1
00:00:15,356 --> 00:00:15,796
Speaker 1: Pushkin.

2
00:00:20,436 --> 00:00:26,436
Speaker 2: AI works amazingly well. It works terrifyingly well even for

3
00:00:26,836 --> 00:00:32,036
Speaker 2: virtual things, for words, for pictures, for videos. This is

4
00:00:32,036 --> 00:00:36,476
Speaker 2: true in large part because of the Internet. The Internet

5
00:00:36,516 --> 00:00:41,876
Speaker 2: provides this wildly abundant, readily available source of words, pictures,

6
00:00:41,916 --> 00:00:47,876
Speaker 2: and videos to train AI models. But there is no analogous,

7
00:00:48,356 --> 00:00:52,356
Speaker 2: wildly abundant, readily available data set for the physical world.

8
00:00:53,116 --> 00:00:57,956
Speaker 2: There is no gargantuan Internet like repository of data that

9
00:00:58,116 --> 00:01:02,836
Speaker 2: describes how things move and bend and break in real

10
00:01:02,956 --> 00:01:06,876
Speaker 2: physical space. And as a result, we do not yet

11
00:01:07,036 --> 00:01:11,996
Speaker 2: have robust AI for the physical But people are working

12
00:01:12,036 --> 00:01:16,116
Speaker 2: on it, and if they succeed, they'll change the way

13
00:01:16,196 --> 00:01:19,596
Speaker 2: the world works, not just the world as it appears

14
00:01:19,596 --> 00:01:23,516
Speaker 2: on our screens, but the actual physical world, the world

15
00:01:23,516 --> 00:01:32,876
Speaker 2: where if you drop something on your foot it hurts.

16
00:01:33,276 --> 00:01:35,636
Speaker 2: I'm Jacob Goldstein and this is What's Your Problem, the

17
00:01:35,716 --> 00:01:37,956
Speaker 2: show where I talk to people who are trying to

18
00:01:38,036 --> 00:01:43,036
Speaker 2: make technological progress. My guest today is Edward Mayer. He's

19
00:01:43,076 --> 00:01:47,436
Speaker 2: the co founder and CEO of Machina Labs. Edward's problem

20
00:01:47,516 --> 00:01:51,596
Speaker 2: is this, how can you use AI to turn robots

21
00:01:51,996 --> 00:01:58,316
Speaker 2: from dumb, inflexible machines into skilled versatile craftsmen. Before he

22
00:01:58,396 --> 00:02:02,396
Speaker 2: started Machina Labs, Edward worked in the rocket ship business,

23
00:02:02,676 --> 00:02:06,516
Speaker 2: first at SpaceX and then at a company called Relativity Space.

24
00:02:07,076 --> 00:02:10,636
Speaker 2: And in the rocket business, the word solw firsthand the

25
00:02:10,756 --> 00:02:14,676
Speaker 2: problems of traditional manufacturing. It's the kind of problem he's

26
00:02:14,716 --> 00:02:17,956
Speaker 2: now trying to solve with AI and robots. It's a

27
00:02:17,956 --> 00:02:24,916
Speaker 2: problem called the rigid factory problem. So I've heard you

28
00:02:24,996 --> 00:02:27,676
Speaker 2: use this phrase that's interesting to me, and it's the

29
00:02:27,796 --> 00:02:32,756
Speaker 2: rigid factory problem. What's the rigid factory problem?

30
00:02:32,996 --> 00:02:36,236
Speaker 1: That main problem with the factories today is that rigidity,

31
00:02:36,316 --> 00:02:39,156
Speaker 1: meaning that if you have to build a physical product,

32
00:02:40,276 --> 00:02:42,916
Speaker 1: you pretty much have to build a factory that's designed

33
00:02:42,956 --> 00:02:45,556
Speaker 1: for it and built for it. There's a lot of

34
00:02:45,596 --> 00:02:49,436
Speaker 1: components that goes into the factory, from machinery all the

35
00:02:49,436 --> 00:02:52,396
Speaker 1: way to the tooling that is required to build products

36
00:02:52,436 --> 00:02:55,916
Speaker 1: that are specifically designed for the geometry for the material

37
00:02:56,036 --> 00:02:58,196
Speaker 1: that you're trying to use. The moment you want to

38
00:02:58,276 --> 00:03:00,836
Speaker 1: change that, you have to change your factory, which is

39
00:03:00,876 --> 00:03:03,516
Speaker 1: a huge investment. You know, I always give an example

40
00:03:03,596 --> 00:03:06,516
Speaker 1: from when I was at SpaceX. You know, you think

41
00:03:06,556 --> 00:03:10,476
Speaker 1: of SpaceX as a very innovative and it is you

42
00:03:10,476 --> 00:03:12,516
Speaker 1: know on the edge of a hardware space in terms

43
00:03:12,516 --> 00:03:17,076
Speaker 1: of innovation in the past twenty four years twenty three

44
00:03:17,156 --> 00:03:20,876
Speaker 1: four years that they have existed, they have two rocket families.

45
00:03:21,236 --> 00:03:24,356
Speaker 1: There's Starship and there's Falcon, right, because at the moment

46
00:03:24,356 --> 00:03:27,236
Speaker 1: you decide on diameter of for example, Falcon nine or

47
00:03:27,276 --> 00:03:29,916
Speaker 1: the Falcon family in general, the diameter of that core,

48
00:03:29,956 --> 00:03:31,756
Speaker 1: it's very hard to change it. A lot of tooling

49
00:03:31,836 --> 00:03:35,516
Speaker 1: and machinery specifically built for that diameter. And that's why

50
00:03:35,516 --> 00:03:37,236
Speaker 1: for Starship they had to start from scratch.

51
00:03:37,636 --> 00:03:40,476
Speaker 2: Start from scratch, meaning like not just design, but like

52
00:03:40,836 --> 00:03:43,476
Speaker 2: the factory itself, like the Factorily had to build a

53
00:03:43,476 --> 00:03:46,436
Speaker 2: whole new factory because they wanted to make a different

54
00:03:46,516 --> 00:03:47,396
Speaker 2: sized rocket.

55
00:03:47,876 --> 00:03:51,076
Speaker 1: Yes, different size, different material, All the tooling has to change,

56
00:03:51,156 --> 00:03:54,876
Speaker 1: right almost almost, Yeah, you have to basically assume building

57
00:03:54,916 --> 00:03:57,516
Speaker 1: from scratch, ground up factory. Why does it need to

58
00:03:57,556 --> 00:03:59,236
Speaker 1: be there for us to build this new product?

59
00:04:00,316 --> 00:04:03,396
Speaker 2: I heard you describe was this from your own experience

60
00:04:03,836 --> 00:04:08,876
Speaker 2: the sort of era at SpaceX when the fact that

61
00:04:08,916 --> 00:04:12,316
Speaker 2: you couldn't make the rocket wider led to all these

62
00:04:12,436 --> 00:04:16,036
Speaker 2: kind of difficult things people were trying to do to

63
00:04:16,036 --> 00:04:18,116
Speaker 2: be like, how can we do all these things under

64
00:04:18,156 --> 00:04:20,716
Speaker 2: this fundamental constraint, Like, can you talk a little bit

65
00:04:20,756 --> 00:04:21,196
Speaker 2: about that.

66
00:04:21,596 --> 00:04:25,636
Speaker 1: Yeah, this is a lot of conversation happening in twenty twelve,

67
00:04:25,716 --> 00:04:30,996
Speaker 1: twenty thirteen, twenty fourteen time when the diameter of the

68
00:04:30,996 --> 00:04:33,676
Speaker 1: Falcon nine could not get any larger, and if you

69
00:04:33,716 --> 00:04:36,356
Speaker 1: look at actually different Falcon versions, the height of that

70
00:04:36,436 --> 00:04:40,036
Speaker 1: vehicle kept going higher, the diameter could not change. So

71
00:04:40,076 --> 00:04:42,796
Speaker 1: it was about where what space can you find to

72
00:04:42,876 --> 00:04:49,116
Speaker 1: put new features and new designs that exist within the vehicle.

73
00:04:49,196 --> 00:04:51,156
Speaker 1: So there was a lot of stuff basically being crammed

74
00:04:51,196 --> 00:04:52,916
Speaker 1: into the space that you have already got.

75
00:04:53,076 --> 00:04:55,556
Speaker 2: So that's true for building rockets. I mean, what are

76
00:04:55,596 --> 00:04:59,836
Speaker 2: some other just you know, different kinds of manufactured products

77
00:04:59,836 --> 00:05:02,036
Speaker 2: where that kind of rigidity is a problem.

78
00:05:02,396 --> 00:05:06,436
Speaker 1: Yeah, I think it is just common almost in all manufacturing.

79
00:05:06,476 --> 00:05:08,836
Speaker 1: That's why this phenomenon. I think it's kind of funny.

80
00:05:08,836 --> 00:05:12,716
Speaker 1: People take it for granted that a thing called economies

81
00:05:12,716 --> 00:05:16,236
Speaker 1: of scale, uh huh, Like people take it for granted

82
00:05:16,276 --> 00:05:18,716
Speaker 1: as if it's rule of nature. It's actually not.

83
00:05:19,236 --> 00:05:21,556
Speaker 2: Just to be clear, it's basically, the more you build

84
00:05:21,556 --> 00:05:23,796
Speaker 2: of a thing, the cheaper each one of those things gets.

85
00:05:23,796 --> 00:05:25,876
Speaker 2: If you build one, it's really expensive. If you build

86
00:05:25,876 --> 00:05:27,236
Speaker 2: a million each one's.

87
00:05:27,116 --> 00:05:29,716
Speaker 1: A lot cheaper, yes, exactly, But then people don't think

88
00:05:29,756 --> 00:05:32,036
Speaker 1: about it's like, oh, okay, intuitively makes sense, but why

89
00:05:32,196 --> 00:05:36,756
Speaker 1: it's actually not that intuitive. It's actually a limitation of technology. Right.

90
00:05:38,836 --> 00:05:41,636
Speaker 1: Why why economies of scale is a thing is because

91
00:05:41,636 --> 00:05:43,796
Speaker 1: you have to make a huge amount of investment to

92
00:05:43,876 --> 00:05:46,156
Speaker 1: make the first thing, and the moment you make the

93
00:05:46,196 --> 00:05:48,436
Speaker 1: second thing and the third thing, then you can break

94
00:05:48,516 --> 00:05:51,876
Speaker 1: even your investment onto more products that you're going to

95
00:05:52,116 --> 00:05:54,276
Speaker 1: come out of it. But that's only true if the

96
00:05:54,316 --> 00:05:57,796
Speaker 1: second product can be built for the first investment. You

97
00:05:57,876 --> 00:05:59,836
Speaker 1: had to turn this concept and say, oh, this is

98
00:05:59,876 --> 00:06:02,076
Speaker 1: a given. This is an axiom of the world that

99
00:06:02,156 --> 00:06:04,676
Speaker 1: economies of scale is a thing, but in reality it

100
00:06:04,716 --> 00:06:08,196
Speaker 1: is a technological challenge. Right. It means that you build

101
00:06:08,196 --> 00:06:10,676
Speaker 1: a car, you ask what application you Once you build

102
00:06:10,676 --> 00:06:14,116
Speaker 1: a factory for a car and all the toolings and

103
00:06:14,196 --> 00:06:17,276
Speaker 1: dies that goes into stamping of the panels of that car,

104
00:06:17,716 --> 00:06:20,596
Speaker 1: that's one hundred and fifty million dollar investment just for stamping.

105
00:06:21,276 --> 00:06:23,116
Speaker 1: And this is a number for example from Tesla. Tesla

106
00:06:23,156 --> 00:06:25,276
Speaker 1: spens one hundred and fifty million dollars in a stamping

107
00:06:25,316 --> 00:06:29,756
Speaker 1: plant they have in Giga factory in Texas, right, and

108
00:06:29,836 --> 00:06:33,076
Speaker 1: that can only make Model Y or Model three right

109
00:06:33,196 --> 00:06:35,956
Speaker 1: the moment you have to change that. That means go

110
00:06:36,036 --> 00:06:40,156
Speaker 1: through every eighty two hundred and thirty sheet metal panels

111
00:06:40,156 --> 00:06:42,436
Speaker 1: that exist on that car and design a new tool

112
00:06:42,476 --> 00:06:44,756
Speaker 1: for it. And each of these tool is going to

113
00:06:44,836 --> 00:06:47,156
Speaker 1: be a few hundred thousand dollars to sometimes a million

114
00:06:47,196 --> 00:06:49,396
Speaker 1: dollars or a million and a half million dollars. And

115
00:06:49,436 --> 00:06:51,756
Speaker 1: you're talking about like eighty t one hundred and thirty

116
00:06:51,796 --> 00:06:52,676
Speaker 1: tools per vehicle.

117
00:06:53,036 --> 00:06:55,956
Speaker 2: And like all you're doing you're not reinventing the car there,

118
00:06:55,996 --> 00:06:58,596
Speaker 2: you're just making a car that's a slightly different shape.

119
00:06:58,716 --> 00:07:01,756
Speaker 1: Basically, yes, maybe you may get sedan, and you're not

120
00:07:02,076 --> 00:07:04,916
Speaker 1: trying to do a slightly longer version, a slightly bigger version.

121
00:07:06,076 --> 00:07:08,476
Speaker 1: And that's why economies a scale of the thing you saying, Okay,

122
00:07:08,516 --> 00:07:11,836
Speaker 1: I made a fact. Now it only pays back if

123
00:07:11,876 --> 00:07:14,556
Speaker 1: I make a million of this car, right, because I

124
00:07:14,596 --> 00:07:16,596
Speaker 1: had to just drop one hundred and fifty million dollars

125
00:07:16,676 --> 00:07:20,116
Speaker 1: on just a stamping plant. So yeah, it's all over manufacturing.

