WEBVTT

1
00:00:01.199 --> 00:00:06.200
<v Speaker 1>Welcome to the Sentient Code, where intelligence is engineered, autonomy

2
00:00:06.280 --> 00:00:10.439
<v Speaker 1>is emerging, and a line between human and machine grows thinner.

3
00:00:10.800 --> 00:00:15.359
<v Speaker 1>Each episode, we decode the algorithms, explore the robotics, and

4
00:00:15.439 --> 00:00:22.640
<v Speaker 1>examine the ideas shaping the future of artificial minds.

5
00:00:23.920 --> 00:00:27.039
<v Speaker 2>Okay, so let's start with the date, just to kind

6
00:00:27.039 --> 00:00:28.199
<v Speaker 2>of anchor ourselves in time.

7
00:00:28.320 --> 00:00:28.879
<v Speaker 3>Good idea.

8
00:00:29.039 --> 00:00:33.479
<v Speaker 2>It is Tuesday, February seventeenth, twenty twenty six. And you know,

9
00:00:33.520 --> 00:00:36.679
<v Speaker 2>for a lot of people today is probably that slow

10
00:00:36.759 --> 00:00:37.679
<v Speaker 2>day back at work.

11
00:00:37.799 --> 00:00:41.000
<v Speaker 3>Oh yeah, nursing the food coma from yesterday's.

12
00:00:40.479 --> 00:00:44.240
<v Speaker 2>Celebrations exactly too many dumplings, too many New Year cakes.

13
00:00:44.679 --> 00:00:46.439
<v Speaker 2>But the date that I think is really going to

14
00:00:46.439 --> 00:00:50.039
<v Speaker 2>be circled in the history books of well of technology. Yeah,

15
00:00:50.119 --> 00:00:51.560
<v Speaker 2>it isn't today, It was yesterday.

16
00:00:51.640 --> 00:00:54.280
<v Speaker 3>It really really feels like the tectonic plates of the

17
00:00:54.399 --> 00:00:57.280
<v Speaker 3>entire tech world shifted while we were all just, you know,

18
00:00:57.359 --> 00:00:59.280
<v Speaker 3>sitting on the couch watching TV last night.

19
00:00:59.280 --> 00:01:02.000
<v Speaker 2>Doesn't It absolutely does. We are, of course, talking about

20
00:01:02.039 --> 00:01:05.200
<v Speaker 2>the day after the lunar New Year celebration. We are

21
00:01:05.239 --> 00:01:09.680
<v Speaker 2>here to unpack the massive, the overwhelming, the just record

22
00:01:09.719 --> 00:01:15.359
<v Speaker 2>breaking twenty twenty six CCTV Spring Festival Gala or tune.

23
00:01:15.200 --> 00:01:17.359
<v Speaker 3>One, which, and we should probably explain this for anyone

24
00:01:17.359 --> 00:01:21.000
<v Speaker 3>who isn't familiar, is the single most watched television event

25
00:01:21.120 --> 00:01:22.359
<v Speaker 3>on the planet period.

26
00:01:22.680 --> 00:01:26.200
<v Speaker 2>It is truly, truly mind boggling. I think for our listeners,

27
00:01:26.280 --> 00:01:29.560
<v Speaker 2>especially maybe in North America or Europe, we always use

28
00:01:29.879 --> 00:01:33.640
<v Speaker 2>the super Bowl as the benchmark for a big TV event.

29
00:01:33.760 --> 00:01:37.159
<v Speaker 3>Sure, the halftime show, the commercials, it's a cultural touchstone,

30
00:01:37.280 --> 00:01:37.799
<v Speaker 3>it is.

31
00:01:37.920 --> 00:01:40.599
<v Speaker 2>But we have to completely recalibrate our scales here. We

32
00:01:40.640 --> 00:01:42.599
<v Speaker 2>need a whole new frame of reference. Oh.

33
00:01:42.599 --> 00:01:45.000
<v Speaker 3>Absolutely, the super Bowl is huge, don't get me wrong.

34
00:01:45.040 --> 00:01:48.280
<v Speaker 3>It's a massive cultural moment in the US. But the

35
00:01:48.319 --> 00:01:51.560
<v Speaker 3>Spring Festival Gala, you are looking at an audience that

36
00:01:51.719 --> 00:01:56.920
<v Speaker 3>regularly consistently hits between five hundred and seven hundred million

37
00:01:57.120 --> 00:01:57.920
<v Speaker 3>live viewers.

38
00:01:57.959 --> 00:02:00.879
<v Speaker 2>That number doesn't even sound real. Five hundred million.

39
00:02:00.920 --> 00:02:03.120
<v Speaker 3>It's half a billion people minimum.

40
00:02:02.640 --> 00:02:05.719
<v Speaker 2>Watching the same program at the exact same time. It's wild.

41
00:02:05.879 --> 00:02:08.360
<v Speaker 3>It's a cultural monolith. It's the background noise of the

42
00:02:08.360 --> 00:02:11.719
<v Speaker 3>new year for nearly what twenty percent of the human population.

43
00:02:11.840 --> 00:02:12.080
<v Speaker 2>Yeah.

44
00:02:12.159 --> 00:02:14.840
<v Speaker 3>Usually it's a mix of pop stars, maybe some comedy sketches.

45
00:02:14.879 --> 00:02:17.680
<v Speaker 3>Traditional dancer is a bit of magic. But this year

46
00:02:18.280 --> 00:02:22.439
<v Speaker 3>the theme wasn't just celebration. It felt like a vision statement.

47
00:02:22.840 --> 00:02:26.879
<v Speaker 2>It was a reveal, a very loud, very very expensive reveal.

48
00:02:27.120 --> 00:02:30.120
<v Speaker 2>Because normally you expect maybe a little cameo from a

49
00:02:30.159 --> 00:02:33.319
<v Speaker 2>piece of tech, a drone swarm in the sky forming

50
00:02:33.400 --> 00:02:33.960
<v Speaker 2>a dragon.

51
00:02:34.159 --> 00:02:36.400
<v Speaker 3>Maybe we've seen that rite a bit of tech spectacle.

52
00:02:36.639 --> 00:02:39.919
<v Speaker 2>But this year the stage was absolutely taken over by

53
00:02:40.039 --> 00:02:43.879
<v Speaker 2>humanoid robots and not just you know, standing in the background.

54
00:02:44.000 --> 00:02:45.080
<v Speaker 2>They were the main character.

55
00:02:45.199 --> 00:02:47.479
<v Speaker 3>It was an unprecedented focus. We aren't talking about a

56
00:02:47.479 --> 00:02:49.719
<v Speaker 3>cute little cameo where a robot waves at the camera

57
00:02:49.759 --> 00:02:52.479
<v Speaker 3>and it never goes ah how neat. This was a

58
00:02:52.560 --> 00:02:57.319
<v Speaker 3>humanoid robotics revolution, broadcast live to the entire world. It

59
00:02:57.400 --> 00:02:59.960
<v Speaker 3>was a statement that the hardware has finally caught up

60
00:03:00.120 --> 00:03:00.759
<v Speaker 3>to the hype.

61
00:03:00.879 --> 00:03:02.840
<v Speaker 2>I mean, I watched it live and I'm still trying

62
00:03:02.840 --> 00:03:06.560
<v Speaker 2>to process the sheer absurdity and the wonder of the visuals.

63
00:03:06.919 --> 00:03:10.120
<v Speaker 2>We had robots performing kung fu, We had robots doing

64
00:03:10.120 --> 00:03:12.560
<v Speaker 2>the drump and fists, which is insane, and we had

65
00:03:12.599 --> 00:03:17.800
<v Speaker 2>them interacting with children on stage live. It was completely.

66
00:03:17.080 --> 00:03:20.680
<v Speaker 3>Surreal and it wasn't chip surrealism either. This wasn't some

67
00:03:20.759 --> 00:03:23.599
<v Speaker 3>creative director just having a wild idea and using like

68
00:03:23.719 --> 00:03:28.240
<v Speaker 3>Cgi or bubbeteers. This was a strategic partnership, a very

69
00:03:28.240 --> 00:03:32.159
<v Speaker 3>public one. It evolved four of the leading startups in

70
00:03:32.159 --> 00:03:36.719
<v Speaker 3>the sector, Unitree, Magic Lab, Galbot and Noatics.

71
00:03:37.120 --> 00:03:39.719
<v Speaker 2>And there was serious money involved here, right, This wasn't

72
00:03:39.759 --> 00:03:40.680
<v Speaker 2>just for the exposure.

73
00:03:40.800 --> 00:03:44.120
<v Speaker 3>Oh no. The reports suggest the partnership deals were worth

74
00:03:44.560 --> 00:03:46.080
<v Speaker 3>around one hundred million YU on.

75
00:03:46.400 --> 00:03:50.080
<v Speaker 2>Which is roughly what fourteen million US dollars About that.

76
00:03:50.199 --> 00:03:53.400
<v Speaker 3>Yeah, but the value goes way way beyond the cash.

77
00:03:53.439 --> 00:03:56.719
<v Speaker 3>This was a signal. It was a statement about embodied AI.

