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 --> 00:00:19.000
<v Speaker 1>examine the ideas shaping the future of artificial minds.

5
00:00:23.920 --> 00:00:27.640
<v Speaker 2>Imagine looking out your living room window on like a

6
00:00:27.679 --> 00:00:28.960
<v Speaker 2>freezing Tuesday morning.

7
00:00:29.039 --> 00:00:31.079
<v Speaker 3>Oh, the worst kind of morning, right, You.

8
00:00:31.160 --> 00:00:33.240
<v Speaker 2>Got these icy front steps, the kind that make you

9
00:00:33.280 --> 00:00:35.039
<v Speaker 2>nervous just walking down to get the mail.

10
00:00:35.119 --> 00:00:37.719
<v Speaker 3>Yeah, you're shuffling so you don't slip exactly.

11
00:00:39.000 --> 00:00:43.439
<v Speaker 2>But suddenly you see this machine navigating your driveway. It

12
00:00:43.479 --> 00:00:46.840
<v Speaker 2>looks almost like a dog, but it's on roller skates,

13
00:00:47.039 --> 00:00:50.359
<v Speaker 2>which is a wild visual, it really is. It smoothly

14
00:00:50.479 --> 00:00:54.960
<v Speaker 2>glides up your driveway, hits the icy stairs, seamlessly shifts

15
00:00:54.960 --> 00:00:58.320
<v Speaker 2>its weight, steps up, and drops a package perfectly on

16
00:00:58.359 --> 00:00:59.119
<v Speaker 2>your welcome app.

17
00:00:59.159 --> 00:01:01.439
<v Speaker 3>And the crazy part is what's happening indoors at the

18
00:01:01.479 --> 00:01:02.479
<v Speaker 3>exact same time.

19
00:01:02.600 --> 00:01:05.959
<v Speaker 2>Yeah, you turn around, still processing what you just saw outside,

20
00:01:06.159 --> 00:01:10.120
<v Speaker 2>and inside your house, a three foot tall bipedal robot

21
00:01:10.400 --> 00:01:13.439
<v Speaker 2>is quietly picking up your kids scattered toys from the

22
00:01:13.519 --> 00:01:15.040
<v Speaker 2>rug and putting them into a bin.

23
00:01:15.200 --> 00:01:17.599
<v Speaker 3>It paints quite a picture honestly, and you know, it's

24
00:01:17.599 --> 00:01:20.879
<v Speaker 3>a picture that forces us to completely rethink our relationship

25
00:01:20.879 --> 00:01:21.680
<v Speaker 3>with the spaces we.

26
00:01:21.680 --> 00:01:23.719
<v Speaker 2>Live in, because this is in the scene from some

27
00:01:23.760 --> 00:01:26.120
<v Speaker 2>sci fi movie set fifty years in the future, is.

28
00:01:26.079 --> 00:01:28.280
<v Speaker 3>It not at all? I mean, this is the exact

29
00:01:28.319 --> 00:01:32.200
<v Speaker 3>reality that Amazon just positioned itself to build right now.

30
00:01:32.359 --> 00:01:34.359
<v Speaker 2>Yeah, in the Spanish just a few days in March

31
00:01:34.439 --> 00:01:40.599
<v Speaker 2>twenty twenty six, Amazon quietly made two absolutely massive acquisitions

32
00:01:40.640 --> 00:01:41.719
<v Speaker 2>in the robotic space.

33
00:01:42.040 --> 00:01:44.879
<v Speaker 3>Massive moves. We're talking about a fundamental shift in how

34
00:01:44.920 --> 00:01:47.040
<v Speaker 3>we interact with technology.

35
00:01:46.439 --> 00:01:51.040
<v Speaker 2>Right moving away from traditional automation like those robotic arms

36
00:01:51.120 --> 00:01:54.799
<v Speaker 2>bolted to a factory floor doing repetitive tasks, and leaping

37
00:01:54.840 --> 00:01:58.040
<v Speaker 2>headfirst into the era of physical AI. Exact right today,

38
00:01:58.079 --> 00:02:00.480
<v Speaker 2>we are going to explore with these two specific moves,

39
00:02:00.519 --> 00:02:04.359
<v Speaker 2>reveal about how the world's biggest e commerce giant is

40
00:02:04.359 --> 00:02:07.560
<v Speaker 2>preparing to take robots out of their highly controlled warehouses

41
00:02:07.719 --> 00:02:10.560
<v Speaker 2>and deploy them into the physical chaos of your everyday life.

42
00:02:10.680 --> 00:02:13.360
<v Speaker 3>And physical chaos is really the perfect way to frame

43
00:02:13.400 --> 00:02:16.400
<v Speaker 3>this challenge. Also well, to understand why these acquisitions are

44
00:02:16.400 --> 00:02:19.199
<v Speaker 3>so pivotal, you have to look at the environment Amazon

45
00:02:19.240 --> 00:02:22.840
<v Speaker 3>is desperately trying to conquer. Inside an Amazon fulfillment center,

46
00:02:22.879 --> 00:02:25.800
<v Speaker 3>the environment is perfectly mapped. It's hyper controlled, right.

47
00:02:25.840 --> 00:02:27.879
<v Speaker 2>The floors are incredibly flat, the lighting.

48
00:02:27.680 --> 00:02:31.960
<v Speaker 3>Is constant exactly, there is zero weather interference, and the

49
00:02:32.080 --> 00:02:36.400
<v Speaker 3>robots know the exact geometric coordinates of every single shelf.

50
00:02:36.479 --> 00:02:38.520
<v Speaker 2>But the real world is messy.

51
00:02:38.400 --> 00:02:43.000
<v Speaker 3>Completely unpredictable. Your front yard, a crack sidewalk, your living room.

