1
00:00:08,320 --> 00:00:08,800
Speaker 1: Pushkin.

2
00:00:16,160 --> 00:00:19,079
Speaker 2: Because I didn't know that sports biomechanics could be a career.

3
00:00:20,280 --> 00:00:22,880
Speaker 2: I decided to go to grad school and initially start

4
00:00:22,960 --> 00:00:25,960
Speaker 2: working on prosthetic limbs. So that was the first two

5
00:00:26,000 --> 00:00:30,920
Speaker 2: years of grad school. Then about two years in I

6
00:00:31,040 --> 00:00:36,640
Speaker 2: discovered baseball pitching biomechanics research. The funny thing that happened

7
00:00:36,640 --> 00:00:40,360
Speaker 2: there was I gave a PhD committee meeting where I

8
00:00:40,400 --> 00:00:44,440
Speaker 2: spent most of the time talking about prosthetic limbs, and

9
00:00:44,479 --> 00:00:46,920
Speaker 2: then I spent the last few minutes as an aside

10
00:00:47,360 --> 00:00:50,520
Speaker 2: on the baseball pitching research I found. And my committee

11
00:00:50,600 --> 00:00:53,760
Speaker 2: was like, Jimmy, the last few minutes were way better

12
00:00:53,800 --> 00:00:55,000
Speaker 2: than the first forty fives.

13
00:00:55,400 --> 00:00:57,040
Speaker 3: So I was like, all right.

14
00:00:57,240 --> 00:01:00,800
Speaker 2: So then they were like, we'll let you do baseball

15
00:01:00,840 --> 00:01:04,080
Speaker 2: pitching biomechanics as your PhD work, but just so you know,

16
00:01:04,319 --> 00:01:06,480
Speaker 2: it might be really hard to have a career doing that.

17
00:01:08,040 --> 00:01:09,480
Speaker 2: But you know, they were like, if you want to

18
00:01:09,520 --> 00:01:10,760
Speaker 2: go for it, you can go for it. And so

19
00:01:10,840 --> 00:01:13,319
Speaker 2: I was like, you know what, Sure, I'll go for it.

20
00:01:19,200 --> 00:01:22,200
Speaker 1: I'm Jacob Goldstein, and this is what's your problem. Today

21
00:01:22,280 --> 00:01:25,639
Speaker 1: we have the second episode in our series about people

22
00:01:25,880 --> 00:01:29,039
Speaker 1: who are working at the frontiers of technology to help

23
00:01:29,080 --> 00:01:33,640
Speaker 1: a lead athletes perform better. My guest today is Jimmy Buffy,

24
00:01:33,959 --> 00:01:36,639
Speaker 1: and as it happened, the concerns of his grad school

25
00:01:36,680 --> 00:01:40,800
Speaker 1: advisors were unfounded. Jimmy has in fact made a career

26
00:01:40,920 --> 00:01:44,920
Speaker 1: out of the biomechanics of pitching in baseball and sports

27
00:01:44,920 --> 00:01:48,560
Speaker 1: biomechanics more broadly. When he finished grad school, he got

28
00:01:48,560 --> 00:01:51,120
Speaker 1: a job with the Los Angeles Dodgers, and he went

29
00:01:51,160 --> 00:01:54,160
Speaker 1: on to co found a company called Reboot Motion that

30
00:01:54,280 --> 00:01:57,360
Speaker 1: works with teams in Major League Baseball and the NBA.

31
00:01:57,840 --> 00:02:01,360
Speaker 1: Jimmy's problem is this, how do you take massive amounts

32
00:02:01,360 --> 00:02:05,400
Speaker 1: of data about how professional athletes move and turn all

33
00:02:05,440 --> 00:02:10,600
Speaker 1: that data into information that actually helps those athletes perform better.

34
00:02:12,240 --> 00:02:15,160
Speaker 1: You end up doing your dissertation research on the on

35
00:02:15,240 --> 00:02:19,480
Speaker 1: the biomechanics of pitching, Yes, of baseball, pitching and baseball. Yeah,

36
00:02:19,560 --> 00:02:22,639
Speaker 1: and and then you get hired by the Dodgers.

37
00:02:23,080 --> 00:02:23,800
Speaker 3: Yeah.

38
00:02:23,960 --> 00:02:27,520
Speaker 2: Yeah, that was That was awesome because I originally didn't again,

39
00:02:27,560 --> 00:02:30,320
Speaker 2: didn't realize that that could be a thing that could happen.

40
00:02:30,800 --> 00:02:33,840
Speaker 1: Well, and it kind of wasn't right, Like, you're kind

41
00:02:33,880 --> 00:02:36,640
Speaker 1: of just coming into this field as it's becoming a

42
00:02:36,680 --> 00:02:38,799
Speaker 1: field where you can get a job where it's a.

43
00:02:38,720 --> 00:02:40,280
Speaker 3: Field, right exactly. Yeah, there.

44
00:02:40,360 --> 00:02:42,280
Speaker 2: I mean the challenge then was there wasn't a lot

45
00:02:42,320 --> 00:02:45,400
Speaker 2: of options for actually even getting the data that you

46
00:02:45,560 --> 00:02:47,079
Speaker 2: need to analyze.

47
00:02:47,800 --> 00:02:51,120
Speaker 1: So ten years ago, like what is the state of

48
00:02:51,160 --> 00:02:54,160
Speaker 1: play in this sort of nasson field that you're in

49
00:02:54,400 --> 00:02:55,200
Speaker 1: helping to create?

50
00:02:56,639 --> 00:03:00,000
Speaker 2: So the field is, I would say, is like sports biomechanic,

51
00:03:00,200 --> 00:03:03,600
Speaker 2: and what that is is being able to analyze the

52
00:03:03,760 --> 00:03:07,519
Speaker 2: movement of athletes for lots of purposes, help them reduce

53
00:03:07,560 --> 00:03:10,720
Speaker 2: injury risk, help them improve performance.

54
00:03:10,680 --> 00:03:14,440
Speaker 1: And to be clear, like folk, sports biomechanics has been

55
00:03:14,480 --> 00:03:16,720
Speaker 1: around forever, right, that's what coaches do. They stay out

56
00:03:16,720 --> 00:03:18,480
Speaker 1: there and they watch it. Yeah, And so it's kind

57
00:03:18,480 --> 00:03:21,519
Speaker 1: of becoming it's becoming more technical, right, the field is

58
00:03:21,560 --> 00:03:22,480
Speaker 1: becoming more technical.

59
00:03:22,600 --> 00:03:22,919
Speaker 3: Yeah.

60
00:03:22,960 --> 00:03:26,799
Speaker 2: So the state of the art relied on what is

61
00:03:26,840 --> 00:03:31,760
Speaker 2: called marker based motion capture, which is where you literally

62
00:03:31,800 --> 00:03:36,080
Speaker 2: put reflective markers like little balls, you stick them all

63
00:03:36,120 --> 00:03:39,440
Speaker 2: over somebody's body. Usually the person has to like strip

64
00:03:39,480 --> 00:03:41,760
Speaker 2: their clothes off because you want the markers like literally

65
00:03:41,840 --> 00:03:45,040
Speaker 2: like on the skin, on the joints, and then you

66
00:03:45,120 --> 00:03:48,400
Speaker 2: have these special cameras that track those markers.

67
00:03:48,760 --> 00:03:50,920
Speaker 1: So you've got like a picture in his underwear with

68
00:03:50,960 --> 00:03:54,000
Speaker 1: a bunch of little metal balls, takes them and they're like,

69
00:03:54,160 --> 00:03:55,800
Speaker 1: just pitch like you always pitch.

70
00:03:55,720 --> 00:03:57,000
Speaker 3: Right, And that's the challenge.

71
00:03:57,160 --> 00:03:59,760
Speaker 2: That's why that wasn't That's why it wasn't very wide

72
00:03:59,800 --> 00:04:02,040
Speaker 2: SPA is a thing people did because it was so

73
00:04:02,240 --> 00:04:04,800
Speaker 2: hard to collect the data because ultimately, what you would

74
00:04:04,840 --> 00:04:08,000
Speaker 2: need is you need the data on how someone is moving.

75
00:04:08,360 --> 00:04:09,280
Speaker 3: You need to track.

76
00:04:09,040 --> 00:04:11,000
Speaker 2: Where their elbow is, where their wrist is, where their

77
00:04:11,080 --> 00:04:13,800
Speaker 2: knees are so that you can analyze it. And the

78
00:04:13,840 --> 00:04:15,960
Speaker 2: state of the art for tracking it was an awful

79
00:04:16,000 --> 00:04:18,240
Speaker 2: experience for the people you were trying to track.

80
00:04:18,720 --> 00:04:22,479
Speaker 1: That presumably would mean they didn't pitch like they usually exactly,

81
00:04:22,480 --> 00:04:25,040
Speaker 1: because they don't usually stand there in their underwear with

82
00:04:25,440 --> 00:04:27,200
Speaker 1: met Were they really in their underwear by the way,

83
00:04:27,240 --> 00:04:28,840
Speaker 1: I'm saying it because it sounds absurd, but is that

84
00:04:28,839 --> 00:04:29,680
Speaker 1: actually what they were doing.

85
00:04:29,760 --> 00:04:32,680
Speaker 2: That's actually you strip down to your your boxers, your

86
00:04:32,680 --> 00:04:34,800
Speaker 2: boxer briefs, and that's it, that's all you're wearing.

87
00:04:35,640 --> 00:04:38,080
Speaker 1: So it basically didn't work, and it basically wasn't very

88
00:04:38,080 --> 00:04:39,560
Speaker 1: widely used as a result.

89
00:04:39,360 --> 00:04:43,320
Speaker 2: Right exactly, Yeah, you yeah, you look at studies in

90
00:04:43,360 --> 00:04:47,279
Speaker 2: that field, and people would be throwing like several miles

91
00:04:47,279 --> 00:04:49,600
Speaker 2: an hour slower than they would be throwing when they

92
00:04:49,640 --> 00:04:55,279
Speaker 2: weren't wearing all that stuff. So what changes computer vision?

93
00:04:56,440 --> 00:05:00,360
Speaker 2: That was the big That was the big inflection point. Now,

94
00:05:00,360 --> 00:05:04,480
Speaker 2: to be fair, like, even when I was finishing my PhD,

95
00:05:04,960 --> 00:05:06,440
Speaker 2: and I'll give them, I'll give them a shout out,

96
00:05:06,480 --> 00:05:10,240
Speaker 2: there was a company that was already like working really

97
00:05:10,279 --> 00:05:13,240
Speaker 2: hard to solve this problem for baseball teams that I

98
00:05:13,480 --> 00:05:16,440
Speaker 2: was getting that I got to be familiar with, called Kinnetrax.

