1
00:00:10,200 --> 00:00:14,280
Speaker 1: Hello, and welcome to another episode of the Odd Lots Podcast.

2
00:00:14,360 --> 00:00:16,080
Speaker 1: I'm Joe Wisenthal.

3
00:00:15,600 --> 00:00:16,639
Speaker 2: And I'm Tracy Alloway.

4
00:00:17,120 --> 00:00:20,000
Speaker 1: Tracy, do you remember a few weeks ago, or maybe

5
00:00:20,000 --> 00:00:22,040
Speaker 1: a couple of months ago now, someone made a like

6
00:00:22,079 --> 00:00:25,840
Speaker 1: a fake AI version of the Odd Lots podcast and

7
00:00:25,880 --> 00:00:27,760
Speaker 1: they like replicated our voice.

8
00:00:28,000 --> 00:00:30,520
Speaker 2: Yes, I think they used chat gpt to do it.

9
00:00:30,600 --> 00:00:33,960
Speaker 2: And they basically asked chat gpt to create a script

10
00:00:34,000 --> 00:00:37,320
Speaker 2: for odd Loots and that they had someone replicate another app

11
00:00:37,400 --> 00:00:38,400
Speaker 2: or something replicated.

12
00:00:38,479 --> 00:00:40,120
Speaker 1: It wasn't a someone like I mean, it was like

13
00:00:40,200 --> 00:00:43,680
Speaker 1: somewhat like other like AI app. It was like, yeah,

14
00:00:43,840 --> 00:00:45,760
Speaker 1: it was pretty it was pretty good. It was like good.

15
00:00:45,840 --> 00:00:48,200
Speaker 1: I mean it clearly like was not us.

16
00:00:48,479 --> 00:00:50,159
Speaker 2: How much did it worry you, Joe?

17
00:00:50,840 --> 00:00:53,319
Speaker 1: I think I'm more of an AI domer worrier than

18
00:00:53,360 --> 00:00:53,639
Speaker 1: you are.

19
00:00:53,960 --> 00:00:54,520
Speaker 3: Interesting.

20
00:00:54,880 --> 00:00:57,840
Speaker 2: I mean, I have concerns, don't get me wrong, But

21
00:00:57,880 --> 00:00:59,520
Speaker 2: what do you think is the worst case scenario? Is

22
00:00:59,520 --> 00:01:00,880
Speaker 2: it robot taking over the world.

23
00:01:01,000 --> 00:01:01,120
Speaker 4: No.

24
00:01:01,200 --> 00:01:04,080
Speaker 1: I think that like in a couple of years, the

25
00:01:04,160 --> 00:01:06,360
Speaker 1: AI will do a really good job of making the

26
00:01:06,360 --> 00:01:09,840
Speaker 1: Odd Lots podcast. And people say, I don't really need

27
00:01:09,880 --> 00:01:11,960
Speaker 1: to listen to Joe and Tracy anymore. The AI is

28
00:01:12,120 --> 00:01:14,920
Speaker 1: I don't know. I'm like, you know, I but I

29
00:01:14,959 --> 00:01:16,400
Speaker 1: feel like maybe that's all a parandoise.

30
00:01:16,520 --> 00:01:19,240
Speaker 2: No, I feel like there is reason for concern. I hope.

31
00:01:19,360 --> 00:01:21,319
Speaker 2: Maybe I'm just being an optimist here. I hope we

32
00:01:21,360 --> 00:01:24,200
Speaker 2: are a few years away from the odd Blots replicat.

33
00:01:24,240 --> 00:01:25,720
Speaker 1: I know what, I want to do this for like

34
00:01:25,760 --> 00:01:27,480
Speaker 1: another thirty years, and I don't want to just have

35
00:01:27,600 --> 00:01:29,480
Speaker 1: like three or four years or a few years.

36
00:01:29,480 --> 00:01:32,080
Speaker 2: That's the optimistic scenario. Is it doing this in thirty

37
00:01:32,160 --> 00:01:32,560
Speaker 2: years time?

38
00:01:32,640 --> 00:01:33,040
Speaker 3: Okay?

39
00:01:33,240 --> 00:01:33,600
Speaker 4: All right?

40
00:01:34,360 --> 00:01:37,720
Speaker 2: AI. Lots of people expressing some worries about.

41
00:01:37,520 --> 00:01:38,920
Speaker 1: It, a million different kind of words.

42
00:01:39,000 --> 00:01:42,199
Speaker 2: Yeah, and we see all these, you know, good ways

43
00:01:42,280 --> 00:01:45,040
Speaker 2: that the technology is being used, but also insidious ways.

44
00:01:45,160 --> 00:01:47,760
Speaker 2: As you pointed out, someone made an Audlots episode. I

45
00:01:47,800 --> 00:01:51,720
Speaker 2: think that was in a in a very fun spirited way,

46
00:01:51,800 --> 00:01:54,520
Speaker 2: a nice way. But there are people who use this

47
00:01:54,560 --> 00:01:57,480
Speaker 2: technology for more nefarious things. Yeah.

48
00:01:57,480 --> 00:01:59,880
Speaker 1: I think there was like a sixty minutes recently where

49
00:01:59,880 --> 00:02:03,640
Speaker 1: like someone was able to like impersonate someone's voice and say, hey,

50
00:02:04,320 --> 00:02:07,360
Speaker 1: you know, you know, you could imagine this scam where

51
00:02:07,400 --> 00:02:09,520
Speaker 1: someone's like kids, say hey, I need money or something

52
00:02:09,560 --> 00:02:13,079
Speaker 1: like that, and they call and it sounds just like them,

53
00:02:13,160 --> 00:02:14,760
Speaker 1: especially in a couple of years when the tech gets

54
00:02:14,800 --> 00:02:17,640
Speaker 1: better and it becomes a bit of a crisis in

55
00:02:17,680 --> 00:02:19,760
Speaker 1: the sense it's like, do I know that a person

56
00:02:19,840 --> 00:02:21,959
Speaker 1: is a person when they talk to me on a zoom,

57
00:02:21,960 --> 00:02:23,480
Speaker 1: when they talk to me on a podcast, when they

58
00:02:23,480 --> 00:02:24,880
Speaker 1: talk to me on a phone call, et cetera.

59
00:02:25,080 --> 00:02:27,760
Speaker 2: Well, even on Twitter on a much more basic level,

60
00:02:27,800 --> 00:02:29,760
Speaker 2: you have all these bots, and I don't know about you,

61
00:02:29,880 --> 00:02:32,200
Speaker 2: but there seems to be a new Tracy Alloway with

62
00:02:32,320 --> 00:02:36,600
Speaker 2: the l as like an I and fooling everyone almost

63
00:02:36,680 --> 00:02:38,520
Speaker 2: like every few weeks, and then you have to go

64
00:02:38,560 --> 00:02:41,800
Speaker 2: through the whole bureaucracy of reporting to Twitter and proving

65
00:02:41,880 --> 00:02:44,720
Speaker 2: that you are in fact the original Tracy alloyway.

66
00:02:44,880 --> 00:02:47,359
Speaker 1: And not to get into much of a tangent, but

67
00:02:47,440 --> 00:02:49,600
Speaker 1: wasn't Elon going to solve this because he said like, oh,

68
00:02:49,600 --> 00:02:52,200
Speaker 1: I'm buying Twitter to solve the bots problem, and I don't,

69
00:02:52,480 --> 00:02:54,120
Speaker 1: as you say, it's not been solved.

70
00:02:54,160 --> 00:02:55,600
Speaker 2: I think that's a whole other episode.

