do think things do get
over hyped, but you know, people

1604
01:20:40,080 --> 01:20:45,000
said cloud was over hyped for a
long time and actually they were

1605
01:20:45,080 --> 01:20:47,160
totally wrong.
It was way under hyped, right?

1606
01:20:47,160 --> 01:20:50,920
Like go look at a Gartner report
from like 2010 about like cloud

1607
01:20:50,920 --> 01:20:53,200
computing, right?
They probably didn't even have a

1608
01:20:53,200 --> 01:20:55,400
quadrant, right?
Like it's just like, but I think

1609
01:20:56,840 --> 01:20:59,600
my intuition, like I think as
you said earlier, and it is hard

1610
01:20:59,600 --> 01:21:01,760
to forecast these things, right?
Especially like 10 years.

1611
01:21:01,760 --> 01:21:04,360
That was a long time.
But from first principles, if

1612
01:21:04,360 --> 01:21:08,440
you could pass the Turing test,
that changes a lot of things

1613
01:21:08,440 --> 01:21:10,440
because the way you sort of
interact and that so that

1614
01:21:10,440 --> 01:21:15,640
interacting with a human or an
AI is similar.

1615
01:21:16,120 --> 01:21:18,040
You can say 1's better than the
other, whatever.

1616
01:21:18,080 --> 01:21:22,280
But I think that changes, yeah,
how entire organization is

1617
01:21:22,280 --> 01:21:23,840
built, right?
Because like, like you said, you

1618
01:21:23,840 --> 01:21:27,840
can manage a bunch of agents
instead of people is also maybe

1619
01:21:27,840 --> 01:21:30,400
managing agents is easier than
managing people, right?

1620
01:21:30,400 --> 01:21:35,480
Like, you know, like, but it's
a, it's a new paradigm.

1621
01:21:35,480 --> 01:21:37,320
And then it's like, of course
things are going to be

1622
01:21:37,320 --> 01:21:39,280
overhyped.
So if you take AI surrogates as

1623
01:21:39,280 --> 01:21:44,560
an example, you see somebody
saying, Hey, you know, you could

1624
01:21:44,560 --> 01:21:46,840
do something 10,000 times
faster.

1625
01:21:46,880 --> 01:21:48,680
You know, there's a big.
Asterisk there that's sort of

1626
01:21:48,680 --> 01:21:51,520
hidden about like well, but
you'd be able to write training

1627
01:21:51,520 --> 01:21:54,960
data and like he gets a nine
9.9% accuracy, yes, but like

1628
01:21:55,600 --> 01:21:58,520
even more training data and then
you know, how does this work

1629
01:21:58,520 --> 01:22:00,120
your organization?
Well, then you need this and

1630
01:22:00,120 --> 01:22:07,040
that and so but that's what I
call example, like sort of over

1631
01:22:07,040 --> 01:22:09,000
hyping.
But then if you look at like the

1632
01:22:09,000 --> 01:22:11,240
real implications, right?
Like now all of a sudden you

1633
01:22:11,280 --> 01:22:14,160
take a bunch of tasks that
engineers were going to do.

1634
01:22:15,600 --> 01:22:19,280
And if you compress this, like,
as we were speaking earlier,

1635
01:22:19,280 --> 01:22:22,840
something from three days into
like an hour or less or like a

1636
01:22:22,840 --> 01:22:26,520
second, right?
It it's a, it's a sort of like

1637
01:22:26,880 --> 01:22:30,000
there's a lot of implications to
how the entire sort of ecosystem

1638
01:22:30,000 --> 01:22:32,320
changes.
I don't think you can challenge

1639
01:22:32,320 --> 01:22:33,840
that, right?
Like, so you can, you can sort

1640
01:22:33,840 --> 01:22:36,560
of take this one data point and
say, like, well, simulations are

1641
01:22:36,560 --> 01:22:39,440
not actually 1000 times faster
today versus like 3 years ago,

1642
01:22:39,440 --> 01:22:41,960
right?
But, you know, on the one hand,

1643
01:22:41,960 --> 01:22:44,960
I don't think you can blame the
marketers because like they're

1644
01:22:44,960 --> 01:22:46,720
just doing their job trying to
grab your attention.

1645
01:22:46,760 --> 01:22:47,880
It's hard to get somebody's
attention.

1646
01:22:48,280 --> 01:22:52,680
Yeah, the like, can something
actually be 1000 times faster,

1647
01:22:52,720 --> 01:22:54,440
like given the right conditions?
Yes, right.

1648
01:22:54,520 --> 01:22:56,600
And then and then it's like,
hey, if you implement this in

1649
01:22:56,600 --> 01:22:58,960
the right way, I think the
benefits are much more

1650
01:22:58,960 --> 01:23:01,040
interesting than 1000 times
faster simulations.

1651
01:23:01,040 --> 01:23:05,440
It's like now you have designers
who can get real time physics

1652
01:23:05,440 --> 01:23:07,560
responses.
Yeah, right.