78
00:03:57.039 --> 00:03:59.960
<v Speaker 2>Embodied AI, that's the phrase I keep hearing this morning.

79
00:04:00.120 --> 00:04:02.680
<v Speaker 2>It sounds so sci fi, but I guess here we are.

80
00:04:02.879 --> 00:04:05.520
<v Speaker 3>Well, it's the idea that artificial intelligence isn't just a

81
00:04:05.639 --> 00:04:08.000
<v Speaker 3>chat bot in your browser anymore. But it's not just

82
00:04:08.039 --> 00:04:10.520
<v Speaker 3>a large language model living in a server farm somewhere

83
00:04:10.520 --> 00:04:12.080
<v Speaker 3>in the desert. It has a body. It has a

84
00:04:12.080 --> 00:04:14.919
<v Speaker 3>body you can move, it can act, it can manipulate

85
00:04:14.960 --> 00:04:16.959
<v Speaker 3>the physical world, and as we saw last night, it

86
00:04:17.000 --> 00:04:21.040
<v Speaker 3>can perform martial arts with frankly terrifying precision.

87
00:04:21.199 --> 00:04:23.160
<v Speaker 2>So the mission of our discussion today, I guess is

88
00:04:23.160 --> 00:04:26.639
<v Speaker 2>to really get into that, to unpack how this single

89
00:04:26.680 --> 00:04:32.000
<v Speaker 2>event managed to blend traditional Chinese culture with this incredibly

90
00:04:32.040 --> 00:04:37.279
<v Speaker 2>cutting edge technology to signal loudly that technological self reliance

91
00:04:37.399 --> 00:04:39.120
<v Speaker 2>isn't a future goal. It's here.

92
00:04:39.399 --> 00:04:42.079
<v Speaker 3>It's happening right now on the world's biggest stage.

93
00:04:42.319 --> 00:04:44.040
<v Speaker 2>Okay, so let's get into the weeds. Let's do a

94
00:04:44.079 --> 00:04:47.680
<v Speaker 2>proper analysis of this, because the star of the show,

95
00:04:47.759 --> 00:04:50.759
<v Speaker 2>the moment that I think just absolutely broke the internet,

96
00:04:51.040 --> 00:04:53.319
<v Speaker 2>was the segment they called Wuviot.

97
00:04:53.639 --> 00:04:56.600
<v Speaker 3>It's a clever name, Wou from Wu Shu or martial

98
00:04:56.680 --> 00:04:57.959
<v Speaker 3>arts and bought.

99
00:04:58.360 --> 00:05:01.879
<v Speaker 2>Simple, simple, effective, but the execution was anything but simple.

100
00:05:01.920 --> 00:05:05.079
<v Speaker 2>This was the flagship segment, and it featured Unitary robotics.

101
00:05:05.160 --> 00:05:07.920
<v Speaker 3>Specifically, they're G one humanoid robots. I mean, they had

102
00:05:07.959 --> 00:05:09.800
<v Speaker 3>a few of the larger H one and H two

103
00:05:09.920 --> 00:05:12.839
<v Speaker 3>variants sprinkled in there for some of the heavy lifting moments,

104
00:05:13.040 --> 00:05:15.600
<v Speaker 3>but the g Ones were the main m the main

105
00:05:15.759 --> 00:05:17.240
<v Speaker 3>chorus line, so to speak.

106
00:05:17.319 --> 00:05:19.319
<v Speaker 2>And the G one is their mass production model, right

107
00:05:19.360 --> 00:05:20.879
<v Speaker 2>the one they're trying to really scale up.

108
00:05:20.920 --> 00:05:23.279
<v Speaker 3>That's the one. This was a showcase for the robot

109
00:05:23.319 --> 00:05:24.240
<v Speaker 3>they intend to sell.

110
00:05:24.079 --> 00:05:26.879
<v Speaker 2>By the thousands, and they weren't alone on stage. This

111
00:05:26.920 --> 00:05:28.800
<v Speaker 2>is the part that gave me anxiety just watching it.

112
00:05:29.120 --> 00:05:32.920
<v Speaker 2>They were performing alongside these incredibly talented kids from the

113
00:05:32.920 --> 00:05:35.439
<v Speaker 2>Hennen Tago Martial Arts School, which.

114
00:05:35.279 --> 00:05:38.279
<v Speaker 3>Just adds layer of complexity that makes me sweat just

115
00:05:38.360 --> 00:05:41.279
<v Speaker 3>thinking about it. From an engineering and a safety perspective,

116
00:05:41.800 --> 00:05:48.199
<v Speaker 3>you have high energy human performers children moving frankly, unpredictably

117
00:05:48.560 --> 00:05:51.480
<v Speaker 3>right next to autonomous machines that are swinging weapons around.

118
00:05:51.879 --> 00:05:55.040
<v Speaker 3>That is, that is a recipe for disaster if your

119
00:05:55.160 --> 00:05:57.160
<v Speaker 3>code isn't absolutely perfect.

120
00:05:57.040 --> 00:05:59.519
<v Speaker 2>And the energy I mean. The official description called it

121
00:05:59.560 --> 00:06:03.519
<v Speaker 2>a fully autonomous humanoid robot cluster Kung Fu performance. Mmm.

122
00:06:04.240 --> 00:06:06.319
<v Speaker 2>Just saying that sentence out loud feels like a mouthful

123
00:06:06.319 --> 00:06:07.319
<v Speaker 2>from a science ficion novel.

124
00:06:07.360 --> 00:06:10.560
<v Speaker 3>It is, but every single word in that sentence matters.

125
00:06:10.600 --> 00:06:14.160
<v Speaker 3>Fully autonomous means no one was driving them with a joystick.

126
00:06:14.360 --> 00:06:17.360
<v Speaker 3>There wasn't a team of people backstage with controllers for

127
00:06:17.399 --> 00:06:18.040
<v Speaker 3>each robot.

128
00:06:18.120 --> 00:06:19.759
<v Speaker 2>They were thinking for themselves in a way.

129
00:06:19.879 --> 00:06:23.199
<v Speaker 3>Yes, they were executing their programming based on real time

130
00:06:23.279 --> 00:06:27.000
<v Speaker 3>sensory input. Cluster means they were communicating with each other.

131
00:06:27.160 --> 00:06:31.000
<v Speaker 3>They were aware of their group positioning, maintaining formation like a.

132
00:06:30.920 --> 00:06:33.519
<v Speaker 2>Flock of birds, but with more spinning kicks exactly.

133
00:06:33.560 --> 00:06:37.160
<v Speaker 3>And kung fu. Oh well, that's the incredible physical challenge

134
00:06:37.199 --> 00:06:38.279
<v Speaker 3>that they set for themselves.

135
00:06:38.480 --> 00:06:41.240
<v Speaker 2>Let's really dig into that physical challenge, because the one

136
00:06:41.279 --> 00:06:43.720
<v Speaker 2>thing that made my jaw hit the floor was the

137
00:06:43.800 --> 00:06:46.079
<v Speaker 2>drinke and fist zuquon.

138
00:06:45.879 --> 00:06:49.079
<v Speaker 3>It is, I would argue, one of the most difficult

139
00:06:49.800 --> 00:06:53.279
<v Speaker 3>martial arts styles for a human to master, let alone

140
00:06:53.319 --> 00:06:54.399
<v Speaker 3>a robot, right.

141
00:06:54.319 --> 00:06:56.759
<v Speaker 2>Because the entire point of drunken fists is that you

142
00:06:56.800 --> 00:06:59.240
<v Speaker 2>look like you're about to fall over. You're stumbling, your

143
00:06:59.240 --> 00:07:02.279
<v Speaker 2>off balance, lurching. It's all about deception.

144
00:07:02.120 --> 00:07:06.079
<v Speaker 3>Exactly, and for a robot, balance is usually the number

145
00:07:06.120 --> 00:07:09.079
<v Speaker 3>one goal. It's the prime directive. You want the center

146
00:07:09.120 --> 00:07:12.720
<v Speaker 3>of gravity to be stable, you want predictable footing, you

147
00:07:12.759 --> 00:07:15.519
<v Speaker 3>want your zero moment point, the point where all forces

148
00:07:15.560 --> 00:07:18.600
<v Speaker 3>are balanced, to be right between your feet. Zuok one

149
00:07:18.680 --> 00:07:20.439
<v Speaker 3>throws all of that out the window.

150
00:07:20.560 --> 00:07:22.399
<v Speaker 2>So how do they even program that? Are they just

151
00:07:22.439 --> 00:07:24.040
<v Speaker 2>telling it to almost fall?

152
00:07:24.199 --> 00:07:28.360
<v Speaker 3>It's more complex than that. It requires programming these incredibly

153
00:07:28.399 --> 00:07:32.759
<v Speaker 3>wobbly off balanced stances. It demands exaggerated swaying. It's like

154
00:07:32.800 --> 00:07:35.040
<v Speaker 3>they have to constantly ride the very edge of their

155
00:07:35.040 --> 00:07:38.959
<v Speaker 3>stability envelope without ever actually losing control. It's a continuous

156
00:07:39.040 --> 00:07:40.240
<v Speaker 3>dynamic balancing act.