71
00:02:55,800 --> 00:03:00,320
Speaker 1: But even with like really rudimentary like human you know,

72
00:03:00,560 --> 00:03:03,520
Speaker 1: fake traces and fake Joe's like, that's the most rudimentary thing,

73
00:03:03,600 --> 00:03:06,000
Speaker 1: just like taking our avatar and like changing like one

74
00:03:06,000 --> 00:03:07,840
Speaker 1: of the l's and trade in L a way to

75
00:03:07,880 --> 00:03:10,760
Speaker 1: like an I or whatever or one like. That's just

76
00:03:10,800 --> 00:03:11,480
Speaker 1: gonna get worse.

77
00:03:13,000 --> 00:03:16,280
Speaker 2: You know what's a really good business model creating a

78
00:03:16,400 --> 00:03:19,800
Speaker 2: problem and then creating the solution to it.

79
00:03:21,240 --> 00:03:23,880
Speaker 1: Yes, that's exactly right. So you have all So we

80
00:03:23,919 --> 00:03:27,440
Speaker 1: have all of these ais, and they're going to like

81
00:03:27,480 --> 00:03:29,120
Speaker 1: replicate us, They're going to look like us, or they're

82
00:03:29,120 --> 00:03:32,040
Speaker 1: going to talk like us. But the good news, in

83
00:03:32,080 --> 00:03:34,280
Speaker 1: a sense good is that others are working out the

84
00:03:34,360 --> 00:03:37,400
Speaker 1: problem of how do you prove that you are who

85
00:03:37,440 --> 00:03:41,400
Speaker 1: you are in a world in which AIS can replicate

86
00:03:41,400 --> 00:03:43,480
Speaker 1: our voices and our faces and our expression.

87
00:03:43,640 --> 00:03:46,320
Speaker 2: Technology can save us from technology?

88
00:03:46,520 --> 00:03:48,440
Speaker 1: Is the thrust of this, right exactly. So we are

89
00:03:48,480 --> 00:03:52,800
Speaker 1: going to be speaking to someone working on the technology

90
00:03:52,840 --> 00:03:57,280
Speaker 1: designed to address this concern in part. And as you say,

91
00:03:57,360 --> 00:03:59,400
Speaker 1: some of the people working on it are some of

92
00:03:59,440 --> 00:04:02,920
Speaker 1: the same peopleeople behind some of the most cutting edge

93
00:04:03,000 --> 00:04:05,880
Speaker 1: AI research that's sort of like blowing everyone's minds in

94
00:04:05,960 --> 00:04:07,680
Speaker 1: both good and scary ways these days.

95
00:04:07,720 --> 00:04:09,520
Speaker 2: So I'm just going to go ahead and lay out

96
00:04:09,560 --> 00:04:12,440
Speaker 2: my priors here, which are that I find this sort

97
00:04:12,440 --> 00:04:16,920
Speaker 2: of instinctively dystopian the topic of this podcast what we're

98
00:04:16,960 --> 00:04:20,200
Speaker 2: going to be discussing. This technology creeps me out a

99
00:04:20,240 --> 00:04:23,080
Speaker 2: little bit. But that said, I don't know that much

100
00:04:23,080 --> 00:04:25,039
Speaker 2: about it. I don't know much about how it works,

101
00:04:25,040 --> 00:04:28,039
Speaker 2: and so I am very interested to hear the bowl

102
00:04:28,120 --> 00:04:30,040
Speaker 2: case on eyeball scanning.

103
00:04:30,560 --> 00:04:32,440
Speaker 1: So you just jump bright. We're going to be talking

104
00:04:32,680 --> 00:04:36,080
Speaker 1: eyeball scanning. We are going to be speaking with Alex Blania.

105
00:04:36,400 --> 00:04:40,159
Speaker 1: He is the co founder of world Coin, an entity

106
00:04:40,200 --> 00:04:43,680
Speaker 1: that was set up also by Sam Altman, who everybody

107
00:04:43,800 --> 00:04:46,680
Speaker 1: a name that everybody knows now because he's the co

108
00:04:46,800 --> 00:04:49,920
Speaker 1: founder or the founder of open ai, which powers chad

109
00:04:49,960 --> 00:04:53,560
Speaker 1: GPT and other tools. Other co founder Max nor Frenstern.

110
00:04:53,800 --> 00:04:56,719
Speaker 1: We're going to be talking about how an eyeball scanner

111
00:04:57,160 --> 00:05:00,760
Speaker 1: can allow us to theoretically prove our human in this

112
00:05:00,920 --> 00:05:04,240
Speaker 1: future AI world. So Alex, thank you so much for

113
00:05:04,600 --> 00:05:05,880
Speaker 1: coming on odd Lots.

114
00:05:06,400 --> 00:05:08,080
Speaker 4: Thank you for having me. I'm excited to talk.

115
00:05:08,880 --> 00:05:12,640
Speaker 1: So I'm trying to think, I mean, like the threat,

116
00:05:13,240 --> 00:05:15,520
Speaker 1: how real is it that you know, in a few

117
00:05:15,600 --> 00:05:21,120
Speaker 1: years do you think not whether AI will be able

118
00:05:21,120 --> 00:05:24,640
Speaker 1: to create a perfect replica of the podcast, but that

119
00:05:25,400 --> 00:05:28,880
Speaker 1: let's say we're having this conversation and I might really

120
00:05:28,920 --> 00:05:32,880
Speaker 1: not know that I am talking to the real Alex

121
00:05:32,960 --> 00:05:37,719
Speaker 1: Blania versus an AI that could perfectly replicate your voice

122
00:05:37,720 --> 00:05:40,400
Speaker 1: and replicate your face. Like is that two years away?

123
00:05:40,440 --> 00:05:43,080
Speaker 1: Is that three years away? Like? How soon is that coming?

124
00:05:44,640 --> 00:05:49,479
Speaker 4: Well? I think partially we are already there. Right the.

125
00:05:51,080 --> 00:05:53,520
Speaker 5: Technology is not completely open source yet, so you don't

126
00:05:53,560 --> 00:05:58,159
Speaker 5: have widely deployed, deployed systems that can't do that, but

127
00:05:58,279 --> 00:05:59,800
Speaker 5: rather you have right now you have like a few

128
00:06:00,080 --> 00:06:02,960
Speaker 5: teas that actually can do things like that or would

129
00:06:03,000 --> 00:06:05,279
Speaker 5: be able to do things like that. But this is

130
00:06:05,320 --> 00:06:08,279
Speaker 5: going to change because of course the progress is increasing.

131
00:06:08,360 --> 00:06:12,440
Speaker 5: So you had it started with Dolly and you had

132
00:06:12,480 --> 00:06:15,440
Speaker 5: stable diffusion and many other things that on the image

133
00:06:15,440 --> 00:06:20,200
Speaker 5: generation side became increasingly impressive. And I think are ai

134
00:06:20,360 --> 00:06:21,880
Speaker 5: R is one of the coolest things of the of

135
00:06:21,920 --> 00:06:25,360
Speaker 5: the last couple of the last year, to be honest.

136
00:06:26,080 --> 00:06:28,360
Speaker 5: So that's on the image generation side of things, and

137
00:06:28,400 --> 00:06:30,080
Speaker 5: then of course you have chat GIPT on the text

138
00:06:30,080 --> 00:06:32,320
Speaker 5: generation side of things. So I think on Twitter, as

139
00:06:32,360 --> 00:06:36,920
Speaker 5: you discussed in the introduction, you are I think already

140
00:06:36,960 --> 00:06:39,720
Speaker 5: there that it could happen that you just chat with

141
00:06:39,760 --> 00:06:41,920
Speaker 5: a neural network and you have absolute no idea that

142
00:06:41,960 --> 00:06:44,200
Speaker 5: you chat with a new network and you might argue

143
00:06:44,240 --> 00:06:47,880
Speaker 5: about odd lots for like an hour and don't realize that.