52
00:02:43.039 --> 00:02:46.560
<v Speaker 3>It is the ultimate engineering hurdle because you just cannot

53
00:02:46.599 --> 00:02:49.240
<v Speaker 3>pre program a robot for a world that changes every

54
00:02:49.280 --> 00:02:49.960
<v Speaker 3>single second.

55
00:02:50.000 --> 00:02:52.400
<v Speaker 2>So let's start right where the package journey currently breaks

56
00:02:52.439 --> 00:02:53.719
<v Speaker 2>down the street curb.

57
00:02:54.039 --> 00:02:57.919
<v Speaker 3>Ah. Yes, the industry calls this the last last.

58
00:02:57.680 --> 00:03:00.599
<v Speaker 2>Mile, and it is an absolute logistics nine nightmare. Right.

59
00:03:00.800 --> 00:03:03.800
<v Speaker 2>I mean Amazon has over a million robots working inside

60
00:03:03.840 --> 00:03:07.000
<v Speaker 2>its massive fulfillment centers right now. Yeah, picking and packing

61
00:03:07.039 --> 00:03:07.960
<v Speaker 2>with terrifying speed.

62
00:03:08.080 --> 00:03:09.199
<v Speaker 3>Terrifying is the right word.

63
00:03:09.280 --> 00:03:11.680
<v Speaker 2>But once that delivery van parks in a dense city

64
00:03:11.719 --> 00:03:14.159
<v Speaker 2>neighborhood or pulls up to a suburban house with a

65
00:03:14.159 --> 00:03:18.680
<v Speaker 2>steep flight of stairs, all that billion dollar automation just stops.

66
00:03:18.800 --> 00:03:19.719
<v Speaker 3>It hits a brick wall.

67
00:03:19.960 --> 00:03:22.879
<v Speaker 2>Yes, suddenly the entire supply chain relies on human labor.

68
00:03:23.080 --> 00:03:25.840
<v Speaker 2>It comes down to a delivery driver physically carrying a

69
00:03:25.879 --> 00:03:26.599
<v Speaker 2>cardboard box.

70
00:03:26.639 --> 00:03:30.439
<v Speaker 3>And from a purely economic perspective, that human bottleneck is

71
00:03:30.599 --> 00:03:34.360
<v Speaker 3>devastating to a company that is just obsessed with mondons.

72
00:03:34.240 --> 00:03:37.400
<v Speaker 2>Because hauling heavy boxes up a fifth floor apartment walk

73
00:03:37.479 --> 00:03:41.879
<v Speaker 2>up or navigating a dark, icy sidewalk is expensive.

74
00:03:42.000 --> 00:03:45.639
<v Speaker 3>It's expensive, it's inefficient, and it's highly prone to injury

75
00:03:45.639 --> 00:03:48.879
<v Speaker 3>and liability. Amazon has been trying to automate this away

76
00:03:48.919 --> 00:03:51.520
<v Speaker 3>for over a decade. Right but right around March nineteen,

77
00:03:51.599 --> 00:03:54.879
<v Speaker 3>twenty twenty six, they didn't just announce some new research initiative.

78
00:03:55.240 --> 00:03:59.639
<v Speaker 3>They bought a working solution. They acquired a Zurich based

79
00:03:59.719 --> 00:04:02.520
<v Speaker 3>robotics startup called rivr.

80
00:04:02.280 --> 00:04:05.759
<v Speaker 2>Rivr, which a lot of people in the industry knew

81
00:04:05.879 --> 00:04:08.599
<v Speaker 2>is Swiss Mile before their rebrand in early twenty twenty five.

82
00:04:08.719 --> 00:04:09.360
<v Speaker 3>Right exactly.

83
00:04:09.439 --> 00:04:12.639
<v Speaker 2>They specialize in autonomous four legged wheeled robots.

84
00:04:13.199 --> 00:04:16.639
<v Speaker 3>Okay, let's untack this. Why build a complex robot dog

85
00:04:16.759 --> 00:04:17.439
<v Speaker 3>on wheels?

86
00:04:17.639 --> 00:04:19.319
<v Speaker 2>It sounds counterintuitive, I know.

87
00:04:19.439 --> 00:04:22.399
<v Speaker 3>It sounds like adding unnecessary moving parts for a tech demo,

88
00:04:22.839 --> 00:04:25.279
<v Speaker 3>Like why not just use delivery drones flying over the

89
00:04:25.319 --> 00:04:28.160
<v Speaker 3>traffic or those little boxy sidewalk coolers we've seen rolling

90
00:04:28.160 --> 00:04:29.920
<v Speaker 3>around college campuses for the last five years.

91
00:04:30.040 --> 00:04:32.839
<v Speaker 2>Well, what's fascinating here is the sheer, mechanistic brilliance of

92
00:04:32.920 --> 00:04:35.600
<v Speaker 2>rivr's hybrid design. You really have to look at the

93
00:04:35.600 --> 00:04:38.560
<v Speaker 2>physics of navigating a human built environment.

94
00:04:38.120 --> 00:04:42.439
<v Speaker 3>Because drones have issues, right, huge issues. Drones are incredibly limited,

95
00:04:42.480 --> 00:04:46.079
<v Speaker 3>They have severe payload restrictions, their batteries drain rapidly in

96
00:04:46.120 --> 00:04:49.319
<v Speaker 3>cold weather. And you know they can't exactly open an

97
00:04:49.319 --> 00:04:53.279
<v Speaker 3>apartment building door or drop a package securely under a

98
00:04:53.319 --> 00:04:54.240
<v Speaker 3>covered porch.

99
00:04:54.160 --> 00:04:56.560
<v Speaker 2>Right, you can't have a drone flying into your building's lobby.

100
00:04:56.639 --> 00:04:58.040
<v Speaker 2>And what about the sidewalk coolers?