1653
01:23:07,560 --> 01:23:11,280
They're pretty accurate.
I think you can take this

1654
01:23:11,280 --> 01:23:15,960
concept of like at a very high
level with AI1 sort of frame

1655
01:23:15,960 --> 01:23:17,960
that I think is important for
organizations to think about is

1656
01:23:17,960 --> 01:23:20,360
like these engineering
organizations are highly, highly

1657
01:23:20,360 --> 01:23:22,360
complex.
And one of the big challenges is

1658
01:23:22,360 --> 01:23:25,640
the people.
So at Boeing, right, you go in

1659
01:23:25,640 --> 01:23:27,480
for like a detailed design
review and they're like, oh, we

1660
01:23:27,480 --> 01:23:32,640
got to call up like this world
famous expert who's like 78

1661
01:23:32,640 --> 01:23:35,680
years old, right?
And they sort of come in and

1662
01:23:35,680 --> 01:23:38,080
they like pontificate and give
you advice on like whether this

1663
01:23:38,080 --> 01:23:39,960
is going to work or not.
But the intelligence of the

1664
01:23:39,960 --> 01:23:42,560
organization gets sort of lost
as the people leave the

1665
01:23:42,560 --> 01:23:44,600
organization.
And so like, the problem is if

1666
01:23:44,600 --> 01:23:48,560
you like lose the best people
for say, building an airplane,

1667
01:23:49,240 --> 01:23:51,880
if that knowledge was actually
just like in their heads and

1668
01:23:51,880 --> 01:23:54,040
like you can kind of see their
work, but like you don't really

1669
01:23:54,040 --> 01:23:56,200
kind of understand exactly how
they did that.

1670
01:23:56,200 --> 01:23:59,040
And then maybe there was like a
aerospace industry goes through

1671
01:23:59,040 --> 01:24:00,720
many cycles.
So there's like a time there's

1672
01:24:00,720 --> 01:24:02,080
like 10 years didn't hire
anybody.

1673
01:24:02,640 --> 01:24:04,720
So there was no like
apprenticeship training, right?

1674
01:24:04,720 --> 01:24:08,120
And so you, you've lost a lot of
the ability of the sort of IP of

1675
01:24:08,120 --> 01:24:11,080
like how to actually build great
airplanes can be solved with

1676
01:24:11,080 --> 01:24:14,360
AIAI can kind of like sort of
again, aggregate a lot of this

1677
01:24:14,360 --> 01:24:18,040
detailed stuff, synthesize it.
It's not always going to be

1678
01:24:18,040 --> 01:24:20,240
perfect or always going to be
right, but it can do a much

1679
01:24:20,240 --> 01:24:23,440
better job in a much shorter
amount of time than any person

1680
01:24:23,440 --> 01:24:25,400
can really do right.
And so if you think of AI as

1681
01:24:25,400 --> 01:24:29,480
ability to do things like that,
it's pretty incredible.

1682
01:24:29,480 --> 01:24:32,560
And then like the, if you just
project forward to even like

1683
01:24:32,560 --> 01:24:35,480
forget about 10 years, like 6
months or like 12 months, right?

1684
01:24:37,280 --> 01:24:39,760
You know, I will say, like when
I saw ChatGPT the first time,

1685
01:24:40,240 --> 01:24:43,040
right?
I'm like, like, interesting toy,

1686
01:24:43,160 --> 01:24:45,480
right?
But I didn't actually think at

1687
01:24:45,480 --> 01:24:48,920
that time, the first time I used
it so I could just get a change.

1688
01:24:48,920 --> 01:24:51,520
I think it was cool that I could
give certain answers, but I

1689
01:24:51,680 --> 01:24:53,600
didn't think it was going to
change my like day-to-day

1690
01:24:53,600 --> 01:24:56,760
workflow.
And like today, I don't think an

1691
01:24:56,760 --> 01:25:00,200
hour goes by, but I'm not like
prompting AI models and and

1692
01:25:00,200 --> 01:25:02,560
running stuff in the background
and all kinds of stuff going on,

1693
01:25:02,560 --> 01:25:05,000
right?
And so like, that is a new way

1694
01:25:05,000 --> 01:25:06,800
of working.
And if you talk to new founders

1695
01:25:06,800 --> 01:25:10,720
that are building new companies,
right, that's how they're doing

1696
01:25:10,720 --> 01:25:14,720
it.
And they're just like I was,

1697
01:25:16,000 --> 01:25:18,160
there's like the the cloud code
founder, right?

1698
01:25:18,840 --> 01:25:23,800
It's kind of showcasing how he's
like does software development.

1699
01:25:24,720 --> 01:25:26,920
He's got like all these
different agents and all this

1700
01:25:26,920 --> 01:25:29,840
stuff, right?
And you know, this cloud code

1701
01:25:29,840 --> 01:25:32,040
thing was just a hack day
project for him like a year ago.