126
00:07:20,156 --> 00:07:22,836
Speaker 1: We abstract this whole concept and gave it the name,

127
00:07:22,916 --> 00:07:24,036
Speaker 1: says economies, of scale.

128
00:07:24,556 --> 00:07:30,556
Speaker 2: Yeah, so you left SpaceX and you went to Relativity Space, right,

129
00:07:30,596 --> 00:07:33,356
Speaker 2: a company that was also in the space business that

130
00:07:33,516 --> 00:07:34,876
Speaker 2: was using three D printing.

131
00:07:34,916 --> 00:07:35,076
Speaker 1: Right.

132
00:07:35,076 --> 00:07:37,396
Speaker 2: That was the idea of the company, which seems like

133
00:07:37,596 --> 00:07:40,556
Speaker 2: an approach to this problem that you're talking about.

134
00:07:40,596 --> 00:07:40,716
Speaker 1: Right.

135
00:07:40,756 --> 00:07:43,996
Speaker 2: An advantage of three D printing is that it is

136
00:07:44,076 --> 00:07:48,156
Speaker 2: much more flexible and less rigid than traditional manufacturing. Right,

137
00:07:48,196 --> 00:07:49,476
Speaker 2: So tell me about that.

138
00:07:50,036 --> 00:07:54,236
Speaker 1: Yeah. So, yeah, we saw this challenge at SpaceX and

139
00:07:54,276 --> 00:07:56,356
Speaker 1: I joined Relativity very early on. I was the fourth

140
00:07:56,396 --> 00:08:01,476
Speaker 1: person on that team. And the goal over there was, Okay,

141
00:08:02,236 --> 00:08:05,316
Speaker 1: let's just think about this fundamentally, can we build a

142
00:08:05,396 --> 00:08:09,396
Speaker 1: rocket that all built with flexible technology and a time?

143
00:08:09,396 --> 00:08:14,476
Speaker 1: Three D printing was that forefront of everybody's minds because

144
00:08:14,556 --> 00:08:18,236
Speaker 1: people were already starting to build that. NASA SpaceX people

145
00:08:18,276 --> 00:08:21,036
Speaker 1: were already starting to build engines out of three D printing,

146
00:08:21,436 --> 00:08:23,916
Speaker 1: and the concept was like, well, that's great, it's very flexible.

147
00:08:24,036 --> 00:08:27,796
Speaker 1: Three D printing has this promise of geometry agnostic, material agnostic.

148
00:08:27,836 --> 00:08:29,396
Speaker 1: You can just feed it a design and can build

149
00:08:29,436 --> 00:08:31,716
Speaker 1: a product for you. And it worked very well with

150
00:08:31,796 --> 00:08:35,316
Speaker 1: rocket engines. I think probably the future all rocket engines

151
00:08:36,516 --> 00:08:39,356
Speaker 1: would be would be three D printed, And the concept was,

152
00:08:39,796 --> 00:08:43,996
Speaker 1: can we take this and scale it to the whole vehicle, right,

153
00:08:44,316 --> 00:08:46,476
Speaker 1: can we build the whole vehicle with a process like

154
00:08:46,476 --> 00:08:50,836
Speaker 1: three D printing so that it is flexible Today, if

155
00:08:50,836 --> 00:08:53,236
Speaker 1: you want to build a rocket with twelve foot diameter,

156
00:08:53,316 --> 00:08:55,636
Speaker 1: we can do it. And then if our calculation changes

157
00:08:55,676 --> 00:08:57,356
Speaker 1: and we wanted to go to another orbit or do

158
00:08:57,356 --> 00:08:59,916
Speaker 1: a different type of emission, then we can change that

159
00:09:00,036 --> 00:09:02,996
Speaker 1: diet twelve diameter to twenty diameter twenty fie.

160
00:09:03,076 --> 00:09:04,796
Speaker 2: Don't have to build a new factory, don't have to

161
00:09:04,796 --> 00:09:07,556
Speaker 2: build new machinery, just change three D printer.

162
00:09:07,716 --> 00:09:12,676
Speaker 1: Yeah, exactly. So that was a concept behind relativity. That's

163
00:09:12,676 --> 00:09:15,316
Speaker 1: a thesis behind relativity, and that was the goal there.

164
00:09:15,396 --> 00:09:17,396
Speaker 1: The goal was, you know, three D print a whole

165
00:09:17,476 --> 00:09:18,796
Speaker 1: rocket so they can be flexible.

166
00:09:19,916 --> 00:09:22,476
Speaker 2: But it hasn't. It hasn't worked at least in the

167
00:09:22,556 --> 00:09:25,276
Speaker 2: kind of maximalist version, right, Like, they just haven't been

168
00:09:25,316 --> 00:09:27,556
Speaker 2: able to do it. They've they've sort of backed off

169
00:09:27,596 --> 00:09:29,996
Speaker 2: of that that big dream, as I understand it.

170
00:09:30,156 --> 00:09:36,396
Speaker 1: Yeah, yeah, So I think the challenge was that three

171
00:09:36,436 --> 00:09:39,276
Speaker 1: D printing is just one process and it's necessarily not

172
00:09:39,396 --> 00:09:44,276
Speaker 1: good for every type of part. You know, manufacturing is

173
00:09:44,356 --> 00:09:47,036
Speaker 1: very versatile. You do different types of geometries, different types

174
00:09:47,076 --> 00:09:50,036
Speaker 1: of material, and three D printing has a very small reach.

175
00:09:50,196 --> 00:09:53,556
Speaker 1: There's certain type of parts like rocket engines, very good fit.

176
00:09:54,036 --> 00:09:58,316
Speaker 1: You're building a tank, maybe not so right. So yeah

177
00:09:58,356 --> 00:10:00,196
Speaker 1: it's good for certain type of parts, but there as a

178
00:10:00,236 --> 00:10:01,876
Speaker 1: whole lot of other parts. Like I said, you know

179
00:10:01,916 --> 00:10:05,196
Speaker 1: you're building a fuel tank, which is basically large sheet

180
00:10:05,276 --> 00:10:09,516
Speaker 1: metal or thin walled structure, then maybe three D printing

181
00:10:09,556 --> 00:10:11,196
Speaker 1: is not as good as a fit because it takes

182
00:10:11,236 --> 00:10:14,196
Speaker 1: a long time, and also because it's thin, you have

183
00:10:14,236 --> 00:10:16,716
Speaker 1: a lot of physical challenges in terms of controlling the

184
00:10:16,756 --> 00:10:22,916
Speaker 1: geometry and the tolerances. So we realize soon that maybe

185
00:10:23,636 --> 00:10:27,356
Speaker 1: other processes are also need to be automated the same

186
00:10:27,356 --> 00:10:29,916
Speaker 1: way three D printing is. We need to have more

187
00:10:30,236 --> 00:10:34,276
Speaker 1: flexible processes that are not just one process, more flexible platforms.

188
00:10:34,276 --> 00:10:36,756
Speaker 1: They can do different types of processes, not just three

189
00:10:36,836 --> 00:10:41,956
Speaker 1: D printing, to be able to cover a whole variety

190
00:10:41,996 --> 00:10:44,436
Speaker 1: of products in a flexible manner, the same way the

191
00:10:44,436 --> 00:10:47,596
Speaker 1: three D printing that's for certain type of products. And

192
00:10:47,636 --> 00:10:50,476
Speaker 1: that was actually the thinking behind MARKETA Labs is that, okay,

193
00:10:50,636 --> 00:10:53,276
Speaker 1: can we step back and say, what do we need

194
00:10:53,316 --> 00:10:56,956
Speaker 1: to build? What is this flexible platform that can do

195
00:10:57,156 --> 00:10:59,556
Speaker 1: three D printing if needed, or it can do sheet

196
00:10:59,596 --> 00:11:01,996
Speaker 1: forming if it's needed, It can do machining if it's needed,

197
00:11:02,916 --> 00:11:06,476
Speaker 1: but chooses the right operation, right flexible operation for the

198
00:11:06,516 --> 00:11:09,436
Speaker 1: right part, but still very agile and doesn't require a

199
00:11:09,476 --> 00:11:11,236
Speaker 1: lot of tooling and it's not inflexible.

200
00:11:12,276 --> 00:11:14,916
Speaker 2: So it's it's sort of zooming out more. It's saying

201
00:11:15,676 --> 00:11:18,276
Speaker 2: three D printing is not going to do everything the

202
00:11:18,276 --> 00:11:21,476
Speaker 2: way manufacturing works now. It's just too rigid, too hard

203
00:11:21,476 --> 00:11:24,876
Speaker 2: to change things, to rely on on scale to make

204
00:11:24,916 --> 00:11:28,316
Speaker 2: the economics work out. So like that's a very big,

205
00:11:28,596 --> 00:11:33,876
Speaker 2: very abstract thought. To start a company, you got to

206
00:11:33,916 --> 00:11:36,196
Speaker 2: make something or you got to make something that makes

207
00:11:36,236 --> 00:11:39,236
Speaker 2: something like what do you what do you actually do?

208
00:11:39,836 --> 00:11:44,756
Speaker 1: Yeah? So it was interesting, right, you know we actually

209
00:11:44,756 --> 00:11:47,956
Speaker 1: the solution was in our past. Right if you look

210
00:11:47,996 --> 00:11:48,716
Speaker 1: at like.

211
00:11:48,636 --> 00:11:50,636
Speaker 2: The lesson in a movie, it's like the Wizard of

212
00:11:50,636 --> 00:11:52,596
Speaker 2: Oz or something exactly.

213
00:11:53,316 --> 00:11:56,876
Speaker 1: If you look at manufacturing, I mean up to Industrial Revolution,

214
00:11:56,996 --> 00:12:00,716
Speaker 1: it was arts and crafts. Right, it was basically humans

215
00:12:01,276 --> 00:12:03,516
Speaker 1: trying to figure out how to conquer nature, right, Like,

216
00:12:03,636 --> 00:12:06,356
Speaker 1: how am I gonna use my hands? On my brains

217
00:12:06,356 --> 00:12:11,156
Speaker 1: and very few primitive tools to deform a product or

218
00:12:11,396 --> 00:12:16,196
Speaker 1: shape a product from raw material. Right and to this state,

219
00:12:16,396 --> 00:12:19,196
Speaker 1: if you are in a very high mixed manufacturing still

220
00:12:19,236 --> 00:12:22,516
Speaker 1: a lot of that creativity exists. There is a person

221
00:12:22,636 --> 00:12:25,396
Speaker 1: at Space IX. His name is Big John. I don't

222
00:12:25,396 --> 00:12:27,956
Speaker 1: think he's there anymore, but there was this guy. It

223
00:12:28,076 --> 00:12:32,236
Speaker 1: was like, you know, a very skilled maker, a craftsman.

224
00:12:32,396 --> 00:12:35,196
Speaker 1: You could figure out how to use simpler tools to

225
00:12:35,196 --> 00:12:37,756
Speaker 1: build different things in a creative way. Maybe it's not

226
00:12:37,796 --> 00:12:41,396
Speaker 1: a repeatable way like you know a stamping works or

227
00:12:41,436 --> 00:12:44,396
Speaker 1: ejection molding works, but you can be flexible. You can

228
00:12:44,396 --> 00:12:46,036
Speaker 1: do different types of things. You can be creative about

229
00:12:46,076 --> 00:12:47,956
Speaker 1: it and do different type of things. So the inspiration

230
00:12:48,076 --> 00:12:52,036
Speaker 1: came from how actually humans used to do manufacturing but

231
00:12:52,196 --> 00:12:56,476
Speaker 1: realized in order to be flexible, you actually need two components.

232
00:12:56,516 --> 00:13:00,516
Speaker 1: You need intelligence and you need set of simple tools

233
00:13:00,716 --> 00:13:03,156
Speaker 1: with a lot of kinematic freedom. Now you can pick

234
00:13:03,196 --> 00:13:05,236
Speaker 1: up those simple tools, and as long as you have

235
00:13:05,236 --> 00:13:07,916
Speaker 1: the intelligence on how to use tools and what sequence

236
00:13:07,956 --> 00:13:11,156
Speaker 1: and what kind of a process how to use those tools,

237
00:13:11,156 --> 00:13:13,796
Speaker 1: you can actually do a whole variety of projects.

238
00:13:14,156 --> 00:13:16,636
Speaker 2: And so when you say kinematic freedom, you basically mean

239
00:13:16,716 --> 00:13:19,956
Speaker 2: like robot arms that can move in lots of different ways.

240
00:13:20,076 --> 00:13:22,876
Speaker 2: Is that practically what kinematic freedom means in this context?

241
00:13:23,036 --> 00:13:26,436
Speaker 1: Yes, basically can apply these tools in a lot with

242
00:13:26,516 --> 00:13:28,596
Speaker 1: a lot of freedom to the material, right the same

243
00:13:28,596 --> 00:13:31,076
Speaker 1: way humans humans do it, right, you know as a human,

244
00:13:31,676 --> 00:13:33,596
Speaker 1: you know, if you think about it, you can pick

245
00:13:33,676 --> 00:13:36,516
Speaker 1: up a welder and weld something, and then you drop

246
00:13:36,556 --> 00:13:38,156
Speaker 1: the welder, and you pick up a drill and you

247
00:13:38,196 --> 00:13:40,116
Speaker 1: put a hole in it, and you drop to drill

248
00:13:40,436 --> 00:13:43,676
Speaker 1: and you pick up a you know, hammer and maybe

249
00:13:43,756 --> 00:13:46,276
Speaker 1: hammer it into shape. So you actually have a few

250
00:13:46,276 --> 00:13:48,196
Speaker 1: set of tools, but you have a lot of good

251
00:13:48,276 --> 00:13:52,396
Speaker 1: kinematic freedom and most importantly, very creative mind too tells

252
00:13:52,436 --> 00:13:55,236
Speaker 1: you how to apply these tools to the material, so

253
00:13:55,236 --> 00:13:58,516
Speaker 1: they can actually get very complex set of products and

254
00:13:58,876 --> 00:13:59,636
Speaker 1: a lot of diversity.