101
00:04:58.360 --> 00:05:01.839
<v Speaker 3>Those little wheeled coolers you make are highly energy efficient

102
00:05:01.920 --> 00:05:05.240
<v Speaker 3>on perfectly flat ground, but the second they hit a

103
00:05:05.279 --> 00:05:09.319
<v Speaker 3>two inch curb, a pothole, or a single step.

104
00:05:09.079 --> 00:05:11.959
<v Speaker 2>They're completely paralyzed exactly.

105
00:05:12.120 --> 00:05:15.040
<v Speaker 3>RAVR solves the physics problem by blending the best of

106
00:05:15.079 --> 00:05:15.759
<v Speaker 3>both worlds.

107
00:05:16.040 --> 00:05:19.519
<v Speaker 2>So it uses wheeled mobility for fast, energy efficient travel

108
00:05:19.560 --> 00:05:23.199
<v Speaker 2>down a paved street, but it can dynamically alter its

109
00:05:23.240 --> 00:05:25.879
<v Speaker 2>locomotion strategy the moment the terrain.

110
00:05:25.800 --> 00:05:29.120
<v Speaker 3>Changes precisely the point when riviar encounters a staircase. It

111
00:05:29.120 --> 00:05:30.920
<v Speaker 3>doesn't just stop and try to roll over it.

112
00:05:31.000 --> 00:05:31.800
<v Speaker 2>That would be a disaster.

113
00:05:31.920 --> 00:05:34.959
<v Speaker 3>Yeah. It actively locks its wheel rotors, effectively turning those

114
00:05:34.959 --> 00:05:39.000
<v Speaker 3>wheels into rigid, high friction feet. Then it actuates its

115
00:05:39.079 --> 00:05:43.839
<v Speaker 3>knee joints to climb, step over obstacles and navigate tight spaces.

116
00:05:43.920 --> 00:05:44.399
<v Speaker 2>Oh wow.

117
00:05:44.480 --> 00:05:47.879
<v Speaker 3>It calculates the friction coefficient of the ice, shifts its

118
00:05:47.920 --> 00:05:51.720
<v Speaker 3>center of mass, and climbs the stairs like a biological quadruped.

119
00:05:51.879 --> 00:05:54.639
<v Speaker 2>It's adapting in real time, and that is really the

120
00:05:54.639 --> 00:05:58.439
<v Speaker 2>core of physical AI. It's not following some pre programmed

121
00:05:58.519 --> 00:06:00.079
<v Speaker 2>track or a GPS waypoint, No.

122
00:06:00.399 --> 00:06:00.839
<v Speaker 3>Not at all.

123
00:06:00.920 --> 00:06:05.079
<v Speaker 2>It relies heavily on proprioceptive sensors to dynamically adjust its

124
00:06:05.079 --> 00:06:08.040
<v Speaker 2>center of gravity. It feels the micro slips on the

125
00:06:08.079 --> 00:06:12.199
<v Speaker 2>ice and instantly fires its motors to adjust its balance.

126
00:06:12.040 --> 00:06:14.040
<v Speaker 3>Just like you would if you felt your shoe slipping

127
00:06:14.079 --> 00:06:17.759
<v Speaker 3>on a slick driveway. It's processing visual data and physical

128
00:06:17.759 --> 00:06:19.920
<v Speaker 3>feedback simultaneously to stay upright.

129
00:06:20.000 --> 00:06:23.879
<v Speaker 2>That real time physics based adaptation is basically the holy

130
00:06:23.920 --> 00:06:28.319
<v Speaker 2>grail of modern robotics, and the business reality surrounding REVR

131
00:06:28.680 --> 00:06:31.319
<v Speaker 2>shows just how seriously the industry takes this approach.

132
00:06:31.399 --> 00:06:32.839
<v Speaker 3>The numbers back it up, for sure.

133
00:06:32.920 --> 00:06:35.879
<v Speaker 2>Yeah, before the acquisition, RIVR had alread raised about twenty

134
00:06:35.879 --> 00:06:39.519
<v Speaker 2>five million dollars, hitting evaluation well over one hundred million dollars.

135
00:06:39.639 --> 00:06:41.920
<v Speaker 3>And if you look at their cab table, Jeff Bezos's

136
00:06:42.000 --> 00:06:46.040
<v Speaker 3>personal investment vehicle BUSUS expeditions, along with the Amazon Industrial

137
00:06:46.079 --> 00:06:48.560
<v Speaker 3>Innovation Fund were already early investors.

138
00:06:48.879 --> 00:06:50.480
<v Speaker 2>Oh, so they had a front row seat to the

139
00:06:50.560 --> 00:06:53.600
<v Speaker 2>underlying code and the field performance for a long time.

140
00:06:53.839 --> 00:06:57.800
<v Speaker 3>Right, this wasn't a spontaneous tech acquisition. RIVR was already

141
00:06:57.879 --> 00:07:00.720
<v Speaker 3>proving its viability in the messy real world.

142
00:07:00.839 --> 00:07:03.759
<v Speaker 2>Yeah. They were running parcel delivery trials with Vho down

143
00:07:03.759 --> 00:07:07.199
<v Speaker 2>in Austin, Texas. They were executing field tests with Swiss

144
00:07:07.199 --> 00:07:09.839
<v Speaker 2>Post and micros Online in Switzerland, and.

145
00:07:09.759 --> 00:07:13.399
<v Speaker 3>Even navigating European city centers doing meal deliveries with just

146
00:07:13.480 --> 00:07:14.680
<v Speaker 3>eattakeaway dot com.

147
00:07:14.959 --> 00:07:17.959
<v Speaker 2>So they were gathering the one resource that AI needs

148
00:07:18.040 --> 00:07:22.720
<v Speaker 2>more than anything else, messy unstructured real world data.