99
00:05:18,279 --> 00:05:21,840
Speaker 2: But yeah, the big inflection point was computer vision, basically

100
00:05:21,960 --> 00:05:26,960
Speaker 2: using artificial intelligence to identify where those joints are in

101
00:05:26,960 --> 00:05:31,760
Speaker 2: a camera image rather than needing to paste those reflective markers.

102
00:05:32,279 --> 00:05:35,320
Speaker 1: So computer vision takes off. You're working at the Dodgers,

103
00:05:35,560 --> 00:05:39,080
Speaker 1: and then eventually in twenty nineteen, right, you leave the

104
00:05:39,120 --> 00:05:43,840
Speaker 1: Dodgers and you start your company a reboot motion, what

105
00:05:43,880 --> 00:05:44,640
Speaker 1: does your company do?

106
00:05:45,240 --> 00:05:49,360
Speaker 2: We do what we call biomechanics as a service, So

107
00:05:49,680 --> 00:05:52,840
Speaker 2: we try to analyze this computer vision data at a

108
00:05:52,960 --> 00:05:56,800
Speaker 2: very large scale to help teams and coaches make use

109
00:05:56,800 --> 00:06:03,200
Speaker 2: of it to help athletes get better. Bas Yeah, that says.

110
00:06:03,040 --> 00:06:06,720
Speaker 1: But bess Yah and and who are your customers?

111
00:06:06,920 --> 00:06:14,239
Speaker 2: Our customers are Major League Baseball teams actually NBA teams,

112
00:06:14,279 --> 00:06:19,520
Speaker 2: so we've gotten into basketball also sort of like league

113
00:06:19,560 --> 00:06:23,839
Speaker 2: wide data providers. So yeah, leagues teams is basically our

114
00:06:23,880 --> 00:06:24,440
Speaker 2: sweet spot.

115
00:06:24,720 --> 00:06:27,760
Speaker 1: So let's talk. Let's talk in like a little more

116
00:06:27,760 --> 00:06:30,839
Speaker 1: detail about about what you actually do, right, tell me

117
00:06:30,880 --> 00:06:32,360
Speaker 1: the story of what you do.

118
00:06:33,600 --> 00:06:37,520
Speaker 2: So, Evan, my co founder, Evan Demchik, he likes to

119
00:06:37,520 --> 00:06:41,920
Speaker 2: call this the biomechanics Trainkay.

120
00:06:41,080 --> 00:06:43,080
Speaker 1: Let's take a ride on the biomechanics trend.

121
00:06:43,720 --> 00:06:45,800
Speaker 2: And we call it that because with the way our

122
00:06:45,839 --> 00:06:49,320
Speaker 2: product works is we let people get on the train

123
00:06:49,760 --> 00:06:51,800
Speaker 2: at whatever stop works for them, and get off the

124
00:06:51,839 --> 00:06:53,640
Speaker 2: train at whatever stop works for them.

125
00:06:53,720 --> 00:06:55,360
Speaker 1: How far are we going to go with this metaphor

126
00:06:55,600 --> 00:06:57,360
Speaker 1: of a little o'b nervous about it?

127
00:06:57,760 --> 00:06:59,160
Speaker 3: That might be as far as we go.

128
00:06:59,320 --> 00:07:03,960
Speaker 1: Okay, good, good, So just tell it to me. Start

129
00:07:04,000 --> 00:07:06,279
Speaker 1: at whatever seems like the beginning of a you know,

130
00:07:07,040 --> 00:07:10,120
Speaker 1: a of an encounter, and so I'd like to understand

131
00:07:10,160 --> 00:07:12,600
Speaker 1: how that works. That's really kind of a way to

132
00:07:12,600 --> 00:07:15,640
Speaker 1: think about it. So let's start with the data, right,

133
00:07:16,040 --> 00:07:19,080
Speaker 1: what is a basic what is the basic thing that's.

134
00:07:18,920 --> 00:07:22,440
Speaker 2: Happening So the very first thing that happens is you

135
00:07:22,520 --> 00:07:27,920
Speaker 2: record videos of the athlete doing the athletic motion. So

136
00:07:27,920 --> 00:07:30,880
Speaker 2: we'll talk about pitching, So you record videos of a

137
00:07:30,920 --> 00:07:34,840
Speaker 2: pitcher pitching. Our product, we actually have implemented our own

138
00:07:34,840 --> 00:07:38,560
Speaker 2: computer vision models, so we can do that if people want.

139
00:07:38,640 --> 00:07:43,480
Speaker 2: But generally speaking, people have systems like Kinetrax is one

140
00:07:43,520 --> 00:07:47,120
Speaker 2: I've mentioned that, Hawkeye is another popular one where there

141
00:07:47,200 --> 00:07:49,600
Speaker 2: is a system in place that has the cameras that

142
00:07:49,680 --> 00:07:52,960
Speaker 2: record the videos and then runs those computer vision models.

143
00:07:53,240 --> 00:07:56,840
Speaker 2: What those computer vision models do is they extract the

144
00:07:56,920 --> 00:08:00,880
Speaker 2: locations in three dimensional space of all of like the

145
00:08:00,960 --> 00:08:03,640
Speaker 2: joint centers. So where's my elbow, where's my wrist, where's

146
00:08:03,640 --> 00:08:07,040
Speaker 2: my knee? In three dimensional space? That's what comes out

147
00:08:07,040 --> 00:08:08,400
Speaker 2: of these computer vision systems.

148
00:08:08,920 --> 00:08:12,080
Speaker 1: Okay, so pretty much everybody has that at this point,

149
00:08:12,120 --> 00:08:16,280
Speaker 1: Like every professional baseball team has that for every pitch

150
00:08:16,320 --> 00:08:17,480
Speaker 1: in every game at this point.

151
00:08:17,600 --> 00:08:18,320
Speaker 3: Yes, exactly.

152
00:08:18,640 --> 00:08:21,520
Speaker 2: Okay, So it's a ton a ton of data. So

153
00:08:21,560 --> 00:08:24,360
Speaker 2: that's another challenge that we've solved, is not only how

154
00:08:24,360 --> 00:08:25,640
Speaker 2: do you do this, but how do you do this

155
00:08:25,680 --> 00:08:27,040
Speaker 2: at a very large scale?

156
00:08:27,280 --> 00:08:30,920
Speaker 1: Okay, so this data, everybody's got it now, and you're not.

157
00:08:31,360 --> 00:08:33,760
Speaker 1: You can sort of process the data, but that's not

158
00:08:33,880 --> 00:08:36,559
Speaker 1: your special sauce. That's not your secret sauce, right, So

159
00:08:36,600 --> 00:08:39,800
Speaker 1: typically they'll send you that that data of like, here's

160
00:08:39,800 --> 00:08:41,360
Speaker 1: all the body points. Here is how they're moving in

161
00:08:41,520 --> 00:08:44,000
Speaker 1: physical space. Then what do you do with it?

162
00:08:44,360 --> 00:08:46,600
Speaker 2: So, yeah, this is where the special sauce comes in.

163
00:08:47,120 --> 00:08:49,320
Speaker 2: So the first step to that is how do you

164
00:08:49,400 --> 00:08:53,160
Speaker 2: turn those key points into a human skeleton. So you've

165
00:08:53,160 --> 00:08:55,439
Speaker 2: got to figure out, like where do the bones connect,

166
00:08:55,520 --> 00:08:58,200
Speaker 2: what sort of degrees of freedom do those bones have,

167
00:08:58,760 --> 00:09:01,840
Speaker 2: So then you can figure out how do those key

168
00:09:01,880 --> 00:09:04,640
Speaker 2: points animate a human skeleton.

169
00:09:05,120 --> 00:09:07,160
Speaker 1: So you're sort of rebuilding it. It's like you start

170
00:09:07,200 --> 00:09:08,880
Speaker 1: with the picture of a person and then you turn

171
00:09:08,880 --> 00:09:10,839
Speaker 1: it into a bunch of data points, and now you've

172
00:09:10,840 --> 00:09:13,160
Speaker 1: got to kind of build the person back up again from.

173
00:09:13,000 --> 00:09:17,120
Speaker 2: The day exactly exactly. So now you have an actual

174
00:09:17,280 --> 00:09:21,240
Speaker 2: human skeleton where the shoulder is rotating, the elbow is flexing,

175
00:09:21,280 --> 00:09:24,240
Speaker 2: the knee is flexing, the hips are rotating. And now

176
00:09:24,280 --> 00:09:27,120
Speaker 2: once you do that, now you can understand that data

177
00:09:27,160 --> 00:09:30,160
Speaker 2: in the context of how the body works. Okay, so

178
00:09:30,240 --> 00:09:34,000
Speaker 2: once we've done that, then we calculate how energy flows

179
00:09:34,040 --> 00:09:37,240
Speaker 2: through the body, we calculate how momentum flows through the body,

180
00:09:37,520 --> 00:09:40,359
Speaker 2: and once we've done that, we can analyze how efficiently

181
00:09:40,360 --> 00:09:43,840
Speaker 2: the athlete is moving. Are they generating energy and momentum

182
00:09:44,080 --> 00:09:47,000
Speaker 2: in the direction that they want to generate in? What

183
00:09:47,160 --> 00:09:50,199
Speaker 2: is that desired direction? So we then we calculate all

184
00:09:50,240 --> 00:09:54,640
Speaker 2: sorts of metrics around movement efficiency and direction. Then once

185
00:09:54,679 --> 00:09:58,319
Speaker 2: we calculate all those metrics, now we can understand how

186
00:09:58,320 --> 00:10:00,720
Speaker 2: those relate to what you're trying to do. Throw the

187
00:10:00,720 --> 00:10:01,960
Speaker 2: ball as hard as possible.

188
00:10:02,040 --> 00:10:05,200
Speaker 1: So so presumably with the picture, what you want to

189
00:10:05,240 --> 00:10:08,640
Speaker 1: optimize for is is having as much of the picture's

190
00:10:09,520 --> 00:10:13,199
Speaker 1: energy of their body go toward making the ball go

191
00:10:13,280 --> 00:10:15,760
Speaker 1: toward home plate. Exactly right, I mean that is that? Yeah,

192
00:10:15,800 --> 00:10:17,800
Speaker 1: they mail optimization problem mail it.

193
00:10:17,840 --> 00:10:20,120
Speaker 2: Yeah, Yeah, that's exactly it. And that's the problem that

194
00:10:20,160 --> 00:10:22,840
Speaker 2: we try to understand. So when we build our sort

195
00:10:22,840 --> 00:10:25,679
Speaker 2: of models regarding like how does a pitcher create efficient

196
00:10:25,679 --> 00:10:28,640
Speaker 2: fastball velocity, one of the most important things that comes

197
00:10:28,679 --> 00:10:31,600
Speaker 2: out of those models is lining up the direction of

198
00:10:31,600 --> 00:10:33,960
Speaker 2: your torsore rotation with the direction of your arm rotation.