255
00:13:59,756 --> 00:14:03,076
Speaker 2: So plainly, instead of big John, you want a robot, right,

256
00:14:03,156 --> 00:14:06,916
Speaker 2: That's where that's the kinematic freedom. The tools are kind

257
00:14:06,916 --> 00:14:09,996
Speaker 2: of like old tools, but optimized for the root. And

258
00:14:10,036 --> 00:14:12,996
Speaker 2: then when you say intelligence, that's the one where it's

259
00:14:13,036 --> 00:14:15,476
Speaker 2: like feels more frontier ish, like does that mean like

260
00:14:15,956 --> 00:14:19,236
Speaker 2: clever engineers figuring out how to automate the robots doesn't

261
00:14:19,236 --> 00:14:20,836
Speaker 2: mean AI? Does it mean both?

262
00:14:21,316 --> 00:14:24,076
Speaker 1: Yeah? So I think yeah, you're basically getting to the

263
00:14:24,076 --> 00:14:27,196
Speaker 1: crux of how do you scale it? Right? You need

264
00:14:27,196 --> 00:14:31,476
Speaker 1: to have those three components, and how does the intelligent piece,

265
00:14:31,516 --> 00:14:34,076
Speaker 1: which is the most important piece, comes into play in

266
00:14:34,116 --> 00:14:37,836
Speaker 1: an automated fashion. So early days we started from basic

267
00:14:37,836 --> 00:14:40,556
Speaker 1: intelligence of humans. But then we had a plan to

268
00:14:40,596 --> 00:14:43,836
Speaker 1: capture data and train AI so that you can replace

269
00:14:44,836 --> 00:14:47,196
Speaker 1: the thinking and the creativity that human had to put in.

270
00:14:47,636 --> 00:14:50,356
Speaker 2: What's the first thing you decided to try and build.

271
00:14:50,356 --> 00:14:52,196
Speaker 2: What's the first sort of problem you want to solve?

272
00:14:52,556 --> 00:14:54,796
Speaker 1: Yeah? I think so. I left relativity in twenty eighteen,

273
00:14:55,156 --> 00:14:57,916
Speaker 1: and the idea when I left relativity was there, right.

274
00:14:57,956 --> 00:15:00,476
Speaker 1: I was like, okay, we need to build basically what

275
00:15:00,556 --> 00:15:03,596
Speaker 1: I had in My mom called it robot craftsmen. Robocraftsman,

276
00:15:03,636 --> 00:15:05,316
Speaker 1: we call it a time. How can you build a

277
00:15:05,436 --> 00:15:07,756
Speaker 1: robot system? To your point, you can pick up different tools,

278
00:15:07,756 --> 00:15:09,476
Speaker 1: has the same king mean, but also have to have

279
00:15:09,476 --> 00:15:13,516
Speaker 1: the intelligence. The challenge is you know you said, in

280
00:15:13,596 --> 00:15:15,796
Speaker 1: order to train these robots with AI, you need to

281
00:15:15,796 --> 00:15:18,076
Speaker 1: have a lot of data. And this is not the

282
00:15:18,156 --> 00:15:19,916
Speaker 1: data you can find on internet.

283
00:15:20,116 --> 00:15:23,476
Speaker 2: Right, this is the AI robotics problem, it seems right,

284
00:15:23,556 --> 00:15:26,676
Speaker 2: like unlike with large language models, like that's why we

285
00:15:26,756 --> 00:15:30,036
Speaker 2: have large language models and not AI robots, right because

286
00:15:30,836 --> 00:15:33,476
Speaker 2: because we have the data just sort of randomly sitting

287
00:15:33,476 --> 00:15:35,516
Speaker 2: around on the Internet, and we don't have that physical

288
00:15:35,516 --> 00:15:37,036
Speaker 2: world data for robots.

289
00:15:36,716 --> 00:15:41,036
Speaker 1: Right exactly. So basically the problem narrowed down into Okay,

290
00:15:41,556 --> 00:15:44,476
Speaker 1: how can I generate enough data? How can I create

291
00:15:44,476 --> 00:15:47,236
Speaker 1: a business that has a sustaining way of generating data

292
00:15:47,636 --> 00:15:49,996
Speaker 1: so I can actually build these models, I can build

293
00:15:49,996 --> 00:15:54,156
Speaker 1: this intelligence for these robots. And the thinking was, Okay,

294
00:15:54,876 --> 00:15:58,516
Speaker 1: I need to create a solution that can scale in

295
00:15:58,556 --> 00:16:03,716
Speaker 1: the industry with limited amount of data and some heuristic.

296
00:16:04,876 --> 00:16:07,116
Speaker 1: But then because it's scaling, we can generate a lot

297
00:16:07,116 --> 00:16:09,036
Speaker 1: of data and it starts building AI mods.

298
00:16:09,436 --> 00:16:12,156
Speaker 2: Right. You need a first thing that you can actually

299
00:16:12,236 --> 00:16:15,996
Speaker 2: do before you really have AI, to generate the data

300
00:16:15,996 --> 00:16:16,876
Speaker 2: that will get you to.

301
00:16:17,596 --> 00:16:22,236
Speaker 1: AI exactly exactly. So we're thinking about, Okay, it needs

302
00:16:22,276 --> 00:16:25,676
Speaker 1: to be a large enough market right where we can

303
00:16:25,716 --> 00:16:28,396
Speaker 1: get mass adoption, and we need to solve a problem

304
00:16:28,476 --> 00:16:30,916
Speaker 1: that's big enough it's ten times at least better than

305
00:16:30,916 --> 00:16:34,236
Speaker 1: the current solution so it can actually get adoption, right.

306
00:16:34,756 --> 00:16:36,876
Speaker 2: Meaning you can't just do something as well, you have

307
00:16:36,916 --> 00:16:38,276
Speaker 2: to do it ten times better.

308
00:16:38,876 --> 00:16:42,156
Speaker 1: Yeah, Because I think what we realize is that through

309
00:16:42,516 --> 00:16:45,116
Speaker 1: the last two companies, if something is not ten times better,

310
00:16:45,236 --> 00:16:49,196
Speaker 1: cannot overcome the inertia that exists in an industry for

311
00:16:49,236 --> 00:16:51,796
Speaker 1: adoption because you know, if you're doing something for the

312
00:16:51,916 --> 00:16:54,116
Speaker 1: same way, and in manufacturing, people have been doing things

313
00:16:54,476 --> 00:16:56,716
Speaker 1: the old way for hundred of years, right.

314
00:16:56,676 --> 00:16:58,836
Speaker 2: Yeah, and it's a risk, right if they're going to

315
00:16:58,876 --> 00:17:01,796
Speaker 2: try working with you, they're immediately taking a risk. And

316
00:17:01,836 --> 00:17:03,636
Speaker 2: if it's only going to be a little better, why

317
00:17:03,636 --> 00:17:04,836
Speaker 2: should I take that risk?

318
00:17:05,276 --> 00:17:08,436
Speaker 1: Exactly? So the idea was, Okay, we need to find

319
00:17:08,476 --> 00:17:11,356
Speaker 1: it large enough market for our first application, and we

320
00:17:11,396 --> 00:17:13,436
Speaker 1: need to have a solution that at least ten times better.

321
00:17:13,756 --> 00:17:15,436
Speaker 1: So that landed us. We actually looked at a lot

322
00:17:15,436 --> 00:17:17,876
Speaker 1: of things, from three D printing to forging to a

323
00:17:17,876 --> 00:17:20,036
Speaker 1: lot of things, and then landed on sheet metal. So

324
00:17:20,076 --> 00:17:23,476
Speaker 1: sheet metal is the largest metal processing sector out of

325
00:17:23,476 --> 00:17:26,316
Speaker 1: all It's a two hundred and eighty billion dollar industry today,

326
00:17:27,236 --> 00:17:30,956
Speaker 1: and forming complex sheet metal shapes is very tool intensive.

327
00:17:31,276 --> 00:17:34,956
Speaker 1: So so what we started to do was, okay, can

328
00:17:35,036 --> 00:17:37,876
Speaker 1: we make our robot craftsman's first operation to be forming

329
00:17:37,916 --> 00:17:40,756
Speaker 1: sheet metal, basically forming sheet metal the same way a

330
00:17:40,796 --> 00:17:43,076
Speaker 1: sheet shaper hammer is a sheet into shape.

331
00:17:43,276 --> 00:17:45,516
Speaker 2: And when I think about sheet metal, I mean I

332
00:17:45,556 --> 00:17:47,276
Speaker 2: don't know anything about sheet metal. I think of like

333
00:17:48,036 --> 00:17:50,556
Speaker 2: I think of cars, I think of planes, right, I

334
00:17:50,596 --> 00:17:53,356
Speaker 2: think of like you know, detroit, like stamping.

335
00:17:53,436 --> 00:17:53,676
Speaker 1: Is that?

336
00:17:53,716 --> 00:17:55,876
Speaker 2: Am I thinking about the right thing? Am I missing

337
00:17:55,996 --> 00:17:57,716
Speaker 2: huge or huge sheet metal universe?

338
00:17:57,756 --> 00:17:57,836
Speaker 1: Like?

339
00:17:57,876 --> 00:17:58,996
Speaker 2: What's the sheet metal universe?

340
00:17:59,156 --> 00:18:01,836
Speaker 1: Yes? So sheep metal almost is everywhere. I think is

341
00:18:01,836 --> 00:18:06,356
Speaker 1: the most common metal part that you see on day

342
00:18:06,356 --> 00:18:08,876
Speaker 1: to day, right, because most of the time we use

343
00:18:08,916 --> 00:18:11,756
Speaker 1: metal to be a container for other things. So it's

344
00:18:11,836 --> 00:18:15,396
Speaker 1: usually a thin metal structure that's formed in complex shape

345
00:18:15,676 --> 00:18:17,916
Speaker 1: to hold something else. Now you know it can be

346
00:18:17,956 --> 00:18:23,556
Speaker 1: from case of a computer. Uh, you know, to a car, right,

347
00:18:23,596 --> 00:18:25,276
Speaker 1: you know you're sitting in a freeway you're in to

348
00:18:25,316 --> 00:18:28,156
Speaker 1: see a sheet metal or to a to a airplane

349
00:18:28,196 --> 00:18:31,316
Speaker 1: you're in a sheet metal can to a rocket body.

350
00:18:31,716 --> 00:18:31,836
Speaker 2: Uh.

351
00:18:32,116 --> 00:18:33,876
Speaker 1: For for a lot of rocket someone will composites with

352
00:18:33,876 --> 00:18:39,636
Speaker 1: a lot of a machine metal. And to agricultural heavy

353
00:18:39,636 --> 00:18:45,476
Speaker 1: equipment machinery you think of combines, tractors to even building equipments.

354
00:18:45,516 --> 00:18:50,396
Speaker 1: You look at your h ract ducts are all sheet metal, right,

355
00:18:50,476 --> 00:18:53,316
Speaker 1: because it just makes sense. It's we mostly use metal

356
00:18:53,316 --> 00:18:56,156
Speaker 1: parts to contain other things, and we give it complex

357
00:18:56,196 --> 00:18:58,756
Speaker 1: shapes and that's where she forming comes into play. So

358
00:18:58,796 --> 00:19:01,796
Speaker 1: you pretty much see it everywhere. But the challenge is

359
00:19:01,796 --> 00:19:04,836
Speaker 1: that in almost in all cases, you have to create tooling.

360
00:19:04,876 --> 00:19:06,316
Speaker 1: It goes back to that first problem. He said, you

361
00:19:06,396 --> 00:19:09,476
Speaker 1: have to create tooling for each of those geometries. And

362
00:19:09,476 --> 00:19:12,996
Speaker 1: that's why you know a Ford needs to make sure

363
00:19:13,116 --> 00:19:15,796
Speaker 1: they can sell a million of an FUN fifty before

364
00:19:15,796 --> 00:19:18,036
Speaker 1: they can invest in a plant that makes a new

365
00:19:18,116 --> 00:19:19,676
Speaker 1: version of FN fifty, right.

366
00:19:19,556 --> 00:19:24,356
Speaker 2: Because you basically have to build a bespoke factory just

367
00:19:24,436 --> 00:19:28,236
Speaker 2: to shape sheet metal in a new way exactly for

368
00:19:28,276 --> 00:19:30,836
Speaker 2: a new geometry, for for a new design. Exactly where

369
00:19:30,876 --> 00:19:34,356
Speaker 2: is that a particular problem? Like where is it? Where

370
00:19:34,396 --> 00:19:36,916
Speaker 2: is that? Where does that acutely bind the fact that

371
00:19:37,156 --> 00:19:39,236
Speaker 2: sheet metal is so hard to do if you're not

372
00:19:39,276 --> 00:19:39,756
Speaker 2: working at.

373
00:19:39,716 --> 00:19:42,716
Speaker 1: Scale so expensive? Yeah, so I think now you're coming

374
00:19:42,716 --> 00:19:45,876
Speaker 1: to the even the third stage of how do you

375
00:19:45,916 --> 00:19:48,476
Speaker 1: scale this technology? You need to first find you know,

376
00:19:48,476 --> 00:19:49,796
Speaker 1: you said you need to be ten ex better we

377
00:19:49,836 --> 00:19:53,036
Speaker 1: need Right, you're in an area that has a lot

378
00:19:53,076 --> 00:19:54,236
Speaker 1: of pain with today's time.

379
00:19:54,316 --> 00:19:56,436
Speaker 2: I was like, oh my god, thank god, you've walked

380
00:19:56,476 --> 00:19:58,236
Speaker 2: through the door. We've been waiting for you.