149
00:07:23.240 --> 00:07:25.759
<v Speaker 3>That data collection is the real story here. When the

150
00:07:25.759 --> 00:07:30.920
<v Speaker 3>acquisition was announced, rivr's CEO Marco Gallonik talked about accelerating

151
00:07:30.959 --> 00:07:35.480
<v Speaker 3>their vision of building general physical AI through doorstep delivery. Wow,

152
00:07:35.639 --> 00:07:39.040
<v Speaker 3>that phrasing unlocks Amazon's and tiremaster plan. They aren't just

153
00:07:39.079 --> 00:07:41.920
<v Speaker 3>building a delivery tool. They are using doorstep delivery as

154
00:07:41.920 --> 00:07:43.199
<v Speaker 3>a trojan horse to.

155
00:07:43.319 --> 00:07:45.800
<v Speaker 2>Subsidize the mapping of the physical world. That is an

156
00:07:45.839 --> 00:07:49.720
<v Speaker 2>incredible insight. It really is, because large language models, like

157
00:07:49.759 --> 00:07:53.639
<v Speaker 2>the ones powering chatbots were trained by screeping billions of

158
00:07:53.680 --> 00:07:57.680
<v Speaker 2>pages of text from the Internet. But physical AI needs

159
00:07:57.680 --> 00:07:59.079
<v Speaker 2>a different kind of data, right.

160
00:07:59.120 --> 00:08:04.199
<v Speaker 3>It needs to learn gravity, friction, wind resistance, and spatial geometry.

161
00:08:04.240 --> 00:08:06.279
<v Speaker 2>You can't just download that from Wikipedia.

162
00:08:06.319 --> 00:08:09.240
<v Speaker 3>You have to experience it. And LM is essentially a

163
00:08:09.279 --> 00:08:12.319
<v Speaker 3>brain in a jar that can only read. Physical AI

164
00:08:12.560 --> 00:08:14.720
<v Speaker 3>is a toddler that has to learn how to walk

165
00:08:14.800 --> 00:08:16.959
<v Speaker 3>by falling down thousands of times.

166
00:08:17.240 --> 00:08:19.680
<v Speaker 2>And collecting that physical data at a global scale is

167
00:08:19.920 --> 00:08:23.720
<v Speaker 2>prohibitively expensive unless you have a business model that pays

168
00:08:23.759 --> 00:08:25.720
<v Speaker 2>for the robots to be out there in the first place.

169
00:08:25.839 --> 00:08:29.720
<v Speaker 3>Exactly, Amazon's delivery network is the only financially viable way

170
00:08:29.759 --> 00:08:32.679
<v Speaker 3>to put millions of data gathering sensors on the streets

171
00:08:32.720 --> 00:08:33.480
<v Speaker 3>every single day.

172
00:08:33.639 --> 00:08:37.879
<v Speaker 2>And Pallo Prijanion, Amazon's VP of Last Mild Delivery Automation,

173
00:08:38.440 --> 00:08:40.759
<v Speaker 2>is already designing the deployment architecture for this.

174
00:08:40.919 --> 00:08:43.039
<v Speaker 3>Yeah, he's talking about hybrid fleets.

175
00:08:42.840 --> 00:08:45.919
<v Speaker 2>Right, So you have a fully autonomous electric van rolling

176
00:08:45.960 --> 00:08:49.120
<v Speaker 2>through a suburban neighborhood. Instead of a human driver getting

177
00:08:49.120 --> 00:08:51.360
<v Speaker 2>out to walk to five different houses, a.

178
00:08:51.279 --> 00:08:54.879
<v Speaker 3>Pack of these quadruped RIVR robots deploys from the back,

179
00:08:55.320 --> 00:08:59.240
<v Speaker 3>scatters to drop packages on five different porches simultaneously, and

180
00:08:59.279 --> 00:09:01.039
<v Speaker 3>then meets the at the end of the block.

181
00:09:01.240 --> 00:09:04.759
<v Speaker 2>It completely obliterates the traditional math of last mile logistics.

182
00:09:04.799 --> 00:09:10.200
<v Speaker 3>Oh absolutely, It lowers fuel costs, minimizes delivery failures, drastically

183
00:09:10.240 --> 00:09:15.519
<v Speaker 3>reduces worker compensation claims, and creates a highly precise contactless

184
00:09:15.600 --> 00:09:16.600
<v Speaker 3>drop off network.

185
00:09:16.759 --> 00:09:19.120
<v Speaker 2>But solving the front steps only gets the box to

186
00:09:19.159 --> 00:09:22.919
<v Speaker 2>the welcome mat. True, the real bottleneck for Amazon's ecosystem

187
00:09:23.240 --> 00:09:26.039
<v Speaker 2>isn't the driveway anymore. It's getting past the dead bolt,

188
00:09:26.200 --> 00:09:29.120
<v Speaker 2>and that requires a completely different type of machine.

189
00:09:29.159 --> 00:09:31.879
<v Speaker 3>Which brings us to their second major move, occurring just

190
00:09:32.039 --> 00:09:36.039
<v Speaker 3>days later. On March twenty fourth, twenty twenty six, Amazon

191
00:09:36.080 --> 00:09:38.279
<v Speaker 3>confirmed the acquisition of Fauna Robotics.

192
00:09:38.480 --> 00:09:41.720
<v Speaker 2>Now, Fauna operates in an entirely different paradigm than.

193
00:09:41.720 --> 00:09:43.279
<v Speaker 3>REVR, completely different.

194
00:09:43.399 --> 00:09:45.600
<v Speaker 2>They're a two year old startup based in New York,

195
00:09:45.919 --> 00:09:48.799
<v Speaker 2>founded by a group of highly specialized former engineers from

196
00:09:48.840 --> 00:09:50.240
<v Speaker 2>Meta and Google right.