157
00:07:40.279 --> 00:07:42.399
<v Speaker 2>And they were doing these sudden drops. I saw robots

158
00:07:42.480 --> 00:07:44.720
<v Speaker 2>just hit the deck and then bounce right back up.

159
00:07:44.800 --> 00:07:47.160
<v Speaker 3>That is the technical marvel right there. Yeah, it's not

160
00:07:47.279 --> 00:07:50.240
<v Speaker 3>just the falling. Anyone can push a robot over. I

161
00:07:50.240 --> 00:07:54.319
<v Speaker 3>could do that, right, It's the recovery. In robotics, we

162
00:07:54.360 --> 00:07:58.240
<v Speaker 3>call this fault recovery algorithms. Usually, if a bipedal robot

163
00:07:58.279 --> 00:08:01.160
<v Speaker 3>falls over, the show is over. It's a fail state.

164
00:08:01.560 --> 00:08:03.800
<v Speaker 3>It needs a crane or a team of engineers in

165
00:08:03.839 --> 00:08:06.879
<v Speaker 3>white coats to come out and reset it. It's embarrassing.

166
00:08:07.240 --> 00:08:10.399
<v Speaker 2>But these things were popping back up like Arnold Schwarzenegger

167
00:08:10.480 --> 00:08:11.879
<v Speaker 2>in a Terminator.

168
00:08:11.360 --> 00:08:14.319
<v Speaker 3>Movie without any help. That is the key. They had

169
00:08:14.319 --> 00:08:17.879
<v Speaker 3>to maintain this real time balance control in chaotic conditions,

170
00:08:18.480 --> 00:08:22.000
<v Speaker 3>fall intentionally as part of the choreography, and then execute

171
00:08:22.040 --> 00:08:25.079
<v Speaker 3>an explosive, powerful recovery to get back on their feet.

172
00:08:25.160 --> 00:08:27.800
<v Speaker 2>So what does that show us technically what's going on

173
00:08:27.879 --> 00:08:28.680
<v Speaker 2>under the hood there.

174
00:08:28.959 --> 00:08:32.559
<v Speaker 3>It demonstrates a level of dynamic stability and proprioception that

175
00:08:32.720 --> 00:08:37.720
<v Speaker 3>is honestly shocking. Proprioception is your body's awareness of itself

176
00:08:37.720 --> 00:08:40.919
<v Speaker 3>in space. It implies the robot has an internal model

177
00:08:40.960 --> 00:08:44.000
<v Speaker 3>of its own body that is incredibly advanced and that

178
00:08:44.080 --> 00:08:47.720
<v Speaker 3>it can use to plan these complex, multi stage movements

179
00:08:47.919 --> 00:08:48.960
<v Speaker 3>in milliseconds.

180
00:08:49.039 --> 00:08:51.039
<v Speaker 2>It's almost like they have an inner ear like humans

181
00:08:51.080 --> 00:08:52.240
<v Speaker 2>do for balance.

182
00:08:52.320 --> 00:08:55.720
<v Speaker 3>In a way they do their imus. Their inertial measurement

183
00:08:55.799 --> 00:08:59.559
<v Speaker 3>units are working overtime. These are the sensors, the gyroscopes

184
00:08:59.559 --> 00:09:03.440
<v Speaker 3>and excelrometers that are processing thousands of data points every second.

185
00:09:03.600 --> 00:09:06.240
<v Speaker 3>To say, okay, my torso is tilting backer at thirty

186
00:09:06.240 --> 00:09:09.039
<v Speaker 3>degrees per second, I'm falling, So engage the knee actuators

187
00:09:09.039 --> 00:09:11.799
<v Speaker 3>at x torque, swinging the left arm forward for a momentum,

188
00:09:11.879 --> 00:09:13.759
<v Speaker 3>and push off the right heel to get back up.

189
00:09:14.120 --> 00:09:15.440
<v Speaker 3>It's a symphony of calculations.

190
00:09:15.519 --> 00:09:18.559
<v Speaker 2>And they weren't just dancing empty handed. They had weapons. This

191
00:09:18.600 --> 00:09:19.519
<v Speaker 2>is another level.

192
00:09:19.279 --> 00:09:24.279
<v Speaker 3>Of crazy broadswords. Yeah, the dies staffs, the gun and nunchucks.

193
00:09:24.480 --> 00:09:27.000
<v Speaker 2>Nunchucks. I can barely use nunchucks without hitting myself on

194
00:09:27.039 --> 00:09:29.159
<v Speaker 2>the face. Seeing robots spin them around with the kids

195
00:09:29.200 --> 00:09:30.200
<v Speaker 2>standing three feet away.

196
00:09:30.440 --> 00:09:34.159
<v Speaker 3>It adds this whole variable of inertia and momentum that

197
00:09:34.200 --> 00:09:37.559
<v Speaker 3>the robot has to calculate in real time. It's not

198
00:09:37.639 --> 00:09:40.559
<v Speaker 3>just its own body anymore. If you swing a heavy staff,

199
00:09:40.799 --> 00:09:43.679
<v Speaker 3>it pulls your body forward. Centrifugal force is.

200
00:09:43.840 --> 00:09:47.080
<v Speaker 2>Very real, so the robot has to actively fight that

201
00:09:47.159 --> 00:09:48.080
<v Speaker 2>force instantly.

202
00:09:48.720 --> 00:09:51.440
<v Speaker 3>It has to compensate for that pull in the exact

203
00:09:51.440 --> 00:09:55.799
<v Speaker 3>opposite direction, or just falls over. It's a constant feedback loop, swinging,

204
00:09:56.159 --> 00:10:00.240
<v Speaker 3>feel the pole, adjust stance, compensate, all while stay in

205
00:10:00.320 --> 00:10:03.399
<v Speaker 3>sync with the music and dozens of other performers.

206
00:10:03.519 --> 00:10:05.840
<v Speaker 2>It really highlights the relevance of this whole segment. Then

207
00:10:05.879 --> 00:10:08.200
<v Speaker 2>it's not just look a robot can dance. It's look

208
00:10:08.399 --> 00:10:11.799
<v Speaker 2>a robot that can handle chaotic, unpredictable physical forces.

209
00:10:11.960 --> 00:10:15.600
<v Speaker 3>Precisely, it's a demonstration of robustness because if it can

210
00:10:15.639 --> 00:10:18.320
<v Speaker 3>handle a nunchuck swing while standing on one leg on

211
00:10:18.320 --> 00:10:21.120
<v Speaker 3>a slippery stage, it can probably handle carrying a heavy

212
00:10:21.159 --> 00:10:24.600
<v Speaker 3>box of groceries over an icy driveway. That's the translation

213
00:10:24.679 --> 00:10:25.399
<v Speaker 3>to the real world.

214
00:10:25.519 --> 00:10:27.320
<v Speaker 2>That makes sense. But then we have to talk about

215
00:10:27.320 --> 00:10:29.759
<v Speaker 2>the acrobatics because they didn't just stand the ground.

216
00:10:29.840 --> 00:10:31.159
<v Speaker 3>No, No, they were airborne.

217
00:10:31.240 --> 00:10:35.360
<v Speaker 2>I saw freestyle table vaulting parkour. A robot just running

218
00:10:35.399 --> 00:10:37.840
<v Speaker 2>at a table, planting its hands and vaulting over it.

219
00:10:37.879 --> 00:10:41.720
<v Speaker 3>And the aerial flips they looked well, they looked impossible

220
00:10:41.720 --> 00:10:44.679
<v Speaker 3>for a machine made of metal, plastic and wires.

221
00:10:44.919 --> 00:10:45.960
<v Speaker 2>How hygrid they getting.

222
00:10:46.080 --> 00:10:49.759
<v Speaker 3>Some of them are doing three meter high aerial flips. Now,

223
00:10:50.440 --> 00:10:52.559
<v Speaker 3>just stop and think about the structural stress on a

224
00:10:52.600 --> 00:10:55.240
<v Speaker 3>machine when it lands from three meters in the air.

225
00:10:55.600 --> 00:10:56.879
<v Speaker 3>That's almost ten feet.

226
00:10:57.120 --> 00:11:00.080
<v Speaker 2>That has to exceed standard structural limits, doesn't it. I

227
00:11:00.080 --> 00:11:02.440
<v Speaker 2>mean my phone screen cracks if I drop at three

228
00:11:02.440 --> 00:11:03.519
<v Speaker 2>feet Usually.

229
00:11:03.360 --> 00:11:06.559
<v Speaker 3>Yes, for most robots, a fall from that height you

230
00:11:06.559 --> 00:11:10.279
<v Speaker 3>would risk snapping the joints, shattering the optical sensors, or

231
00:11:10.360 --> 00:11:13.120
<v Speaker 3>stripping the gears and the harmonic drives that control the limbs.

232
00:11:13.519 --> 00:11:17.440
<v Speaker 3>But these units, they took the impact, they absorbed the shock,

233
00:11:17.679 --> 00:11:18.519
<v Speaker 3>and they kept moving.

234
00:11:18.799 --> 00:11:23.879
<v Speaker 2>I saw continuous single legg flips and two step wall assisted.

235
00:11:23.440 --> 00:11:25.879
<v Speaker 3>Backflips, the kind of stuff you see in an action movie.