144
00:06:48,120 --> 00:06:51,520
Speaker 2: I'm just thinking, like half the people I argue with

145
00:06:51,560 --> 00:06:52,839
Speaker 2: on Twitter probably are.

146
00:06:53,040 --> 00:06:54,280
Speaker 1: Bots neural networks.

147
00:06:54,480 --> 00:06:58,039
Speaker 2: Yeah, well, maybe just to step back in time a

148
00:06:58,040 --> 00:07:00,280
Speaker 2: little bit, I mean, world coin has been around for

149
00:07:00,400 --> 00:07:03,760
Speaker 2: a few years now, and to some degree, as Joe mentioned,

150
00:07:03,839 --> 00:07:07,599
Speaker 2: you guys kind of anticipated this threat. What was the

151
00:07:07,760 --> 00:07:12,240
Speaker 2: sort of aha moment that made this all come together, Like,

152
00:07:12,480 --> 00:07:16,240
Speaker 2: how did this actually develop into a company, and what

153
00:07:16,480 --> 00:07:20,200
Speaker 2: was the use case thesis when you were just setting.

154
00:07:19,960 --> 00:07:24,440
Speaker 5: Out, Well, we honestly did not. We didn't start off

155
00:07:24,480 --> 00:07:28,800
Speaker 5: to solve a threat, all right, that is not brief

156
00:07:28,840 --> 00:07:32,080
Speaker 5: history in time. Sure, back then, I did research.

157
00:07:32,720 --> 00:07:34,600
Speaker 4: So we started the company a little bit over three

158
00:07:34,680 --> 00:07:35,200
Speaker 4: years ago.

159
00:07:35,800 --> 00:07:40,160
Speaker 5: And I was doing research in deep learning applied to

160
00:07:40,240 --> 00:07:44,520
Speaker 5: quantum physics. So I basically predicted quantum system of neural

161
00:07:44,600 --> 00:07:47,440
Speaker 5: networks and that's I think it's a very promising field

162
00:07:47,440 --> 00:07:50,920
Speaker 5: in physics. And Sam was already working on the idea

163
00:07:51,640 --> 00:07:54,120
Speaker 5: overall kind together with Max, so I think they've already

164
00:07:54,160 --> 00:07:55,960
Speaker 5: been on it for like six months. But of course

165
00:07:56,280 --> 00:07:58,480
Speaker 5: Sam had a full time job at opening Ie. Max

166
00:07:58,520 --> 00:08:00,480
Speaker 5: had a full time job back then, so it was

167
00:08:00,520 --> 00:08:03,880
Speaker 5: all like in an early idea face. Then I got

168
00:08:03,880 --> 00:08:09,240
Speaker 5: an email from from Max back then to Poop with

169
00:08:09,320 --> 00:08:12,800
Speaker 5: like an early paper describing the vision of world cooin

170
00:08:13,680 --> 00:08:15,520
Speaker 5: and that it should come by and talk. And then

171
00:08:15,960 --> 00:08:18,520
Speaker 5: at a couple of interree years with Max uh one

172
00:08:18,560 --> 00:08:20,160
Speaker 5: with Sam, we spent a lot of time together and

173
00:08:20,560 --> 00:08:23,440
Speaker 5: actually decided to.

174
00:08:22,560 --> 00:08:23,280
Speaker 4: Proceed, but.

175
00:08:25,200 --> 00:08:27,440
Speaker 5: To go back to the initial proposition, and I think

176
00:08:27,440 --> 00:08:30,640
Speaker 5: that is very important. It's not about solving a threat,

177
00:08:30,680 --> 00:08:34,839
Speaker 5: but rather how I mean, Sam was already breaking in

178
00:08:34,880 --> 00:08:37,240
Speaker 5: open EIE and this was three years ago, so back

179
00:08:37,280 --> 00:08:41,520
Speaker 5: then it was not yet an accepted topic that AI

180
00:08:41,679 --> 00:08:43,240
Speaker 5: is going to happen and it's going to be as

181
00:08:43,240 --> 00:08:44,280
Speaker 5: powerful as today.

182
00:08:44,320 --> 00:08:45,280
Speaker 4: It was like a very.

183
00:08:46,040 --> 00:08:50,120
Speaker 5: Niche topic I think, although open of course was already

184
00:08:50,200 --> 00:08:53,760
Speaker 5: quite famous entity. But of course Sam was motivated by

185
00:08:53,760 --> 00:08:56,160
Speaker 5: the idea of AI and what it's going to do

186
00:08:56,200 --> 00:08:57,320
Speaker 5: to the world and it's going to be this very

187
00:08:57,320 --> 00:09:01,839
Speaker 5: powerful technology. And so the initial proposition of Rocline was

188
00:09:01,880 --> 00:09:06,600
Speaker 5: really what is the societal infrastructure that we can build

189
00:09:06,600 --> 00:09:10,280
Speaker 5: that makes sure that AI is for the better of

190
00:09:10,400 --> 00:09:13,559
Speaker 5: all of society versus a few and so there's a

191
00:09:13,679 --> 00:09:15,839
Speaker 5: there's a many things actually that come together here, like

192
00:09:15,880 --> 00:09:19,480
Speaker 5: proof of personhood is what you just described, is the

193
00:09:19,520 --> 00:09:22,040
Speaker 5: ability to prove that you're actually human being on the Internet,

194
00:09:22,040 --> 00:09:25,679
Speaker 5: which is basically just a much more capable identity system.

195
00:09:25,720 --> 00:09:27,360
Speaker 4: But it goes as far as.

196
00:09:28,880 --> 00:09:32,200
Speaker 5: We think eventually UBI will become a thing, and it

197
00:09:32,240 --> 00:09:34,240
Speaker 5: actually is like a big chance if you have like

198
00:09:34,240 --> 00:09:36,640
Speaker 5: a very very powerful technology and it's not about kind

199
00:09:36,640 --> 00:09:39,560
Speaker 5: of political redistribution, but you have a very centralizing technology

200
00:09:40,200 --> 00:09:42,720
Speaker 5: that you want to share the upside with the world with.

201
00:09:43,760 --> 00:09:46,800
Speaker 5: And there was another motivation, kind of Rokin can actually

202
00:09:46,800 --> 00:09:48,920
Speaker 5: be the infrastructure to do that, because you have a

203
00:09:49,000 --> 00:09:51,760
Speaker 5: very capable identity system, you have an economic system that

204
00:09:51,800 --> 00:09:55,960
Speaker 5: connects everyone. So there were like a couple of very

205
00:09:55,960 --> 00:09:59,360
Speaker 5: ambitious reasons why we think something like that. So basically

206
00:09:59,360 --> 00:10:05,880
Speaker 5: creating and bootstrapping the largest identity and financial network that

207
00:10:06,040 --> 00:10:12,160
Speaker 5: is actually open and privacy preserving versus relying on centralized

208
00:10:12,679 --> 00:10:15,559
Speaker 5: government systems that might not work in large parts of

209
00:10:15,600 --> 00:10:19,760
Speaker 5: the world is a very important preposition for the coming decades.

210
00:10:20,160 --> 00:10:23,800
Speaker 1: So let's just jump right to it. If you're listening

211
00:10:23,920 --> 00:10:26,360
Speaker 1: people who are listening to this podcast, you might want

212
00:10:26,400 --> 00:10:29,040
Speaker 1: to actually watch the video version. We're recording this on

213
00:10:29,200 --> 00:10:32,880
Speaker 1: video as well. People just the audio butt here on

214
00:10:32,960 --> 00:10:35,719
Speaker 1: the set. I'm just gonna I was gonna wait a

215
00:10:35,760 --> 00:10:38,640
Speaker 1: little while before unveiling it, but this is I'm holding

216
00:10:38,640 --> 00:10:38,960
Speaker 1: in my.

217
00:10:38,880 --> 00:10:41,199
Speaker 2: Hand like we need extra music, a big.