199
00:10:34,400 --> 00:10:37,439
Speaker 1: The pitching bution is crazy complex, right, Like they're they

200
00:10:37,520 --> 00:10:39,640
Speaker 1: kick their leg up and they got their front arms

201
00:10:39,679 --> 00:10:42,200
Speaker 1: doing something and their back arms going back like a

202
00:10:42,280 --> 00:10:46,040
Speaker 1: lot is happening. Yeah, the sort of platonic ideal is

203
00:10:46,120 --> 00:10:50,160
Speaker 1: every little millimeter of every motion is going toward maximizing

204
00:10:50,160 --> 00:10:52,000
Speaker 1: the energy of the ball going toward home plate.

205
00:10:52,160 --> 00:10:56,000
Speaker 2: Yes, and not just not just maximizing like in a vacuum,

206
00:10:56,160 --> 00:10:59,640
Speaker 2: but doing it in the most efficient way, because if

207
00:10:59,640 --> 00:11:02,560
Speaker 2: you just sort of maximize in a vacuum, maybe you're

208
00:11:02,559 --> 00:11:04,680
Speaker 2: transferring that energy in a way that hurts your elbow.

209
00:11:04,840 --> 00:11:07,040
Speaker 2: You're transferring that energy in a way that hurts your shoulder.

210
00:11:07,520 --> 00:11:09,280
Speaker 2: So not only do we figure out how can a

211
00:11:09,280 --> 00:11:11,880
Speaker 2: pitcher maximize it, but we try to figure out how

212
00:11:11,920 --> 00:11:14,360
Speaker 2: they can have that energy and momentum go in a

213
00:11:14,440 --> 00:11:18,800
Speaker 2: direction that doesn't hurt their joints, So try to have

214
00:11:18,920 --> 00:11:21,960
Speaker 2: them throw the ball a little harder while also reducing

215
00:11:21,960 --> 00:11:23,160
Speaker 2: their injury risk a little bit.

216
00:11:23,559 --> 00:11:30,640
Speaker 1: So you're whatever doing the math at reboot HQ, and

217
00:11:30,679 --> 00:11:33,319
Speaker 1: then what are you sending back to the.

218
00:11:33,320 --> 00:11:38,199
Speaker 2: Team, So some teams so it sort of so again,

219
00:11:38,440 --> 00:11:39,760
Speaker 2: all right, I told you that was the end of

220
00:11:39,760 --> 00:11:41,160
Speaker 2: the biomechanics train analogy.

221
00:11:41,720 --> 00:11:43,320
Speaker 3: I'm going to bring it back for a brief second.

222
00:11:43,840 --> 00:11:44,679
Speaker 1: Okay, I'm ready.

223
00:11:45,800 --> 00:11:48,160
Speaker 2: So we go all the way to building a report

224
00:11:48,240 --> 00:11:51,200
Speaker 2: that has a bunch of suggestions. So there's a report

225
00:11:51,240 --> 00:11:53,719
Speaker 2: that's like, this is how efficient you are. You can

226
00:11:53,840 --> 00:11:55,599
Speaker 2: like tilt your torso a little bit to be a

227
00:11:55,640 --> 00:11:57,439
Speaker 2: little bit more efficient. You can tilt your arm a

228
00:11:57,440 --> 00:11:59,280
Speaker 2: little bit to be a little bit more efficient. So

229
00:11:59,320 --> 00:12:02,440
Speaker 2: we go all the way to generating a report that's

230
00:12:02,480 --> 00:12:03,840
Speaker 2: like the final stop on the train.

231
00:12:04,480 --> 00:12:06,400
Speaker 1: And that's a report for for one pitcher.

232
00:12:06,679 --> 00:12:09,120
Speaker 3: Yeah, that's a report on a pitcher for whatever.

233
00:12:09,480 --> 00:12:13,200
Speaker 1: And and in a way that's like the nerdiest it's

234
00:12:13,280 --> 00:12:15,360
Speaker 1: what a pitching coach does, but just in a way

235
00:12:15,440 --> 00:12:16,120
Speaker 1: nerdier way.

236
00:12:16,360 --> 00:12:16,600
Speaker 3: Yeah.

237
00:12:16,640 --> 00:12:20,120
Speaker 2: Well, so we try to sort of like give the

238
00:12:20,160 --> 00:12:23,320
Speaker 2: coaches superpowers, you know, even though that you know, the

239
00:12:23,360 --> 00:12:25,800
Speaker 2: coaches like can look at an athlete and understand a

240
00:12:25,840 --> 00:12:29,120
Speaker 2: lot about the athlete just by looking at them. We

241
00:12:29,160 --> 00:12:32,920
Speaker 2: try to make a report that can really amplify what

242
00:12:33,000 --> 00:12:36,160
Speaker 2: the coach is already doing, maybe help them discover some

243
00:12:36,240 --> 00:12:39,560
Speaker 2: things they weren't thinking about, or measure some things that

244
00:12:39,600 --> 00:12:41,600
Speaker 2: they were thinking about, but now they can track those

245
00:12:41,640 --> 00:12:44,040
Speaker 2: a little bit easier. So that's you know, that's the

246
00:12:44,160 --> 00:12:47,880
Speaker 2: ultimate thing that we produces a report that can sort

247
00:12:47,880 --> 00:12:49,719
Speaker 2: of like we say, give a coach superpowers.

248
00:12:51,760 --> 00:12:53,960
Speaker 1: Where was the trained metaphor doing that? Is there an

249
00:12:53,960 --> 00:12:57,000
Speaker 1: earlier station of disembarkation exactly?

250
00:12:57,760 --> 00:13:01,600
Speaker 2: So a lot of teams now are hiring people with

251
00:13:01,679 --> 00:13:06,080
Speaker 2: biomechanics and analytics backgrounds, so rather than just use our

252
00:13:06,200 --> 00:13:09,000
Speaker 2: reports out of the box, they want to build their

253
00:13:09,040 --> 00:13:11,440
Speaker 2: own reports and their own statistical models and their own

254
00:13:11,480 --> 00:13:14,559
Speaker 2: AI models. So we also get let people get off

255
00:13:14,600 --> 00:13:17,040
Speaker 2: the train a little bit earlier and build whatever they

256
00:13:17,120 --> 00:13:19,120
Speaker 2: want on top of the data that we're generating.

257
00:13:19,360 --> 00:13:22,880
Speaker 1: Are they not going to disintermediate you once those people

258
00:13:22,880 --> 00:13:25,400
Speaker 1: are there? Does that reduce the value you provide to

259
00:13:25,440 --> 00:13:26,439
Speaker 1: the team.

260
00:13:26,120 --> 00:13:28,880
Speaker 2: Now because we still have to process all that data.

261
00:13:29,520 --> 00:13:32,720
Speaker 1: Do you have some like IP or like why can't

262
00:13:32,920 --> 00:13:35,400
Speaker 1: somebody like you who works for a team just do

263
00:13:35,480 --> 00:13:36,040
Speaker 1: that without you?

264
00:13:36,160 --> 00:13:39,720
Speaker 2: That is a great question. We answered this question all

265
00:13:39,760 --> 00:13:44,920
Speaker 2: the time. Is because it's a very complex engineering problem.

266
00:13:44,960 --> 00:13:47,440
Speaker 2: Not only to do all the math that the physics

267
00:13:47,440 --> 00:13:50,520
Speaker 2: based math, to calculate the energy, calculate the momentum, like

268
00:13:50,600 --> 00:13:54,000
Speaker 2: all of that math is really hard, but also to

269
00:13:54,080 --> 00:13:57,959
Speaker 2: do it at a very large scale. So someone like

270
00:13:58,040 --> 00:14:00,760
Speaker 2: me in grad school learned how to do that on

271
00:14:00,800 --> 00:14:03,880
Speaker 2: a sample size. You know, my PhD was actually really

272
00:14:03,920 --> 00:14:06,439
Speaker 2: just one pitch, but lots of people do it on

273
00:14:06,520 --> 00:14:10,040
Speaker 2: like ten pitches or maybe one hundred pitches, So we

274
00:14:10,200 --> 00:14:16,760
Speaker 2: do it on several thousand pitches like and swings like

275
00:14:16,880 --> 00:14:17,520
Speaker 2: every morning.

276
00:14:18,240 --> 00:14:18,480
Speaker 3: You know.

277
00:14:18,559 --> 00:14:23,600
Speaker 2: There's just every team has like seven affiliates, so there's

278
00:14:23,920 --> 00:14:26,120
Speaker 2: you know, one hundred and fifty games every day that

279
00:14:26,160 --> 00:14:26,560
Speaker 2: need to be.

280
00:14:26,680 --> 00:14:28,640
Speaker 1: So you're doing this on farm teams as well.

281
00:14:28,720 --> 00:14:29,600
Speaker 3: Yeah, exactly.

282
00:14:29,960 --> 00:14:32,000
Speaker 2: So not only did we solve the problem of doing

283
00:14:32,000 --> 00:14:34,960
Speaker 2: it for like one pitcher in a way that's really actionable,

284
00:14:35,320 --> 00:14:38,240
Speaker 2: but we solved the problem doing it for every game

285
00:14:38,560 --> 00:14:41,840
Speaker 2: every day, you know, so that you have the data

286
00:14:41,880 --> 00:14:42,720
Speaker 2: when you wake up.

287
00:14:44,720 --> 00:14:46,920
Speaker 1: So you guys are doing it at scale. The answer

288
00:14:46,920 --> 00:14:48,760
Speaker 1: to why it is that that there is in fact

289
00:14:48,800 --> 00:14:51,680
Speaker 1: an economy of scale, a benefit of scale.

290
00:14:51,480 --> 00:14:53,640
Speaker 3: Yes, which you have, Yes, exactly, Yeah. Yeah.

291
00:14:54,600 --> 00:14:57,960
Speaker 1: What's what's a specific example of a thing that a

292
00:14:58,040 --> 00:15:01,200
Speaker 1: coach might tell a pitcher in response to your report

293
00:15:01,240 --> 00:15:04,040
Speaker 1: to try and get them to throw whatever differently?

294
00:15:05,440 --> 00:15:06,400
Speaker 3: A really.