236
00:11:25.960 --> 00:11:28.440
<v Speaker 2>Yeah. And the speed they weren't moving in that slow,

237
00:11:28.639 --> 00:11:31.279
<v Speaker 2>deliberate I'm thinking about my next step motion like those

238
00:11:31.279 --> 00:11:33.000
<v Speaker 2>old robot videos we used to laugh at.

239
00:11:33.279 --> 00:11:36.720
<v Speaker 3>No, this was fluid and fast. They were clocking running

240
00:11:36.720 --> 00:11:39.559
<v Speaker 3>speeds up to four meters per second, which is about

241
00:11:39.960 --> 00:11:43.600
<v Speaker 3>about fourteen kilometers per hour or around nine miles per hour.

242
00:11:43.960 --> 00:11:45.960
<v Speaker 3>That is a respectable human sprint.

243
00:11:46.120 --> 00:11:49.240
<v Speaker 2>And then there was that break dancing move, the airflare.

244
00:11:49.360 --> 00:11:52.480
<v Speaker 3>Was it the seven point five rotation airflare Grand skin.

245
00:11:53.080 --> 00:11:55.440
<v Speaker 2>I don't even know what that means technically, but visually

246
00:11:55.480 --> 00:11:57.759
<v Speaker 2>it was just a blur of a robot spinning on

247
00:11:57.799 --> 00:11:58.200
<v Speaker 2>its hands.

248
00:11:58.200 --> 00:12:01.279
<v Speaker 3>It's an incredibly difficult power move and break dancing. The

249
00:12:01.279 --> 00:12:04.519
<v Speaker 3>body spins horizontally, almost parallel to the floor while you

250
00:12:04.559 --> 00:12:06.840
<v Speaker 3>balance on your hands, switching from one hand to the

251
00:12:06.919 --> 00:12:09.120
<v Speaker 3>other rapidly teeth the momentum.

252
00:12:09.519 --> 00:12:13.600
<v Speaker 2>So to do seven and a half rotations continuously.

253
00:12:13.000 --> 00:12:16.240
<v Speaker 3>It requires a battery output and a joint torque density

254
00:12:16.240 --> 00:12:19.279
<v Speaker 3>that is absolutely cutting edge. It's a massive stress test

255
00:12:19.320 --> 00:12:22.120
<v Speaker 3>for the battery management system and the cooling systems as

256
00:12:22.200 --> 00:12:24.320
<v Speaker 3>much as it is for the motors. That move has

257
00:12:24.360 --> 00:12:28.320
<v Speaker 3>never been achieved live by a humanoid cluster before ever.

258
00:12:28.639 --> 00:12:30.399
<v Speaker 2>And cluster is the word that brings us to the

259
00:12:30.440 --> 00:12:33.080
<v Speaker 2>next big point here. It wasn't just one super robot

260
00:12:33.120 --> 00:12:35.600
<v Speaker 2>showing off in the spotlight. It was dozens of them.

261
00:12:35.720 --> 00:12:37.919
<v Speaker 3>This is where we get into the realm of swarm intelligence.

262
00:12:38.440 --> 00:12:41.600
<v Speaker 3>And this to me is almost more impressive than a

263
00:12:41.639 --> 00:12:44.120
<v Speaker 3>single robot's acrobatics.

264
00:12:43.559 --> 00:12:46.200
<v Speaker 2>Right, Because if you have one robot doing a backflip,

265
00:12:46.240 --> 00:12:49.559
<v Speaker 2>that's incredible engineering. But if you have twenty robots doing

266
00:12:49.600 --> 00:12:52.320
<v Speaker 2>backflips next to each other without crashing into each other,

267
00:12:52.840 --> 00:12:54.080
<v Speaker 2>that's a whole different problem.

268
00:12:54.120 --> 00:12:57.159
<v Speaker 3>It's a logistics and a communication problem. They were moving

269
00:12:57.159 --> 00:13:00.320
<v Speaker 3>in these perfectly synchronized formations, and we have to remit umber.

270
00:13:00.399 --> 00:13:03.879
<v Speaker 3>They aren't on rails, they aren't following magnets in the floor.

271
00:13:03.960 --> 00:13:07.519
<v Speaker 2>They're using closed loop AI perception. What does that mean?

272
00:13:07.679 --> 00:13:11.080
<v Speaker 3>In simple terms, it means they are constantly seeing the

273
00:13:11.120 --> 00:13:13.720
<v Speaker 3>world around them and reacting to it. The loop is

274
00:13:14.159 --> 00:13:18.919
<v Speaker 3>see think, act, repeat over and over hundreds of times.

275
00:13:18.639 --> 00:13:20.879
<v Speaker 2>A second, so they are literally seeing each other.

276
00:13:21.039 --> 00:13:24.559
<v Speaker 3>Yes, they're equipped with a sensor suite, primarily using triangular

277
00:13:24.639 --> 00:13:27.759
<v Speaker 3>light ar and vision systems cameras to map the stage,

278
00:13:27.879 --> 00:13:30.600
<v Speaker 3>to map the position of their neighbors, and most importantly,

279
00:13:30.720 --> 00:13:32.919
<v Speaker 3>to map the human performers in real time.

280
00:13:33.000 --> 00:13:35.840
<v Speaker 2>Which brings me back to the safety question, because again,

281
00:13:35.960 --> 00:13:38.840
<v Speaker 2>we have robots swinging sticks and nunchucks and they are

282
00:13:39.000 --> 00:13:41.039
<v Speaker 2>at times inches away from children.

283
00:13:41.240 --> 00:13:44.440
<v Speaker 3>It's a terrifying prospect for a safety officer. But it's

284
00:13:44.480 --> 00:13:47.720
<v Speaker 3>a massive flex for the engineers who program them. They

285
00:13:47.759 --> 00:13:51.200
<v Speaker 3>use a technique called reinforcement learning to achieve over ninety

286
00:13:51.200 --> 00:13:52.559
<v Speaker 3>percent motion accuracy.

287
00:13:52.679 --> 00:13:55.240
<v Speaker 2>Reinforcement learning. We hear that term a lot with AI,

288
00:13:55.399 --> 00:13:58.879
<v Speaker 2>like chat rept but how does it apply to physical

289
00:13:58.960 --> 00:14:00.320
<v Speaker 2>legs and arms.

290
00:14:00.559 --> 00:14:02.960
<v Speaker 3>Well, think of it like training a dog, but digitally

291
00:14:03.159 --> 00:14:06.120
<v Speaker 3>and millions of times faster. You give the AI a

292
00:14:06.159 --> 00:14:08.960
<v Speaker 3>goal in a computer simulation. Do this kung fu move,

293
00:14:09.279 --> 00:14:11.080
<v Speaker 3>but don't let the end of your staff get within

294
00:14:11.159 --> 00:14:14.399
<v Speaker 3>twelve inches of this moving child shaped object, and you

295
00:14:14.480 --> 00:14:16.840
<v Speaker 3>reward it when it gets it right and penalize it

296
00:14:16.879 --> 00:14:17.519
<v Speaker 3>when it fails.

297
00:14:17.840 --> 00:14:20.639
<v Speaker 2>So it practice is in a video game, essentially a very.

298
00:14:20.600 --> 00:14:24.360
<v Speaker 3>Very realistic video game. They ran these simulations millions, maybe

299
00:14:24.360 --> 00:14:27.200
<v Speaker 3>billions of times. We call this process sim too real

300
00:14:27.200 --> 00:14:30.080
<v Speaker 3>simulation to reality. They let the AI figure out the

301
00:14:30.120 --> 00:14:33.200
<v Speaker 3>optimal movements on its own through trial and error, so.

302
00:14:33.120 --> 00:14:35.519
<v Speaker 2>By the time they got to the Gallas stage, the

303
00:14:35.639 --> 00:14:39.320
<v Speaker 2>robots had effectively practiced this routine more times than any

304
00:14:39.399 --> 00:14:40.799
<v Speaker 2>human being ever could.

305
00:14:40.679 --> 00:14:44.600
<v Speaker 3>Exactly, and the system is dynamic. It allows for real

306
00:14:44.679 --> 00:14:48.000
<v Speaker 3>time spacing adjustments. If a human performer is a few

307
00:14:48.039 --> 00:14:50.600
<v Speaker 3>inches off their mark, which they will be because they're human.

308
00:14:50.720 --> 00:14:55.000
<v Speaker 3>Of course, the robust perception system detects it and instantly

309
00:14:55.039 --> 00:14:58.120
<v Speaker 3>adjusts its swing or its step to avoid a collision.

310
00:14:58.480 --> 00:15:03.559
<v Speaker 3>Its dynamic safety, pre program safety. It's adapting on the fly.

311
00:15:03.840 --> 00:15:07.000
<v Speaker 2>That is just incredible. And they even had sparring sequences, right,

312
00:15:07.000 --> 00:15:10.159
<v Speaker 2>It wasn't just dancing side by side. They were actually fighting.

313
00:15:10.200 --> 00:15:14.120
<v Speaker 3>They called it the Louis Fists sequences. Yes, robots dueling

314
00:15:14.159 --> 00:15:16.879
<v Speaker 3>with the young performers. It was choreygraphed, of course, but

315
00:15:16.919 --> 00:15:20.120
<v Speaker 3>it required that same real time perception to work.