197
00:09:50.120 --> 00:09:53.360
<v Speaker 3>And unlike the RIVR team, Fauna's fifty person staff is

198
00:09:53.360 --> 00:09:54.879
<v Speaker 3>staying right there in New York.

199
00:09:55.200 --> 00:09:59.639
<v Speaker 2>Their focus isn't industrial logistics, package weight capacities, or rugged

200
00:09:59.639 --> 00:10:05.200
<v Speaker 2>street delivery. Their singular focus is consumer humanoid robots designed

201
00:10:05.240 --> 00:10:05.559
<v Speaker 2>for the.

202
00:10:05.519 --> 00:10:10.799
<v Speaker 3>Home, specifically kid sized humanoids. Their flagship robot is named Sprout.

203
00:10:11.320 --> 00:10:14.039
<v Speaker 2>To give you a visual Sprout stands exactly three feet

204
00:10:14.120 --> 00:10:17.240
<v Speaker 2>six inches tall, and it weighs fifty nine pounds, And that.

205
00:10:17.360 --> 00:10:21.000
<v Speaker 3>Fifty nine pounds specification is arguably the most important data

206
00:10:21.039 --> 00:10:22.799
<v Speaker 3>point about the entire company.

207
00:10:22.919 --> 00:10:25.399
<v Speaker 2>Here's where it gets really interesting. It sounds like they

208
00:10:25.559 --> 00:10:30.279
<v Speaker 2>essentially built an incredibly polite, highly capable four year old child.

209
00:10:30.360 --> 00:10:32.240
<v Speaker 3>That is the perfect analogy.

210
00:10:32.039 --> 00:10:34.879
<v Speaker 2>Because when we usually think of humanoid robots, our minds

211
00:10:34.919 --> 00:10:39.559
<v Speaker 2>immediately jump to those massive, heavy duty industrial humanoids. We

212
00:10:39.639 --> 00:10:42.679
<v Speaker 2>think of the Boston Dynamics robots doing parkre or the

213
00:10:42.759 --> 00:10:46.200
<v Speaker 2>Towering Figure robots designed to lift engine blocks and BMW

214
00:10:46.320 --> 00:10:47.360
<v Speaker 2>manufacturing plants.

215
00:10:47.440 --> 00:10:50.759
<v Speaker 3>And those industrial robots are breathtaking feats of engineering, but

216
00:10:50.840 --> 00:10:53.279
<v Speaker 3>they are heavy, they use high torque motors, and frankly,

217
00:10:53.320 --> 00:10:55.639
<v Speaker 3>they are incredibly dangerous if you share a space with.

218
00:10:55.679 --> 00:10:57.559
<v Speaker 2>Um, yeah, I wouldn't want one in my kitchen. Right.

219
00:10:57.960 --> 00:11:00.799
<v Speaker 3>If a two hundred pound metal machine walking around your

220
00:11:00.799 --> 00:11:03.480
<v Speaker 3>living room loses its balance and falls on your toddler,

221
00:11:03.879 --> 00:11:06.639
<v Speaker 3>it is a catastrophic hospital level event.

222
00:11:06.879 --> 00:11:09.240
<v Speaker 2>So that fifty nine pound weight limit isn't just about

223
00:11:09.240 --> 00:11:11.960
<v Speaker 2>carrying capacity. It's a strict liability calculation.

224
00:11:12.200 --> 00:11:16.440
<v Speaker 3>It's a complete shift and form factor to ensure psychological

225
00:11:16.519 --> 00:11:21.240
<v Speaker 3>and physical safety. Sprout is deliberately designed to be approachable.

226
00:11:21.399 --> 00:11:23.879
<v Speaker 2>So if a fifty nine pound robot accidentally bumps into

227
00:11:23.879 --> 00:11:25.799
<v Speaker 2>you in the kitchen or trips over a dog toy

228
00:11:25.799 --> 00:11:27.720
<v Speaker 2>and falls, it's just a minor annoyance.

229
00:11:27.919 --> 00:11:31.000
<v Speaker 3>It's like a Golden retriever bumping your leg. It relies

230
00:11:31.080 --> 00:11:35.360
<v Speaker 3>on compliant actuators, motors that act more like biological muscles

231
00:11:35.360 --> 00:11:36.759
<v Speaker 3>with built in springiness.

232
00:11:36.799 --> 00:11:40.080
<v Speaker 2>Oh so they literally give way upon unexpected impact instead

233
00:11:40.080 --> 00:11:42.240
<v Speaker 2>of rigidly forcing their way through an obstacle.

234
00:11:42.399 --> 00:11:45.480
<v Speaker 3>Exactly. That compliance is what makes it viable for the

235
00:11:45.519 --> 00:11:49.159
<v Speaker 3>family ecosystem. Makes sense because Sprout is bipedal, it inherently

236
00:11:49.159 --> 00:11:52.879
<v Speaker 3>faces massive balance challenges, but that bipedalism allows it to

237
00:11:52.960 --> 00:11:55.159
<v Speaker 3>navigate human homes seamlessly.

238
00:11:55.279 --> 00:11:58.600
<v Speaker 2>It can walk up carpeted stairs, naturally grip objects with

239
00:11:58.720 --> 00:12:01.440
<v Speaker 2>human like hands, pick up dropped pantry goods, and.

240
00:12:01.480 --> 00:12:05.679
<v Speaker 3>Even engage in natural social interactions or dancing. It's positioned

241
00:12:05.720 --> 00:12:09.440
<v Speaker 3>as a versatile home companion capable of light chores.

242
00:12:09.320 --> 00:12:11.799
<v Speaker 2>And for early developer units, it was priced around fifty

243
00:12:11.799 --> 00:12:12.879
<v Speaker 2>thousand dollars.

244
00:12:12.679 --> 00:12:17.159
<v Speaker 3>Which sounds steep, but for advanced bipedal robotics capable of

245
00:12:17.200 --> 00:12:22.000
<v Speaker 3>dexterous manipulation, that is surprisingly accessible for research partners.