218
00:10:41,160 --> 00:10:44,440
Speaker 1: Orb It feels about six pounds. I don't know how

219
00:10:44,760 --> 00:10:47,360
Speaker 1: heavy is this, Alex, I'm sure you know the exact weight.

220
00:10:48,360 --> 00:10:50,640
Speaker 4: It's a little bit over one kilogram.

221
00:10:51,120 --> 00:10:53,400
Speaker 1: Oh okay, so I was wrong on So that's like

222
00:10:53,480 --> 00:10:58,440
Speaker 1: just on like about two and a half pounds, Okay, okay,

223
00:10:58,520 --> 00:11:00,840
Speaker 1: that's fine, about two and have found and it's an

224
00:11:00,880 --> 00:11:03,400
Speaker 1: eyeball scanner, and so here it is, and it's not

225
00:11:03,440 --> 00:11:06,240
Speaker 1: working right now, or we're not scanning right now, so

226
00:11:06,280 --> 00:11:09,160
Speaker 1: I can like tilt this in our face. It's not scanning.

227
00:11:09,360 --> 00:11:12,040
Speaker 1: But let's like talk about what this piece of hardware

228
00:11:12,160 --> 00:11:13,920
Speaker 1: that I'm actually holding it is that I'm going to

229
00:11:13,960 --> 00:11:16,800
Speaker 1: set it on the table in front of me. There's

230
00:11:16,800 --> 00:11:18,120
Speaker 1: an eyeball scanner. What does it do?

231
00:11:19,240 --> 00:11:23,000
Speaker 5: So the show to answer is one it will start

232
00:11:23,040 --> 00:11:25,200
Speaker 5: with shortened than the longer answer. The short answer is

233
00:11:25,240 --> 00:11:29,199
Speaker 5: that it first ensures that whatever is interacting with the device,

234
00:11:29,240 --> 00:11:32,400
Speaker 5: is an actual human being, So okay, not a display,

235
00:11:33,320 --> 00:11:36,480
Speaker 5: not a photograph of an eyeball, not a photograph or

236
00:11:37,080 --> 00:11:40,160
Speaker 5: like more sophisticated optical attacks, because there's like many.

237
00:11:39,960 --> 00:11:41,040
Speaker 4: Things that you can do here.

238
00:11:41,720 --> 00:11:44,000
Speaker 5: So that's the first step checking that whatever we see

239
00:11:44,000 --> 00:11:47,040
Speaker 5: is an actual human being. And then the second piece

240
00:11:47,800 --> 00:11:52,120
Speaker 5: is it takes an image of my eye, calculates a

241
00:11:52,240 --> 00:11:57,040
Speaker 5: unique embedding out of this picture. Everything happens on a device,

242
00:11:57,920 --> 00:12:01,720
Speaker 5: and that embedding then signed by the device, and that's

243
00:12:01,840 --> 00:12:05,040
Speaker 5: then the only thing that actually leaves the device and

244
00:12:06,320 --> 00:12:10,240
Speaker 5: gets compared against all other users. And if that uniqueness

245
00:12:10,320 --> 00:12:14,480
Speaker 5: check is successful, then me that I just verified with

246
00:12:14,480 --> 00:12:17,720
Speaker 5: an ORB I then can later do a zero knowledge

247
00:12:17,760 --> 00:12:21,880
Speaker 5: proof that I am included in that set. And if

248
00:12:21,920 --> 00:12:23,880
Speaker 5: you're not familiar with your knowledge proves, what it allows

249
00:12:23,920 --> 00:12:27,160
Speaker 5: you to do is it lets you prove something without

250
00:12:27,200 --> 00:12:32,520
Speaker 5: actually revealing the underlying information. So what that enables us

251
00:12:32,600 --> 00:12:36,560
Speaker 5: to do is it builds a self like essentially a

252
00:12:36,600 --> 00:12:41,040
Speaker 5: self custodial identity network where the data control is with

253
00:12:41,080 --> 00:12:44,080
Speaker 5: the user and there's no centralized entity that actually has

254
00:12:44,200 --> 00:12:47,000
Speaker 5: the information of who you are, where you're from. All

255
00:12:47,000 --> 00:12:50,199
Speaker 5: of these things but rather this person is a unique

256
00:12:50,280 --> 00:12:55,160
Speaker 5: human being and has verified revocan before, while the user

257
00:12:55,520 --> 00:12:58,320
Speaker 5: could decide to attach more information to that. So that's

258
00:12:58,360 --> 00:13:00,600
Speaker 5: kind of the short of it. I think we should

259
00:13:00,640 --> 00:13:02,560
Speaker 5: talk much more about the why and for a second,

260
00:13:02,559 --> 00:13:03,960
Speaker 5: because that's not trivial.

261
00:13:04,000 --> 00:13:05,079
Speaker 4: But yeah, that's it.

262
00:13:05,520 --> 00:13:08,319
Speaker 2: Yeah, maybe just to help us get to the why,

263
00:13:08,440 --> 00:13:11,480
Speaker 2: you could talk about what would be a practical use

264
00:13:11,559 --> 00:13:14,920
Speaker 2: case for this technology. Because if I think about something

265
00:13:15,040 --> 00:13:18,600
Speaker 2: like you know, you mentioned UBI universal Basic income. If

266
00:13:18,640 --> 00:13:21,880
Speaker 2: I think a government starts some sort of UBI program,

267
00:13:21,920 --> 00:13:25,600
Speaker 2: then I would assume that they identify people by something

268
00:13:25,679 --> 00:13:28,760
Speaker 2: like a social Security number or whatever the equivalent might

269
00:13:28,800 --> 00:13:32,160
Speaker 2: be in that country. So what does this do that

270
00:13:32,240 --> 00:13:35,120
Speaker 2: those traditional systems of identity.

271
00:13:34,679 --> 00:13:36,760
Speaker 4: Do not well?

272
00:13:36,800 --> 00:13:40,640
Speaker 5: So when you start with traditional identity, the really hard

273
00:13:40,679 --> 00:13:44,679
Speaker 5: thing for governments is actually making sure everyone only has

274
00:13:44,720 --> 00:13:48,960
Speaker 5: one password, and that is with an identity is called deduplication.

275
00:13:49,640 --> 00:13:53,160
Speaker 5: It's a really hard problem because all the systems, all

276
00:13:53,160 --> 00:13:55,800
Speaker 5: the technologies you use usually in your life, let's say

277
00:13:55,840 --> 00:13:59,360
Speaker 5: your iPhone Phase ID. What that actually does is just

278
00:13:59,520 --> 00:14:04,080
Speaker 5: it re authenticates you, so it realizes I'm the same

279
00:14:04,120 --> 00:14:06,240
Speaker 5: person again logging into that phone. And that's a one

280
00:14:06,240 --> 00:14:09,079
Speaker 5: to one comparison, and that is fairly easy to do.

281
00:14:10,160 --> 00:14:12,200
Speaker 5: What is really hard to do is making sure okay,

282
00:14:12,320 --> 00:14:15,360
Speaker 5: Alex did not sign up compared to a billion other people.

283
00:14:16,280 --> 00:14:20,360
Speaker 5: And so that's why many governments actually have biometric systems

284
00:14:20,360 --> 00:14:24,320
Speaker 5: in place, because otherwise also their Social Security number system

285
00:14:24,320 --> 00:14:27,880
Speaker 5: would completely break. But that's not rolled out globally. So

286
00:14:27,960 --> 00:14:31,920
Speaker 5: in fact, a good mental model is about fifty percent

287
00:14:31,920 --> 00:14:35,560
Speaker 5: of the global population actually have a digitally verifiable identity.

288
00:14:36,120 --> 00:14:37,920
Speaker 4: But that's a huge number of people.