316
00:15:20.360 --> 00:15:24.600
<v Speaker 2>It was a showcase of harmony, I guess, between man

317
00:15:24.639 --> 00:15:27.279
<v Speaker 2>and machine in a very confined, very active space.

318
00:15:27.519 --> 00:15:29.840
<v Speaker 3>That's the thematic takeaway they were going for. It sends

319
00:15:29.840 --> 00:15:32.600
<v Speaker 3>a pretty strong message. We are safe, we are precise,

320
00:15:32.679 --> 00:15:34.120
<v Speaker 3>we can work with you, not against you.

321
00:15:34.399 --> 00:15:37.559
<v Speaker 2>It's an attempt to erase that classic sci fi fear

322
00:15:37.679 --> 00:15:41.080
<v Speaker 2>that robots are dangerous, clumsy machines that are going to

323
00:15:41.159 --> 00:15:42.039
<v Speaker 2>trip and crush you.

324
00:15:42.279 --> 00:15:45.879
<v Speaker 3>Exactly. It's a very powerful image. But speaking of powerful images,

325
00:15:46.519 --> 00:15:48.440
<v Speaker 3>we absolutely have to talk about the Monkey King.

326
00:15:48.559 --> 00:15:51.200
<v Speaker 2>Oh man, the monkey King.

327
00:15:51.320 --> 00:15:53.759
<v Speaker 3>This is where the cultural fusion peaked. It was just

328
00:15:53.799 --> 00:15:56.679
<v Speaker 3>a masterstroke of stagecraft and symbolism.

329
00:15:56.960 --> 00:15:59.480
<v Speaker 2>For those who might not know, the Monkey King's Sun

330
00:15:59.559 --> 00:16:03.600
<v Speaker 2>Wukong is like the ultimate superhero of Chinese mythology from

331
00:16:03.639 --> 00:16:05.480
<v Speaker 2>the classic novel Journey to the West.

332
00:16:05.559 --> 00:16:10.960
<v Speaker 3>He represents agility, cleverness, rebellion, and incredible power, and typically

333
00:16:10.960 --> 00:16:12.519
<v Speaker 3>in the year of the Horse. You might not expect

334
00:16:12.519 --> 00:16:15.159
<v Speaker 3>the Monkey King to take center stage, but he is

335
00:16:15.200 --> 00:16:17.440
<v Speaker 3>such an icon of movement and transformation.

336
00:16:17.639 --> 00:16:20.600
<v Speaker 2>And they dressed up a Unitree H two model. This

337
00:16:20.639 --> 00:16:22.720
<v Speaker 2>is the big one, the one point eight meter tall,

338
00:16:22.799 --> 00:16:26.360
<v Speaker 2>heavily armored looking robot in full Monkey King armor.

339
00:16:26.440 --> 00:16:29.840
<v Speaker 3>It was visually so striking, the ornate armor, the powerful stance.

340
00:16:30.240 --> 00:16:32.399
<v Speaker 3>But the genius part, the part that made everyone gassed,

341
00:16:32.720 --> 00:16:33.759
<v Speaker 3>was the cloud.

342
00:16:33.440 --> 00:16:36.240
<v Speaker 2>The Somersault cloud. In the myths, the Monkey King can

343
00:16:36.240 --> 00:16:38.799
<v Speaker 2>fly by riding on a magical cloud. So how did

344
00:16:38.799 --> 00:16:40.080
<v Speaker 2>they pull that off with a robot.

345
00:16:40.320 --> 00:16:43.960
<v Speaker 3>They used unit Trees B to W quadruped robot dogs,

346
00:16:44.399 --> 00:16:45.320
<v Speaker 3>the four legged.

347
00:16:45.080 --> 00:16:49.120
<v Speaker 2>Ones the robot dogs. Of course, Boston Dynamics made them famous,

348
00:16:49.120 --> 00:16:50.399
<v Speaker 2>but Unitrey makes them.

349
00:16:50.240 --> 00:16:53.120
<v Speaker 3>Too, and they are very very good at it. They

350
00:16:53.159 --> 00:16:55.480
<v Speaker 3>basically covered a small pack of these B to W

351
00:16:55.639 --> 00:16:58.840
<v Speaker 3>dogs in cloud like props and the humanoid Monkey King

352
00:16:58.919 --> 00:17:01.440
<v Speaker 3>robot stood on top of them as they trotted around

353
00:17:01.440 --> 00:17:04.920
<v Speaker 3>the stage. It was a robot writing other robots.

354
00:17:04.599 --> 00:17:07.000
<v Speaker 2>Delivering New Year blessings from the sky.

355
00:17:07.359 --> 00:17:09.880
<v Speaker 3>It was incredible and it appeared at both the main

356
00:17:09.960 --> 00:17:12.799
<v Speaker 3>Beijing venue and the u Wu venue. It frames the

357
00:17:12.920 --> 00:17:16.039
<v Speaker 3>robot not as some foreign alien invader, but as a

358
00:17:16.359 --> 00:17:17.319
<v Speaker 3>cultural preserver.

359
00:17:17.720 --> 00:17:20.480
<v Speaker 2>That's a fascinating phrase, cultural preserver. I hadn't thought of

360
00:17:20.519 --> 00:17:21.000
<v Speaker 2>it like that.

361
00:17:21.079 --> 00:17:24.119
<v Speaker 3>It's one hundred percent intentional. By having the robot learn

362
00:17:24.200 --> 00:17:27.200
<v Speaker 3>kung fu, learn the traditions of the Shallon Temple, embody

363
00:17:27.240 --> 00:17:30.079
<v Speaker 3>the Monkey King, the message is that this new technology

364
00:17:30.079 --> 00:17:32.640
<v Speaker 3>isn't here to replace our culture, It's here to learn

365
00:17:32.640 --> 00:17:34.559
<v Speaker 3>from it, to carry it forward in a new form.

366
00:17:35.039 --> 00:17:38.319
<v Speaker 3>It frames the AI as a student of human history,

367
00:17:38.599 --> 00:17:39.599
<v Speaker 3>not its replacement.

368
00:17:39.880 --> 00:17:42.279
<v Speaker 2>I love that it's such a smart way to position

369
00:17:42.359 --> 00:17:45.519
<v Speaker 2>it now. Unitree was clearly the star, but they weren't

370
00:17:45.559 --> 00:17:47.640
<v Speaker 2>the only ones at the party. We mentioned a few

371
00:17:47.640 --> 00:17:51.440
<v Speaker 2>other companies at the top. Let's talk about Noahtics Robotics, right.

372
00:17:51.559 --> 00:17:55.279
<v Speaker 3>Noadics took a very very different approach. While Unitary was

373
00:17:55.319 --> 00:17:58.640
<v Speaker 3>out there doing backflips and swinging swords, no It's went

374
00:17:58.680 --> 00:18:00.759
<v Speaker 3>for the heart strings. They went for the social angle.

375
00:18:00.799 --> 00:18:04.279
<v Speaker 2>They were in that comedy sketch, the one called Grandma's Favorite.

376
00:18:03.960 --> 00:18:08.160
<v Speaker 3>Which is a classic Spring Festival Gala trope, the family

377
00:18:08.200 --> 00:18:10.680
<v Speaker 3>comedy sketch. You know, there's always a misunderstanding a big

378
00:18:10.720 --> 00:18:14.160
<v Speaker 3>family dinner. It's very relatable. And Noahdicgs brought out their

379
00:18:14.200 --> 00:18:17.720
<v Speaker 3>Boomy and two and e one models, along with some

380
00:18:18.039 --> 00:18:19.680
<v Speaker 3>custom biotic models.

381
00:18:19.359 --> 00:18:21.920
<v Speaker 2>And these weren't fighting robots, they were acting. They were

382
00:18:21.960 --> 00:18:23.119
<v Speaker 2>part of the family exactly.

383
00:18:23.160 --> 00:18:27.359
<v Speaker 3>The focus here was entirely on social presence, on gesture recognition,

384
00:18:27.920 --> 00:18:29.759
<v Speaker 3>on humorous timing, which.

385
00:18:29.519 --> 00:18:32.440
<v Speaker 2>Is really really hard. Comedy is all about timing. If

386
00:18:32.440 --> 00:18:35.160
<v Speaker 2>the robot pauses for a second too long or delivers

387
00:18:35.200 --> 00:18:38.039
<v Speaker 2>a line too quickly, the entire joke just dies.

388
00:18:38.720 --> 00:18:41.480
<v Speaker 3>And they pulled it off. It shows a completely different

389
00:18:41.559 --> 00:18:45.559
<v Speaker 3>kind of intelligence. It's not just physical agility, it's social agility.

390
00:18:46.039 --> 00:18:50.119
<v Speaker 3>They were interacting naturally with human actors in these everyday

391
00:18:50.119 --> 00:18:54.039
<v Speaker 3>home scenarios. It paints a picture of a future where

392
00:18:54.200 --> 00:18:57.759
<v Speaker 3>robots are helpers in the home, companions for the elderly,

393
00:18:58.000 --> 00:18:59.920
<v Speaker 3>not just soldiers or factory workers.