246
00:12:22.039 --> 00:12:25.120
<v Speaker 2>And Amazon has a long history of taking expensive hardware

247
00:12:25.480 --> 00:12:28.879
<v Speaker 2>like the first Kindle or early echo speakers, and driving

248
00:12:28.879 --> 00:12:33.159
<v Speaker 2>the manufacturing costs down drastically to achieve consumer scale. Oh,

249
00:12:33.360 --> 00:12:37.360
<v Speaker 2>for sure, this represents a paradigm shifting leap from Amazon's

250
00:12:37.399 --> 00:12:41.919
<v Speaker 2>previous attempts at home robotics. I'm thinking specifically about Astro.

251
00:12:42.320 --> 00:12:45.159
<v Speaker 3>Astro is the perfect comparison to show how far we've come.

252
00:12:45.480 --> 00:12:48.039
<v Speaker 3>Amazon released Astro to the public a few years ago,

253
00:12:48.399 --> 00:12:52.360
<v Speaker 3>but functionally, Astro was essentially an alepplish tablet screen glued

254
00:12:52.440 --> 00:12:54.039
<v Speaker 3>to a roomabase.

255
00:12:53.639 --> 00:12:56.200
<v Speaker 2>Right, it was just a rolling companion with cameras.

256
00:12:56.240 --> 00:12:58.200
<v Speaker 3>It couldn't pick up a drop towel, It couldn't reach

257
00:12:58.200 --> 00:13:00.039
<v Speaker 3>a countertop to wipe it down. It couldn't put a

258
00:13:00.080 --> 00:13:01.480
<v Speaker 3>box of cereal back on the shelf.

259
00:13:01.519 --> 00:13:06.480
<v Speaker 2>But Fauna brings actual bipedal locomotion and dextrous physical manipulation

260
00:13:06.639 --> 00:13:07.279
<v Speaker 2>into the home.

261
00:13:07.639 --> 00:13:10.679
<v Speaker 3>It crosses the threshold from a machine that merely observes

262
00:13:10.679 --> 00:13:14.960
<v Speaker 3>and listens into a machine that actively physically participates in

263
00:13:15.000 --> 00:13:16.440
<v Speaker 3>the upkeep of your household.

264
00:13:16.720 --> 00:13:20.639
<v Speaker 2>So what does this all mean. We've got rivr conquering

265
00:13:20.639 --> 00:13:23.799
<v Speaker 2>the unpredictable physics of the outdoors, and we've got Fauna

266
00:13:23.840 --> 00:13:27.000
<v Speaker 2>sprout engineer to safely manipulate the intimate indoors.

267
00:13:27.120 --> 00:13:28.159
<v Speaker 3>It's a massive push.

268
00:13:28.759 --> 00:13:31.639
<v Speaker 2>Why is Amazon dropping massive amounts of capital to acquire

269
00:13:31.679 --> 00:13:35.360
<v Speaker 2>these two very specific capabilities in the exact same week.

270
00:13:35.600 --> 00:13:38.600
<v Speaker 3>Well, if we connect this to the bigger picture, this

271
00:13:38.720 --> 00:13:42.639
<v Speaker 3>aggressive land grab is about the ultimate convergence of artificial

272
00:13:42.679 --> 00:13:44.480
<v Speaker 3>intelligence and physical robotics.

273
00:13:44.519 --> 00:13:47.240
<v Speaker 2>Because for the last few years, the entire AI boom

274
00:13:47.279 --> 00:13:48.600
<v Speaker 2>has been trapped behind glass.

275
00:13:48.919 --> 00:13:52.960
<v Speaker 3>Exactly, it's been generative chatbots on your phone, screen writing emails,

276
00:13:53.039 --> 00:13:56.919
<v Speaker 3>or creating digital art. But the true multi trillion dollar

277
00:13:56.960 --> 00:13:59.559
<v Speaker 3>frontier is embodied intelligence.

278
00:13:59.159 --> 00:14:02.440
<v Speaker 2>An AI brand map to physical motors that can manipulate

279
00:14:02.440 --> 00:14:03.039
<v Speaker 2>the real world.

280
00:14:03.200 --> 00:14:06.879
<v Speaker 3>Yes, Amazon is racing against Elon Musk's Tesla with their

281
00:14:06.919 --> 00:14:10.279
<v Speaker 3>optimist humanoid program and a multitude of incredibly well funded

282
00:14:10.360 --> 00:14:13.639
<v Speaker 3>Chinese firms to completely dominate the physical AI space.

283
00:14:14.000 --> 00:14:17.279
<v Speaker 2>They realize that the company that builds the physical infrastructure

284
00:14:17.279 --> 00:14:20.399
<v Speaker 2>of your life owns the ultimate consumer platform.

285
00:14:20.519 --> 00:14:22.960
<v Speaker 3>They don't just want to be the digital storefront where

286
00:14:22.960 --> 00:14:24.039
<v Speaker 3>you buy paper towels.

287
00:14:24.240 --> 00:14:26.639
<v Speaker 2>They want to own the machine that brings the paper

288
00:14:26.679 --> 00:14:30.159
<v Speaker 2>towels up your driveway and the machine that physically places

289
00:14:30.200 --> 00:14:31.759
<v Speaker 2>the role in your kitchen pantry.

290
00:14:32.200 --> 00:14:35.480
<v Speaker 3>But as inevitable as this corporate strategy seems, we have

291
00:14:35.559 --> 00:14:38.919
<v Speaker 3>to look at the massive real world friction. This isn't

292
00:14:38.960 --> 00:14:40.240
<v Speaker 3>going to be a seamless rollout.