1702
01:25:32,120 --> 01:25:34,480
And now it's like at I think 400
million run rate, right?

1703
01:25:34,480 --> 01:25:38,400
So like this is no joke, right?
And that's, I think that same

1704
01:25:38,400 --> 01:25:41,040
sort of shift is going to happen
in for engineers and scientists,

1705
01:25:41,040 --> 01:25:42,680
right?
Like the exact timing of these

1706
01:25:42,680 --> 01:25:46,640
things is always really hard to
predict, but same thing with

1707
01:25:46,640 --> 01:25:50,680
like cloud HPC, right?
Like I think the the it's easy

1708
01:25:50,680 --> 01:25:53,600
to be right about sort of the
secular trends, if you will, at

1709
01:25:53,600 --> 01:25:55,560
least from my perspective,
right, where it's like you're

1710
01:25:55,560 --> 01:25:57,640
going to get more processor
fragmentation, like Moore's law

1711
01:25:57,640 --> 01:25:59,880
is going to kind of slow down
and like got to go to

1712
01:25:59,880 --> 01:26:01,480
specialized processors and all
this stuff, right?

1713
01:26:01,480 --> 01:26:03,200
If you want to just get more
advanced computing, just all

1714
01:26:03,200 --> 01:26:04,080
these things are going to
happen.

1715
01:26:04,200 --> 01:26:07,080
It's very hard to say like at
what exact point in time it's

1716
01:26:07,080 --> 01:26:10,680
good this like big shift or
whatever and sort of the Overton

1717
01:26:10,680 --> 01:26:12,920
window of what's acceptable to
people will shift, right.

1718
01:26:14,360 --> 01:26:16,240
But you could be right about the
trends, right?

1719
01:26:16,280 --> 01:26:18,600
And so same thing with AII think
you can kind of see where this

1720
01:26:18,600 --> 01:26:21,880
is going.
And I think then the timing is

1721
01:26:21,880 --> 01:26:24,200
really hard.
I would say all the AI

1722
01:26:24,200 --> 01:26:27,280
predictions I would have made in
the last 12 months would have

1723
01:26:27,280 --> 01:26:29,880
been like, if they were about
something that happened in the

1724
01:26:29,880 --> 01:26:32,760
last 12 months, they would have,
I would have predicted them over

1725
01:26:32,760 --> 01:26:35,160
longer time horizons than they
actually happened over, right?

1726
01:26:35,280 --> 01:26:37,640
Like, and so if everything is
just happening much faster, I

1727
01:26:37,640 --> 01:26:40,040
almost don't even trust my own
intuition and forecasting too

1728
01:26:40,040 --> 01:26:41,360
much, right?
And so like what?

1729
01:26:41,360 --> 01:26:45,360
Well, what can you do to prepare
for that is like, I think you

1730
01:26:45,360 --> 01:26:48,200
got to really lean in to like
that feature.

1731
01:26:48,240 --> 01:26:51,720
And so like, even if right now,
like an example is like you can

1732
01:26:51,720 --> 01:26:55,760
do CAD and CAE modeling through
natural language prompting

1733
01:26:56,600 --> 01:26:59,160
through MCP, right?
So like this concept of model

1734
01:26:59,160 --> 01:27:01,360
context protocol, think of it
like an API, right?

1735
01:27:01,360 --> 01:27:05,240
It's sort of like connect like a
piece of software with like

1736
01:27:05,240 --> 01:27:07,800
your, your favorite AI tooling
and it can interact.

1737
01:27:07,800 --> 01:27:10,080
So we have a rescale MCP.
You can interact with it.

1738
01:27:10,120 --> 01:27:12,480
You can spin up jobs and do all
the all the things you'd want to

1739
01:27:12,480 --> 01:27:14,560
do in the user interface, But
you can now do it through an

1740
01:27:14,560 --> 01:27:16,360
LLM.
But then you can connect to LLM

1741
01:27:16,360 --> 01:27:19,280
to like a cab tool, right.
And you say, hey, like, you

1742
01:27:19,280 --> 01:27:21,120
know, change the angle of this
like windshield.

1743
01:27:21,120 --> 01:27:23,400
You're the author of the driver
ML data set.

1744
01:27:23,800 --> 01:27:25,760
Like create all these models.
Like you probably did it

1745
01:27:25,760 --> 01:27:28,520
manually.
And so like now you just yeah,

1746
01:27:29,480 --> 01:27:33,400
so now you just say, hey, you
know, like building this design

1747
01:27:33,400 --> 01:27:36,800
space here's like the
parameters, right, to generate

1748
01:27:36,800 --> 01:27:39,640
all the cat.
It can do that today.

1749
01:27:39,680 --> 01:27:42,920
Now your cat designer is going
to say, Oh yeah, it can do that.

1750
01:27:42,920 --> 01:27:45,040
But look, look at the
discontinuity over there.