289
00:14:37,680 --> 00:14:40,400
Speaker 5: That just don't have it right, And so I think

290
00:14:40,600 --> 00:14:45,840
Speaker 5: United States is pretty good, or at least decent. It's

291
00:14:45,880 --> 00:14:50,120
Speaker 5: actually not a really good at any solution. Europe in

292
00:14:50,240 --> 00:14:56,880
Speaker 5: parts is pretty capable. The strongest is India, although there's

293
00:14:57,000 --> 00:15:00,240
Speaker 5: definitely privacy concerns here because they don't use you knowledge

294
00:15:00,280 --> 00:15:01,120
Speaker 5: proofs and things like that.

295
00:15:01,160 --> 00:15:02,800
Speaker 4: But for example, India.

296
00:15:03,000 --> 00:15:06,080
Speaker 5: They have a program called at HOR that also used

297
00:15:06,880 --> 00:15:09,680
Speaker 5: ours recognition to kind of roll out to the global

298
00:15:09,720 --> 00:15:12,160
Speaker 5: population and it really accelerated the whole economy like crazy.

299
00:15:12,280 --> 00:15:14,600
Speaker 5: So they had at HOR and then they build UPI

300
00:15:14,720 --> 00:15:17,000
Speaker 5: on Top, which is a payment rail, and so whenever

301
00:15:17,040 --> 00:15:18,960
Speaker 5: you talk to people that are in that kind of

302
00:15:19,560 --> 00:15:22,480
Speaker 5: area of technology, everyone always brings up India as the

303
00:15:22,520 --> 00:15:26,480
Speaker 5: most prominent example of like what identity plus a strong

304
00:15:26,520 --> 00:15:29,600
Speaker 5: financial rail can actually do to accelerate an economy.

305
00:15:30,320 --> 00:15:32,160
Speaker 4: So, and this is what roll coin is doing. Basically,

306
00:15:32,240 --> 00:15:32,800
Speaker 4: roll coin.

307
00:15:32,760 --> 00:15:38,400
Speaker 5: Is is able to create a d duplication set, so

308
00:15:38,720 --> 00:15:40,960
Speaker 5: it provides me with the ability to prove that I'm

309
00:15:41,000 --> 00:15:44,640
Speaker 5: actually unique against everyone else. But then also there's a

310
00:15:44,640 --> 00:15:47,160
Speaker 5: world ID, which is an identity protocol on top of that.

311
00:15:48,040 --> 00:15:49,840
Speaker 4: And let's start with the basics.

312
00:15:49,920 --> 00:15:51,640
Speaker 5: So the first thing I could do is I could

313
00:15:51,800 --> 00:15:54,840
Speaker 5: log into Twitter and I could say I don't already

314
00:15:54,840 --> 00:15:58,640
Speaker 5: have a Twitter account without actually really my name, And

315
00:15:58,680 --> 00:16:00,920
Speaker 5: then a few things a little bit forward. You could

316
00:16:00,920 --> 00:16:04,040
Speaker 5: even cryptographically sign every tweet you make, so that you

317
00:16:04,480 --> 00:16:06,440
Speaker 5: kind of lay a trace of like, okay, this is

318
00:16:06,480 --> 00:16:10,200
Speaker 5: actually Alex versus anyone else. And I think this is

319
00:16:10,240 --> 00:16:14,280
Speaker 5: really where this is going to go, right, because this

320
00:16:14,720 --> 00:16:19,920
Speaker 5: progress around AI is definitely an exponential So we will

321
00:16:19,920 --> 00:16:21,960
Speaker 5: get to a place where you will help You will

322
00:16:21,960 --> 00:16:24,240
Speaker 5: have to authenticate pretty much everything you do on the internet.

323
00:16:25,080 --> 00:16:25,880
Speaker 4: So that's the basic.

324
00:16:25,960 --> 00:16:28,920
Speaker 5: But then with rold eddy, which is a protocol that

325
00:16:29,040 --> 00:16:33,880
Speaker 5: lets you add verifiable credentials and many other things, with

326
00:16:33,920 --> 00:16:36,080
Speaker 5: an identity space, you can actually bootstop a whole identity

327
00:16:36,120 --> 00:16:36,760
Speaker 5: system on top.

328
00:16:37,480 --> 00:16:39,960
Speaker 4: So this was like a quick grant. But no, it's good.

329
00:16:40,520 --> 00:16:44,040
Speaker 2: So it's not that it's proving that you are, you know,

330
00:16:44,120 --> 00:16:46,720
Speaker 2: a Joe Wisenthal or a Tracy Alloway. It's proving that

331
00:16:46,760 --> 00:16:51,040
Speaker 2: there is only one of you without necessarily the ID

332
00:16:51,240 --> 00:16:52,800
Speaker 2: attached or the name attached.

333
00:16:53,720 --> 00:16:55,960
Speaker 4: That is exactly so that that is the I.

334
00:16:56,520 --> 00:16:59,600
Speaker 5: That is the whole point because then but then what

335
00:16:59,680 --> 00:17:04,600
Speaker 5: we're id lets you do so if you're not into crypto,

336
00:17:04,840 --> 00:17:08,720
Speaker 5: but this whole idea of self custody is very important,

337
00:17:08,880 --> 00:17:10,160
Speaker 5: right of course, I'm pretty sure.

338
00:17:10,080 --> 00:17:10,800
Speaker 4: It came up before.

339
00:17:12,560 --> 00:17:15,679
Speaker 5: What that lets you do is you basically you verify

340
00:17:15,720 --> 00:17:19,679
Speaker 5: for well, can you receive your world Eddy? And then

341
00:17:19,720 --> 00:17:22,000
Speaker 5: it might be that for some applications you actually want

342
00:17:22,000 --> 00:17:26,040
Speaker 5: to prove that you're over eighteen or you're Joe viisatal.

343
00:17:26,800 --> 00:17:30,119
Speaker 5: Then the world Eddie protocol, which is literally just like

344
00:17:30,119 --> 00:17:33,000
Speaker 5: an open source protocol and every developer, every company could

345
00:17:33,040 --> 00:17:36,120
Speaker 5: implement with that, You as a user reader can then

346
00:17:36,160 --> 00:17:39,760
Speaker 5: decide to also attach additional information to your world Eddie,

347
00:17:39,800 --> 00:17:42,239
Speaker 5: so that is my name or that is my can

348
00:17:42,320 --> 00:17:45,400
Speaker 5: we see information and then you can decide to reveel

349
00:17:45,480 --> 00:17:47,840
Speaker 5: that to some applications with zon knowledge proofs as well,

350
00:17:48,240 --> 00:17:51,360
Speaker 5: but no one other than you actually has that information.

351
00:17:51,840 --> 00:17:55,960
Speaker 1: So I think, God, I've I have like hundreds of questions.

352
00:17:56,000 --> 00:17:57,800
Speaker 1: There are so many different ways. So you know, one

353
00:17:57,840 --> 00:18:01,399
Speaker 1: basic but one idea. You keep mentioning zero knowledge proven

354
00:18:01,520 --> 00:18:03,399
Speaker 1: just to sort of like the basic idea, And this

355
00:18:03,480 --> 00:18:06,320
Speaker 1: is the shortest. Like you know, I go to websites,

356
00:18:06,359 --> 00:18:09,760
Speaker 1: sometimes government websites, you entering your social security number. But

357
00:18:09,800 --> 00:18:12,560
Speaker 1: the idea is with zero knowledge proof, it would be

358
00:18:13,040 --> 00:18:16,200
Speaker 1: I prove to the government that I know my social

359
00:18:16,200 --> 00:18:19,480
Speaker 1: Security number without telling the government what my social security

360
00:18:19,520 --> 00:18:20,000
Speaker 1: number is.