295
00:15:08,080 --> 00:15:12,720
Speaker 2: Common low hanging fruit type of piece of feedback that

296
00:15:12,840 --> 00:15:16,120
Speaker 2: often comes out of the reports is how a pitcher

297
00:15:16,400 --> 00:15:21,040
Speaker 2: is using their lead arm in a typical pitching motion,

298
00:15:21,320 --> 00:15:23,720
Speaker 2: the pitcher will reach forward with their lead arm.

299
00:15:24,160 --> 00:15:25,600
Speaker 1: So, just to be clear, the lead arm is the

300
00:15:25,680 --> 00:15:27,520
Speaker 1: arm that is not holding the ball.

301
00:15:27,400 --> 00:15:29,080
Speaker 3: Right right, yeah, the arm in front of you.

302
00:15:29,200 --> 00:15:29,480
Speaker 1: Yeah.

303
00:15:29,680 --> 00:15:30,440
Speaker 3: Yeah.

304
00:15:30,480 --> 00:15:32,720
Speaker 2: A pitcher will reach forward with that lead arm while

305
00:15:32,760 --> 00:15:35,600
Speaker 2: the rear arm is holding the baseball, and they'll rotate

306
00:15:35,800 --> 00:15:38,960
Speaker 2: that lead arm really hard, and that's the thing that

307
00:15:39,040 --> 00:15:43,640
Speaker 2: kind of initiates the torso rotation. So a very common

308
00:15:43,720 --> 00:15:47,040
Speaker 2: flaw that we see is if a pitcher has a

309
00:15:47,200 --> 00:15:51,400
Speaker 2: very vertical pitching arm, they're pitching the ball way over

310
00:15:51,440 --> 00:15:54,520
Speaker 2: the top of their head, but their lead arm when

311
00:15:54,560 --> 00:15:56,760
Speaker 2: they pull it through, when they swing it through, they

312
00:15:56,800 --> 00:15:58,400
Speaker 2: swing it in a very flat.

313
00:15:58,160 --> 00:16:02,040
Speaker 1: Plane as horizontal.

314
00:16:01,480 --> 00:16:05,520
Speaker 2: Horizontal, right, yeah, That is not a very efficient plane

315
00:16:05,560 --> 00:16:07,320
Speaker 2: to use when you're throwing the ball on a very

316
00:16:07,400 --> 00:16:10,560
Speaker 2: vertical plane. So a very common low hang piece of

317
00:16:10,560 --> 00:16:15,280
Speaker 2: fruit feedback that comes out of the reports is having

318
00:16:15,320 --> 00:16:18,960
Speaker 2: pitchers just try to rotate their lead arm pull with

319
00:16:19,000 --> 00:16:21,840
Speaker 2: their lead arm in a more vertical plane to better

320
00:16:21,920 --> 00:16:24,200
Speaker 2: match what their torso is doing, a better match with

321
00:16:24,280 --> 00:16:25,240
Speaker 2: their pitching arm is doing.

322
00:16:26,320 --> 00:16:28,160
Speaker 1: That's a good one. I feel like that one's so

323
00:16:28,240 --> 00:16:30,360
Speaker 1: simple that you don't want it to get out that

324
00:16:30,480 --> 00:16:31,840
Speaker 1: everybody ell to start looking at it.

325
00:16:31,920 --> 00:16:34,440
Speaker 2: No, I mean really, I mean like it's this has

326
00:16:34,480 --> 00:16:37,080
Speaker 2: happened when I went to talk to a team and

327
00:16:37,120 --> 00:16:40,080
Speaker 2: we talked about some pieces, you know, some low hanging fruit,

328
00:16:40,120 --> 00:16:43,120
Speaker 2: and they're like, okay, great, we'll take the lead arm thing,

329
00:16:43,160 --> 00:16:45,800
Speaker 2: will implement it everywhere. And I'm like, well, what about

330
00:16:45,840 --> 00:16:48,160
Speaker 2: like the other ten pages of the report.

331
00:16:49,800 --> 00:16:50,760
Speaker 3: Good with the lead arm thing.

332
00:16:52,200 --> 00:16:56,240
Speaker 1: Thanks by so for you. The end of the train

333
00:16:56,600 --> 00:16:59,840
Speaker 1: is the report. But that report goes to the coach, right,

334
00:17:00,000 --> 00:17:03,680
Speaker 1: and so presumably the meaningful change hasn't happened yet, right,

335
00:17:03,680 --> 00:17:06,040
Speaker 1: It has to somehow get from the coach to the picture.

336
00:17:06,720 --> 00:17:08,840
Speaker 1: And like, I know that piece of it is not

337
00:17:09,640 --> 00:17:11,679
Speaker 1: your business now, but it was kind of your business

338
00:17:11,680 --> 00:17:13,959
Speaker 1: when you were at the Dodgers. Presumably you're familiar with

339
00:17:14,000 --> 00:17:16,240
Speaker 1: it now, Like, how does that piece of it work?

340
00:17:16,320 --> 00:17:18,120
Speaker 1: Is it like the pitching coach is like reading from

341
00:17:18,160 --> 00:17:20,320
Speaker 1: the report to the picture. I imagine not, but I

342
00:17:20,359 --> 00:17:20,679
Speaker 1: don't know.

343
00:17:20,960 --> 00:17:27,480
Speaker 2: No, definitely, definite, definitely not. Even at the Dodgers, my

344
00:17:27,680 --> 00:17:32,000
Speaker 2: role was not being the one to coach the players.

345
00:17:32,200 --> 00:17:34,800
Speaker 1: It's like, whatever you do, don't talk to the pictures. Man,

346
00:17:34,960 --> 00:17:35,879
Speaker 1: go back to your computer.

347
00:17:36,960 --> 00:17:39,760
Speaker 2: Yeah no, no, I mean thankfully I got to be

348
00:17:40,400 --> 00:17:43,280
Speaker 2: in the in the room as you know, dark Walld,

349
00:17:43,320 --> 00:17:47,439
Speaker 2: the interaction is happening. But I think that is the

350
00:17:47,440 --> 00:17:52,439
Speaker 2: the art of coaching that is so important is understanding

351
00:17:52,960 --> 00:17:55,960
Speaker 2: the picture and how the pitcher thinks about themselves and

352
00:17:56,080 --> 00:18:00,080
Speaker 2: giving the right feedback to have the picture do the

353
00:18:00,119 --> 00:18:02,240
Speaker 2: thing that you that you want them to do.

354
00:18:02,800 --> 00:18:06,000
Speaker 1: Uh huh. Knowing how to talk to a player in

355
00:18:06,040 --> 00:18:09,360
Speaker 1: a way that is not generic. Presumably different pitchers need

356
00:18:09,400 --> 00:18:12,680
Speaker 1: to hear different things, even if the outcome is the same, right.

357
00:18:12,760 --> 00:18:15,399
Speaker 2: I mean, there's a there's a classic like debate in

358
00:18:15,520 --> 00:18:19,200
Speaker 2: baseball of like do you swing down on the ball

359
00:18:19,320 --> 00:18:22,280
Speaker 2: or do you swing with an uppercut? And in reality,

360
00:18:22,440 --> 00:18:25,760
Speaker 2: like the batpath is an arc, the path goes down

361
00:18:25,800 --> 00:18:28,440
Speaker 2: and then the path goes up. But some coach, some

362
00:18:28,480 --> 00:18:32,120
Speaker 2: players respond better when a coach will say swing down

363
00:18:32,160 --> 00:18:34,960
Speaker 2: on the ball, and some players spawn better when a

364
00:18:35,000 --> 00:18:37,040
Speaker 2: coach will say, you know, have get a little bit

365
00:18:37,040 --> 00:18:39,640
Speaker 2: more uppercut, do your swing. But in reality, you know,

366
00:18:39,960 --> 00:18:41,760
Speaker 2: the bat goes down and the bat goes up. So

367
00:18:41,800 --> 00:18:44,639
Speaker 2: it's like really understanding what is the player what helps

368
00:18:44,640 --> 00:18:45,520
Speaker 2: the player the most?

369
00:18:45,960 --> 00:18:48,320
Speaker 1: So I heard you use this phrase in another interview

370
00:18:48,680 --> 00:18:50,600
Speaker 1: that I think is kind of what you're talking about here,

371
00:18:50,600 --> 00:18:54,159
Speaker 1: And it's feel versus real? What is that? What is

372
00:18:54,160 --> 00:18:55,080
Speaker 1: feel versus real?

373
00:18:55,320 --> 00:18:59,480
Speaker 2: So it's exactly what we're talking about. Is sometimes with

374
00:18:59,560 --> 00:19:03,080
Speaker 2: the apt let feels like they're doing is not.

375
00:19:03,359 --> 00:19:05,000
Speaker 3: What's actually happening.

376
00:19:05,200 --> 00:19:08,480
Speaker 2: But if you understand what the athlete feels like they're doing,

377
00:19:08,920 --> 00:19:12,720
Speaker 2: you can give feedback that interacts with how they're feeling.

378
00:19:12,960 --> 00:19:16,000
Speaker 2: So if they feel like they're swinging down on the

379
00:19:16,040 --> 00:19:18,880
Speaker 2: ball and you give them feedback related to that, even

380
00:19:18,880 --> 00:19:21,560
Speaker 2: if they're actually swinging with an uppercut, you know, the

381
00:19:21,640 --> 00:19:24,119
Speaker 2: important thing is like how do they feel and how

382
00:19:24,160 --> 00:19:26,520
Speaker 2: do you give them feedback related to how they feel,

383
00:19:26,760 --> 00:19:30,400
Speaker 2: which then impacts is real. But it's understanding the interplay

384
00:19:30,400 --> 00:19:31,600
Speaker 2: between the feel and the real.

385
00:19:32,920 --> 00:19:36,840
Speaker 1: Yeah, it's wild that, like there is there is a

386
00:19:36,920 --> 00:19:39,800
Speaker 1: human being doing a thing, and then there's this huge

387
00:19:39,920 --> 00:19:44,560
Speaker 1: industrial machinery of your company and the team and all

388
00:19:44,600 --> 00:19:47,639
Speaker 1: of these scientists and all these cameras and computers that

389
00:19:47,680 --> 00:19:50,640
Speaker 1: are trying to get this human being to change their

390
00:19:50,640 --> 00:19:52,320
Speaker 1: behavior in a very subtle way.

391
00:19:52,440 --> 00:19:52,960
Speaker 3: Yeah, exactly.

392
00:19:53,000 --> 00:19:57,560
Speaker 1: It's a behavior that is in many ways intuitive right,

393
00:19:57,600 --> 00:20:00,520
Speaker 1: it's sort of partly conscious but partly intuitive him like

394
00:20:00,560 --> 00:20:05,480
Speaker 1: there's a really interesting human being at the center.

395
00:20:05,200 --> 00:20:05,840
Speaker 3: Of all of this.