1751
01:27:45,040 --> 01:27:46,840
Like because of this thing and
like whatever.

1752
01:27:46,840 --> 01:27:49,760
Like it's not good enough yet to
like 100% replace it.

1753
01:27:49,760 --> 01:27:52,080
But the whole point is like not
to 100% replace.

1754
01:27:52,600 --> 01:27:55,600
The point is now you can Gen.
cab like nobody's business,

1755
01:27:55,600 --> 01:27:57,960
right?
Like it's as easy as a prompt

1756
01:27:58,640 --> 01:28:00,280
that is just going to get way
better, right?

1757
01:28:00,280 --> 01:28:02,920
Like, so if, if we're talking
about like, is this going to be

1758
01:28:02,920 --> 01:28:05,960
good enough to do the like
aerospace level carbon fiber

1759
01:28:05,960 --> 01:28:08,840
layup design work?
Absolutely at some point, right?

1760
01:28:09,040 --> 01:28:12,360
Is that in like 1 month or is
that in 12 months?

1761
01:28:12,360 --> 01:28:14,680
Or is that in two years?
I don't know, but it's getting

1762
01:28:14,680 --> 01:28:17,480
really, really good, right?
And so those curves are really

1763
01:28:17,480 --> 01:28:19,880
fast and the adoption curves of
the of the technology are

1764
01:28:19,880 --> 01:28:21,440
usually pretty slow in
enterprise, right?

1765
01:28:21,440 --> 01:28:23,760
So like just because it's
possible doesn't mean people

1766
01:28:23,760 --> 01:28:29,160
adopt it, But you know, these
industry pressures are real,

1767
01:28:29,160 --> 01:28:31,800
right?
Like I, I do think like if you

1768
01:28:31,800 --> 01:28:36,080
can just kind of do innovation a
lot faster in automotive, in

1769
01:28:36,080 --> 01:28:41,720
aerospace, in like semiconductor
and life sciences, they're all

1770
01:28:41,720 --> 01:28:45,040
very big R&D spenders, right.
So if you just get a lot more

1771
01:28:45,040 --> 01:28:47,400
leverage out of that investment,
you know, that's a, that's a

1772
01:28:47,400 --> 01:28:49,840
huge leverage for society, I
think.

1773
01:28:50,880 --> 01:28:54,320
So maybe a final question, more
of a forward-looking question or

1774
01:28:54,320 --> 01:28:56,840
advice question.
Given all that we've said, if

1775
01:28:56,840 --> 01:29:02,440
you are a young founder, OK,
you're coming out of the first

1776
01:29:02,440 --> 01:29:06,200
job or a PhD and you, you know,
you think, what problem can I

1777
01:29:06,200 --> 01:29:09,520
try and tackle without giving
away anything that you you're

1778
01:29:09,520 --> 01:29:12,080
working on so you can't get too
good around.

1779
01:29:12,080 --> 01:29:16,200
So I guess what would be like
the biggest unsolved challenge

1780
01:29:16,200 --> 01:29:20,360
that you think the next startup
should try to tackle?

1781
01:29:20,360 --> 01:29:22,000
Yeah, there's a lot of big
challenges out there.

1782
01:29:22,960 --> 01:29:25,640
I think this is going to sound a
little bit self-serving, but if

1783
01:29:25,640 --> 01:29:27,720
you're amazing, you can always
come work at Rescale.

1784
01:29:27,720 --> 01:29:32,080
My e-mail is yours@rescale.com.
But I think this look, all the

1785
01:29:32,080 --> 01:29:34,000
things we've discussed, so
they're all happening right now,

1786
01:29:34,120 --> 01:29:35,680
right?
And so like I say, so where's

1787
01:29:35,680 --> 01:29:37,360
like the sort of next challenge
lie?

1788
01:29:38,320 --> 01:29:41,440
I do think it's in this kind of
like very buzzwordy word like

1789
01:29:41,440 --> 01:29:43,920
digital twin.
But like there is this, you

1790
01:29:43,920 --> 01:29:46,320
know, this concept of digital
twin, right, which is the the

1791
01:29:46,320 --> 01:29:49,160
equivalent of like what's
happening in the real world in a

1792
01:29:49,160 --> 01:29:53,240
digital form.
That concept is super important.

1793
01:29:54,760 --> 01:29:59,120
And this sort of SIM to reel
gap, like real being reality,

1794
01:29:59,160 --> 01:30:00,240
right?
And the SIM being the kind of

1795
01:30:00,240 --> 01:30:03,840
the simulation of that reality,
the more you can close that gap,

1796
01:30:03,920 --> 01:30:07,600
the more powerful products you
can build, right?

1797
01:30:07,600 --> 01:30:11,720
So like a good example I think
is I don't know if you've been

1798
01:30:11,720 --> 01:30:13,160
to San Francisco taking like a
Waymo.