293
00:14:40.399 --> 00:14:43.759
<v Speaker 2>Far from it. There are towering regulatory and physical barriers

294
00:14:43.799 --> 00:14:47.879
<v Speaker 2>Amazon has to overcome before RIVR and sprout are normalized.

295
00:14:48.039 --> 00:14:50.639
<v Speaker 3>Oh, the physical deployment is going to be brutal. Let's

296
00:14:50.679 --> 00:14:54.639
<v Speaker 3>look at the outside world with RIVR. Deploying fleets of legged,

297
00:14:54.799 --> 00:14:59.039
<v Speaker 3>rolling autonomous robots on public sidewalks is a municipal, regulatory

298
00:14:59.080 --> 00:15:00.120
<v Speaker 3>and nightmare.

299
00:14:59.799 --> 00:15:03.240
<v Speaker 2>Right, because every single city has different ordinances about what

300
00:15:03.320 --> 00:15:06.679
<v Speaker 2>can share the pavement with pedestrians, wheelchairs and strollers.

301
00:15:06.799 --> 00:15:10.360
<v Speaker 3>Then you have the brutal physics of battery chemistry. Battery

302
00:15:10.399 --> 00:15:14.000
<v Speaker 3>life drains exponentially faster when a robot is constantly calculating

303
00:15:14.039 --> 00:15:18.000
<v Speaker 3>balance and actuating hydraulic or electric joints to climb snowy

304
00:15:18.039 --> 00:15:20.519
<v Speaker 3>stairs compared to just rolling on flat ground.

305
00:15:20.840 --> 00:15:23.679
<v Speaker 2>Not to mention the environmental wear and tear you have.

306
00:15:24.159 --> 00:15:28.799
<v Speaker 2>Weather resistant sensitive computing electronics and heavy rain or road

307
00:15:28.840 --> 00:15:30.559
<v Speaker 2>salt do not mix well.

308
00:15:30.840 --> 00:15:31.720
<v Speaker 3>They really don't.

309
00:15:32.039 --> 00:15:36.879
<v Speaker 2>The maintenance costs of keeping millions of complex robotic joints lubricated, calibrated,

310
00:15:36.919 --> 00:15:40.720
<v Speaker 2>and functional across the globe is a staggering logistical hurdle

311
00:15:40.759 --> 00:15:41.480
<v Speaker 2>on its own.

312
00:15:41.440 --> 00:15:43.480
<v Speaker 3>And when you shift to the home front with Sprout,

313
00:15:43.600 --> 00:15:48.000
<v Speaker 3>the engineering hurdles are quickly overshadowed by deeply entrenched social

314
00:15:48.039 --> 00:15:49.519
<v Speaker 3>and psychological barriers.

315
00:15:49.720 --> 00:15:51.759
<v Speaker 2>You mean, the public acceptance hurt exactly.

316
00:15:51.840 --> 00:15:54.799
<v Speaker 3>Even if Amazon gets the cost down to five thousand dollars,

317
00:15:55.159 --> 00:15:58.039
<v Speaker 3>is the average consumer ready to pay that for a humanoid?

318
00:15:58.519 --> 00:16:01.960
<v Speaker 2>Furthermore, even with com client actuators and a fifty nine

319
00:16:02.000 --> 00:16:04.840
<v Speaker 2>pound weight limit, building trust is going to take years.

320
00:16:05.200 --> 00:16:07.559
<v Speaker 2>Our parents truly going to trust a walking machine around

321
00:16:07.559 --> 00:16:09.399
<v Speaker 2>an unpredictable, fragile toddler.

322
00:16:09.480 --> 00:16:10.480
<v Speaker 3>It's a huge ask.

323
00:16:10.639 --> 00:16:12.879
<v Speaker 2>If you're listening to this and thinking I'm never letting

324
00:16:12.879 --> 00:16:16.320
<v Speaker 2>a robot with cameras walk around my bedroom, well, Amazon

325
00:16:16.399 --> 00:16:18.960
<v Speaker 2>actually knows that that's the biggest friction point.

326
00:16:19.159 --> 00:16:22.440
<v Speaker 3>Astro already raised massive privacy eyebrows. But Sprout is an

327
00:16:22.639 --> 00:16:24.240
<v Speaker 3>entirely different level of intrusion.

328
00:16:24.519 --> 00:16:28.720
<v Speaker 2>It is an Internet connected physical entity equipped with multiple

329
00:16:28.799 --> 00:16:31.679
<v Speaker 2>high definition cameras, depth sensors, and microphones.

330
00:16:32.279 --> 00:16:35.159
<v Speaker 3>It has the physical capability to walk into your bedroom,

331
00:16:35.679 --> 00:16:39.440
<v Speaker 3>open your closet doors, and geometrically map every inch of

332
00:16:39.480 --> 00:16:40.799
<v Speaker 3>your most private spaces.

333
00:16:41.000 --> 00:16:43.279
<v Speaker 2>The level of trust to consumer has to extend to

334
00:16:43.360 --> 00:16:47.039
<v Speaker 2>a massive corporate entity to allow that kind of access

335
00:16:47.039 --> 00:16:49.159
<v Speaker 2>into their home is unprecedented.

336
00:16:49.240 --> 00:16:53.039
<v Speaker 3>It's an astronomical leap of faith. But Amazon's deep pockets,

337
00:16:53.080 --> 00:16:56.919
<v Speaker 3>their vast aws, server infrastructure for computing, the AI models,

338
00:16:57.279 --> 00:17:01.159
<v Speaker 3>and their unmatched global logistics expertise give them a unique

339
00:17:01.200 --> 00:17:05.000
<v Speaker 3>advantage to power through these physical and social barriers over

340
00:17:05.039 --> 00:17:05.880
<v Speaker 3>the next decade.

341
00:17:05.920 --> 00:17:07.440
<v Speaker 2>They're playing a very long game here.