381
00:19:58,636 --> 00:20:02,436
Speaker 1: Yeah, So end up being very much defense in airspace. Right,

382
00:20:02,516 --> 00:20:06,756
Speaker 1: So think of you know, think of our military for example,

383
00:20:07,196 --> 00:20:12,836
Speaker 1: right today, they we have fifty sixty different weapon system

384
00:20:13,116 --> 00:20:15,956
Speaker 1: or defense systems you can basically think of like aircrafts

385
00:20:17,236 --> 00:20:19,916
Speaker 1: that they're maintaining. And some of these systems have been

386
00:20:19,916 --> 00:20:23,036
Speaker 1: built from sixty seventy eighty years ago, like think of

387
00:20:23,116 --> 00:20:26,156
Speaker 1: B fifty two C one thirty like World War.

388
00:20:25,996 --> 00:20:27,916
Speaker 2: Two planes still flying kind.

389
00:20:27,756 --> 00:20:31,836
Speaker 1: Of still flying, yes, exactly, and they have like you know,

390
00:20:31,996 --> 00:20:35,076
Speaker 1: thirty of one, fifty of another, one hundred of another one.

391
00:20:35,396 --> 00:20:38,076
Speaker 1: And these things get break down, right, and unlike a

392
00:20:38,116 --> 00:20:41,996
Speaker 1: Ford factory, there is no factory for seventy different products

393
00:20:42,036 --> 00:20:43,156
Speaker 1: that they're carrying.

394
00:20:42,876 --> 00:20:46,156
Speaker 2: Right, and presumably the factory they built in nineteen forty

395
00:20:46,196 --> 00:20:48,076
Speaker 2: one to build this plane doesn't.

396
00:20:47,796 --> 00:20:50,716
Speaker 1: Exist any It doesn't exact. Even the vendor might completely

397
00:20:50,756 --> 00:20:54,356
Speaker 1: have disappeared, right, that made that misspecific component. So they're

398
00:20:54,396 --> 00:20:57,796
Speaker 1: constantly battling with this challenge of an aircraft goes down,

399
00:20:58,036 --> 00:20:59,876
Speaker 1: how can I fix it? How can I find the part?

400
00:20:59,916 --> 00:21:02,276
Speaker 1: And there are thousands of parts in each of these aircrafts, right,

401
00:21:02,316 --> 00:21:04,436
Speaker 1: so any of them can go down, and that's a

402
00:21:04,516 --> 00:21:06,836
Speaker 1: huge challenge. I mean, if you look at you know,

403
00:21:06,916 --> 00:21:11,436
Speaker 1: government of a government accountabilit The office put this report out,

404
00:21:11,756 --> 00:21:13,116
Speaker 1: I think it was a couple of years ago or

405
00:21:13,156 --> 00:21:16,676
Speaker 1: a year ago about how ready each weapon system is

406
00:21:16,716 --> 00:21:19,436
Speaker 1: to defend the United States. Out of the forty eight

407
00:21:19,476 --> 00:21:22,836
Speaker 1: to forty nine weapon systems they look into only one,

408
00:21:23,196 --> 00:21:28,196
Speaker 1: only one in the past eleven years, every year was ready, right.

409
00:21:29,076 --> 00:21:31,996
Speaker 1: I think only top four had like at least half

410
00:21:32,076 --> 00:21:35,236
Speaker 1: of the years ready right. So that means in most

411
00:21:35,316 --> 00:21:38,036
Speaker 1: years these weapons are not ready to fight, like.

412
00:21:37,996 --> 00:21:39,156
Speaker 2: They're waiting for parts.

413
00:21:39,196 --> 00:21:42,236
Speaker 1: They're waiting for parts. Something is broken, something is damaged,

414
00:21:42,596 --> 00:21:44,756
Speaker 1: and we used to go deeper. Some of these components

415
00:21:44,756 --> 00:21:47,156
Speaker 1: take four years to be replaced. So if a plane

416
00:21:47,156 --> 00:21:49,636
Speaker 1: gets damaged, it needs to sit on the ground for

417
00:21:49,676 --> 00:21:51,996
Speaker 1: four years before it can be it can be replaced,

418
00:21:52,076 --> 00:21:55,196
Speaker 1: and the cost of replacement is building another factory basically,

419
00:21:55,436 --> 00:21:57,396
Speaker 1: So some of these parts, and think of it, a

420
00:21:57,476 --> 00:22:00,316
Speaker 1: landing gear door that goes on a plane will cost

421
00:22:00,316 --> 00:22:02,556
Speaker 1: them eight hundred thousand dollars for example, because they have.

422
00:22:02,516 --> 00:22:05,796
Speaker 2: To go make it because it's bespoke essentially, like buying

423
00:22:05,836 --> 00:22:08,316
Speaker 2: a bespoke suit or something. It's just like it's gonna

424
00:22:08,356 --> 00:22:08,796
Speaker 2: cost a lot.

425
00:22:08,876 --> 00:22:11,556
Speaker 1: Yeah, yeah, So the idea started there. I think that

426
00:22:11,636 --> 00:22:14,596
Speaker 1: was one of our first customers. Can we make defense

427
00:22:14,636 --> 00:22:19,036
Speaker 1: manufacturing more agile? Directly affects our national readiness for military

428
00:22:19,076 --> 00:22:22,156
Speaker 1: conflict and it's a huge problem. But then you know,

429
00:22:22,276 --> 00:22:24,996
Speaker 1: even in a broader sense, any defense product or aerospace

430
00:22:25,076 --> 00:22:28,916
Speaker 1: product usually has very low volume but high mix of products.

431
00:22:29,196 --> 00:22:31,516
Speaker 1: You know, even you know, you're building a missile, you

432
00:22:31,556 --> 00:22:33,396
Speaker 1: make like, you know, a few thousand a year, and

433
00:22:33,436 --> 00:22:35,796
Speaker 1: you might make five, six seven different versions of right,

434
00:22:35,796 --> 00:22:38,396
Speaker 1: So it's very unlike cars, where you know, you make

435
00:22:38,436 --> 00:22:40,156
Speaker 1: a million of the same car over and over all.

436
00:22:40,716 --> 00:22:44,116
Speaker 1: So that ends up being our first application, which we've

437
00:22:44,116 --> 00:22:47,636
Speaker 1: got a lot of traction with. But but you know,

438
00:22:47,676 --> 00:22:50,156
Speaker 1: even outside of that, you know, you look at companies

439
00:22:50,316 --> 00:22:53,796
Speaker 1: like Caterpillar, like John Deere's of the world. These folks

440
00:22:53,836 --> 00:22:55,556
Speaker 1: also are in the same book. You know, they make

441
00:22:55,756 --> 00:22:58,676
Speaker 1: two hundred combines, right, but they need to support them

442
00:22:58,676 --> 00:23:01,316
Speaker 1: in the field. And these folks have the exactly same problem, right,

443
00:23:01,396 --> 00:23:05,396
Speaker 1: you know, do I need to run a large factory

444
00:23:05,956 --> 00:23:08,156
Speaker 1: to support all these models at all time, and that's

445
00:23:08,196 --> 00:23:11,716
Speaker 1: will be very expensive to support, like one hundred vehicles out.

446
00:23:11,556 --> 00:23:17,436
Speaker 2: There still to come on the show. We'll talk about

447
00:23:17,436 --> 00:23:20,836
Speaker 2: the future of AI and robotics at Mocking the Labs

448
00:23:20,876 --> 00:23:33,396
Speaker 2: and beyond. And so you got the right market. Now

449
00:23:33,396 --> 00:23:35,556
Speaker 2: you've got to make a thing. You got to figure

450
00:23:35,556 --> 00:23:37,916
Speaker 2: out how to actually do the thing, how to make

451
00:23:37,916 --> 00:23:40,956
Speaker 2: your idea come true? Like how does that work?

452
00:23:41,436 --> 00:23:44,796
Speaker 1: So the idea originally was can we get rid of

453
00:23:44,796 --> 00:23:47,436
Speaker 1: a die right and do it the same way a

454
00:23:47,516 --> 00:23:50,876
Speaker 1: sheet shaper forms a sheet of metal? And what does

455
00:23:50,876 --> 00:23:52,876
Speaker 1: a sheet chaper do? Sheet chaper get starts from a

456
00:23:52,916 --> 00:23:56,516
Speaker 1: flat sheet of metal and it slowly hammers it into shape.

457
00:23:57,116 --> 00:23:59,076
Speaker 1: So what we wanted to do was have a robot

458
00:23:59,196 --> 00:24:03,116
Speaker 1: do that, right, have robotic system basically do that incremental

459
00:24:03,156 --> 00:24:04,956
Speaker 1: deformation into shape. We call it romophor.

460
00:24:05,716 --> 00:24:07,516
Speaker 2: So you're sort of bending it, right, I mean you're

461
00:24:07,516 --> 00:24:09,996
Speaker 2: hammering it. It's sort of like if you take a whatever,

462
00:24:10,076 --> 00:24:11,916
Speaker 2: cut open a luminum can and kind of bend it

463
00:24:11,956 --> 00:24:15,676
Speaker 2: into shape. Like that's a version of what's happening here, right, Yes,

464
00:24:16,116 --> 00:24:17,996
Speaker 2: exactly complicated way, yeah.

465
00:24:17,716 --> 00:24:20,996
Speaker 1: Exactly, you're right, I mean the same way a potter

466
00:24:22,036 --> 00:24:24,436
Speaker 1: forms a clay bowl. That's basically what our robots do.

467
00:24:24,476 --> 00:24:26,796
Speaker 1: They start from a flash sheet of metal and slowly

468
00:24:26,836 --> 00:24:28,876
Speaker 1: deforming in the shape the same way a potter, which

469
00:24:28,996 --> 00:24:31,556
Speaker 1: form it clay bowl or a sheet shape of hammets

470
00:24:31,556 --> 00:24:33,876
Speaker 1: and hammers a sheet into shape. Yeah, so I've seen

471
00:24:33,916 --> 00:24:34,116
Speaker 1: it right.

472
00:24:34,156 --> 00:24:35,876
Speaker 2: So there will be a sheet of metal like hanging

473
00:24:36,076 --> 00:24:39,156
Speaker 2: hanging up in whatever above the ground. And then you

474
00:24:39,156 --> 00:24:42,316
Speaker 2: have a robot arm on either side, right, like one

475
00:24:42,356 --> 00:24:44,916
Speaker 2: on one side, one on the other. And then what.

476
00:24:44,956 --> 00:24:48,036
Speaker 1: Happens Basically the robots come together from both sides the

477
00:24:48,036 --> 00:24:51,996
Speaker 1: sheet and they pinch the sheet in a certain way

478
00:24:52,196 --> 00:24:57,636
Speaker 1: so that that location that they're pinching slightly stretches and deforms. Right.

479
00:24:57,876 --> 00:25:02,636
Speaker 1: And if you start applying this pinching all over the

480
00:25:02,636 --> 00:25:05,796
Speaker 1: sheet and incrementally, you slowly start to form it into

481
00:25:05,836 --> 00:25:09,116
Speaker 1: a shape. Right. So instead of traditionally would use it

482
00:25:09,276 --> 00:25:12,236
Speaker 1: die and with sheer pressure of the press pushing the

483
00:25:12,236 --> 00:25:14,116
Speaker 1: sheet against the dye to give it a shape. Now

484
00:25:14,116 --> 00:25:17,196
Speaker 1: the robots are like a craftsman, like a trades person

485
00:25:17,516 --> 00:25:20,716
Speaker 1: coming in to slowly deform the sheet into shape by

486
00:25:20,796 --> 00:25:23,116
Speaker 1: just applying pressure. So one robot is pushing it the

487
00:25:23,156 --> 00:25:26,756
Speaker 1: other robot is supporting it, and by applying a pinch

488
00:25:26,836 --> 00:25:29,916
Speaker 1: you slightly stretch the material and you form it into

489
00:25:29,956 --> 00:25:30,476
Speaker 1: a shape.

490
00:25:30,716 --> 00:25:33,876
Speaker 2: So I mean, the way you describe it, it makes

491
00:25:33,956 --> 00:25:38,996
Speaker 2: sense and it sounds easy. I'm sure it wasn't easy,

492
00:25:39,276 --> 00:25:43,876
Speaker 2: Like were there things that just didn't work for a while.

493
00:25:44,236 --> 00:25:47,196
Speaker 1: So you should have been here when the first time

494
00:25:47,236 --> 00:25:50,236
Speaker 1: we actually tried to form a part, the part looked

495
00:25:50,316 --> 00:25:53,156
Speaker 1: like it was like a ghost of the geometry that

496
00:25:53,196 --> 00:25:55,076
Speaker 1: they wanted to make, and actually in the end it

497
00:25:55,156 --> 00:25:58,556
Speaker 1: tore right. So think about it. You have this very

498
00:25:58,596 --> 00:26:03,036
Speaker 1: flimsy sheet applying pressure to it, and if features apply

499
00:26:03,116 --> 00:26:08,036
Speaker 1: pressure slightly wrong right, it can potentially tear it. It

500
00:26:08,076 --> 00:26:10,356
Speaker 1: can form it into a different shape. And also the

501
00:26:10,396 --> 00:26:12,556
Speaker 1: whole sheet is moving the whole time you're trying to

502
00:26:12,596 --> 00:26:15,316
Speaker 1: move to form it. The whole sheet is moving because

503
00:26:15,316 --> 00:26:18,636
Speaker 1: it's very flimsy. It's not a rigid structure, right. So

504
00:26:19,276 --> 00:26:22,116
Speaker 1: the main challenge was how do you get this accurate? Right?