1799
01:30:15,000 --> 01:30:16,520
Oh yeah, that freaks me out.
Yeah, yeah.

1800
01:30:16,960 --> 01:30:20,400
Yeah, Yeah, I, I think Waymo is
like the coolest thing now.

1801
01:30:20,400 --> 01:30:24,160
Why is Waymo so impressive?
I remember there were self

1802
01:30:24,160 --> 01:30:29,480
driving vehicles in San
Francisco about roughly 10 years

1803
01:30:29,480 --> 01:30:31,440
before.
Like Waymo kind of really went

1804
01:30:31,440 --> 01:30:34,880
live for like the sort of
private data slash real people,

1805
01:30:34,880 --> 01:30:40,480
right, consumers that last mile
of like whether it's regulatory

1806
01:30:40,480 --> 01:30:42,880
or like getting the software to
the right level, etcetera.

1807
01:30:44,400 --> 01:30:46,480
Took a long time, the way longer
than I expected.

1808
01:30:46,520 --> 01:30:48,680
At that time I would have said
like, oh, in a year, if they're

1809
01:30:48,680 --> 01:30:50,600
testing it right now, like in a
year, this is going to be ready,

1810
01:30:50,600 --> 01:30:51,840
right?
And you sort of know all the

1811
01:30:51,840 --> 01:30:53,720
technology already works.
Like you already know, self

1812
01:30:53,720 --> 01:30:55,640
driving kind of works.
There's obviously a lot of

1813
01:30:55,640 --> 01:30:58,560
safety and things like that.
But that I think that's a great

1814
01:30:58,560 --> 01:31:03,600
example of like Waymo is able to
simulate, right, what's

1815
01:31:03,600 --> 01:31:05,840
happening in the real world very
effectively.

1816
01:31:06,240 --> 01:31:10,640
Like locally on the edge uses
enormous amount of AI, right?

1817
01:31:10,760 --> 01:31:13,080
Enormous amount of training
data, right?

1818
01:31:13,320 --> 01:31:17,200
And it, it sort of solves this
transportation problem in a way

1819
01:31:17,200 --> 01:31:25,080
that's like awesome, right?
Like it's, you know, I and I

1820
01:31:25,080 --> 01:31:28,960
think that's a great example of
like what the future looks like.

1821
01:31:28,960 --> 01:31:30,880
Then then you have to ask
yourself like, OK, well, what

1822
01:31:30,880 --> 01:31:33,200
are the next set of problems?
Well, robots is an obvious one,

1823
01:31:34,080 --> 01:31:35,880
right?
So like if you can solve this

1824
01:31:35,880 --> 01:31:38,360
sort of simulation problem for
robots, I don't know if you've

1825
01:31:38,360 --> 01:31:41,240
watched these robots like
folding laundry, but it's like.

1826
01:31:41,800 --> 01:31:44,320
I kind of wait.
You know, it's it's yeah, but

1827
01:31:44,320 --> 01:31:47,240
it's it's not pretty when you
watch them walk or fold laundry

1828
01:31:47,240 --> 01:31:49,080
or some of these robot Olympics
that are going on.

1829
01:31:49,200 --> 01:31:52,440
We got a ways to go that said,
like, we all understand physics,

1830
01:31:52,440 --> 01:31:53,560
right?
So like, this is actually a

1831
01:31:53,560 --> 01:31:55,680
solved problem.
Like we we know how robots

1832
01:31:55,680 --> 01:31:57,800
should operate in a sort of real
world, right?

1833
01:31:57,800 --> 01:31:59,320
And I think videos are amazing
work and.

1834
01:31:59,320 --> 01:32:02,240
Like sort of the.
Developing the software and SDKS

1835
01:32:02,240 --> 01:32:03,880
for like the virtual world for
this, right?

1836
01:32:03,880 --> 01:32:05,880
And so you can sort of train
these models in a virtual way,

1837
01:32:06,000 --> 01:32:07,920
but ultimately you want to kind
of get that match to real,

1838
01:32:07,920 --> 01:32:08,920
right?
Like somebody's going to build

1839
01:32:08,920 --> 01:32:10,440
all these robots.
These robots are going to go do

1840
01:32:10,440 --> 01:32:11,720
all the things you don't want to
do at home.

1841
01:32:12,280 --> 01:32:14,920
And that's solving those types
of problems.

1842
01:32:14,960 --> 01:32:17,720
If you're like sort of have a
simulation background, right, is

1843
01:32:18,640 --> 01:32:20,400
I think that's the next
frontier, right?

1844
01:32:20,400 --> 01:32:23,240
And you already have this
example of Waymo, right?