396
00:20:06,160 --> 00:20:09,080
Speaker 2: Yeah, And I think that's also why it's so important

397
00:20:09,600 --> 00:20:12,720
Speaker 2: to have the coach in the loop, because they understand

398
00:20:13,160 --> 00:20:15,800
Speaker 2: the human being even beyond just like the feel versus

399
00:20:15,800 --> 00:20:19,240
Speaker 2: real aspect of like did the athlete not get a

400
00:20:19,240 --> 00:20:21,560
Speaker 2: good night's sleep last night, then maybe today is not

401
00:20:21,760 --> 00:20:24,000
Speaker 2: a good day to give them feedback. Are they going

402
00:20:24,040 --> 00:20:27,040
Speaker 2: through challenges, you know, with a significant other, are they

403
00:20:27,080 --> 00:20:30,040
Speaker 2: going through challenges in other ways? Or today are they

404
00:20:30,080 --> 00:20:31,920
Speaker 2: really fired up? So today is a really good day

405
00:20:31,920 --> 00:20:34,879
Speaker 2: to give them feedback. It's like understanding the human being.

406
00:20:34,920 --> 00:20:38,000
Speaker 2: I'll give him a shout out. Connor McGuinness is the

407
00:20:38,000 --> 00:20:41,439
Speaker 2: assistant pitching coach right now for the Dodgers, and he

408
00:20:41,640 --> 00:20:44,280
Speaker 2: was so good at this. He would always talk about

409
00:20:45,040 --> 00:20:49,919
Speaker 2: he never felt comfortable truly coaching a player until the

410
00:20:49,960 --> 00:20:52,159
Speaker 2: player trusted him, until he felt like he had a

411
00:20:52,160 --> 00:20:54,760
Speaker 2: good enough relationship with the player with the player would

412
00:20:54,880 --> 00:20:57,239
Speaker 2: feel comfortable taking his feedback.

413
00:20:57,280 --> 00:20:59,560
Speaker 1: Like start with the player as a human beings, Start

414
00:20:59,600 --> 00:21:01,760
Speaker 1: with the player and get to get to.

415
00:21:02,400 --> 00:21:04,119
Speaker 2: And this is why I think it's going to be

416
00:21:04,200 --> 00:21:09,920
Speaker 2: really really difficult to have a product that goes direct

417
00:21:10,080 --> 00:21:12,919
Speaker 2: to a player. We've dabbled with that, you know, and

418
00:21:13,000 --> 00:21:15,679
Speaker 2: lots of companies have dabbled with film yourself with your

419
00:21:15,680 --> 00:21:19,680
Speaker 2: iPhone and you get some feedback, you know, on your movement.

420
00:21:20,280 --> 00:21:23,800
Speaker 2: But I think there's so much subjective stuff that goes

421
00:21:23,800 --> 00:21:26,600
Speaker 2: into what we're talking about, how does the athlete move

422
00:21:26,720 --> 00:21:29,280
Speaker 2: and how do they feel like they're moving, that I

423
00:21:29,280 --> 00:21:31,000
Speaker 2: think it's going to be really hard to solve the

424
00:21:31,119 --> 00:21:34,080
Speaker 2: challenge of giving an athlete feedback without a coach.

425
00:21:36,800 --> 00:21:40,120
Speaker 1: In a minute, Jimmy's work in the NBA and why

426
00:21:40,160 --> 00:21:43,480
Speaker 1: figuring out how to help NBA players shoot better is

427
00:21:43,520 --> 00:21:55,159
Speaker 1: actually a really hard problem. What are you doing in basketball?

428
00:21:55,960 --> 00:21:59,280
Speaker 2: Basketball is very very cool because it's a slightly different

429
00:21:59,760 --> 00:22:05,120
Speaker 2: channe In baseball, like we've been talking about, the challenge

430
00:22:05,119 --> 00:22:10,200
Speaker 2: for a pitcher is mostly just maximizing efficiency, or even

431
00:22:10,240 --> 00:22:12,600
Speaker 2: for a hitter, you know, they're reacting to the pitcher,

432
00:22:13,040 --> 00:22:15,399
Speaker 2: but they're still trying to swing as hard as they

433
00:22:15,440 --> 00:22:17,880
Speaker 2: can and hit the ball as far as they can.

434
00:22:18,359 --> 00:22:20,760
Speaker 2: So basketball is very different because you got a target.

435
00:22:22,280 --> 00:22:24,479
Speaker 2: You're not trying to do this thing as hard as

436
00:22:24,560 --> 00:22:25,520
Speaker 2: humanly possible.

437
00:22:25,880 --> 00:22:28,280
Speaker 1: The target is the hoop. The targets to be clear, the.

438
00:22:28,200 --> 00:22:29,160
Speaker 3: Target is the hoop.

439
00:22:29,359 --> 00:22:33,880
Speaker 2: So that's a really interesting kind of like motor control

440
00:22:34,080 --> 00:22:37,679
Speaker 2: problem of no matter where you are on the court,

441
00:22:38,240 --> 00:22:41,040
Speaker 2: how good are you at getting the ball to do

442
00:22:41,080 --> 00:22:44,120
Speaker 2: what you want it to do. We've been finding that

443
00:22:44,160 --> 00:22:49,600
Speaker 2: there seems to be lots of trade offs that good

444
00:22:49,600 --> 00:22:53,000
Speaker 2: shooters are making regarding like when do they release the

445
00:22:53,000 --> 00:22:55,879
Speaker 2: ball in the course of their jump, how high of

446
00:22:55,920 --> 00:22:58,600
Speaker 2: an arc do they use when they release the ball,

447
00:22:59,080 --> 00:23:01,639
Speaker 2: how high they have their release point when they.

448
00:23:01,520 --> 00:23:02,160
Speaker 3: Release the ball.

449
00:23:02,400 --> 00:23:05,360
Speaker 2: So there's all sorts of trade offs that these shooters

450
00:23:05,400 --> 00:23:09,240
Speaker 2: seem to be making related to how good they are

451
00:23:09,440 --> 00:23:13,360
Speaker 2: at controlling their own jump, controlling their own velocity, where

452
00:23:13,359 --> 00:23:15,520
Speaker 2: they are on the court, how close the nearest defender is.

453
00:23:16,040 --> 00:23:17,880
Speaker 2: It's it's a very different problem.

454
00:23:18,240 --> 00:23:21,440
Speaker 1: Sounds way harder, sounds way way harder, is that right

455
00:23:22,040 --> 00:23:25,000
Speaker 1: for you? Harder for you to sort of solve to

456
00:23:25,600 --> 00:23:27,800
Speaker 1: write a report that says, do this different thing and

457
00:23:27,840 --> 00:23:30,120
Speaker 1: you'll hit a higher percentage of your shots.

458
00:23:29,840 --> 00:23:33,320
Speaker 2: Right exactly? Yeah, And which makes it which makes it fun.

459
00:23:33,400 --> 00:23:34,960
Speaker 2: I love solving hard podcasts.

460
00:23:35,640 --> 00:23:37,840
Speaker 1: What have you solved in basketball so far?

461
00:23:38,280 --> 00:23:40,480
Speaker 2: The big the first challenge that we had to solve

462
00:23:40,600 --> 00:23:44,399
Speaker 2: was the scale challenge in basketball. There isn't quite as

463
00:23:44,520 --> 00:23:48,720
Speaker 2: much there aren't quite as many games in basketball, but

464
00:23:49,119 --> 00:23:52,920
Speaker 2: the data is a lot harder to parse because the

465
00:23:53,359 --> 00:23:57,000
Speaker 2: events aren't as distinct. It's not like it's not like

466
00:23:57,280 --> 00:23:59,479
Speaker 2: this is a shot, I mean only for free throws.

467
00:23:59,560 --> 00:24:01,400
Speaker 1: It's a more continuous.

468
00:24:00,880 --> 00:24:03,760
Speaker 2: Continuous motion. So the first big challenge was how do

469
00:24:03,800 --> 00:24:06,840
Speaker 2: we even just isolate the events that we care about.

470
00:24:06,800 --> 00:24:09,200
Speaker 1: Like what is time? When does a shot begin? When

471
00:24:09,240 --> 00:24:10,840
Speaker 1: does a time equals zero for shot?

472
00:24:11,000 --> 00:24:15,520
Speaker 2: But kind of debatable, right exactly. So that, yeah, that

473
00:24:15,600 --> 00:24:17,120
Speaker 2: was the first big challenge that we had to solve.

474
00:24:17,240 --> 00:24:21,160
Speaker 2: Was just sort of like the data engineering challenge. And now,

475
00:24:21,240 --> 00:24:26,880
Speaker 2: like I said, our reports are more than telling you,

476
00:24:26,880 --> 00:24:29,240
Speaker 2: you know, how to be more or less efficient. It's

477
00:24:29,240 --> 00:24:32,280
Speaker 2: trying to surface the trade offs that you're making. Where

478
00:24:32,280 --> 00:24:34,639
Speaker 2: in your jump are you are you releasing the ball,

479
00:24:34,920 --> 00:24:37,480
Speaker 2: how high is your release point? What kind of arc

480
00:24:37,600 --> 00:24:40,160
Speaker 2: are you using? And how does that compare to other

481
00:24:40,320 --> 00:24:41,320
Speaker 2: arcs you could be using?

482
00:24:41,600 --> 00:24:43,280
Speaker 1: What like what what are the trade offs?

483
00:24:44,560 --> 00:24:50,120
Speaker 2: A really really interesting trade off to me is how

484
00:24:50,200 --> 00:24:53,119
Speaker 2: much arc are you putting on the basketball? Not just

485
00:24:53,200 --> 00:24:57,000
Speaker 2: to like evade a defender. But there's a trade off

486
00:24:57,040 --> 00:25:00,800
Speaker 2: where if you shoot the ball at a highhigher arc,

487
00:25:01,680 --> 00:25:04,560
Speaker 2: you have to use more velocity to get the ball

488
00:25:04,560 --> 00:25:06,520
Speaker 2: to go all the way to the rim.

489
00:25:07,119 --> 00:25:10,640
Speaker 1: Right, because it's going to travel farther in total in space.

490
00:25:10,600 --> 00:25:15,480
Speaker 2: Right, And if you are not the most coordinated human,

491
00:25:16,080 --> 00:25:18,919
Speaker 2: it might be harder for you to add more velocity

492
00:25:19,840 --> 00:25:22,399
Speaker 2: in a really in a really precise way. So the

493
00:25:22,440 --> 00:25:25,440
Speaker 2: more arc you have, the more prone you can be

494
00:25:25,600 --> 00:25:29,480
Speaker 2: to what we would call like velocity errors overshooting undershooting.