361
00:18:21,080 --> 00:18:23,160
Speaker 5: Well, I mean that that one is actually tricky because

362
00:18:23,520 --> 00:18:25,720
Speaker 5: in that cause you would reveal the social security number.

363
00:18:25,760 --> 00:18:28,240
Speaker 5: What is a much more what is a much better

364
00:18:28,359 --> 00:18:33,680
Speaker 5: way to explain it is a website asks you are

365
00:18:33,680 --> 00:18:35,639
Speaker 5: you on any kinds of sanctional list?

366
00:18:36,520 --> 00:18:36,720
Speaker 2: Right?

367
00:18:37,200 --> 00:18:37,480
Speaker 4: Right?

368
00:18:38,560 --> 00:18:42,399
Speaker 5: Or are you over eighteen? Or are you actually a

369
00:18:42,440 --> 00:18:43,080
Speaker 5: US citizen?

370
00:18:43,960 --> 00:18:44,240
Speaker 4: Right?

371
00:18:44,359 --> 00:18:47,600
Speaker 5: And then these things you could prove of zon knowledge

372
00:18:47,960 --> 00:18:49,800
Speaker 5: And then basically what comes out of this is like

373
00:18:50,600 --> 00:18:54,639
Speaker 5: approve runs on your phone that you're not included in

374
00:18:54,680 --> 00:18:57,280
Speaker 5: a sanctions list. Let's say and that the only thing

375
00:18:57,320 --> 00:18:59,359
Speaker 5: that comes out and ends up with the website is

376
00:18:59,520 --> 00:19:02,920
Speaker 5: yes or no. But the website does not know what

377
00:19:03,080 --> 00:19:05,399
Speaker 5: is the social security number you check against it.

378
00:19:22,160 --> 00:19:24,719
Speaker 1: So then the other question I have, and the obvious

379
00:19:24,760 --> 00:19:29,160
Speaker 1: thing is look, scanning your eyeball it's gonna make anyone anxious,

380
00:19:29,200 --> 00:19:31,960
Speaker 1: and I'm like totally upward even I'm like, you know,

381
00:19:32,000 --> 00:19:37,040
Speaker 1: a little bit anxious about it. How do you establish

382
00:19:37,680 --> 00:19:41,359
Speaker 1: so there's no unlike say clear at the airport, which

383
00:19:41,359 --> 00:19:43,560
Speaker 1: also uses an eyeball scan so that you can go

384
00:19:43,640 --> 00:19:48,520
Speaker 1: buy security, there's no centralized database of our irises, And

385
00:19:48,560 --> 00:19:50,800
Speaker 1: how do you sort of like establish or prove to

386
00:19:50,880 --> 00:19:53,960
Speaker 1: people that there's no database Because if my understanding, there

387
00:19:54,000 --> 00:19:56,679
Speaker 1: has been reporting that at least at some point in

388
00:19:56,760 --> 00:20:00,359
Speaker 1: the process of the development of world Cli and as

389
00:20:00,359 --> 00:20:02,720
Speaker 1: you've grown as a startup, that at one point or

390
00:20:02,760 --> 00:20:05,400
Speaker 1: at least currently there was a Irish database.

391
00:20:06,200 --> 00:20:08,600
Speaker 4: Okay, so there there's many things here. Yeah, we've got

392
00:20:08,600 --> 00:20:08,960
Speaker 4: to start.

393
00:20:09,000 --> 00:20:12,919
Speaker 5: I'm going to start with explaining how to develop comfort

394
00:20:12,920 --> 00:20:16,960
Speaker 5: around the idea and why why it's necessary. So, first

395
00:20:16,960 --> 00:20:20,399
Speaker 5: of all, I hated the idea of building the door,

396
00:20:20,560 --> 00:20:20,879
Speaker 5: like when we.

397
00:20:20,880 --> 00:20:22,280
Speaker 4: Actually started working a rockin like it.

398
00:20:22,359 --> 00:20:26,399
Speaker 5: It was absolutely horrific because it's We've been four people

399
00:20:26,480 --> 00:20:29,760
Speaker 5: sitting in a smart, small apartment in San Francisco and

400
00:20:30,040 --> 00:20:34,560
Speaker 5: the idea of building hydred devices that take biometric data

401
00:20:34,600 --> 00:20:37,240
Speaker 5: and enroll theomagic globally was a horrific idea back then.

402
00:20:37,280 --> 00:20:41,640
Speaker 5: One because it's very complicated expensive, And then yeah.

403
00:20:41,440 --> 00:20:43,760
Speaker 2: I imagine pitching that one to VCS would be kind

404
00:20:43,760 --> 00:20:44,560
Speaker 2: of tricky.

405
00:20:44,320 --> 00:20:47,240
Speaker 4: Right, Oh, it was, it was. It was horrific.

406
00:20:47,320 --> 00:20:49,159
Speaker 5: It was in the middle of it was also like,

407
00:20:49,320 --> 00:20:52,840
Speaker 5: I mean rock kind is also a crypto project, and

408
00:20:53,440 --> 00:20:56,199
Speaker 5: well I want to ask you. Yeah, we started like

409
00:20:56,240 --> 00:20:59,160
Speaker 5: in the middle of a bear market. So I remember

410
00:20:59,240 --> 00:21:01,280
Speaker 5: back then when we talked the first syvestors, They're like, okay,

411
00:21:01,320 --> 00:21:03,879
Speaker 5: like other than etheroreum or bridcoin, nothing will exist, Like

412
00:21:03,920 --> 00:21:06,760
Speaker 5: all of this nonsense. What do you talk about AI? Also,

413
00:21:06,920 --> 00:21:09,440
Speaker 5: no one cares about that. Also like what are you

414
00:21:09,440 --> 00:21:09,960
Speaker 5: even doing?

415
00:21:10,000 --> 00:21:10,119
Speaker 4: So?

416
00:21:10,600 --> 00:21:16,160
Speaker 5: But again, so neither me or anyone else was really

417
00:21:16,200 --> 00:21:21,600
Speaker 5: excited about building a hyder device. But really we spent

418
00:21:21,760 --> 00:21:25,919
Speaker 5: almost a year doing research in pretty much everything you

419
00:21:25,920 --> 00:21:28,879
Speaker 5: could do to solve the problem, right, and solving the

420
00:21:28,960 --> 00:21:32,639
Speaker 5: problem is okay, issuing a privacy preserving identity that works globally,

421
00:21:33,160 --> 00:21:36,000
Speaker 5: not only in the United States, not only in Europe.

422
00:21:37,200 --> 00:21:41,439
Speaker 5: What do we have to do? And there's basically like

423
00:21:41,520 --> 00:21:45,600
Speaker 5: three big answers to this. One is you use government

424
00:21:45,720 --> 00:21:49,359
Speaker 5: KYC right, which might be a sufficient answer for some

425
00:21:49,359 --> 00:21:51,399
Speaker 5: parts of the world, but it was pretty shocking to

426
00:21:51,440 --> 00:21:55,600
Speaker 5: realize for how few right, So kind of that very

427
00:21:55,680 --> 00:22:01,520
Speaker 5: quickly disqualified. Second is what people call of trust, So

428
00:22:01,720 --> 00:22:03,720
Speaker 5: this whole idea that you just like you build reputation

429
00:22:03,840 --> 00:22:06,560
Speaker 5: between people, and that also, like in Shure, it just

430
00:22:06,600 --> 00:22:07,120
Speaker 5: didn't work.

431
00:22:07,320 --> 00:22:08,800
Speaker 4: And the second and then the third.