394
00:19:00.039 --> 00:19:01.559
<v Speaker 2>A much more gentle vision of.

395
00:19:01.519 --> 00:19:03.359
<v Speaker 3>The future, much more approachable one.

396
00:19:03.519 --> 00:19:04.759
<v Speaker 2>And then we had Magic Lab.

397
00:19:04.880 --> 00:19:07.799
<v Speaker 3>Magic Lab brought the groove. They were all about dance

398
00:19:07.799 --> 00:19:10.559
<v Speaker 3>and precision. They had their Magic bought Gen one and

399
00:19:10.680 --> 00:19:11.680
<v Speaker 3>Z one models on something.

400
00:19:11.680 --> 00:19:14.480
<v Speaker 2>So you're in that big musical number right the song

401
00:19:14.519 --> 00:19:15.519
<v Speaker 2>We Are Made in China.

402
00:19:15.960 --> 00:19:18.200
<v Speaker 3>Very subtle song title, very subtle.

403
00:19:18.200 --> 00:19:19.599
<v Speaker 2>Couldn't miss the message there, but.

404
00:19:19.599 --> 00:19:23.559
<v Speaker 3>It was effective. They started with these basic, almost robotic poses,

405
00:19:24.079 --> 00:19:26.400
<v Speaker 3>but then as the music swelled, they joined the human

406
00:19:26.480 --> 00:19:29.640
<v Speaker 3>dancers in a fully synchronized routine. And there was one

407
00:19:29.640 --> 00:19:33.480
<v Speaker 3>specific moment the Z one robot doing a three hundred

408
00:19:33.480 --> 00:19:35.079
<v Speaker 3>and sixty degree Thomas rotation.

409
00:19:35.200 --> 00:19:36.880
<v Speaker 2>That's a gymnastics move, it is.

410
00:19:37.079 --> 00:19:39.000
<v Speaker 3>It's that move you see on the pommel horse or

411
00:19:39.079 --> 00:19:41.799
<v Speaker 3>on the floor in gymnastics where they spin their whole

412
00:19:41.839 --> 00:19:45.640
<v Speaker 3>body around with their legs flared out, balanced on their hands. Again,

413
00:19:45.680 --> 00:19:49.519
<v Speaker 3>it requires massive core strength and balance. It's showing that

414
00:19:49.559 --> 00:19:52.519
<v Speaker 3>these robots have the range of motion and power of

415
00:19:52.559 --> 00:19:53.880
<v Speaker 3>an elite athlete.

416
00:19:54.079 --> 00:19:56.519
<v Speaker 2>And Galbit was there too. What was their contribution?

417
00:19:57.359 --> 00:19:59.880
<v Speaker 3>Galbut was kind of the reliable backbone of the operation.

418
00:20:00.200 --> 00:20:02.680
<v Speaker 3>They contributed a number of general purpose models and some

419
00:20:02.799 --> 00:20:06.480
<v Speaker 3>G one variants that helped facilitate the group dynamics across

420
00:20:06.599 --> 00:20:10.759
<v Speaker 3>various sketches and performances. Their big technical highlight is really

421
00:20:10.799 --> 00:20:13.079
<v Speaker 3>their focus on in house hardware.

422
00:20:12.759 --> 00:20:14.359
<v Speaker 2>Meaning they build their own parts.

423
00:20:14.440 --> 00:20:19.160
<v Speaker 3>Exactly, specifically their joint modules and their dexterous hands. They're

424
00:20:19.279 --> 00:20:21.119
<v Speaker 3>very proud of the fact that they don't have to

425
00:20:21.160 --> 00:20:23.200
<v Speaker 3>import those critical components.

426
00:20:22.839 --> 00:20:25.880
<v Speaker 2>And dexter's hands are the holy grail of robotics, aren't they.

427
00:20:25.920 --> 00:20:29.359
<v Speaker 3>They really are. It's relatively easy to make a robot

428
00:20:29.400 --> 00:20:31.359
<v Speaker 3>walk now, making you pick up an egg or a

429
00:20:31.400 --> 00:20:35.759
<v Speaker 3>strawberry without queshing it, that's incredibly hard. So Galbot was

430
00:20:35.759 --> 00:20:38.559
<v Speaker 3>there to show off that fine motor control, that dexterity

431
00:20:38.799 --> 00:20:40.160
<v Speaker 3>and tying all this together.

432
00:20:40.200 --> 00:20:43.119
<v Speaker 2>There was a software element too, we saw ByteDance was involved,

433
00:20:43.519 --> 00:20:44.839
<v Speaker 2>the company behind.

434
00:20:44.559 --> 00:20:48.400
<v Speaker 3>TikTok Yes, their Dubao AI chatbot. They integrated it for

435
00:20:48.480 --> 00:20:49.839
<v Speaker 3>the dialogue segments of the show.

436
00:20:49.960 --> 00:20:52.240
<v Speaker 2>So when the robots were speaking in the sketches, or

437
00:20:52.240 --> 00:20:54.440
<v Speaker 2>when the human characters were interacting with an AI on

438
00:20:54.480 --> 00:20:55.640
<v Speaker 2>a screen, that.

439
00:20:55.680 --> 00:20:58.400
<v Speaker 3>Was Dubao correct. And that's a critical piece of the puzzle.

440
00:20:58.519 --> 00:21:01.359
<v Speaker 3>It links the physical body the robot to the brain

441
00:21:01.839 --> 00:21:04.880
<v Speaker 3>of the genitive AI ecosystem that's exploding right now. It

442
00:21:04.960 --> 00:21:08.240
<v Speaker 3>shows the complete package the body of a unitary robot

443
00:21:08.240 --> 00:21:10.480
<v Speaker 3>with the mind of a large language model like Dubao.

444
00:21:10.720 --> 00:21:13.960
<v Speaker 2>You know. Seeing all this, this incredible display of capability,

445
00:21:13.960 --> 00:21:16.400
<v Speaker 2>I have to ask, haven't we seen this before? I

446
00:21:16.400 --> 00:21:18.880
<v Speaker 2>have this vague memory that Unitree did something last year?

447
00:21:18.920 --> 00:21:19.440
<v Speaker 2>Am I wrong?

448
00:21:19.519 --> 00:21:22.119
<v Speaker 3>No, You're not wrong. They did, But the comparison between

449
00:21:22.200 --> 00:21:25.000
<v Speaker 3>last year and this year is, well, it's night and day.

450
00:21:25.400 --> 00:21:29.519
<v Speaker 3>It is the single most stunning example of accelerated progress

451
00:21:29.559 --> 00:21:30.240
<v Speaker 3>I have ever seen.

452
00:21:30.279 --> 00:21:32.240
<v Speaker 2>Okay, so let's go back to twenty twenty five. Paint

453
00:21:32.240 --> 00:21:34.759
<v Speaker 2>me a picture. What did that performance look like?

454
00:21:35.119 --> 00:21:38.079
<v Speaker 3>So in twenty twenty five, Unitree had sixteen of their

455
00:21:38.160 --> 00:21:41.400
<v Speaker 3>humanoids on the Gallas stage doing a Yang folk dance.

456
00:21:41.759 --> 00:21:43.720
<v Speaker 2>Yo, that's the one with the colorful costumes on the.

457
00:21:43.680 --> 00:21:47.839
<v Speaker 3>Handkerchief exactly, twirling handkerchiefs and honestly and honestly for the

458
00:21:47.880 --> 00:21:50.319
<v Speaker 3>time for twenty twenty five. It was impressive. It was

459
00:21:50.319 --> 00:21:52.880
<v Speaker 3>the first time a humanoid robot cluster had done anything

460
00:21:53.000 --> 00:21:55.880
<v Speaker 3>like that on such a big stage. But it was

461
00:21:56.039 --> 00:21:59.039
<v Speaker 3>visibly scripted. Oh so, the robots were a bit wobbly,

462
00:21:59.279 --> 00:22:02.720
<v Speaker 3>their movements stiff, a little jerky. It looked like machines.

463
00:22:02.759 --> 00:22:07.039
<v Speaker 3>Following a very strict, pre programmed set of instructions. There

464
00:22:07.119 --> 00:22:09.799
<v Speaker 3>was no adaptability. If one of them had tripped or

465
00:22:09.839 --> 00:22:12.440
<v Speaker 3>been bumped, it would have been a disaster for little formation.

466
00:22:12.920 --> 00:22:15.359
<v Speaker 2>So impressive, but fragile.

467
00:22:15.079 --> 00:22:16.799
<v Speaker 3>Very fragile. And this year, this.

468
00:22:16.839 --> 00:22:20.079
<v Speaker 2>Year, we saw real time reactivity. We saw adaptive behaviors,

469
00:22:20.079 --> 00:22:23.559
<v Speaker 2>we saw fault recovery from falls, we saw extreme acrobatics.

470
00:22:23.880 --> 00:22:27.799
<v Speaker 2>We went from wobbly handkerchief twirl to seven point five

471
00:22:27.960 --> 00:22:30.799
<v Speaker 2>rotation backspin while avoiding children in twelve months.