1845
01:32:23,600 --> 01:32:28,680
And to me, like, I think if you
really the closing the SIM to

1846
01:32:28,680 --> 01:32:31,000
reel gap is possible for
anything you're passionate

1847
01:32:31,000 --> 01:32:34,360
about, you can do for airplanes,
you can do for robots, you can

1848
01:32:34,360 --> 01:32:36,440
do for Earth, right?
Like there's this Earth model

1849
01:32:36,680 --> 01:32:38,160
again.
NVIDIA has done an amazing job

1850
01:32:38,400 --> 01:32:41,720
developing this weather
simulation, you know,

1851
01:32:41,920 --> 01:32:47,000
horrendously difficult problem,
but worth solving, right?

1852
01:32:47,000 --> 01:32:50,640
Like if you solve weather
prediction, you know, a little

1853
01:32:50,640 --> 01:32:54,680
bit better or even like say 10X
better saves a lot of lives,

1854
01:32:54,680 --> 01:32:55,760
right?
It makes it makes a big

1855
01:32:55,760 --> 01:32:58,640
difference, right?
And yeah, I think those are

1856
01:32:59,040 --> 01:33:02,120
meaningful missions.
You know, talk a little bit

1857
01:33:02,120 --> 01:33:04,640
about starting a company.
You join a company, Do you, do

1858
01:33:04,640 --> 01:33:06,880
you stay in academia?
It's I think you have to really

1859
01:33:06,880 --> 01:33:09,880
think for yourself what what are
you motivated by, right?

1860
01:33:10,040 --> 01:33:11,680
Like, like, what do you really
want to do, right?

1861
01:33:13,040 --> 01:33:16,800
I personally would encourage
people to generally like veer

1862
01:33:16,800 --> 01:33:20,000
towards where they can make the
biggest impact and going to

1863
01:33:20,000 --> 01:33:23,680
learn the most right.
And so can be a start up, can be

1864
01:33:23,680 --> 01:33:27,360
a big company, could be many
different sort of platforms can

1865
01:33:27,360 --> 01:33:32,000
be in university.
But I I do think what's what we

1866
01:33:32,000 --> 01:33:35,600
need more of is like the people
to really have the courage to

1867
01:33:35,600 --> 01:33:38,600
kind of, I think, solve the
problems that are really on the

1868
01:33:38,600 --> 01:33:41,880
edge, but nobody solved before.
The problem with academia in my

1869
01:33:41,880 --> 01:33:44,520
view is like I spent a lot of
time in school, right?

1870
01:33:44,520 --> 01:33:49,280
A lot too many degrees.
And it's sort of like, I think

1871
01:33:49,280 --> 01:33:52,800
the challenge with school is
that you start competing on a

1872
01:33:52,800 --> 01:33:56,320
dimension that's that's a little
bit removed from like the

1873
01:33:56,320 --> 01:33:59,600
practical reality and you start
competing.

1874
01:33:59,600 --> 01:34:01,720
I'm like, you know, can we solve
this equation more efficiently,

1875
01:34:01,720 --> 01:34:04,880
whether or not Boeing like uses
this and tops this or anybody

1876
01:34:04,880 --> 01:34:08,480
else does, as long as this other
really smart person thinks it's

1877
01:34:08,480 --> 01:34:10,040
smart, right?
And I get a lot of references,

1878
01:34:10,040 --> 01:34:11,720
like I win the game, so to
speak, right?

1879
01:34:11,800 --> 01:34:14,040
But, you know, like it's, I
think it's like, I think it's

1880
01:34:14,320 --> 01:34:18,360
Kissinger quote, which is like
the battles in academia are so

1881
01:34:18,360 --> 01:34:20,080
fierce because the stakes are so
small.

1882
01:34:20,080 --> 01:34:24,400
So, you know, I I think that's a
yeah, it's, it's, it's

1883
01:34:24,440 --> 01:34:26,840
unfortunate because you see a
lot of really great talent,

1884
01:34:27,040 --> 01:34:28,800
right?
And, you know, I was part of

1885
01:34:28,800 --> 01:34:31,360
this as well myself, right?
You know, Peter Thiel has a good

1886
01:34:31,360 --> 01:34:34,200
view on this as well.
I think where it's like it's

1887
01:34:34,200 --> 01:34:37,680
really elite students who are
super smart and they keep

1888
01:34:37,680 --> 01:34:39,080
climbing this sort of academic
letter.

1889
01:34:39,080 --> 01:34:40,000
Why?
Because these are like the

1890
01:34:40,000 --> 01:34:42,120
badges you want to collect.
I went to Stanford, I went to

1891
01:34:42,120 --> 01:34:44,000
Harvard.
I get to this and that right

1892
01:34:44,280 --> 01:34:46,360
have all the sort of
certifications, but that's sort

1893
01:34:46,360 --> 01:34:49,080
of like the status game and like
signaling to other people,

1894
01:34:49,120 --> 01:34:49,960
right?
I think.

1895
01:34:49,960 --> 01:34:53,920
And the problem is, you know,
he, what he talks about is that

1896
01:34:53,920 --> 01:34:57,640
like the sort of the in that
process, the sort of dreams get

1897
01:34:57,640 --> 01:35:01,560
stomped out of you, right?
And that's his view.