432
00:22:08,600 --> 00:22:14,760
Speaker 5: Is biometrics and everything that surrounds that whole field. And

433
00:22:14,880 --> 00:22:19,840
Speaker 5: within biometrics, we then build prototypes for pretty much everything

434
00:22:19,840 --> 00:22:23,800
Speaker 5: you could imagine. We build palm scanners, we build kind

435
00:22:23,800 --> 00:22:27,320
Speaker 5: of a phase ID like implementation of it, we build

436
00:22:27,480 --> 00:22:29,800
Speaker 5: fingerprint scanners, all these things. But the short of it

437
00:22:29,920 --> 00:22:33,680
Speaker 5: is to solve this D duplication problem. So this uniqueness check,

438
00:22:33,960 --> 00:22:38,920
Speaker 5: you need a lot of entropy, so information about each user.

439
00:22:39,320 --> 00:22:43,600
Speaker 5: If you don't have that, your error rate explodes exponentially.

440
00:22:43,920 --> 00:22:47,399
Speaker 5: And so simple example, if you would use a face

441
00:22:47,480 --> 00:22:50,800
Speaker 5: ID to solve the same problem, after tens of millions

442
00:22:50,840 --> 00:22:54,199
Speaker 5: of users, you would have to reject everyone, so you

443
00:22:54,200 --> 00:22:56,200
Speaker 5: don't have a constant errate. You don't have like five

444
00:22:56,240 --> 00:22:59,400
Speaker 5: percent will constantly reject, but you just hit a wall

445
00:22:59,440 --> 00:23:02,240
Speaker 5: and your system breaks, and.

446
00:23:02,320 --> 00:23:03,280
Speaker 4: So face does work.

447
00:23:03,760 --> 00:23:07,000
Speaker 5: Same is true for fingerprints, and the only thing that

448
00:23:07,040 --> 00:23:10,520
Speaker 5: really works is and is proven to work as harvest recognition.

449
00:23:10,640 --> 00:23:11,880
Speaker 4: So that's why we ended up there.

450
00:23:12,280 --> 00:23:14,119
Speaker 5: And then even with an iris recognition, we even had

451
00:23:14,160 --> 00:23:15,760
Speaker 5: to build our own hardware. We had to build our

452
00:23:15,800 --> 00:23:21,080
Speaker 5: own lens, which sounds absolutely ridiculous, but we kind of

453
00:23:21,119 --> 00:23:22,919
Speaker 5: had to build this whole device from the ground up

454
00:23:22,960 --> 00:23:23,760
Speaker 5: to make that happen.

455
00:23:24,280 --> 00:23:26,760
Speaker 4: So much to the why and.

456
00:23:26,880 --> 00:23:29,240
Speaker 5: How do we get people are like, how do we

457
00:23:29,320 --> 00:23:34,600
Speaker 5: think about this whole headline of comfort around this? There's

458
00:23:34,640 --> 00:23:38,720
Speaker 5: a couple of things. One actually built a privacy preserving

459
00:23:38,760 --> 00:23:42,600
Speaker 5: system that goes way further than what you're used to,

460
00:23:42,920 --> 00:23:45,080
Speaker 5: all right, And that's I think really what we did

461
00:23:45,480 --> 00:23:49,359
Speaker 5: because there's no information about you that you actually reveal

462
00:23:49,400 --> 00:23:51,040
Speaker 5: by using world coin, and I think.

463
00:23:50,880 --> 00:23:52,840
Speaker 4: That's pretty remarkable.

464
00:23:52,840 --> 00:23:55,720
Speaker 5: It's definitely countentitive because you're like, okay, we use eyeball

465
00:23:55,720 --> 00:23:59,600
Speaker 5: scanning and yeah, so this is headline one, headline two,

466
00:24:00,480 --> 00:24:05,280
Speaker 5: and that's a journey for sure is open sourcing everything, right,

467
00:24:05,359 --> 00:24:09,080
Speaker 5: So this, I think that's one of the core parts

468
00:24:09,080 --> 00:24:13,919
Speaker 5: of the crypto ethos is just you don't have to trust,

469
00:24:14,040 --> 00:24:18,320
Speaker 5: but rather you can verify and we the hardware is

470
00:24:18,320 --> 00:24:22,159
Speaker 5: completely open source already, the software to meaningful parts, the

471
00:24:22,200 --> 00:24:26,639
Speaker 5: protocol completely and so everything will be open source in

472
00:24:26,680 --> 00:24:31,439
Speaker 5: the coming months. And so that's that's answer two. And

473
00:24:31,480 --> 00:24:32,520
Speaker 5: then three.

474
00:24:33,600 --> 00:24:36,960
Speaker 4: It actually is way less of a thing than you

475
00:24:36,960 --> 00:24:37,440
Speaker 4: would think.

476
00:24:37,600 --> 00:24:40,320
Speaker 5: So we have now one point seven million users, roughly

477
00:24:41,280 --> 00:24:43,480
Speaker 5: six hundred thousand of them that use the app on

478
00:24:43,600 --> 00:24:44,480
Speaker 5: a monthly basis.

479
00:24:45,560 --> 00:24:47,480
Speaker 4: I travel the world very frequently.

480
00:24:47,520 --> 00:24:49,160
Speaker 5: I talk to a lot of users on the ground

481
00:24:49,160 --> 00:24:51,560
Speaker 5: and just like try to understand what's kind of what's

482
00:24:51,640 --> 00:24:53,359
Speaker 5: what's your feeling, what do you what do you understand?

483
00:24:53,400 --> 00:24:54,240
Speaker 4: What you not do understand?

484
00:24:54,280 --> 00:25:00,719
Speaker 5: And it is very much Yeah, like broadly speakingly speaking,

485
00:25:00,840 --> 00:25:04,200
Speaker 5: this is way less of a concern, which obviously does

486
00:25:04,240 --> 00:25:07,080
Speaker 5: not lure our bar but it's the case.

487
00:25:07,960 --> 00:25:10,159
Speaker 2: So I have two things on that, and I think

488
00:25:10,160 --> 00:25:12,560
Speaker 2: they're sort of related. But one why does open source

489
00:25:12,560 --> 00:25:16,879
Speaker 2: seem to be such a thing an AI in particular?

490
00:25:17,040 --> 00:25:22,640
Speaker 2: And then secondly, what is the crypto role in this project?

491
00:25:22,720 --> 00:25:25,520
Speaker 2: Because yeah, you know we mentioned world coin. I can

492
00:25:25,560 --> 00:25:28,520
Speaker 2: only imagine what it was like pitching two vcs in

493
00:25:28,560 --> 00:25:32,159
Speaker 2: the midst of crypto bear market saying we want to

494
00:25:32,240 --> 00:25:35,040
Speaker 2: scan everyone's eyeballs and oh, by the way, there's also

495
00:25:35,119 --> 00:25:37,720
Speaker 2: like a crypto element to this as well, a token.

496
00:25:38,200 --> 00:25:39,760
Speaker 2: Why have that component of it?

497
00:25:41,080 --> 00:25:43,640
Speaker 5: I mean, at this point I'm actually not an AI

498
00:25:43,680 --> 00:25:46,080
Speaker 5: expert anymore because I think I'm basically two years out

499
00:25:46,080 --> 00:25:49,359
Speaker 5: of the field. Went it too crypto, which is an

500
00:25:49,400 --> 00:25:52,440
Speaker 5: interesting journey also, but as.

501
00:25:52,359 --> 00:25:54,720
Speaker 2: Soon you'll be going back into AI probably that's what

502
00:25:54,800 --> 00:25:56,200
Speaker 2: all the crypto people are doing.

503
00:25:56,040 --> 00:25:59,480
Speaker 4: Now right, Well, no.

504
00:26:01,160 --> 00:26:01,440
Speaker 2: Joking.