495
00:25:29,880 --> 00:25:33,240
Speaker 2: The advantage you get when you create more arc is

496
00:25:33,240 --> 00:25:35,760
Speaker 2: if you if you imagine the ball coming down from

497
00:25:35,760 --> 00:25:38,119
Speaker 2: that arc and the angle the angle with which it

498
00:25:38,160 --> 00:25:41,200
Speaker 2: approaches the rim, it approaches at a steeper angle, which

499
00:25:41,240 --> 00:25:44,199
Speaker 2: means you literally have more rim to aim at.

500
00:25:44,560 --> 00:25:46,440
Speaker 3: So if you're hot, so if you shoot.

501
00:25:46,160 --> 00:25:49,000
Speaker 2: Higher, you know this is a steph Curry thing. He

502
00:25:49,040 --> 00:25:52,760
Speaker 2: has a really high arc. Presumably this is the hypothesis,

503
00:25:52,800 --> 00:25:55,120
Speaker 2: because he's one of the most coordinated humans on the.

504
00:25:55,040 --> 00:25:59,480
Speaker 1: Planet, so he sea.

505
00:25:59,359 --> 00:26:01,680
Speaker 2: So he can use a really high arc because he's

506
00:26:01,760 --> 00:26:05,000
Speaker 2: really good at controlling his velocity output. Oh huh, so

507
00:26:05,040 --> 00:26:07,680
Speaker 2: he can really dial in how hard he releases the ball,

508
00:26:07,960 --> 00:26:10,160
Speaker 2: which means he gets the advantage of having more rimmed

509
00:26:10,160 --> 00:26:13,240
Speaker 2: to aim at, where as somebody who's bad at controlling

510
00:26:13,280 --> 00:26:16,040
Speaker 2: their velocity output when they try to aim higher, they'll

511
00:26:16,080 --> 00:26:17,399
Speaker 2: just overshoot an undershoot.

512
00:26:17,720 --> 00:26:20,880
Speaker 1: So is the notion, and this is something of an oversimplification,

513
00:26:21,040 --> 00:26:26,080
Speaker 1: but that for any given level of coordination, there is

514
00:26:26,119 --> 00:26:29,199
Speaker 1: some optimal arc, and the more coordinated you are, the

515
00:26:29,280 --> 00:26:33,280
Speaker 1: higher the optimal arc would be for you setting aside defense.

516
00:26:33,760 --> 00:26:35,280
Speaker 3: Maybe that's the hypothesis.

517
00:26:35,800 --> 00:26:38,879
Speaker 1: Yeah, I mean that seems like where what you were

518
00:26:38,920 --> 00:26:40,200
Speaker 1: saying goes exactly.

519
00:26:40,280 --> 00:26:42,680
Speaker 2: Yeah, that's the hypothesis. So that's what we're trying to

520
00:26:42,720 --> 00:26:43,160
Speaker 2: look into.

521
00:26:43,840 --> 00:26:46,960
Speaker 1: Okay, I'll be curious to see what you figure out.

522
00:26:47,600 --> 00:26:50,080
Speaker 1: Are free throws easier? Did you think of starting with

523
00:26:50,119 --> 00:26:50,679
Speaker 1: free throws?

524
00:26:51,200 --> 00:26:52,320
Speaker 3: Yeah?

525
00:26:52,520 --> 00:26:56,080
Speaker 2: Right, right now. Honestly, the phase we're at in basketball

526
00:26:57,960 --> 00:27:01,480
Speaker 2: is mostly just collecting a lot of data. So you know,

527
00:27:01,520 --> 00:27:04,159
Speaker 2: a starting pitcher will you know, throw ninety pitches in

528
00:27:04,200 --> 00:27:05,639
Speaker 2: a game, and now you have a sample size of

529
00:27:05,960 --> 00:27:08,840
Speaker 2: ninety pitches that are all the same, whereas a basketball

530
00:27:08,840 --> 00:27:11,320
Speaker 2: shooter only has a couple of free throws in a game.

531
00:27:12,160 --> 00:27:15,280
Speaker 2: You know, we're working on ways with teams of collecting

532
00:27:15,400 --> 00:27:18,320
Speaker 2: data in a practice setting, so getting a lot of

533
00:27:18,320 --> 00:27:22,320
Speaker 2: this data in a bigger chunks. But really at this

534
00:27:22,359 --> 00:27:24,520
Speaker 2: point it's collecting a lot of data so we can

535
00:27:24,600 --> 00:27:27,320
Speaker 2: do some of this research to explore some of these hypotheses.

536
00:27:27,640 --> 00:27:31,240
Speaker 1: Huh. So it seems like in basketball you're where you

537
00:27:31,280 --> 00:27:33,120
Speaker 1: were ten years ago or something in.

538
00:27:33,040 --> 00:27:34,840
Speaker 3: Baseball, right exactly.

539
00:27:35,440 --> 00:27:40,520
Speaker 1: So basketball is sort of one kind of frontier. It

540
00:27:40,560 --> 00:27:42,480
Speaker 1: seems like one thing you're trying to figure out and

541
00:27:42,520 --> 00:27:44,800
Speaker 1: haven't really cracked yet. What are some of the other

542
00:27:45,800 --> 00:27:48,359
Speaker 1: frontiers the other things you're figuring out, whether it's in

543
00:27:48,440 --> 00:27:51,520
Speaker 1: baseball or in the fundamental technology or whatever. What are

544
00:27:51,560 --> 00:27:52,000
Speaker 1: you working on?

545
00:27:54,000 --> 00:27:59,159
Speaker 2: We want to try to have a computer vision be

546
00:27:59,400 --> 00:28:04,639
Speaker 2: even more accessible. You know, there's been a lot and

547
00:28:04,680 --> 00:28:07,240
Speaker 2: a lot, a lot a lot of improvements over the

548
00:28:07,320 --> 00:28:11,080
Speaker 2: last ten years in computer vision where you can do

549
00:28:11,200 --> 00:28:17,520
Speaker 2: really good motion capture with just your iPhone, but still

550
00:28:17,560 --> 00:28:25,320
Speaker 2: for certain specialized movements, pitching, shooting, things like that, there's

551
00:28:25,320 --> 00:28:27,640
Speaker 2: still have little ways to go to get really really

552
00:28:27,640 --> 00:28:30,720
Speaker 2: good data straight from your iPhone. So that's one of

553
00:28:30,760 --> 00:28:35,160
Speaker 2: the frontiers, is like continuing to try to help understand

554
00:28:35,160 --> 00:28:39,640
Speaker 2: how to make computer vision more accessible. Another one is

555
00:28:40,640 --> 00:28:46,160
Speaker 2: more fitness based analysis. You know, it's interesting to think

556
00:28:46,200 --> 00:28:50,040
Speaker 2: about companies like Mirror and companies that have tried to

557
00:28:50,080 --> 00:28:55,080
Speaker 2: give people feedback and like a fitness environment, But how

558
00:28:55,120 --> 00:28:58,080
Speaker 2: do you give someone or how do you give like

559
00:28:58,120 --> 00:29:00,840
Speaker 2: a strength and conditioning coach good feedback that can be

560
00:29:00,960 --> 00:29:03,920
Speaker 2: used in a weight room setting. Yeah, and then continue

561
00:29:03,960 --> 00:29:07,680
Speaker 2: to explore other emotions and other sports, Like football is

562
00:29:07,680 --> 00:29:12,080
Speaker 2: a really interesting one because a lot of football is

563
00:29:12,360 --> 00:29:17,080
Speaker 2: interacting with other humans. Indeed, yeah, I mean that's obvious,

564
00:29:17,800 --> 00:29:21,680
Speaker 2: but like, how do you get a takeaway from like

565
00:29:21,760 --> 00:29:23,120
Speaker 2: two linemen interacting?

566
00:29:24,000 --> 00:29:26,440
Speaker 1: Ye, things, but that one I wonder. I mean, it

567
00:29:26,760 --> 00:29:29,560
Speaker 1: seems like if you think of the line, you know,

568
00:29:29,920 --> 00:29:33,400
Speaker 1: in football, it's so they're so on top of each other,

569
00:29:33,480 --> 00:29:35,360
Speaker 1: the defensive line, of the defensive line, that I feel

570
00:29:35,360 --> 00:29:37,320
Speaker 1: like vision might not be what you want. You might

571
00:29:37,360 --> 00:29:40,720
Speaker 1: want sensors, right, you might want pressure sensors in the

572
00:29:40,840 --> 00:29:43,040
Speaker 1: lineman's clothes or something. I don't know, I'm just making

573
00:29:43,080 --> 00:29:45,080
Speaker 1: that up, but like it seems like that might be

574
00:29:45,120 --> 00:29:47,000
Speaker 1: more useful, just because it's hard to see what's going

575
00:29:47,000 --> 00:29:47,680
Speaker 1: on in the line.

576
00:29:47,760 --> 00:29:48,040
Speaker 3: Yeah.

577
00:29:48,160 --> 00:29:52,920
Speaker 2: Yeah, yeah, except professional athletes don't like wearing random.

578
00:29:52,640 --> 00:29:57,120
Speaker 1: Things, well are aren't people trying to make sensors like

579
00:29:57,400 --> 00:29:59,400
Speaker 1: woven into the clothes. I mean, I feel like there

580
00:29:59,400 --> 00:30:02,080
Speaker 1: are ways you would just be putting on your jersey

581
00:30:02,160 --> 00:30:03,960
Speaker 1: or putting on your paths or whatever and they would

582
00:30:03,960 --> 00:30:04,920
Speaker 1: have the sensors built in.

583
00:30:05,040 --> 00:30:06,160
Speaker 3: Yeah. Yeah, one hundred percent.

584
00:30:06,200 --> 00:30:09,840
Speaker 2: There's lots of really cool technology of that is you know,

585
00:30:10,600 --> 00:30:13,760
Speaker 2: you know, microscopic sensors that are just woven, woven into

586
00:30:13,760 --> 00:30:15,480
Speaker 2: clothing for sure.

587
00:30:16,440 --> 00:30:20,040
Speaker 1: So so let's talk for a second more about the

588
00:30:20,080 --> 00:30:23,080
Speaker 1: consumer side. You sort of touched on it and moved on.

589
00:30:23,160 --> 00:30:25,960
Speaker 1: I mean, is that are you actively working on that

590
00:30:26,120 --> 00:30:28,320
Speaker 1: or is that just like yeah, it kind of seems interesting,

591
00:30:28,360 --> 00:30:30,640
Speaker 1: but not for now. We're too busy, like what's what's

592
00:30:30,640 --> 00:30:32,120
Speaker 1: happening on the consumer side.