1898
01:35:01,840 --> 01:35:03,880
I think that's, that's, that's
pretty correct.

1899
01:35:04,280 --> 01:35:06,160
I also think there's a
specialization that happens

1900
01:35:06,160 --> 01:35:08,160
because like, OK, you want to be
best in the world at something,

1901
01:35:08,200 --> 01:35:09,840
you know, narrow, narrow,
narrow, narrow, right?

1902
01:35:10,040 --> 01:35:12,920
But great innovation actually
often happens by, you know,

1903
01:35:12,920 --> 01:35:16,000
taking concepts from 1 field,
applying it in another, right?

1904
01:35:16,000 --> 01:35:18,360
Like, like it's actually like
the deep mind folks that sort of

1905
01:35:18,360 --> 01:35:21,040
like mesh, graphnet, etcetera.
Like that's why we have AI

1906
01:35:21,040 --> 01:35:24,400
service and they're not all like
physics experts, right?

1907
01:35:24,440 --> 01:35:26,840
Like, in fact, you often need
somebody from a domain who's

1908
01:35:26,840 --> 01:35:29,080
like naive enough to be like,
hey, let's just try this thing.

1909
01:35:30,880 --> 01:35:37,560
And so I do think you want
people to follow like a mission

1910
01:35:37,560 --> 01:35:40,680
and sort of a like, like the
where they're passionate about,

1911
01:35:40,680 --> 01:35:42,160
right?
It's like, if you're your 20s,

1912
01:35:42,160 --> 01:35:44,360
like, like you probably have
enough life experience to kind

1913
01:35:44,360 --> 01:35:47,920
of know what, what you like.
And there's so much more

1914
01:35:47,920 --> 01:35:50,560
possible than people think they
are capable of themselves,

1915
01:35:50,920 --> 01:35:52,360
right?
Like I personally would have

1916
01:35:52,360 --> 01:35:54,880
never thought when I was like an
engineer, simulation engineer

1917
01:35:54,880 --> 01:35:58,720
working at Boeing doing my
little thing, right, that I

1918
01:35:58,720 --> 01:36:02,120
could kind of break out, start
your own company, get these like

1919
01:36:02,200 --> 01:36:04,480
like maybe they sit across the
table from like some of the

1920
01:36:04,480 --> 01:36:08,760
smartest people in Silicon
Valley and together kind of like

1921
01:36:08,760 --> 01:36:10,520
it like build an amazing
company, right?

1922
01:36:10,520 --> 01:36:14,360
And I think it takes you of
course need the ambition, but

1923
01:36:14,360 --> 01:36:18,040
most of all you need the courage
and the the willingness to kind

1924
01:36:18,040 --> 01:36:21,360
of persevere, right?
Many people give up, like, I

1925
01:36:21,360 --> 01:36:24,680
don't know if you see this, but
like, you know, there's a lot of

1926
01:36:24,680 --> 01:36:26,600
people who apply to jobs that
rescale and I see a lot of

1927
01:36:26,600 --> 01:36:30,120
resumes and like you see a lot
of this, like, hey, 1 1/2 years

1928
01:36:30,120 --> 01:36:32,400
here, 1 1/2 years there,
etcetera.

1929
01:36:32,400 --> 01:36:33,680
Right?
I, I think you got to kind of

1930
01:36:33,680 --> 01:36:35,840
find that right, problem that
you're really passionate about

1931
01:36:35,840 --> 01:36:42,760
and then like really pursue that
with, with all the energy you

1932
01:36:42,760 --> 01:36:44,560
have.
There's a good essay written by

1933
01:36:44,560 --> 01:36:48,280
Paul Graham called, it's called
great work or something like

1934
01:36:48,280 --> 01:36:50,160
that.
It's like how to do great work.

1935
01:36:50,960 --> 01:36:54,320
It's kind of a long essay, so it
might take a while to read, but

1936
01:36:54,320 --> 01:36:57,400
one of the things he talks about
is like, you know, you want to

1937
01:36:57,400 --> 01:37:00,400
be on the edges, like sort of
the he thinks of like knowledge

1938
01:37:00,400 --> 01:37:02,920
as this sort of like tree and
these like fractals basically.

1939
01:37:02,920 --> 01:37:04,520
And you're sort of want to be on
the edge, right?

1940
01:37:04,520 --> 01:37:06,760
And then you want to kind of see
where the gaps are.

1941
01:37:07,440 --> 01:37:10,200
And often I, I think I'm not
sure if he talks about an essay,

1942
01:37:10,200 --> 01:37:14,080
but in my view, it's like if you
take kind of the lessons from a

1943
01:37:14,080 --> 01:37:15,800
certain field applying in
another.