505
00:26:22,196 --> 00:26:23,676
Speaker 1: How do you get this process accurate? How do you

506
00:26:23,676 --> 00:26:27,116
Speaker 1: get accuracy? And the idea was what does the robot

507
00:26:27,156 --> 00:26:30,436
Speaker 1: need to do given all of these chaotic nature of

508
00:26:30,476 --> 00:26:32,436
Speaker 1: the process where the sheet moves and if you apply

509
00:26:32,476 --> 00:26:35,116
Speaker 1: it too much pressure, it will deform in a bad

510
00:26:35,116 --> 00:26:38,276
Speaker 1: way or in my tear. If you're probably not enough pressure,

511
00:26:38,316 --> 00:26:40,916
Speaker 1: it might just not form. So how do you come

512
00:26:40,996 --> 00:26:43,436
Speaker 1: up with the right set of robot movements and process

513
00:26:43,436 --> 00:26:46,556
Speaker 1: parameters to form the part? And that was the problem

514
00:26:46,596 --> 00:26:48,956
Speaker 1: we want to solve with AI right, but we didn't

515
00:26:48,996 --> 00:26:51,676
Speaker 1: have the data right right in the beginning. Right. The

516
00:26:52,036 --> 00:26:56,076
Speaker 1: idea was that if I form enough parts with this process,

517
00:26:56,196 --> 00:26:58,796
Speaker 1: and I can capture all the data throughout the process,

518
00:26:58,796 --> 00:27:01,356
Speaker 1: where did the robot go, how much pressure did it apply,

519
00:27:01,636 --> 00:27:04,796
Speaker 1: and what was the resulting geometry? That can start building

520
00:27:04,796 --> 00:27:07,676
Speaker 1: a model that says that correlates the inputs to the outputs,

521
00:27:07,996 --> 00:27:09,996
Speaker 1: and I can explore this and say, okay, in order

522
00:27:10,036 --> 00:27:12,436
Speaker 1: to get to the right output, I need these inputs.

523
00:27:12,836 --> 00:27:15,076
Speaker 1: But we didn't have them in the beginning. So the

524
00:27:15,116 --> 00:27:18,076
Speaker 1: idea was two things. One was maybe we can simulate

525
00:27:18,116 --> 00:27:22,156
Speaker 1: the data right and very early on we started doing

526
00:27:22,156 --> 00:27:24,876
Speaker 1: some simulation, physics based simulation, and we soon realized in

527
00:27:24,956 --> 00:27:28,076
Speaker 1: order to get an accurate result, the simulations are going

528
00:27:28,156 --> 00:27:31,956
Speaker 1: to be very computationally intensive. A simulation of a part

529
00:27:31,996 --> 00:27:35,636
Speaker 1: that took only fifteen minutes to form took us one

530
00:27:35,676 --> 00:27:39,236
Speaker 1: week on twenty seven core machine. Wow. Right, so okay,

531
00:27:39,476 --> 00:27:42,996
Speaker 1: stimulation not only is not accurate, it takes forever. So

532
00:27:43,036 --> 00:27:46,076
Speaker 1: we realized, okay, so that's not the right route. The

533
00:27:46,156 --> 00:27:48,316
Speaker 1: right route was like, Okay, we can also form a

534
00:27:48,356 --> 00:27:51,516
Speaker 1: lot of parts and gather the data. But in order

535
00:27:51,556 --> 00:27:53,236
Speaker 1: to do that we go back to that same problem.

536
00:27:53,276 --> 00:27:54,876
Speaker 1: We need to have a scale. We need to have

537
00:27:54,916 --> 00:27:56,836
Speaker 1: a lot of these machines for these parts and get

538
00:27:56,836 --> 00:27:57,276
Speaker 1: that data.

539
00:27:57,316 --> 00:27:59,716
Speaker 2: I mean, one of the big AI insights of the

540
00:27:59,796 --> 00:28:03,556
Speaker 2: last whatever decade is like, you need a ton of data,

541
00:28:03,596 --> 00:28:06,996
Speaker 2: which is easy if it's words, but hard if it's metal. Right.

542
00:28:07,116 --> 00:28:11,596
Speaker 1: Yes, we ended up doing was created a hybrid model.

543
00:28:11,836 --> 00:28:15,276
Speaker 1: We said, okay, what if we keep the humans in

544
00:28:15,316 --> 00:28:18,196
Speaker 1: the loop, so the human can give an instruction initially

545
00:28:18,236 --> 00:28:21,836
Speaker 1: based on herostricts, and then we look at the data

546
00:28:21,916 --> 00:28:25,636
Speaker 1: and human can adjust and then iterate on that. But

547
00:28:25,796 --> 00:28:29,036
Speaker 1: while we are capturing all these data, and over time,

548
00:28:29,196 --> 00:28:31,556
Speaker 1: as we're capturing the data, we start building the models

549
00:28:31,876 --> 00:28:35,716
Speaker 1: that will help the human do less trials. Right. It's

550
00:28:35,716 --> 00:28:38,516
Speaker 1: basically guided reinforcement learning, right, and a humans are actually

551
00:28:38,516 --> 00:28:41,436
Speaker 1: guiding it where to go, but it's exploring those areas

552
00:28:41,516 --> 00:28:43,916
Speaker 1: but after a while, once we started forming south thousands

553
00:28:43,916 --> 00:28:47,236
Speaker 1: of parts, then you can start feeding this data into model.

554
00:28:47,236 --> 00:28:50,756
Speaker 1: Then the model will be like, okay, human, you don't

555
00:28:50,756 --> 00:28:52,676
Speaker 1: need to do twenty five different trials. Now you can

556
00:28:52,716 --> 00:28:54,236
Speaker 1: do with five trials, you're going to get to the

557
00:28:54,276 --> 00:28:56,356
Speaker 1: right place, which is actually the number we are at

558
00:28:56,436 --> 00:28:56,876
Speaker 1: right now.

559
00:28:56,956 --> 00:29:00,396
Speaker 2: And that's happening in the physical world largely those iterations

560
00:29:00,436 --> 00:29:02,636
Speaker 2: like you're trying a piece of metal and it's bad

561
00:29:02,676 --> 00:29:04,516
Speaker 2: and it tears, and you do another piece of metal

562
00:29:04,516 --> 00:29:06,556
Speaker 2: and it's a little less bad and eventually.

563
00:29:06,196 --> 00:29:09,836
Speaker 1: Exactly exactly, and that initially would take twenty five parts,

564
00:29:10,476 --> 00:29:13,156
Speaker 1: like you know, before we find a recipe for that design.

565
00:29:13,716 --> 00:29:16,396
Speaker 1: But twenty five parts still was better than traditional alternative.

566
00:29:16,516 --> 00:29:18,356
Speaker 2: When you say twenty five parts, I mean twenty five

567
00:29:18,396 --> 00:29:21,196
Speaker 2: tries twenty five pieces of metal before you make the

568
00:29:21,236 --> 00:29:22,916
Speaker 2: part the right way exactly.

569
00:29:23,356 --> 00:29:25,556
Speaker 1: And that was like, you know, they would sit down

570
00:29:25,596 --> 00:29:27,316
Speaker 1: basically twenty five days in a row, so in a

571
00:29:27,356 --> 00:29:30,796
Speaker 1: month they could actually define a recipe where traditionally making

572
00:29:30,796 --> 00:29:32,716
Speaker 1: a mold would take at least three four months. Right,

573
00:29:32,756 --> 00:29:36,156
Speaker 1: So we were still better. But then now with over time,

574
00:29:36,196 --> 00:29:38,716
Speaker 1: when we generated the data and now the model can

575
00:29:38,796 --> 00:29:43,116
Speaker 1: tell the engineer, okay, maybe you want to choose these parameters,

576
00:29:43,276 --> 00:29:46,396
Speaker 1: is now becoming an advisor with down to five trials.

577
00:29:46,556 --> 00:29:48,436
Speaker 1: In five trials, we can actually get to the right

578
00:29:48,476 --> 00:29:50,756
Speaker 1: part and then hopefully in the future we get to

579
00:29:50,796 --> 00:29:53,076
Speaker 1: a point where you know, the machine will tell the

580
00:29:53,156 --> 00:29:55,316
Speaker 1: romans what to do and the human can be completely

581
00:29:55,316 --> 00:29:57,676
Speaker 1: out of the loop. Yeah. But the idea was like,

582
00:29:57,716 --> 00:29:59,556
Speaker 1: how do you kind of create that hybrid model that's

583
00:29:59,556 --> 00:30:03,116
Speaker 1: efficient so that we can generate the data until the

584
00:30:03,156 --> 00:30:05,596
Speaker 1: model is good enough to do the job itself.

585
00:30:05,956 --> 00:30:09,076
Speaker 2: And you find that the data is sort of generalizable,

586
00:30:09,436 --> 00:30:11,996
Speaker 2: I mean clearly, like making one kind of part makes

587
00:30:12,076 --> 00:30:16,076
Speaker 2: the model the AI smarter about making another kind of part.

588
00:30:16,516 --> 00:30:18,956
Speaker 1: Yes, you know, yeah it is. It's kind of interesting.

589
00:30:18,996 --> 00:30:21,236
Speaker 1: I think people don't think about it. I used to

590
00:30:21,316 --> 00:30:23,796
Speaker 1: do sheet shaping by hand, right, That was one of

591
00:30:23,796 --> 00:30:25,956
Speaker 1: the hobbies I had. I was working with this shop

592
00:30:25,956 --> 00:30:29,036
Speaker 1: in Pomona that we were actually hammer sheets into shape,

593
00:30:29,356 --> 00:30:31,036
Speaker 1: and we used to say, you know, if you spent

594
00:30:31,156 --> 00:30:33,516
Speaker 1: five years doing it, you're really good. You get really

595
00:30:33,516 --> 00:30:36,236
Speaker 1: good at it. I was used to think, you know, okay,

596
00:30:36,236 --> 00:30:38,356
Speaker 1: after five years of doing this, yes, you have this

597
00:30:38,396 --> 00:30:40,716
Speaker 1: intuitive understanding of you look at the sheet and be like, okay,

598
00:30:40,756 --> 00:30:42,596
Speaker 1: this this place needs to be hammered more. This place

599
00:30:42,636 --> 00:30:44,836
Speaker 1: needs to be hammered. It was, it was, it was.

600
00:30:44,876 --> 00:30:46,876
Speaker 1: It was intuitive. It was like you couldn't explains why

601
00:30:46,876 --> 00:30:49,636
Speaker 1: you're thinking this need to happen. There was no physical explanation.

602
00:30:49,996 --> 00:30:52,356
Speaker 1: None of these people who were she shaping got PhDs

603
00:30:52,356 --> 00:30:55,236
Speaker 1: in material science. Yeah, they just learned over time seeing

604
00:30:55,236 --> 00:30:58,436
Speaker 1: the pattern of how the sheet formed. Yes, craftsmanship, that's

605
00:30:58,516 --> 00:31:02,036
Speaker 1: craftsmanship right. Yeah, but really reminded me of Okay, these

606
00:31:02,036 --> 00:31:04,716
Speaker 1: people can know how to do it, but without really

607
00:31:04,716 --> 00:31:07,276
Speaker 1: being able to explain it, to do it for five years.

608
00:31:07,316 --> 00:31:09,076
Speaker 2: It's that kind of tacit knowledge.

609
00:31:09,316 --> 00:31:13,236
Speaker 1: Yeah, and reminded me of the same challenge we had

610
00:31:14,116 --> 00:31:16,716
Speaker 1: early machine learning challenge where they were like, okay, a

611
00:31:16,796 --> 00:31:19,476
Speaker 1: human can look at two pictures and say, okay, this

612
00:31:19,556 --> 00:31:21,916
Speaker 1: is a cat and this is a dog. Something happens

613
00:31:21,916 --> 00:31:23,436
Speaker 1: in their brain that knows which is a cat, but

614
00:31:23,436 --> 00:31:25,836
Speaker 1: they cannot really define why they're calling this cat and

615
00:31:25,916 --> 00:31:28,036
Speaker 1: this sort of dog. So that was where it starts

616
00:31:28,036 --> 00:31:29,996
Speaker 1: to click for me. If I can capture enough data,

617
00:31:30,356 --> 00:31:34,476
Speaker 1: five years worth of data right of a human, then

618
00:31:34,556 --> 00:31:36,556
Speaker 1: I should be able to get to a very good

619
00:31:36,636 --> 00:31:40,036
Speaker 1: sheet shaper, right, And you know it's funny. Back at

620
00:31:40,036 --> 00:31:42,116
Speaker 1: the end, I was like, okay, humans are you know,

621
00:31:42,156 --> 00:31:45,076
Speaker 1: receiving x amount of megabytes a second? Okay, how five

622
00:31:45,116 --> 00:31:47,796
Speaker 1: years worse of data? Is that much? So roughly, I

623
00:31:47,796 --> 00:31:50,236
Speaker 1: think once we get a certain amount of data, I

624
00:31:50,236 --> 00:31:53,636
Speaker 1: think we have enough data to be able to basically

625
00:31:53,676 --> 00:31:56,596
Speaker 1: replace a like not replace replace the mentality or the

626
00:31:56,636 --> 00:31:58,316
Speaker 1: model that the sheet shaper has in their mind.

627
00:31:59,236 --> 00:32:02,716
Speaker 2: So how how how many years of kind of human

628
00:32:02,956 --> 00:32:06,556
Speaker 2: level craftsmen sheet shaping data does the model have at

629
00:32:06,556 --> 00:32:07,076
Speaker 2: this point?

630
00:32:07,316 --> 00:32:09,556
Speaker 1: Yeah? No, so I think rastam I check? One year ago?

631
00:32:09,596 --> 00:32:11,436
Speaker 1: I checked around, we were like three fourths of the

632
00:32:11,436 --> 00:32:13,076
Speaker 1: way there in terms of the data that we have

633
00:32:14,236 --> 00:32:16,036
Speaker 1: for just she shaping. Right. So once we get to

634
00:32:16,276 --> 00:32:17,996
Speaker 1: I think full, I think and these at that point

635
00:32:18,036 --> 00:32:19,916
Speaker 1: we have no excuse. We have enough data. The model

636
00:32:19,916 --> 00:32:21,636
Speaker 1: should be good. We just need to figure out how

637
00:32:22,476 --> 00:32:24,516
Speaker 1: why it's not. Maybe far from it is.