593
00:30:33,120 --> 00:30:37,640
Speaker 2: We're not actively tackling in the consumer side right now.

594
00:30:37,720 --> 00:30:40,560
Speaker 2: And that the reason why we started with professional sports

595
00:30:40,640 --> 00:30:42,920
Speaker 2: is one it's because like it's what I know, you know,

596
00:30:42,960 --> 00:30:47,600
Speaker 2: I work for the Dodgers, but also because the value

597
00:30:47,680 --> 00:30:52,920
Speaker 2: proposition of what we're doing is so impactful. You know, yeah,

598
00:30:53,000 --> 00:30:54,920
Speaker 2: you know, you understand, you try to understand, like the

599
00:30:54,960 --> 00:30:58,120
Speaker 2: relationship between people have tried to put numbers on this,

600
00:30:58,280 --> 00:31:01,040
Speaker 2: it's the people have estimated, it's in the millions of dollars.

601
00:31:01,080 --> 00:31:03,920
Speaker 2: But the value of adding one mile an hour of

602
00:31:04,240 --> 00:31:08,080
Speaker 2: fastball velocity to a picture, you know, people have valued that,

603
00:31:08,360 --> 00:31:10,360
Speaker 2: We have valued that in millions of dollars.

604
00:31:10,880 --> 00:31:13,160
Speaker 1: Well, sure you will tell you millions of dollars, But

605
00:31:13,560 --> 00:31:16,800
Speaker 1: what is the like, I don't know what's a what's

606
00:31:16,800 --> 00:31:18,960
Speaker 1: a what's a top picture make? These days? I don't

607
00:31:18,960 --> 00:31:19,680
Speaker 1: even know anymore.

608
00:31:20,640 --> 00:31:21,960
Speaker 3: I mean, you want to talk about show.

609
00:31:23,680 --> 00:31:28,320
Speaker 1: Half a billion, hundreds of millions, half a billion? Yeah,

610
00:31:28,360 --> 00:31:32,480
Speaker 1: so so right, so so the marginal benefit has a

611
00:31:32,600 --> 00:31:35,480
Speaker 1: very large value. Yah a million dollars against one hundred

612
00:31:35,520 --> 00:31:37,800
Speaker 1: million dollars, like one percent better. If they're making a

613
00:31:37,880 --> 00:31:40,240
Speaker 1: hundred million dollars, it's worth a million dollars presumably.

614
00:31:40,400 --> 00:31:43,200
Speaker 2: Yeah, And we start to get the pro sports teams

615
00:31:43,200 --> 00:31:44,200
Speaker 2: to believe.

616
00:31:43,840 --> 00:31:48,120
Speaker 1: That, well, how many how many pro sports teams are

617
00:31:48,200 --> 00:31:50,360
Speaker 1: paying you at this point, about uh.

618
00:31:51,320 --> 00:31:54,480
Speaker 2: Close to ten in Major League Baseball and a couple

619
00:31:54,520 --> 00:31:55,040
Speaker 2: in the NBA.

620
00:31:55,120 --> 00:31:56,080
Speaker 3: The NBA is a lot newer.

621
00:31:57,240 --> 00:31:59,520
Speaker 1: And how big is the field? Like, what's sort of

622
00:31:59,560 --> 00:32:03,360
Speaker 1: the state of play in the field of I don't

623
00:32:03,360 --> 00:32:06,360
Speaker 1: even know what to say. Biomechanics as a service seems

624
00:32:06,360 --> 00:32:09,160
Speaker 1: like a niche construction of it. But what would you

625
00:32:09,200 --> 00:32:12,680
Speaker 1: say sports analytics? I mean, I guess that's not quite right.

626
00:32:12,720 --> 00:32:14,640
Speaker 1: What how do you construct the broader field?

627
00:32:15,840 --> 00:32:18,920
Speaker 2: Sports analytics is definitely a part of it. The data

628
00:32:18,920 --> 00:32:25,080
Speaker 2: that we provide is novel or is different than like

629
00:32:25,200 --> 00:32:28,400
Speaker 2: traditional sports analytics. So we don't really have a lot

630
00:32:28,440 --> 00:32:37,480
Speaker 2: of companies as competitors. Honestly, our biggest competitors are teams

631
00:32:38,280 --> 00:32:40,440
Speaker 2: wanting to try to do this type of thing internally.

632
00:32:40,760 --> 00:32:42,880
Speaker 2: How you know, hire a bunch of data engineers, hire

633
00:32:42,880 --> 00:32:46,560
Speaker 2: a bunch of software engineers, hire people with biomechanics backgrounds,

634
00:32:46,600 --> 00:32:49,320
Speaker 2: and try to build out these pipeline processing pipelines themselves.

635
00:32:50,080 --> 00:32:53,760
Speaker 1: Does every team have like a somebody with a PhD

636
00:32:53,880 --> 00:32:55,720
Speaker 1: in biomechanics working for them now?

637
00:32:55,760 --> 00:32:57,240
Speaker 3: In baseball? Yeah?

638
00:32:57,440 --> 00:32:57,800
Speaker 1: Wow?

639
00:32:58,440 --> 00:33:01,600
Speaker 2: In basketball and not Yeah, but they're starting to.

640
00:33:02,320 --> 00:33:06,120
Speaker 1: So I know people, I mean, I think in general

641
00:33:06,160 --> 00:33:08,440
Speaker 1: baseball fans like to complain, but so one of the

642
00:33:08,440 --> 00:33:11,880
Speaker 1: things they've complained about lately is the way analytics more

643
00:33:11,920 --> 00:33:15,520
Speaker 1: broadly made the game more boring, right, like the shift

644
00:33:16,800 --> 00:33:21,120
Speaker 1: and changing pictures more frequently and whatever else people complain about.

645
00:33:21,720 --> 00:33:27,520
Speaker 1: Are you do you fit into that at all? Uh?

646
00:33:28,400 --> 00:33:35,440
Speaker 3: That's a good that's a good question. Some people.

647
00:33:35,680 --> 00:33:38,680
Speaker 2: Some people complain about how hard pictures are throwing these

648
00:33:38,760 --> 00:33:43,000
Speaker 2: days because it creates more strikeouts, and people think strikeouts

649
00:33:43,000 --> 00:33:43,440
Speaker 2: are boring.

650
00:33:45,360 --> 00:33:47,200
Speaker 3: So maybe, yeah, so.

651
00:33:47,200 --> 00:33:50,600
Speaker 1: Maybe you need to get better at helping hitters. Really

652
00:33:50,680 --> 00:33:53,240
Speaker 1: you're out that way, you can even at back up, all.

653
00:33:53,200 --> 00:33:55,560
Speaker 2: Right, And that's what we talked about takeaways four Hitters

654
00:33:55,640 --> 00:33:58,160
Speaker 2: are harder because they're reacting to the picture.

655
00:33:59,000 --> 00:34:02,640
Speaker 1: So if you think about the field, what your company

656
00:34:02,720 --> 00:34:05,600
Speaker 1: say in five years, yeah, whatever is your kind of

657
00:34:05,680 --> 00:34:09,640
Speaker 1: medium term future that you think about, what is the

658
00:34:09,680 --> 00:34:12,680
Speaker 1: company in the sort of the world, the sports world

659
00:34:12,680 --> 00:34:15,560
Speaker 1: that you're interacting with look like at that time and

660
00:34:15,640 --> 00:34:17,359
Speaker 1: say whatever, five years, ten years.

661
00:34:18,800 --> 00:34:23,799
Speaker 2: What I hope is that what we are trying to

662
00:34:23,880 --> 00:34:32,680
Speaker 2: foster is an environment where coaches have really incredible tools

663
00:34:32,680 --> 00:34:35,920
Speaker 2: at their disposal to understand how an athlete moves. So,

664
00:34:35,960 --> 00:34:38,359
Speaker 2: you know, we talk about the very beginning where sort

665
00:34:38,400 --> 00:34:40,840
Speaker 2: of the still more or less the state of the

666
00:34:40,960 --> 00:34:43,560
Speaker 2: art is a coach just looks at an athlete, watches

667
00:34:43,680 --> 00:34:46,200
Speaker 2: video of an athlete, and tries to give the athlete

668
00:34:46,239 --> 00:34:48,840
Speaker 2: feedback regarding what they see on the video, what they

669
00:34:48,880 --> 00:34:52,880
Speaker 2: see with their eyes. We hope that the standard becomes

670
00:34:54,040 --> 00:34:59,399
Speaker 2: you use an analytical tool to help you understand how

671
00:34:59,440 --> 00:35:02,920
Speaker 2: the athlete is moving and to really like level up

672
00:35:02,960 --> 00:35:06,480
Speaker 2: your coaching because now you have this objective information about

673
00:35:06,480 --> 00:35:09,120
Speaker 2: how the athlete is moving. I kind of I kind

674
00:35:09,120 --> 00:35:13,960
Speaker 2: of make the analogy related to just radar guns. Before

675
00:35:14,080 --> 00:35:17,040
Speaker 2: radar guns were a thing, a coach would just like

676
00:35:17,160 --> 00:35:18,919
Speaker 2: look at a picture and be like, oh, that looks

677
00:35:18,920 --> 00:35:22,600
Speaker 2: pretty fast, you know, make an adjustment, and I think

678
00:35:22,640 --> 00:35:27,239
Speaker 2: that looks a little faster. But then radar guns came out,

679
00:35:27,239 --> 00:35:30,120
Speaker 2: and you could actually measure how fast the picture was throwing,

680
00:35:30,400 --> 00:35:33,040
Speaker 2: and you could actually measure if the picture is getting

681
00:35:33,440 --> 00:35:35,280
Speaker 2: throwing the ball harder based on your feedback.

682
00:35:35,800 --> 00:35:39,160
Speaker 1: Yeah, and now you're you're doing that but in a

683
00:35:40,560 --> 00:35:41,920
Speaker 1: in a way more complex way.

684
00:35:41,960 --> 00:35:42,360
Speaker 3: Exactly.

685
00:35:42,520 --> 00:35:50,000
Speaker 1: Yeah, we'll be back in a minute with the lightning round.

686
00:36:01,640 --> 00:36:03,319
Speaker 1: Let's do the lightning rounds. Let's start with a few

687
00:36:03,320 --> 00:36:07,840
Speaker 1: baseball questions. Who's the most underrated picture of all time?

688
00:36:08,800 --> 00:36:16,840
Speaker 2: Whoa underrated picture of all time? Oh my goodness, I

689
00:36:16,840 --> 00:36:19,640
Speaker 2: don't know if I can give you one that's like

690
00:36:20,280 --> 00:36:22,600
Speaker 2: all time, because I don't know if that's fair. I'm

691
00:36:22,640 --> 00:36:24,959
Speaker 2: sure I'm not thinking of everybody of.