342
00:17:07.559 --> 00:17:10.960
<v Speaker 3>They are meticulously building a continuous robotic chain of custody.

343
00:17:11.400 --> 00:17:14.880
<v Speaker 3>Think about it. The warehouse, robotic arms, pack the box.

344
00:17:14.880 --> 00:17:18.720
<v Speaker 2>The REVR quadruped scales, your icy steps to deliver.

345
00:17:18.519 --> 00:17:21.359
<v Speaker 3>The box, and the fauna humanoid brings the box inside

346
00:17:21.400 --> 00:17:23.039
<v Speaker 3>and physically puts the contents away.

347
00:17:23.079 --> 00:17:26.759
<v Speaker 2>It's an absolute masterclass in vertical integration. Riveyard tams the

348
00:17:26.839 --> 00:17:30.000
<v Speaker 2>rugged outdoors and fauna masters the delicate indoors.

349
00:17:30.079 --> 00:17:32.759
<v Speaker 3>This raises an important question, though, when you look at

350
00:17:32.759 --> 00:17:35.680
<v Speaker 3>how seamlessly these machines are designed to blend into our

351
00:17:35.759 --> 00:17:39.720
<v Speaker 3>daily routines, it forces us to look past the impressive

352
00:17:39.720 --> 00:17:44.359
<v Speaker 3>engineering and deeply consider the human element. Also, we are

353
00:17:44.400 --> 00:17:49.920
<v Speaker 3>talking about introducing moving intelligent entities into human social dynamics, right.

354
00:17:50.359 --> 00:17:53.920
<v Speaker 3>Think about Sprout. If a kid's size robot becomes as

355
00:17:53.960 --> 00:17:56.880
<v Speaker 3>common in a house as a dishwasher, we aren't just

356
00:17:56.920 --> 00:18:00.720
<v Speaker 3>buying appliances anymore. Imagine a child growing up with Sprout.

357
00:18:01.000 --> 00:18:02.240
<v Speaker 2>Yeah, that's a wild thought.

358
00:18:02.640 --> 00:18:07.279
<v Speaker 3>What happens to human psychological development when a child's primary playmate, tutor,

359
00:18:07.359 --> 00:18:12.119
<v Speaker 3>and chore assistant is an ever patient, entirely unfeeling, corporate

360
00:18:12.200 --> 00:18:13.400
<v Speaker 3>owned humanoid.

361
00:18:13.599 --> 00:18:15.880
<v Speaker 2>We are talking about a machine that never gets angry,

362
00:18:16.359 --> 00:18:20.960
<v Speaker 2>never sets emotional boundaries, never exhibits fatigue, and perfectly caters

363
00:18:21.000 --> 00:18:22.400
<v Speaker 2>to their requests exactly.

364
00:18:22.440 --> 00:18:25.400
<v Speaker 3>We are fundamentally altering the social and emotional environment in

365
00:18:25.440 --> 00:18:28.160
<v Speaker 3>which our children will develop their baseline empathy.

366
00:18:28.200 --> 00:18:29.880
<v Speaker 2>Wait, let me push back on that a bit, because

367
00:18:29.880 --> 00:18:31.960
<v Speaker 2>I think there's another side to that. Coint go for it.

368
00:18:32.039 --> 00:18:35.839
<v Speaker 2>We already use iPads and algorithm driven YouTube feeds as

369
00:18:36.160 --> 00:18:40.279
<v Speaker 2>digital babysitters, which isolate kids behind a screen. An embodied

370
00:18:40.279 --> 00:18:43.400
<v Speaker 2>AI like Sprout could physically interact with them in the

371
00:18:43.440 --> 00:18:47.599
<v Speaker 2>real world. An infinitely patient tutor that never gets frustrated,

372
00:18:47.960 --> 00:18:53.559
<v Speaker 2>unlike an overworked tired parent, might actually provide customized, stress

373
00:18:53.599 --> 00:18:58.359
<v Speaker 2>free early childhood education that humans simply can't match.

374
00:18:58.599 --> 00:19:00.240
<v Speaker 3>That's a really interesting point.

375
00:19:00.319 --> 00:19:03.279
<v Speaker 2>It could be a massive upgrade for cognitive development, even

376
00:19:03.319 --> 00:19:05.279
<v Speaker 2>if the emotional dynamic is strange.

377
00:19:05.480 --> 00:19:09.480
<v Speaker 3>That is a fair counterpoint. It removes the friction of

378
00:19:09.559 --> 00:19:13.759
<v Speaker 3>human fatigue from education. But whether it's a utopian educational

379
00:19:13.799 --> 00:19:17.680
<v Speaker 3>tool or a concerning disruption to human empathy, the core

380
00:19:17.680 --> 00:19:21.799
<v Speaker 3>reality remains, which is the era of widespread embodied artificial

381
00:19:21.839 --> 00:19:24.640
<v Speaker 3>intelligence isn't just arriving faster than we thought. It's going

382
00:19:24.720 --> 00:19:27.119
<v Speaker 3>to change our psychological landscape or ways we haven't even

383
00:19:27.160 --> 00:19:27.880
<v Speaker 3>begun to measure.

384
00:19:28.000 --> 00:19:30.319
<v Speaker 2>It's going to redefine what a household even looks like.

385
00:19:30.759 --> 00:19:32.960
<v Speaker 2>Keep an eye on your front steps. Everyone yeah. The

386
00:19:33.000 --> 00:19:35.160
<v Speaker 2>next time you see a dog on roller skates braving

387
00:19:35.200 --> 00:19:38.440
<v Speaker 2>the ice to bring you a package, remember that's just

388
00:19:38.519 --> 00:19:40.759
<v Speaker 2>the tip of the spear. Thanks for joining us, and

389
00:19:40.799 --> 00:19:41.720
<v Speaker 2>we'll catch you on the next one.