638
00:32:24,876 --> 00:32:29,836
Speaker 2: It is interesting to analogize it to like human craftsmanship, right,

639
00:32:29,876 --> 00:32:31,676
Speaker 2: And I mean even if you want to zoom out

640
00:32:31,716 --> 00:32:34,756
Speaker 2: even more, the like fifty year history of AI, where

641
00:32:34,796 --> 00:32:36,716
Speaker 2: first everybody was like, oh, you just got to teach

642
00:32:36,716 --> 00:32:39,396
Speaker 2: the machine all the rules for to use your example,

643
00:32:39,476 --> 00:32:41,076
Speaker 2: like what's a cat and what's a dog? But then

644
00:32:41,116 --> 00:32:43,916
Speaker 2: you realize it's actually wildly hard to make a list

645
00:32:43,956 --> 00:32:47,436
Speaker 2: of rules that can reliably distinguish a cat from a dog.

646
00:32:48,116 --> 00:32:51,956
Speaker 2: And the weird thing that has happened in AI is like, oh,

647
00:32:51,996 --> 00:32:55,756
Speaker 2: you don't actually have to make a list. You just

648
00:32:55,836 --> 00:32:58,596
Speaker 2: need like image that you just need like a giant

649
00:32:58,636 --> 00:33:03,156
Speaker 2: database of images and a giant neural network and you

650
00:33:03,276 --> 00:33:06,276
Speaker 2: just throw it at it like and say figure it out,

651
00:33:06,316 --> 00:33:09,036
Speaker 2: and it figures it out, and you're sort of doing that.

652
00:33:09,436 --> 00:33:11,396
Speaker 2: But for shaping metal.

653
00:33:11,356 --> 00:33:13,436
Speaker 1: For metal, and then the only challenge was, like you know,

654
00:33:13,636 --> 00:33:17,636
Speaker 1: cats and dogs pictures were Internet and sheet metal forming

655
00:33:17,716 --> 00:33:20,196
Speaker 1: data wasn't. And so that's that was an additional problem

656
00:33:20,236 --> 00:33:21,636
Speaker 1: we have to solve, as you pointed out, which is

657
00:33:21,676 --> 00:33:23,196
Speaker 1: a big problem in physical AI.

658
00:33:23,516 --> 00:33:25,116
Speaker 2: So I want to talk a little bit more about

659
00:33:26,236 --> 00:33:31,596
Speaker 2: AI and robotics. Jansen Wong has been talking about it,

660
00:33:31,676 --> 00:33:34,796
Speaker 2: as I'm sure you know in video and videos vc

661
00:33:35,116 --> 00:33:38,196
Speaker 2: ARM as an investor in your company, other people are

662
00:33:38,236 --> 00:33:40,356
Speaker 2: working on what you're working on. I mean, I'm curious

663
00:33:41,076 --> 00:33:43,356
Speaker 2: what does the sort of AI and robotics path look

664
00:33:43,476 --> 00:33:45,396
Speaker 2: like to you? For the next few years, and what like,

665
00:33:45,916 --> 00:33:48,116
Speaker 2: what do you understand about it now that you didn't

666
00:33:48,196 --> 00:33:51,036
Speaker 2: understand whatever five years ago? Like what what have you

667
00:33:51,116 --> 00:33:53,916
Speaker 2: really come to realize by working on it all the time?

668
00:33:54,516 --> 00:33:57,396
Speaker 1: I think the biggest problem for physical AI is data

669
00:33:57,436 --> 00:34:01,436
Speaker 1: generation for now to train models. So we need to

670
00:34:01,596 --> 00:34:04,076
Speaker 1: either there's two things need to happen. Either new types

671
00:34:04,116 --> 00:34:07,436
Speaker 1: of models needs to be created, new architectures, new new

672
00:34:07,476 --> 00:34:10,996
Speaker 1: algorithms basically, which I'm sure it's going to happen that

673
00:34:11,116 --> 00:34:15,476
Speaker 1: can learn more with less data basically and the same

674
00:34:15,516 --> 00:34:17,636
Speaker 1: way humans kind of learn more with less data. Right.

675
00:34:18,156 --> 00:34:20,916
Speaker 1: But at the same time, I think, you know, we

676
00:34:21,156 --> 00:34:25,356
Speaker 1: only exposed our models to categorically to ten percent of

677
00:34:25,516 --> 00:34:28,756
Speaker 1: type of data that humans receive. You know, you think

678
00:34:28,796 --> 00:34:32,476
Speaker 1: about you know, human intrictions. You and I are now talking,

679
00:34:33,196 --> 00:34:35,356
Speaker 1: if it was AI, AI is probably only listening to

680
00:34:35,396 --> 00:34:38,716
Speaker 1: the words we're saying, right, But that's only ten percent

681
00:34:38,716 --> 00:34:41,076
Speaker 1: of communication. I can see your lips moving, I can

682
00:34:41,116 --> 00:34:43,596
Speaker 1: see your eyebrows moving. I can see like maybe you're

683
00:34:43,596 --> 00:34:45,876
Speaker 1: folding your arms and okay, I know that like okay,

684
00:34:45,876 --> 00:34:48,596
Speaker 1: maybe there's all these ninety percent of the signals are

685
00:34:48,596 --> 00:34:52,116
Speaker 1: not captured. That that's used for learning. You know, you

686
00:34:52,156 --> 00:34:56,316
Speaker 1: look at if you ask chat GPT or Dolly or

687
00:34:56,596 --> 00:34:58,636
Speaker 1: you know any of the you know, even even you

688
00:34:58,676 --> 00:35:02,636
Speaker 1: know Grock say okay, draw me a clock that is

689
00:35:02,796 --> 00:35:06,196
Speaker 1: shows five thirty. It cannot show you draw you a clock.

690
00:35:06,236 --> 00:35:07,836
Speaker 1: It will draw you a clock, but it doesn't show

691
00:35:07,876 --> 00:35:10,516
Speaker 1: five thirty. Actually, most the time it shows ten ten.

692
00:35:10,876 --> 00:35:14,436
Speaker 2: Ten ten, because that's where watchhands, like analog watchhands look good.

693
00:35:14,516 --> 00:35:17,196
Speaker 1: Right, it's a nice little v because those are all

694
00:35:17,236 --> 00:35:19,516
Speaker 1: the images that they're seen on internet because they watch it.

695
00:35:19,516 --> 00:35:22,156
Speaker 2: It's almost always ten ten. It's the classic watch photo.

696
00:35:22,556 --> 00:35:24,636
Speaker 1: It's like five thirty is also ten ten, because.

697
00:35:24,516 --> 00:35:28,956
Speaker 2: It's always ten ten right to a generative AI, it's

698
00:35:28,996 --> 00:35:30,396
Speaker 2: always ten ten somewhere.

699
00:35:30,716 --> 00:35:33,316
Speaker 1: So I think, but that humans, you know, receive this

700
00:35:33,476 --> 00:35:35,516
Speaker 1: data of movement. When you grow up you look at

701
00:35:35,516 --> 00:35:37,796
Speaker 1: the clock on the wall as a kid, you're like, okay,

702
00:35:37,836 --> 00:35:39,636
Speaker 1: now I intuitively get it. I think I know what's

703
00:35:39,676 --> 00:35:41,596
Speaker 1: going on, so I can actually make it work. So

704
00:35:42,476 --> 00:35:44,116
Speaker 1: even though we train it a lot of data, I

705
00:35:44,156 --> 00:35:46,796
Speaker 1: don't think we trained it on the right categorically right

706
00:35:46,876 --> 00:35:50,836
Speaker 1: data yet right to get all the intuitive understanding that

707
00:35:50,916 --> 00:35:53,756
Speaker 1: we have today. So I think we have a data

708
00:35:53,756 --> 00:35:56,356
Speaker 1: problem and that exists the physical AI. So I think

709
00:35:56,396 --> 00:35:58,516
Speaker 1: the applications will win. There's a lot of people are

710
00:35:58,516 --> 00:36:01,076
Speaker 1: working in this. I think the applications will win who

711
00:36:01,116 --> 00:36:04,916
Speaker 1: can either synthetically generate that data or they can actually

712
00:36:05,236 --> 00:36:07,916
Speaker 1: scale in the physical world in a way where they

713
00:36:07,916 --> 00:36:12,076
Speaker 1: can actually generate the day for themselves. But the scaling

714
00:36:12,116 --> 00:36:14,836
Speaker 1: needs to happen with less data, and I think that

715
00:36:14,916 --> 00:36:17,196
Speaker 1: was That's why I'm like, for example, like very bullish

716
00:36:17,196 --> 00:36:19,876
Speaker 1: on manufacturing. So I think the data is going to

717
00:36:19,876 --> 00:36:21,636
Speaker 1: be the biggest challenge. And I think, you know, in

718
00:36:21,716 --> 00:36:25,356
Speaker 1: order for us to massively change this space, we need

719
00:36:25,396 --> 00:36:26,876
Speaker 1: to be able to get to the data. I don't

720
00:36:26,916 --> 00:36:30,756
Speaker 1: think algorithms is a bottleneck there yet. It's just a

721
00:36:30,836 --> 00:36:31,476
Speaker 1: data for us.

722
00:36:31,516 --> 00:36:36,316
Speaker 2: And is it just a matter of people doing what

723
00:36:36,356 --> 00:36:39,636
Speaker 2: you're doing and like finding little wedge places to start

724
00:36:39,676 --> 00:36:41,676
Speaker 2: and having people sort of hold the hand of the

725
00:36:41,716 --> 00:36:46,276
Speaker 2: model and training up the models. I mean that seems

726
00:36:46,316 --> 00:36:48,876
Speaker 2: slow on a certain level, like not you know, obviously

727
00:36:48,956 --> 00:36:52,236
Speaker 2: it's working for you, but like, is there some kind

728
00:36:52,236 --> 00:36:55,316
Speaker 2: of breakthrough move people can make? Can you put sensors

729
00:36:55,356 --> 00:36:58,396
Speaker 2: somewhere in the world to you know, train AI without

730
00:36:58,476 --> 00:37:01,516
Speaker 2: having to you know, have a human stand next to

731
00:37:01,596 --> 00:37:04,276
Speaker 2: it as it messes up one piece of sheet metal

732
00:37:04,316 --> 00:37:04,836
Speaker 2: after another.

733
00:37:05,076 --> 00:37:09,356
Speaker 1: Yeah, I think I think that there is there's another path,

734
00:37:09,396 --> 00:37:15,156
Speaker 1: which is simulation path. Make physics based simulations faster and

735
00:37:15,356 --> 00:37:17,316
Speaker 1: kind of learn. Let the robots just go play in

736
00:37:17,356 --> 00:37:19,956
Speaker 1: a digital playground as opposed to deploy it in real role,

737
00:37:20,236 --> 00:37:22,276
Speaker 1: and that becomes a computation problem. And then you know,

738
00:37:22,276 --> 00:37:23,956
Speaker 1: as long as you have enough computation, you can train

739
00:37:24,036 --> 00:37:26,596
Speaker 1: to robots. But I think, you know, I think you

740
00:37:26,636 --> 00:37:28,756
Speaker 1: know the good examples that we have had such success

741
00:37:28,796 --> 00:37:31,796
Speaker 1: so far as like autonomous cars, right, did the same

742
00:37:31,836 --> 00:37:34,196
Speaker 1: thing we were doing, but in the car like Okay, Tesla,

743
00:37:35,276 --> 00:37:37,676
Speaker 1: you know, deploy the fleet of robots that are capturing

744
00:37:37,756 --> 00:37:40,716
Speaker 1: data still be driven by humans, but the data can

745
00:37:40,716 --> 00:37:42,396
Speaker 1: be used later on to kind of automate it.

746
00:37:42,916 --> 00:37:46,356
Speaker 2: I mean, that's an interesting case because it has been

747
00:37:46,556 --> 00:37:51,636
Speaker 2: much harder clearly than many people thought. Maybe most people thought, right, Like,

748
00:37:51,916 --> 00:37:54,476
Speaker 2: I know, that's a particular instance where you're really worried

749
00:37:54,476 --> 00:37:59,116
Speaker 2: about edge cases. I don't know, is autonomous cars like

750
00:37:59,276 --> 00:38:01,196
Speaker 2: a good model or not. It seems complicated.

751
00:38:01,196 --> 00:38:03,596
Speaker 1: I think the model of capturing data is there, but

752
00:38:03,636 --> 00:38:06,716
Speaker 1: then the the task at hand is very hard. Yeah, right,

753
00:38:07,236 --> 00:38:10,436
Speaker 1: so I think that's the challenge, right so where it

754
00:38:10,436 --> 00:38:13,476
Speaker 1: says like with us, it's still much more structured environment,

755
00:38:13,556 --> 00:38:15,236
Speaker 1: And I think that's that was the thinking we're thinking.

756
00:38:15,236 --> 00:38:17,756
Speaker 1: I think the hardest problem right now in physical AI

757
00:38:17,956 --> 00:38:20,676
Speaker 1: is finding the business model of how do you scale

758
00:38:20,756 --> 00:38:24,316
Speaker 1: data capture without requiring billions of dollars in investment?

759
00:38:24,556 --> 00:38:25,916
Speaker 2: So what do you make in today?

760
00:38:26,236 --> 00:38:29,676
Speaker 1: I imagine you know, so last time I checked in

761
00:38:29,716 --> 00:38:32,916
Speaker 1: the facility, one four of the sales are working on

762
00:38:32,956 --> 00:38:33,996
Speaker 1: a defense application.