692
00:36:24,880 --> 00:36:26,440
Speaker 1: Your lifetime, of your lifetime.

693
00:36:26,440 --> 00:36:29,880
Speaker 3: I grew up a hardcore Red Sox fan.

694
00:36:29,960 --> 00:36:33,520
Speaker 2: I grew up in Rhode Island, and the first one

695
00:36:33,520 --> 00:36:37,120
Speaker 2: that comes to mind mostly in the Red Sox atmosphere,

696
00:36:37,160 --> 00:36:39,280
Speaker 2: but I wonder if you could make a broader argument

697
00:36:39,400 --> 00:36:43,759
Speaker 2: was Tim Wakefield, who was a knuckleball pitcher for the

698
00:36:43,800 --> 00:36:48,600
Speaker 2: Red Sox and what made amazing physics amazing physics and aerodynamics,

699
00:36:49,160 --> 00:36:53,400
Speaker 2: And the reason being is he just did so many

700
00:36:53,400 --> 00:36:55,840
Speaker 2: things for the Red Sox, ate up so many innings

701
00:36:56,120 --> 00:37:00,959
Speaker 2: and was so effective closing, starting, whatever, But he never

702
00:37:01,200 --> 00:37:05,120
Speaker 2: got incredible recognition because it was a knuckleball that was going,

703
00:37:05,160 --> 00:37:06,719
Speaker 2: you know, fifty five sixty miles an hour.

704
00:37:07,280 --> 00:37:09,239
Speaker 1: What's one thing you would change about baseball to make

705
00:37:09,280 --> 00:37:10,080
Speaker 1: it more popular?

706
00:37:13,000 --> 00:37:17,240
Speaker 2: I do feel like actually the changes that are being

707
00:37:17,280 --> 00:37:22,239
Speaker 2: made are good ones, in like reducing the amount of downtime.

708
00:37:23,080 --> 00:37:25,719
Speaker 1: A pitch clock, it's one, is that the way you're

709
00:37:25,719 --> 00:37:26,120
Speaker 1: thinking of?

710
00:37:26,239 --> 00:37:29,000
Speaker 2: Or yeah, yeah, yeah, exactly, yeah, a pitch clock. The

711
00:37:29,080 --> 00:37:32,160
Speaker 2: thing that's challenging for me as a biomechanist is when

712
00:37:32,160 --> 00:37:34,680
Speaker 2: you reduced the amount of time that a pitcher has

713
00:37:35,160 --> 00:37:39,840
Speaker 2: to throw, you theoretically could introduce more fatigue, which theoretically

714
00:37:39,880 --> 00:37:42,480
Speaker 2: also introduces more injury risk. So this is something that

715
00:37:42,520 --> 00:37:46,839
Speaker 2: we've been thinking about. Is like, So for me, while

716
00:37:46,840 --> 00:37:50,160
Speaker 2: I like that change that baseball is making to make

717
00:37:50,200 --> 00:37:53,279
Speaker 2: the game, to speed the game up, I think the

718
00:37:53,320 --> 00:37:57,319
Speaker 2: pitchers also need to train a little bit differently. Oh,

719
00:37:57,440 --> 00:38:01,600
Speaker 2: that's to be able to better withstand the shorter rest time.

720
00:38:03,560 --> 00:38:07,799
Speaker 1: Are there aspects of your work and the changes you've

721
00:38:07,840 --> 00:38:11,799
Speaker 1: seen over the course of your career that illuminate sort

722
00:38:11,840 --> 00:38:15,400
Speaker 1: of broader changes in computer vision and AI more generally.

723
00:38:16,560 --> 00:38:19,239
Speaker 2: Yeah, In particular, the most important one has been the

724
00:38:19,280 --> 00:38:23,279
Speaker 2: improvement in computer vision, which because computer vision at its

725
00:38:23,360 --> 00:38:28,000
Speaker 2: core is artificial and intelligence neural networks, and as those

726
00:38:28,719 --> 00:38:31,239
Speaker 2: as that that technology has gotten better and better, you know,

727
00:38:31,280 --> 00:38:33,799
Speaker 2: the latest and greatest people always talk about like the

728
00:38:33,880 --> 00:38:38,000
Speaker 2: Transformer model really changed AI. I mean that changed computer

729
00:38:38,080 --> 00:38:41,160
Speaker 2: vision too, Like lots of the modern more modern like

730
00:38:41,200 --> 00:38:44,640
Speaker 2: computer vision models are based on transformers.

731
00:38:44,400 --> 00:38:46,560
Speaker 1: And which, to be clear to the Transformer model is

732
00:38:47,440 --> 00:38:50,360
Speaker 1: what gives us chat GPT, right, The T in GPT

733
00:38:50,520 --> 00:38:53,520
Speaker 1: is transformed, right, So how has it affected the computer vision.

734
00:38:53,320 --> 00:38:58,040
Speaker 2: Side making the models more accurate and more efficient? It

735
00:38:58,200 --> 00:38:59,600
Speaker 2: used to be you know, a couple of years ago

736
00:38:59,640 --> 00:39:02,239
Speaker 2: when I try to run a computer vision model on

737
00:39:02,280 --> 00:39:05,399
Speaker 2: my laptop, like it could take an hour just to

738
00:39:05,440 --> 00:39:09,839
Speaker 2: like analyze one pitch, one video, and now it takes

739
00:39:09,840 --> 00:39:11,000
Speaker 2: a matter of seconds.

740
00:39:12,960 --> 00:39:15,920
Speaker 1: And so what is the what is the bigger implication

741
00:39:16,040 --> 00:39:18,480
Speaker 1: of that? Beyond your beyond your work?

742
00:39:19,880 --> 00:39:24,200
Speaker 2: More and more data is available regarding how people move,

743
00:39:25,320 --> 00:39:27,319
Speaker 2: you know, like when when I first started, it was

744
00:39:27,360 --> 00:39:31,480
Speaker 2: really hard to have data in a baseball game. Now

745
00:39:31,600 --> 00:39:35,000
Speaker 2: every major league game, every minor league game, every NBA game,

746
00:39:35,800 --> 00:39:38,680
Speaker 2: maybe every G League game, every w NBA game, you know,

747
00:39:39,040 --> 00:39:44,279
Speaker 2: every single uh basketball and baseball game, you know, more

748
00:39:44,360 --> 00:39:48,200
Speaker 2: or less is now like being recorded with computer vision

749
00:39:48,239 --> 00:39:50,600
Speaker 2: to get the three dimensional data about how people are moving.

750
00:39:51,000 --> 00:39:54,080
Speaker 2: Just lots and lots of data on how people move.

751
00:39:54,120 --> 00:39:58,120
Speaker 2: And this is really impacting. I think lots of fields.

752
00:39:58,520 --> 00:40:02,440
Speaker 2: In particular, I think like self driving cars, robots that

753
00:40:02,480 --> 00:40:05,680
Speaker 2: are meant to like interact with the world, all rely

754
00:40:05,800 --> 00:40:08,360
Speaker 2: on computer vision models. I mean, I think one of

755
00:40:08,400 --> 00:40:11,600
Speaker 2: the coolest things about how cars do this sort of

756
00:40:11,640 --> 00:40:15,680
Speaker 2: thing is not only do they have to understand where

757
00:40:15,719 --> 00:40:20,560
Speaker 2: a person is now, but they're really cool models that

758
00:40:21,800 --> 00:40:25,200
Speaker 2: they take where a person is now and the last

759
00:40:25,320 --> 00:40:28,160
Speaker 2: like ten seconds of what that person did and try

760
00:40:28,160 --> 00:40:30,680
Speaker 2: to predict all the different things that the person might do.

761
00:40:30,960 --> 00:40:34,120
Speaker 2: They might run across the street, they might jump out

762
00:40:34,160 --> 00:40:36,120
Speaker 2: of the way, they might jump forward, you know, they

763
00:40:36,200 --> 00:40:37,520
Speaker 2: might run and chase a soccer ball.

764
00:40:38,760 --> 00:40:41,399
Speaker 1: I mean, my sense is all of the hard edge

765
00:40:41,440 --> 00:40:44,439
Speaker 1: cases in self driving cars. Basically, the reason we don't

766
00:40:44,480 --> 00:40:48,200
Speaker 1: truly have self driving cars yet is because people are

767
00:40:48,239 --> 00:40:50,800
Speaker 1: so hard to understate. Yeah, right, Like if the world

768
00:40:50,920 --> 00:40:53,840
Speaker 1: was all self driving cars, then it would be a

769
00:40:53,880 --> 00:40:56,960
Speaker 1: solved problem, right, Like the machine could understand what other

770
00:40:57,000 --> 00:40:59,959
Speaker 1: machines are going to do. But people, human drivers, human

771
00:41:00,040 --> 00:41:05,480
Speaker 1: pedestrians are strange and very hard for machines to understand.

772
00:41:06,000 --> 00:41:07,040
Speaker 3: Yeah.

773
00:41:07,120 --> 00:41:09,560
Speaker 1: Yeah, it's like they need a coach. It's like the

774
00:41:09,600 --> 00:41:18,160
Speaker 1: coaching problem again. We found your SoundCloud. Oh yeah, so,

775
00:41:18,400 --> 00:41:22,640
Speaker 1: Jimmy Buffett fans are called parrot heads. What are Jimmy

776
00:41:22,680 --> 00:41:23,880
Speaker 1: Buffy fans called?

777
00:41:26,480 --> 00:41:28,759
Speaker 3: I never thought about it.

778
00:41:31,600 --> 00:41:35,760
Speaker 1: You have a song called Let's have some Fruit parentheses

779
00:41:36,040 --> 00:41:40,400
Speaker 1: the fruit song is fruit a metaphor.

780
00:41:41,239 --> 00:41:43,680
Speaker 3: Leave it up to your imagination. Fair.

781
00:41:45,560 --> 00:41:49,120
Speaker 1: Jimmy Buffy is the co founder and CEO of Reboot

782
00:41:49,160 --> 00:42:06,040
Speaker 1: Motion Swanson swe Today's show was produced by Gabriel Hunter Cheng.

783
00:42:06,360 --> 00:42:09,719
Speaker 1: It was edited by Lyddy Jean Kott and engineered by

784
00:42:09,760 --> 00:42:13,320
Speaker 1: Sarah Bruguer. You can email us at problem at Pushkin

785
00:42:13,440 --> 00:42:16,640
Speaker 1: dot Fm. I'm Jacob Goldstein and we'll be back next

786
00:42:16,640 --> 00:42:18,960
Speaker 1: week with another episode of What's Your Problems