1944
01:37:15,880 --> 01:37:18,680
That's where I've seen like
amazing breakthroughs like that

1945
01:37:18,680 --> 01:37:22,600
rescale and like also beyond
like in places like Boeing and

1946
01:37:22,600 --> 01:37:24,840
other places.
And that requires like the

1947
01:37:24,840 --> 01:37:26,960
willingness to kind of learn
these different things, right?

1948
01:37:26,960 --> 01:37:28,840
And, and sort of apply that
curiosity in different ways.

1949
01:37:28,840 --> 01:37:33,440
So, you know, for somebody who
is just graduating with their,

1950
01:37:33,480 --> 01:37:36,480
you know, master's or PhD or
something like I would look to

1951
01:37:36,480 --> 01:37:38,400
the in short, I would go
somewhere where I think you're

1952
01:37:38,400 --> 01:37:40,920
going to learn the most or make
the biggest impact.

1953
01:37:42,720 --> 01:37:46,320
What is hard is that there's a
lot of like sidle things that

1954
01:37:46,320 --> 01:37:49,880
will pressure you to do other
things, right, Like maybe go

1955
01:37:49,880 --> 01:37:52,160
work for the company that has
the brand that your parents will

1956
01:37:52,160 --> 01:37:54,640
be proud of you for, right?
Like it's a natural thing,

1957
01:37:54,640 --> 01:37:56,560
right?
Like maybe go, you know, like,

1958
01:37:56,560 --> 01:37:58,760
for example, if you just start a
company, it's like, well, like

1959
01:37:59,080 --> 01:38:00,960
now you're unemployed is a
different view of the same

1960
01:38:00,960 --> 01:38:02,960
thing, right?
Like, so I think that some of

1961
01:38:02,960 --> 01:38:06,960
these things are hard, but but
you know, I do think people

1962
01:38:07,040 --> 01:38:09,640
could take much more risks than
they usually think they can,

1963
01:38:09,800 --> 01:38:12,200
right?
And the challenge is a little

1964
01:38:12,200 --> 01:38:15,120
bit as if you find that out late
in life, there are less

1965
01:38:15,120 --> 01:38:17,520
opportunities.
Like we, we all have just time,

1966
01:38:17,640 --> 01:38:19,400
right?
And, and that's the sort of the,

1967
01:38:19,640 --> 01:38:21,720
the great equalizer.
You know, people spend a lot of

1968
01:38:21,720 --> 01:38:23,480
time doing work.
I certainly do.

1969
01:38:23,480 --> 01:38:27,640
I think it's like the that time
that you spent, it should be

1970
01:38:27,640 --> 01:38:28,760
something really meaningful to
you.

1971
01:38:28,760 --> 01:38:30,440
I can't just be a mean student,
right?

1972
01:38:30,520 --> 01:38:32,960
Like I love spending time with
my kids and family and those

1973
01:38:32,960 --> 01:38:35,960
things too, right?
But like, if you're a motivated

1974
01:38:35,960 --> 01:38:37,600
individual, you're probably
going to spend a lot of time

1975
01:38:37,600 --> 01:38:39,080
working, right?
And so that whatever that work

1976
01:38:39,080 --> 01:38:43,040
is, it should be, you know, I
wouldn't compromise that too

1977
01:38:43,040 --> 01:38:44,960
much for, say, like a better
salary.

1978
01:38:44,960 --> 01:38:48,720
You're like a mission you're not
so excited about, right?

1979
01:38:49,280 --> 01:38:51,840
A lot of big tech companies that
have, you know, some missions

1980
01:38:51,840 --> 01:38:53,520
are certainly more interesting
than others, right?

1981
01:38:54,400 --> 01:38:57,440
And I think applying yourself to
kind of make an impact, you

1982
01:38:57,440 --> 01:38:59,960
know, move society forward is
super important.

1983
01:39:00,600 --> 01:39:03,400
Yeah, wise words and thank you
so much for this.

1984
01:39:03,400 --> 01:39:07,160
I love this conversation and I
think what we need to do is

1985
01:39:07,160 --> 01:39:12,400
schedule in like a few years
time and see how close we were,

1986
01:39:12,720 --> 01:39:15,760
whether we're all using AI
engineers or whether we were

1987
01:39:15,760 --> 01:39:17,200
wrong.
I'm sure something completely

1988
01:39:17,200 --> 01:39:19,920
different will have come around
that nine of us predicted, but

1989
01:39:19,920 --> 01:39:22,360
it'd be a good thing to see.
So with with that, thank you

1990
01:39:22,360 --> 01:39:24,440
Joris, really, really
appreciated your time.

1991
01:39:25,200 --> 01:39:26,840
Thank you, Neil.
Thanks for the opportunity.

1992
01:39:26,840 --> 01:39:29,720
And yeah, let's let's put it on
the calendar and pass it back

1993
01:39:29,720 --> 01:39:31,360
up.
Sounds good.

1994
01:39:31,520 --> 01:39:32,600
All right.
Cheers.

1995
01:39:32,680 --> 01:39:34,000
Thank you.
See you.