763
00:38:35,436 --> 00:38:37,316
Speaker 2: Is it secret? Can you tell me what it is?

764
00:38:38,076 --> 00:38:41,876
Speaker 1: It's a missile? And two of them were working on

765
00:38:41,876 --> 00:38:45,076
Speaker 1: an aerospace application. This is components of an aircraft or

766
00:38:45,076 --> 00:38:48,716
Speaker 1: a drawing. And one of them, as an interesting one,

767
00:38:48,796 --> 00:38:51,756
Speaker 1: was working on an architectural component, which is a roof

768
00:38:51,836 --> 00:38:56,756
Speaker 1: tile for a specific building that's used by the Department

769
00:38:56,796 --> 00:38:58,716
Speaker 1: of the by Bureau of Water Recognition.

770
00:38:58,916 --> 00:39:00,436
Speaker 2: Oh, I was I was going to say, what is it?

771
00:39:00,516 --> 00:39:04,676
Speaker 2: Something like Frank Gary, like nightmare weirdo metal park.

772
00:39:05,516 --> 00:39:07,396
Speaker 1: Oh, those we have had those in the past two

773
00:39:07,476 --> 00:39:10,756
Speaker 1: but this one is actually very practical. Well, it's this building.

774
00:39:10,756 --> 00:39:14,036
Speaker 1: It's actually very interesting. Exactly these buildings, these large industrial

775
00:39:14,076 --> 00:39:16,996
Speaker 1: buildings that built they built in the sixties or fifties,

776
00:39:17,636 --> 00:39:19,716
Speaker 1: and they use these type of roof tiles that the

777
00:39:19,796 --> 00:39:23,596
Speaker 1: manufacturer doesn't exist anymore. And anybody else who they went

778
00:39:23,636 --> 00:39:26,916
Speaker 1: to quoted them hundreds of thousand dollars to make those tiles,

779
00:39:26,916 --> 00:39:28,276
Speaker 1: and we're like, oh no, we can make it for you.

780
00:39:29,476 --> 00:39:31,156
Speaker 1: But also that show is kind of the diversity. I mean,

781
00:39:31,156 --> 00:39:32,476
Speaker 1: like like I say, in the morning, we have like

782
00:39:32,476 --> 00:39:37,716
Speaker 1: aerospace parts. In the afternoon, roof tiles for a industrial complex,

783
00:39:37,796 --> 00:39:39,636
Speaker 1: for you know, for a dam.

784
00:39:40,076 --> 00:39:43,596
Speaker 2: Now you're in the sheet metal business. I know you're large.

785
00:39:43,676 --> 00:39:47,676
Speaker 2: Dream is much larger than that, right, but like what

786
00:39:48,596 --> 00:39:51,996
Speaker 2: like that, tell me where you are now? Tell me

787
00:39:51,996 --> 00:39:53,436
Speaker 2: where you are now? Like what are you doing? What

788
00:39:53,436 --> 00:39:55,996
Speaker 2: are you selling? And then kind of what's the next

789
00:39:56,396 --> 00:39:56,996
Speaker 2: big step.

790
00:39:57,396 --> 00:39:59,396
Speaker 1: So some of our systems are now operating out in

791
00:39:59,396 --> 00:40:03,996
Speaker 1: the wild and working for the customers. And but I

792
00:40:03,996 --> 00:40:07,556
Speaker 1: think the next phase of growth for us is getting

793
00:40:07,636 --> 00:40:09,996
Speaker 1: into each of these applications and own more of the

794
00:40:10,076 --> 00:40:13,276
Speaker 1: process so we can teach the robocraftsmen the future processes

795
00:40:13,356 --> 00:40:15,716
Speaker 1: not just sheet for me, but also maybe how to

796
00:40:15,716 --> 00:40:17,756
Speaker 1: assemble it, how to weld it, how do you surface

797
00:40:17,756 --> 00:40:21,076
Speaker 1: finish it right. So what we are doing now in

798
00:40:21,116 --> 00:40:24,716
Speaker 1: the next phase is actually instead of selling parts or

799
00:40:24,716 --> 00:40:29,156
Speaker 1: components or systems, we're actually saying, Okay, can we get

800
00:40:29,196 --> 00:40:33,796
Speaker 1: this robocraftsman to actually build you a subassembly or a

801
00:40:33,836 --> 00:40:36,756
Speaker 1: full product, not just a component of it, but a

802
00:40:36,796 --> 00:40:39,596
Speaker 1: full product. So that's something we're describing with folks. Can

803
00:40:39,596 --> 00:40:43,676
Speaker 1: we have the robocraftsmen build the full drone for you?

804
00:40:43,716 --> 00:40:45,636
Speaker 1: Can we have the robocrafts and build you a full

805
00:40:45,636 --> 00:40:49,556
Speaker 1: missile as opposed to just build missile you know missile scans.

806
00:40:49,876 --> 00:40:53,316
Speaker 2: Is there that seems like a leap? Is there not

807
00:40:53,396 --> 00:40:54,756
Speaker 2: an intermediate step?

808
00:40:55,276 --> 00:40:58,596
Speaker 1: Like yes? Yes? So I mean how we're doing it

809
00:40:58,676 --> 00:41:01,156
Speaker 1: is we're gradually stepping into it right the same way

810
00:41:01,196 --> 00:41:03,596
Speaker 1: she metal was our first application. So we're putting a

811
00:41:03,636 --> 00:41:08,196
Speaker 1: facility that maybe makes drones, but the main component that

812
00:41:08,236 --> 00:41:11,916
Speaker 1: we automate today is sheet for me, which is the bottleneck.

813
00:41:12,436 --> 00:41:14,756
Speaker 1: And then we do the welding in a traditional way

814
00:41:14,916 --> 00:41:17,156
Speaker 1: on the same robots, but we actually instruct them to

815
00:41:17,196 --> 00:41:19,716
Speaker 1: do it.

816
00:41:18,356 --> 00:41:21,716
Speaker 2: So that way, the robot is kind of back where

817
00:41:21,716 --> 00:41:24,196
Speaker 2: it was on sheet metal five years ago, but it's

818
00:41:24,276 --> 00:41:26,276
Speaker 2: learning how to weld now exactly.

819
00:41:26,716 --> 00:41:29,076
Speaker 1: I used to work in a you know, a shop

820
00:41:29,116 --> 00:41:31,636
Speaker 1: that we will do custom cars, build the custom cars

821
00:41:31,636 --> 00:41:33,836
Speaker 1: with hand, and so it was also near near and

822
00:41:33,836 --> 00:41:37,556
Speaker 1: dear to my heart. So what we realize is that

823
00:41:37,556 --> 00:41:41,556
Speaker 1: with our technology, for the first time, we can actually

824
00:41:41,676 --> 00:41:45,156
Speaker 1: enable a product that didn't exist in automotive, meaning that

825
00:41:45,676 --> 00:41:48,516
Speaker 1: instead of buying a car that's mass produced and every

826
00:41:48,516 --> 00:41:50,796
Speaker 1: single one of them look the same, you can now

827
00:41:50,956 --> 00:41:54,236
Speaker 1: let the customer design a custom car for them. You know,

828
00:41:54,316 --> 00:41:55,716
Speaker 1: right now, if you go buy a car, you can

829
00:41:55,796 --> 00:41:58,396
Speaker 1: you have options of what the what the seat color

830
00:41:58,396 --> 00:42:01,276
Speaker 1: would be, or maybe the color of the car would be,

831
00:42:01,316 --> 00:42:04,116
Speaker 1: and what some trim options. But you can't really choose

832
00:42:04,156 --> 00:42:06,396
Speaker 1: the design of your car. You can't say, oh, I

833
00:42:06,396 --> 00:42:08,516
Speaker 1: want a different hood and I want a different fender

834
00:42:09,236 --> 00:42:11,716
Speaker 1: because going to the back same problem you have to

835
00:42:11,716 --> 00:42:15,756
Speaker 1: make tooling and mold for the vender of certain designs.

836
00:42:15,756 --> 00:42:18,356
Speaker 1: It cannot easily change it. So with our technology you can.

837
00:42:18,556 --> 00:42:22,796
Speaker 1: So what we started doing was like, okay, applying this

838
00:42:23,116 --> 00:42:26,636
Speaker 1: freedom that this technology provides to now automotive is the

839
00:42:26,796 --> 00:42:28,756
Speaker 1: ability of the customers to be able to go to

840
00:42:28,796 --> 00:42:33,116
Speaker 1: a website design a fully customized car for themselves. It

841
00:42:33,196 --> 00:42:36,076
Speaker 1: can be either from already design panels round car designer

842
00:42:36,476 --> 00:42:40,236
Speaker 1: or adding a specific customer customizations they want to do,

843
00:42:40,316 --> 00:42:42,396
Speaker 1: for example, logo of their company to their door of

844
00:42:42,436 --> 00:42:44,636
Speaker 1: the car or the hood of the car, and actually

845
00:42:44,636 --> 00:42:49,116
Speaker 1: get a completely unique car right manufactured for them. And

846
00:42:49,156 --> 00:42:52,036
Speaker 1: we're actually working with this with some of our automotive

847
00:42:52,036 --> 00:42:55,516
Speaker 1: partners Automotive Aims as well. Right, we actually showed some

848
00:42:55,556 --> 00:42:58,556
Speaker 1: of this work in the biggest aftermarket show in the

849
00:42:58,636 --> 00:43:02,116
Speaker 1: United States is called SEMA with our partner Toyota. So

850
00:43:02,596 --> 00:43:05,156
Speaker 1: I think this is going to be, in my opinion,

851
00:43:05,476 --> 00:43:08,356
Speaker 1: one of the new product categories in automotive. We have

852
00:43:08,476 --> 00:43:10,996
Speaker 1: had a time of his cars, we have had you know,

853
00:43:11,156 --> 00:43:13,596
Speaker 1: electric cars, and I think now for the first time,

854
00:43:13,636 --> 00:43:17,116
Speaker 1: with technologies like ours, you can have custom to order cars,

855
00:43:17,276 --> 00:43:19,636
Speaker 1: like cars that are like, you know, the same way

856
00:43:19,676 --> 00:43:21,316
Speaker 1: you choose what T shirt you wear and your T

857
00:43:21,436 --> 00:43:23,556
Speaker 1: shirt is different than mine. We also don't have to

858
00:43:23,636 --> 00:43:26,356
Speaker 1: drive the same you know model S or you know

859
00:43:26,476 --> 00:43:29,076
Speaker 1: Model three. We can actually have our own customized Model

860
00:43:29,076 --> 00:43:30,036
Speaker 1: three is and modelesses.

861
00:43:30,676 --> 00:43:33,276
Speaker 2: So what's the I mean is that the if you

862
00:43:33,316 --> 00:43:36,636
Speaker 2: think sort of long term for Makeina is like that

863
00:43:36,676 --> 00:43:38,676
Speaker 2: what you think about like give me the give me

864
00:43:38,716 --> 00:43:43,196
Speaker 2: the five year vision yeah, or ten year or whatever.

865
00:43:43,556 --> 00:43:46,876
Speaker 1: Yeah. So I think the long term motivation behind our

866
00:43:46,876 --> 00:43:51,996
Speaker 1: company is can you grant this democratization of ideas for

867
00:43:52,036 --> 00:43:54,116
Speaker 1: people who want to build anything? Right? Can I express

868
00:43:54,156 --> 00:43:56,956
Speaker 1: myself if I'm a builder, can I go build something

869
00:43:56,956 --> 00:43:58,916
Speaker 1: with not having to build a factory for So that's

870
00:43:58,996 --> 00:44:02,836
Speaker 1: really the long term goal. So I imagine in the

871
00:44:02,876 --> 00:44:07,356
Speaker 1: next five to ten years, you can as a designer,

872
00:44:07,396 --> 00:44:09,596
Speaker 1: somebody who has an idea, you can go to a website,

873
00:44:09,916 --> 00:44:13,996
Speaker 1: get guided through your ideas on how to make and

874
00:44:14,076 --> 00:44:17,156
Speaker 1: design a physical product, hit a button and say, okay,

875
00:44:17,196 --> 00:44:22,676
Speaker 1: I want twenty of these, and I want in Chatsworth, California,

876
00:44:22,716 --> 00:44:25,596
Speaker 1: and the right facility programs the right number of robots

877
00:44:25,876 --> 00:44:28,916
Speaker 1: to actually do those operations without any hardware or investment

878
00:44:29,236 --> 00:44:31,156
Speaker 1: that needs to be made for those of specific parts

879
00:44:31,596 --> 00:44:33,316
Speaker 1: and ship it to you two days later in the

880
00:44:33,396 --> 00:44:37,716
Speaker 1: right location. That is the future we're building towards cars

881
00:44:37,796 --> 00:44:41,076
Speaker 1: is just you know, one of the products that could

882
00:44:41,076 --> 00:44:43,116
Speaker 1: be built. But I imagine that you know this technology

883
00:44:43,156 --> 00:44:45,436
Speaker 1: or technology like these, technologies like these can be used

884
00:44:45,476 --> 00:44:48,276
Speaker 1: to do the myriad of designs. I think the moment

885
00:44:48,476 --> 00:44:52,276
Speaker 1: you open up this possibility of any designs could be

886
00:44:52,316 --> 00:44:54,556
Speaker 1: a reality. I think so many things will be created

887
00:44:54,596 --> 00:44:57,316
Speaker 1: that we're not even thinking of right now. You know

888
00:44:57,396 --> 00:45:00,356
Speaker 1: the fact that we have cars today and they all

889
00:45:00,396 --> 00:45:03,316
Speaker 1: look the same as limitation of technology. But the moment

890
00:45:03,356 --> 00:45:07,156
Speaker 1: you can open up this creativity of turning ideas into

891
00:45:07,156 --> 00:45:11,116
Speaker 1: physical reality without a without an initial investment or huge

892
00:45:11,156 --> 00:45:13,156
Speaker 1: barrier to entry, then I think we're going to have

893
00:45:13,196 --> 00:45:15,836
Speaker 1: all kinds of drones, all kinds of satellites, all kinds

894
00:45:15,876 --> 00:45:18,356
Speaker 1: of rockets, all kinds of cars that you're going to

895
00:45:18,356 --> 00:45:22,076
Speaker 1: be this like you know, Cambrian explosion of different designs

896
00:45:22,116 --> 00:45:23,636
Speaker 1: that's going to come into our world. And I think

897
00:45:23,636 --> 00:45:25,676
Speaker 1: that's what future is about. The future is about you

898
00:45:25,676 --> 00:45:27,196
Speaker 1: know what I call it, Like future is custom Like

899
00:45:27,276 --> 00:45:29,956
Speaker 1: future is about being able to make these all these

900
00:45:29,996 --> 00:45:33,716
Speaker 1: ideas in reality. We had this explosion happening in digital world. Yeah,

901
00:45:33,756 --> 00:45:36,636
Speaker 1: you know, now we have even models generating images and

902
00:45:37,156 --> 00:45:41,116
Speaker 1: videos and there's this you know, explosion of different ideas

903
00:45:41,116 --> 00:45:45,276
Speaker 1: and content being created using the technology. But the link

904
00:45:45,356 --> 00:45:47,636
Speaker 1: is broken to the physical world and the physical work

905
00:45:47,676 --> 00:45:49,596
Speaker 1: is still pretty uniform because it's very hard to make

906
00:45:49,636 --> 00:45:52,756
Speaker 1: things in the physical world. Can we bridge that gap?