472
00:22:30.880 --> 00:22:32.039
<v Speaker 3>It's insane leap.

473
00:22:32.359 --> 00:22:36.720
<v Speaker 2>So how how is that even possible? How does an

474
00:22:36.880 --> 00:22:40.359
<v Speaker 2>entire industry move that fast in just one year? More's

475
00:22:40.400 --> 00:22:43.279
<v Speaker 2>law doesn't usually apply to hardware and physical machines like this.

476
00:22:43.640 --> 00:22:46.279
<v Speaker 3>It's the result of a perfect storm, a confluence of

477
00:22:46.319 --> 00:22:50.680
<v Speaker 3>factors all hitting at once. First, you have massive and

478
00:22:50.759 --> 00:22:54.799
<v Speaker 3>I mean massive government subsidies pouring into the robotic sector

479
00:22:55.039 --> 00:22:56.319
<v Speaker 3>as a national priority.

480
00:22:56.400 --> 00:22:57.640
<v Speaker 2>So the money is there.

481
00:22:57.440 --> 00:23:00.240
<v Speaker 3>The money is there in huge amounts. Second, you have

482
00:23:00.400 --> 00:23:03.839
<v Speaker 3>aggressive talent poaching. You have the best and brightest engineers

483
00:23:03.880 --> 00:23:07.519
<v Speaker 3>moving from big tech firms and advanced automotive companies into

484
00:23:07.559 --> 00:23:10.200
<v Speaker 3>these robotic startups, bringing their expertise with them.

485
00:23:10.279 --> 00:23:12.160
<v Speaker 2>And what about the supply chain that's always been a

486
00:23:12.160 --> 00:23:13.200
<v Speaker 2>bottleneck for hardware.

487
00:23:13.319 --> 00:23:15.759
<v Speaker 3>That's the hidden hero in all of this. The domestic

488
00:23:15.799 --> 00:23:19.799
<v Speaker 3>supply chain for high performance actuators, for precision sensors, for

489
00:23:20.240 --> 00:23:24.799
<v Speaker 3>energy dense batteries. It has matured incredibly fast in China,

490
00:23:25.319 --> 00:23:28.799
<v Speaker 3>is becoming like the smartphone supply chain a decade ago, efficient,

491
00:23:29.079 --> 00:23:31.960
<v Speaker 3>high quality, and crucially cheap.

492
00:23:31.799 --> 00:23:34.720
<v Speaker 2>So they can build and test new prototypes much faster.

493
00:23:34.880 --> 00:23:37.839
<v Speaker 3>They're iterating on hardware as fast as other companies iterate

494
00:23:37.880 --> 00:23:40.519
<v Speaker 3>on software. And then you add the final ingredient, this

495
00:23:40.680 --> 00:23:45.359
<v Speaker 3>ferocious private sector competition. These four companies we've talked about Unitry, Magic,

496
00:23:45.440 --> 00:23:48.079
<v Speaker 3>lab galbut Nodics, they are at a flat out race

497
00:23:48.119 --> 00:23:49.039
<v Speaker 3>against each other.

498
00:23:49.200 --> 00:23:52.200
<v Speaker 2>And that competition is driving innovation.

499
00:23:52.440 --> 00:23:54.559
<v Speaker 3>It's a frenzy and the result is what we saw

500
00:23:54.680 --> 00:23:59.039
<v Speaker 3>last night, a quantum leap in capability. This is not

501
00:23:59.160 --> 00:24:02.079
<v Speaker 3>an incremental update like a new camera on an iPhone.

502
00:24:02.200 --> 00:24:04.079
<v Speaker 3>This is like going from a black and white flip

503
00:24:04.079 --> 00:24:06.799
<v Speaker 3>phone to a modern smartphone in a single year.

504
00:24:06.960 --> 00:24:11.440
<v Speaker 2>It's staggering, and predictably the world noticed. I mean, we

505
00:24:11.480 --> 00:24:13.960
<v Speaker 2>talked about the viewership numbers, but the social media reaction

506
00:24:14.240 --> 00:24:16.079
<v Speaker 2>was just instant and global.

507
00:24:16.160 --> 00:24:18.559
<v Speaker 3>It completely exploded on Weebo and dw in. Of course

508
00:24:18.799 --> 00:24:21.480
<v Speaker 3>that was a given, but it broke containment and went

509
00:24:21.519 --> 00:24:26.240
<v Speaker 3>everywhere internationally. YouTube, x, Instagram read it. The clips were

510
00:24:26.279 --> 00:24:28.599
<v Speaker 3>being shared and re shared everywhere within hours.

511
00:24:29.000 --> 00:24:31.240
<v Speaker 2>What were the headlines saying? What was the general sentiment?

512
00:24:31.400 --> 00:24:35.839
<v Speaker 3>Dazzling, insane, historic, a glimpse into twenty thirty. People were

513
00:24:35.880 --> 00:24:37.799
<v Speaker 3>just floored. I think for a lot of people who

514
00:24:37.839 --> 00:24:41.160
<v Speaker 3>have seen the slower, more cautious progress in the West,

515
00:24:41.279 --> 00:24:43.240
<v Speaker 3>this looked like it came out of nowhere. It broke

516
00:24:43.279 --> 00:24:45.920
<v Speaker 3>through a lot of the skepticism about humanoid robots.

517
00:24:46.079 --> 00:24:48.599
<v Speaker 2>And I heard a certain tech billionaire, someone who has

518
00:24:48.640 --> 00:24:50.559
<v Speaker 2>his own robot project, had something to say.

519
00:24:50.839 --> 00:24:55.519
<v Speaker 3>Yes, Elon Musk. He chimed in on x reiterating something

520
00:24:55.519 --> 00:24:59.880
<v Speaker 3>he said before that China, and he specifically name dropped

521
00:24:59.920 --> 00:25:04.039
<v Speaker 3>U is the top competitor to Tesla's Optimus robot.

522
00:25:04.400 --> 00:25:07.279
<v Speaker 2>That is high praise from a direct competitor and also

523
00:25:07.440 --> 00:25:09.480
<v Speaker 2>probably a bit of a warning shot to his own team.

524
00:25:09.519 --> 00:25:13.279
<v Speaker 3>I would imagine, well, absolutely, it validates the entire industry.

525
00:25:13.759 --> 00:25:17.160
<v Speaker 3>When the biggest player in the game publicly acknowledges the competition,

526
00:25:17.640 --> 00:25:19.759
<v Speaker 3>you know, it's real. It signals that this isn't a

527
00:25:19.799 --> 00:25:22.359
<v Speaker 3>regional phenomenon anymore. This is a global race.

528
00:25:22.799 --> 00:25:25.599
<v Speaker 2>But here is where it gets really really interesting for me.

529
00:25:25.720 --> 00:25:28.519
<v Speaker 2>This is the part that signals a real shift. Usually,

530
00:25:28.599 --> 00:25:31.480
<v Speaker 2>when you see high tech concept cars or futuristic robots,

531
00:25:31.519 --> 00:25:33.200
<v Speaker 2>you think, Okay, that's cool, maybe i'll see one in

532
00:25:33.240 --> 00:25:35.680
<v Speaker 2>ten years. But this wasn't just a show. It was

533
00:25:36.079 --> 00:25:37.319
<v Speaker 2>in effect a commercial.

534
00:25:37.400 --> 00:25:39.759
<v Speaker 3>It was the ultimate shop now TV moment.

535
00:25:40.039 --> 00:25:43.720
<v Speaker 2>Jd dot Com, the massive Chinese e commerce site, actually

536
00:25:43.759 --> 00:25:46.200
<v Speaker 2>listed the models that were featured during the gala.

537
00:25:46.240 --> 00:25:49.119
<v Speaker 3>They did. As the robots were dancing and fighting on screen,

538
00:25:49.480 --> 00:25:51.440
<v Speaker 3>a button popped up in the corner of the app

539
00:25:51.759 --> 00:25:53.240
<v Speaker 3>by now, that.

540
00:25:53.279 --> 00:25:55.640
<v Speaker 2>Is just brilliant marketing. And guess what happened.

541
00:25:55.799 --> 00:25:58.640
<v Speaker 3>Let me guess. They didn't stay on the virtual shelf

542
00:25:58.680 --> 00:25:59.200
<v Speaker 3>for very long.

543
00:25:59.279 --> 00:26:02.799
<v Speaker 2>They sold out within minutes, minutes, minutes. And these are

544
00:26:02.799 --> 00:26:04.480
<v Speaker 2>not cheap toys, are they. We're not talking about a

545
00:26:04.519 --> 00:26:05.359
<v Speaker 2>couple hundred.

546
00:26:05.119 --> 00:26:08.119
<v Speaker 3>Dollars, no, not at all. The high end Galbit units,

547
00:26:08.119 --> 00:26:10.720
<v Speaker 3>for example, we're selling for around six hundred and thirty

548
00:26:10.799 --> 00:26:11.480
<v Speaker 3>thousand u.

549
00:26:11.400 --> 00:26:14.720
<v Speaker 2>On, which is nearly ninety thousand US.

550
00:26:14.519 --> 00:26:17.599
<v Speaker 3>Dollars, almost ninety thousand dollars and people were buying them

551
00:26:18.079 --> 00:26:20.240
<v Speaker 3>like they were the hot new concert tickets.

552
00:26:20.359 --> 00:26:23.319
<v Speaker 2>Who is buying a ninety thousand dollars robot.

553
00:26:23.039 --> 00:26:28.279
<v Speaker 3>Early adopters, research institutions, wealthy tech enthusiasts. But it shows

554
00:26:28.319 --> 00:26:30.519
<v Speaker 3>that there is a market hunger for this. It's not

555
00:26:30.720 --> 00:26:36.559
<v Speaker 3>just idle curiosity anymore. It's actual commercial demand. Unitary has

556
00:26:36.599 --> 00:26:40.039
<v Speaker 3>publicly stated they aim to ship twenty thousand units in

557
00:26:40.119 --> 00:26:41.640
<v Speaker 3>twenty twenty six alone.

558
00:26:41.880 --> 00:26:44.960
<v Speaker 2>Twenty thousand humanoid robots. That is a lot of metal

559
00:26:44.960 --> 00:26:45.599
<v Speaker 2>walking around.

560
00:26:45.640 --> 00:26:47.920
<v Speaker 3>It's a massive scaling up. It's the moment you move

561
00:26:47.960 --> 00:26:52.039
<v Speaker 3>from lab prototype to consumer electronic. It's a fundamental shift

562
00:26:52.039 --> 00:26:52.680
<v Speaker 3>in the industry.

563
00:26:52.799 --> 00:26:54.599
<v Speaker 2>So if we zoom all the way out, we have

564
00:26:54.640 --> 00:26:56.920
<v Speaker 2>to look at the strategic implications here. Why do this?

565
00:26:57.039 --> 00:26:59.640
<v Speaker 2>Why put this incredible display on the Spring Festival galat,

566
00:26:59.640 --> 00:27:01.680
<v Speaker 2>the most watch show on Earth? Why make it such

567
00:27:01.680 --> 00:27:02.160
<v Speaker 2>a big deal.

568
00:27:02.400 --> 00:27:04.880
<v Speaker 3>It's never just about entertainment on a stage like that.

569
00:27:05.000 --> 00:27:07.440
<v Speaker 2>Is it never? There's always a message.

570
00:27:07.799 --> 00:27:12.400
<v Speaker 3>This serves as state orchestrated messaging. It's propaganda in the

571
00:27:12.440 --> 00:27:15.480
<v Speaker 3>most literal sense of the word, propagating an idea, and

572
00:27:15.519 --> 00:27:19.640
<v Speaker 3>it's a message to two audiences simultaneously, to the domestic audience,

573
00:27:19.880 --> 00:27:23.039
<v Speaker 3>it's a source of immense national pride. To the global audience,

574
00:27:23.079 --> 00:27:27.359
<v Speaker 3>it's a declaration China aims to dominate the global humanoid market.

575
00:27:27.519 --> 00:27:31.759
<v Speaker 2>I've seen the projections for that market. They're astronomical trillions

576
00:27:31.759 --> 00:27:33.359
<v Speaker 2>in value by twenty fifty.

577
00:27:33.400 --> 00:27:36.559
<v Speaker 3>Hundreds of millions of units deployed. The goal is to

578
00:27:36.559 --> 00:27:40.599
<v Speaker 3>make the phrase made in China synonymous with advanced robotics.

579
00:27:40.799 --> 00:27:44.160
<v Speaker 3>It serves multiple purposes at once. It addresses future labor

580
00:27:44.160 --> 00:27:47.680
<v Speaker 3>shortages from an aging population, It boosts natural pride, and

581
00:27:47.759 --> 00:27:51.440
<v Speaker 3>perhaps most importantly, it fundamentally shifts the public perception of

582
00:27:51.480 --> 00:27:52.279
<v Speaker 3>this technology.

583
00:27:52.519 --> 00:27:55.039
<v Speaker 2>That's the key, right there, isn't it shifting perception?

584
00:27:55.480 --> 00:27:59.000
<v Speaker 3>It moves robots from the realm of scary sci fi

585
00:27:59.160 --> 00:28:04.319
<v Speaker 3>killer to inevitable, helpful appliance. When you see a robot

586
00:28:04.359 --> 00:28:07.640
<v Speaker 3>interacting gently with a grandmother in a comedy sketch, or

587
00:28:07.720 --> 00:28:11.880
<v Speaker 3>performing ancient martial arts alongside a child, you stop feeling

588
00:28:11.920 --> 00:28:15.519
<v Speaker 3>it and you start accepting it. It normalizes the technology

589
00:28:15.839 --> 00:28:16.920
<v Speaker 3>on a massive scale.

590
00:28:17.079 --> 00:28:19.799
<v Speaker 2>There are always the concerns though, the dual use question

591
00:28:19.880 --> 00:28:20.720
<v Speaker 2>always comes up.

592
00:28:20.680 --> 00:28:24.319
<v Speaker 3>Of course, and analysts and defense circles are absolutely looking

593
00:28:24.319 --> 00:28:26.920
<v Speaker 3>at this and thinking about the line between an industrial

594
00:28:26.960 --> 00:28:30.400
<v Speaker 3>application and potential military application. A robot that can do

595
00:28:30.519 --> 00:28:35.039
<v Speaker 3>parkour can navigate a complex battlefield. That's a reality. But

596
00:28:35.079 --> 00:28:38.000
<v Speaker 3>the gala was very, very careful to focus strictly on

597
00:28:38.119 --> 00:28:39.079
<v Speaker 3>positive harmony.

598
00:28:39.200 --> 00:28:42.279
<v Speaker 2>It was all about culture, family, tradition and entertainment, a.

599
00:28:42.319 --> 00:28:46.640
<v Speaker 3>Very soft, culturally resonant glove over a very very powerful

600
00:28:46.680 --> 00:28:50.039
<v Speaker 3>iron hand of technological capability. It's about showing that these

601
00:28:50.119 --> 00:28:53.079
<v Speaker 3>robots can integrate into society, not just disrupt it.

602
00:28:53.279 --> 00:28:55.839
<v Speaker 2>So what does this all mean. Then We've seen the robots,

603
00:28:55.880 --> 00:28:58.200
<v Speaker 2>we've analyzed the tech, we've looked at the sales figures

604
00:28:58.200 --> 00:29:00.640
<v Speaker 2>and the global reaction. What's the fine I'll take away

605
00:29:00.839 --> 00:29:01.799
<v Speaker 2>The twenty twenty.

606
00:29:01.519 --> 00:29:06.079
<v Speaker 3>Six Spring Festival Gala was a bold, unambiguous national statement.

607
00:29:07.119 --> 00:29:09.839
<v Speaker 3>China is not just catching up in physical AI, it

608
00:29:09.920 --> 00:29:13.559
<v Speaker 3>is sprinting ahead, and they were doing it by blending

609
00:29:13.599 --> 00:29:16.880
<v Speaker 3>their ancient heritage with their future ambitions in a way

610
00:29:16.920 --> 00:29:18.960
<v Speaker 3>that is incredibly savvy and powerful.

611
00:29:19.119 --> 00:29:20.599
<v Speaker 2>It really does feel like we are standing on the

612
00:29:20.680 --> 00:29:24.759
<v Speaker 2>edge of a cliff looking over at a completely new landscape.

613
00:29:24.119 --> 00:29:26.519
<v Speaker 3>And we're not just looking. I think last night proves

614
00:29:26.519 --> 00:29:28.480
<v Speaker 3>we are jumping off whether we're ready or not.

615
00:29:28.960 --> 00:29:31.880
<v Speaker 2>So I want to leave our listeners with one final thought, Timlover,

616
00:29:32.039 --> 00:29:33.799
<v Speaker 2>we just spent a lot of time talking about the

617
00:29:33.839 --> 00:29:36.279
<v Speaker 2>incredible leap from twenty twenty five to twenty twenty six

618
00:29:36.680 --> 00:29:41.480
<v Speaker 2>from wobbly handkerchief dances to fully autonomous weapon wielding backflips,

619
00:29:41.720 --> 00:29:44.359
<v Speaker 2>from scripted movements to adaptive swarms, a.

620
00:29:44.319 --> 00:29:46.279
<v Speaker 3>Massive, almost unbelievable acceleration.

621
00:29:46.440 --> 00:29:48.799
<v Speaker 2>If that is what happened in just one year, what

622
00:29:49.000 --> 00:29:51.319
<v Speaker 2>on earth will the stage look like in twenty twenty seven? And,

623
00:29:51.359 --> 00:29:54.039
<v Speaker 2>maybe more importantly, as these cultural preservers move from the

624
00:29:54.079 --> 00:29:57.319
<v Speaker 2>stage and into our factories, into our hospitals, into our homes,

625
00:29:58.119 --> 00:30:01.119
<v Speaker 2>are we as a society for the speed at which

626
00:30:01.160 --> 00:30:03.400
<v Speaker 2>science fiction is becoming a household appliance.

627
00:30:03.920 --> 00:30:05.920
<v Speaker 3>That is the question we all need to start answering,

628
00:30:06.960 --> 00:30:07.559
<v Speaker 3>and fast.