907
00:45:52,916 --> 00:45:56,236
Speaker 1: Can we connect the digital world of creation to physical

908
00:45:56,316 --> 00:45:59,396
Speaker 1: world of creation and create the same variety in the

909
00:45:59,396 --> 00:46:01,996
Speaker 1: physical world as we have in the digital world. I

910
00:46:02,036 --> 00:46:03,876
Speaker 1: think that's the goal in our company.

911
00:46:06,956 --> 00:46:08,916
Speaker 2: We'll be back in a minute with the light Year Round.

912
00:46:17,156 --> 00:46:22,516
Speaker 2: Let's finish with the Lightning Round. Do you drive a

913
00:46:22,556 --> 00:46:23,996
Speaker 2: customized car?

914
00:46:24,436 --> 00:46:28,036
Speaker 1: I don't actually yet. Well, if I am, what have I.

915
00:46:27,996 --> 00:46:30,436
Speaker 2: Seen on your Instagram? What's that truck you keep posting

916
00:46:30,476 --> 00:46:31,316
Speaker 2: on your Instagram?

917
00:46:31,356 --> 00:46:33,396
Speaker 1: So I so I have a truck that's customized. I

918
00:46:33,396 --> 00:46:36,636
Speaker 1: don't drive it around as much, but maybe this year

919
00:46:36,676 --> 00:46:38,516
Speaker 1: I'll start taking it out. This year I've been you know,

920
00:46:38,516 --> 00:46:41,556
Speaker 1: we have been kind of stelf about it, talking about it,

921
00:46:41,716 --> 00:46:43,316
Speaker 1: but we haven't talked about it in a big way

922
00:46:43,356 --> 00:46:45,076
Speaker 1: because we have a big release coming soon.

923
00:46:45,556 --> 00:46:48,036
Speaker 2: I mean, you're literally posting it on Instagram. It's not

924
00:46:48,116 --> 00:46:51,116
Speaker 2: that stuff. Tell me about Tell me about that truck

925
00:46:51,156 --> 00:46:53,196
Speaker 2: you keep posting on Instagram? What's going on with that?

926
00:46:53,316 --> 00:46:55,116
Speaker 1: So it's a truck is fully the full body is

927
00:46:55,116 --> 00:46:55,996
Speaker 1: fully customized.

928
00:46:56,396 --> 00:46:58,836
Speaker 2: It says anvil in the back when you post it?

929
00:46:58,876 --> 00:46:59,756
Speaker 2: Is it called anvil?

930
00:46:59,836 --> 00:47:01,916
Speaker 1: Dumb question and call it and call it anvil? I

931
00:47:01,956 --> 00:47:04,476
Speaker 1: think it's the idea was actually the shape design of

932
00:47:04,516 --> 00:47:06,236
Speaker 1: it was inspired by amil. If you look at the

933
00:47:06,236 --> 00:47:09,516
Speaker 1: front fender, it actually looks like a the front bumper

934
00:47:09,556 --> 00:47:11,876
Speaker 1: looks like an anvil. But also the idea is that, like,

935
00:47:11,916 --> 00:47:14,756
Speaker 1: you know, we're actually forming shits on an anvil. So yeah,

936
00:47:14,796 --> 00:47:15,676
Speaker 1: it was very fitting.

937
00:47:15,916 --> 00:47:18,036
Speaker 2: Tell me about that truck, like, tell just tell me

938
00:47:18,076 --> 00:47:18,756
Speaker 2: what's it look like.

939
00:47:18,876 --> 00:47:20,716
Speaker 1: Yeah, so for example, like you know, we put a

940
00:47:20,716 --> 00:47:23,276
Speaker 1: lot of form and sharp edges in the in the hood. Right.

941
00:47:24,276 --> 00:47:27,156
Speaker 1: Most vehicles have a very hard time if you look

942
00:47:27,196 --> 00:47:29,756
Speaker 1: at most of the you know, hood of the vehicles,

943
00:47:30,116 --> 00:47:32,636
Speaker 1: they are very smooth because it's very hard to actually

944
00:47:32,636 --> 00:47:35,436
Speaker 1: put sharp angles in the hood. So if you look

945
00:47:35,476 --> 00:47:37,436
Speaker 1: at this truck, this truck has a lot of angles,

946
00:47:37,436 --> 00:47:40,636
Speaker 1: a lot of sharp detail right in the hood. Right,

947
00:47:41,156 --> 00:47:43,556
Speaker 1: And and and that's very expressive of the type of

948
00:47:43,636 --> 00:47:46,756
Speaker 1: person for example, that I am, right, I like things

949
00:47:46,756 --> 00:47:50,116
Speaker 1: that are edgy, and and that truck is certainly edgy. Right.

950
00:47:51,836 --> 00:47:55,236
Speaker 1: It's bare metal, right, you know, there is no blemishes

951
00:47:55,276 --> 00:47:59,596
Speaker 1: being hiden and hidden under the under the vehicle. You know,

952
00:47:59,636 --> 00:48:02,036
Speaker 1: a lot of people when cyber Truck came out, we

953
00:48:02,236 --> 00:48:05,316
Speaker 1: got very excited about you know, oh, it's bare metal.

954
00:48:05,356 --> 00:48:06,596
Speaker 1: It looks like a metal, but then there was no

955
00:48:06,676 --> 00:48:08,756
Speaker 1: form in it because it's actually very hard to make

956
00:48:08,756 --> 00:48:12,196
Speaker 1: it for metal look nice. And so that's one of

957
00:48:12,196 --> 00:48:14,356
Speaker 1: the things we wanted to show. We want to show that, Okay,

958
00:48:14,396 --> 00:48:16,676
Speaker 1: you can actually have a form metal with a lot

959
00:48:16,676 --> 00:48:18,756
Speaker 1: of detail in it and still keep it bare metal

960
00:48:18,796 --> 00:48:22,236
Speaker 1: because it will look nice. Right, So, yeah, a lot

961
00:48:22,276 --> 00:48:24,756
Speaker 1: of design features of it. For me kind of represents

962
00:48:24,756 --> 00:48:27,436
Speaker 1: the type of personality and character that I have. But

963
00:48:27,516 --> 00:48:29,716
Speaker 1: I think that's how every car should be. You know,

964
00:48:29,716 --> 00:48:31,556
Speaker 1: people should be able to have that freedom to choose

965
00:48:31,596 --> 00:48:32,676
Speaker 1: what their cars look like.

966
00:48:34,876 --> 00:48:36,516
Speaker 2: How many skull tattoos do you have?

967
00:48:38,316 --> 00:48:46,036
Speaker 1: I've got three? Why it's it's so, yeah, it's an

968
00:48:46,036 --> 00:48:50,556
Speaker 1: interesting thing. So a skull for me represents kind of

969
00:48:50,596 --> 00:48:54,276
Speaker 1: and it's an abstract for death of ego. So I

970
00:48:54,276 --> 00:48:57,076
Speaker 1: have a tattoo on my thumb which is a skull

971
00:48:57,196 --> 00:49:01,156
Speaker 1: that's holding a microphone to his ears. And this was

972
00:49:01,196 --> 00:49:03,556
Speaker 1: a time where you know, I felt like, you know,

973
00:49:03,636 --> 00:49:05,276
Speaker 1: I had a good platform and I could talk a

974
00:49:05,276 --> 00:49:07,996
Speaker 1: lot and people would listen. But then I realized I

975
00:49:08,036 --> 00:49:10,436
Speaker 1: should yes, that's right, but I should maybe keep them

976
00:49:10,716 --> 00:49:13,436
Speaker 1: keep the mic close to my ears and also listen

977
00:49:13,476 --> 00:49:17,676
Speaker 1: as opposed to talk all the time, right, So I think.

978
00:49:17,956 --> 00:49:21,116
Speaker 2: Skull microphones don't work that way for the record, but.

979
00:49:21,116 --> 00:49:24,876
Speaker 1: I like it as a metaphor exactly. But I think

980
00:49:24,876 --> 00:49:27,716
Speaker 1: the idea of is around, you know, kind of reminders

981
00:49:27,716 --> 00:49:29,116
Speaker 1: of you can see a lot of my tattoos on

982
00:49:29,156 --> 00:49:31,676
Speaker 1: my on my hands, so it's a really reminder for

983
00:49:31,756 --> 00:49:36,196
Speaker 1: myself to know that, you know, be present and make

984
00:49:36,236 --> 00:49:38,676
Speaker 1: sure that you know you're not involved with your ego

985
00:49:38,756 --> 00:49:41,196
Speaker 1: too much and you can see others people's perspective.

986
00:49:41,836 --> 00:49:45,476
Speaker 2: Is there any tension between ego death and custom cars?

987
00:49:46,996 --> 00:49:50,316
Speaker 1: Tension between ego death and custom cars? I don't know.

988
00:49:50,436 --> 00:49:52,996
Speaker 2: I'm just playing, but like you know, custom car kind

989
00:49:52,996 --> 00:49:55,636
Speaker 2: of seems like, hey, look at me, I'm special, and

990
00:49:55,716 --> 00:49:58,476
Speaker 2: ego death seems like, oh, don't look at me, I'm

991
00:49:58,476 --> 00:49:59,116
Speaker 2: not so special.

992
00:49:59,316 --> 00:50:01,996
Speaker 1: Yeah, no, I think I think the difference is I think, yeah,

993
00:50:01,996 --> 00:50:04,636
Speaker 1: if you have attachment to your custom card, then maybe

994
00:50:04,996 --> 00:50:07,556
Speaker 1: there's tension. But I more think of it in terms

995
00:50:07,556 --> 00:50:10,476
Speaker 1: of expression. Right. You know, you can be an artist,

996
00:50:10,636 --> 00:50:12,996
Speaker 1: you can. You can you can design your home the

997
00:50:13,076 --> 00:50:16,356
Speaker 1: way it expresses you. You can design the theme of

998
00:50:16,396 --> 00:50:18,756
Speaker 1: your podcast the way it expresses you. You can design

999
00:50:18,836 --> 00:50:21,596
Speaker 1: your car. Also, the way it expresses you. I think

1000
00:50:21,636 --> 00:50:23,676
Speaker 1: it's leus so about oh look at me, I'm special.

1001
00:50:23,676 --> 00:50:26,636
Speaker 1: It's more like, here's my expression to the world for

1002
00:50:26,676 --> 00:50:30,516
Speaker 1: the people to see. But I think that expressiveness is

1003
00:50:30,596 --> 00:50:32,716
Speaker 1: it is pretty amazing. I think that's uniquely one of

1004
00:50:32,756 --> 00:50:34,876
Speaker 1: the unique things about humans that like, you know, we

1005
00:50:34,876 --> 00:50:37,196
Speaker 1: we I think all we do when we come to

1006
00:50:37,196 --> 00:50:40,596
Speaker 1: this world is expressing ourselves right, expressing uself through our work,

1007
00:50:40,716 --> 00:50:44,956
Speaker 1: expressing through ourselves through our relationships. And if you can

1008
00:50:45,076 --> 00:50:48,436
Speaker 1: enable people to express themselves better better, I think that's great.

1009
00:50:49,156 --> 00:50:51,596
Speaker 1: But if you get attached to your expressions and your

1010
00:50:51,636 --> 00:50:54,036
Speaker 1: ideas and your thoughts and think, oh, I'm better than

1011
00:50:54,036 --> 00:50:56,956
Speaker 1: everybody else, and I think that that becomes that becomes

1012
00:50:56,956 --> 00:50:58,516
Speaker 1: a little bit of an ego driven trip.

1013
00:51:05,996 --> 00:51:09,676
Speaker 2: Edward Mayer is the co founder and CEO of Mocking Labs.

1014
00:51:10,356 --> 00:51:13,636
Speaker 2: Today's show was produced by Gabriel Hunter Cheang. It was

1015
00:51:13,876 --> 00:51:17,356
Speaker 2: edited by Lyddy jeene Kott and engineered by Sarah Bruginner.

1016
00:51:17,836 --> 00:51:20,996
Speaker 2: You can email us at problem at Pushkin dot FM.

1017
00:51:21,556 --> 00:51:23,876
Speaker 2: I'm Jacob Goldstein and we'll be back next week with

1018
00:51:23,956 --> 00:51:32,916
Speaker 2: another episode of What's Your Problem.