505
00:26:03,280 --> 00:26:05,520
Speaker 5: So why is open sourcing such a thing? It is

506
00:26:05,600 --> 00:26:10,760
Speaker 5: basically around AI specifically. It is that you have a

507
00:26:10,880 --> 00:26:16,840
Speaker 5: technology that is probably as foundational as electricity, as literally

508
00:26:17,480 --> 00:26:21,600
Speaker 5: like it's it's still I think, incomprehensible to people. What

509
00:26:21,680 --> 00:26:23,600
Speaker 5: a big deal this is, right, This is like one

510
00:26:23,600 --> 00:26:27,240
Speaker 5: of the biggest paradigm shifts in technology that happened ever.

511
00:26:28,080 --> 00:26:31,959
Speaker 5: And there is different approaches of how to deal with

512
00:26:32,000 --> 00:26:36,880
Speaker 5: this and all the underlying risks, and basically two different

513
00:26:37,840 --> 00:26:40,400
Speaker 5: lines of thought is one as the open source everything.

514
00:26:40,920 --> 00:26:43,560
Speaker 5: Everyone should be able to ratify and understand how everything

515
00:26:43,600 --> 00:26:46,680
Speaker 5: is happening, and then the other is, well, you actually

516
00:26:46,720 --> 00:26:48,399
Speaker 5: kind of have to lock down these models because if

517
00:26:48,440 --> 00:26:51,880
Speaker 5: it ends up in the wrong hands, then malicious actors

518
00:26:51,920 --> 00:26:54,720
Speaker 5: have a lot of power. And I think, I mean

519
00:26:54,760 --> 00:26:58,720
Speaker 5: OPENINGI is definitely the kind of in the center of

520
00:26:58,720 --> 00:27:01,760
Speaker 5: discourse because it started with open sourcing and then the

521
00:27:01,760 --> 00:27:06,800
Speaker 5: team realized throughout the years that actually this whole premise

522
00:27:06,840 --> 00:27:11,200
Speaker 5: of open sourcing is probably a terrible idea. So that's

523
00:27:11,200 --> 00:27:14,480
Speaker 5: why it's a discussion discussion with an AI. Within crypto,

524
00:27:15,040 --> 00:27:19,200
Speaker 5: it is just that the whole thesis of crypto is

525
00:27:19,240 --> 00:27:22,280
Speaker 5: that you build actual protocols, not companies, So you build

526
00:27:22,520 --> 00:27:26,560
Speaker 5: systems that are not dependent on a small group of people,

527
00:27:27,440 --> 00:27:32,639
Speaker 5: can be completely verified and can run over decades without

528
00:27:32,680 --> 00:27:36,000
Speaker 5: being interrupted by a living and that's open sources, just

529
00:27:36,040 --> 00:27:40,040
Speaker 5: like the core of this, together with decentralization. So I

530
00:27:40,040 --> 00:27:42,800
Speaker 5: think that's a I actually do think with an AI

531
00:27:42,880 --> 00:27:45,480
Speaker 5: it's it's a very overblown discussion.

532
00:27:46,400 --> 00:27:49,200
Speaker 4: I think within crypto it's the core thesis almost.

533
00:27:50,119 --> 00:27:53,879
Speaker 2: And so the crypto element with world coin, as I

534
00:27:54,000 --> 00:27:57,840
Speaker 2: understand it, is to like insert a degree of I

535
00:27:57,840 --> 00:28:01,359
Speaker 2: guess community control over that technology. Right, Like there's some

536
00:28:01,400 --> 00:28:04,320
Speaker 2: talk about the tokens coming with voting rights and things

537
00:28:04,359 --> 00:28:04,720
Speaker 2: like that.

538
00:28:05,520 --> 00:28:06,560
Speaker 4: There are so many things there.

539
00:28:06,600 --> 00:28:13,679
Speaker 5: Actually, one is I think crypto, and that's something that

540
00:28:13,800 --> 00:28:16,560
Speaker 5: is if you're not kind of deeply in the space,

541
00:28:16,640 --> 00:28:19,640
Speaker 5: that is hard to kind of realize what a big

542
00:28:19,680 --> 00:28:23,560
Speaker 5: deal is. But it is this idea that tokens build

543
00:28:23,640 --> 00:28:27,960
Speaker 5: business models for networks almost right, So you don't know,

544
00:28:28,080 --> 00:28:30,200
Speaker 5: you don't have a company, but rather you have a

545
00:28:31,040 --> 00:28:34,800
Speaker 5: you have a like let's let's let's say ethereum. Etheroreum

546
00:28:34,840 --> 00:28:38,360
Speaker 5: is a very powerful protocol, but it has kind of

547
00:28:38,400 --> 00:28:40,600
Speaker 5: an underlying business model. But it's not that Vitaly gets

548
00:28:40,680 --> 00:28:44,080
Speaker 5: rich by that. It is his fees on the system.

549
00:28:44,360 --> 00:28:47,280
Speaker 5: He well, he did, but because he was early, not

550
00:28:47,400 --> 00:28:49,520
Speaker 5: because now he still makes money. Right, It's not that

551
00:28:49,560 --> 00:28:51,800
Speaker 5: the company is printing money and then Vitaly gets rich

552
00:28:51,840 --> 00:28:54,480
Speaker 5: every month, but rather you have a token and you

553
00:28:54,560 --> 00:28:57,680
Speaker 5: have a fee structure, and as the utility increases, the

554
00:28:57,720 --> 00:29:01,080
Speaker 5: network can basically sustain itself, right, so it can sustain

555
00:29:01,120 --> 00:29:03,960
Speaker 5: its security, it can sustain its operation and things like that,

556
00:29:04,640 --> 00:29:06,680
Speaker 5: and so.

557
00:29:06,880 --> 00:29:10,120
Speaker 4: You basically find a way how you fund a.

558
00:29:10,120 --> 00:29:14,280
Speaker 5: Decentralized operation and can it just let it sustain over

559
00:29:14,320 --> 00:29:18,360
Speaker 5: time and that is one of the big core things

560
00:29:18,400 --> 00:29:22,960
Speaker 5: here is you build very like when rol con works,

561
00:29:23,000 --> 00:29:28,080
Speaker 5: it's really foundational infrastructure. It's I think, without being overconfident,

562
00:29:28,160 --> 00:29:30,160
Speaker 5: I think it's it is a very big deal because

563
00:29:30,160 --> 00:29:33,320
Speaker 5: it's a very foundational piece of technology. It just doesn't

564
00:29:33,320 --> 00:29:36,480
Speaker 5: exist right now. That should not be in the hands

565
00:29:36,520 --> 00:29:40,280
Speaker 5: off a few people, but rather it should be governed

566
00:29:40,320 --> 00:29:42,880
Speaker 5: by a wider group of people, and it should actually

567
00:29:42,920 --> 00:29:46,480
Speaker 5: be decentralized. So also that it cannot break just if

568
00:29:46,480 --> 00:29:48,680
Speaker 5: I'm in a bad mood, Let's say that would be depth.

569
00:29:48,800 --> 00:29:51,160
Speaker 2: That would be so instead of having like a centralized

570
00:29:51,200 --> 00:29:55,560
Speaker 2: actor who controls the ORB, which is a sentence I

571
00:29:55,560 --> 00:29:59,480
Speaker 2: didn't think I would say today you have this decentralized

572
00:29:59,520 --> 00:30:04,640
Speaker 2: group of investors slash stakeholders. Is that how it works?

573
00:30:05,600 --> 00:30:06,880
Speaker 3: Yep, got it.

574
00:30:07,000 --> 00:30:09,120
Speaker 5: This is like the sparks now like a whole other

575
00:30:09,360 --> 00:30:13,440
Speaker 5: part of conversation, but basically how it works is so

576
00:30:13,480 --> 00:30:17,040
Speaker 5: the ORB is an open source heard of advice. Right now,

577
00:30:18,240 --> 00:30:20,680
Speaker 5: the company that I'm a CEO of Tools for Humanity

578
00:30:21,600 --> 00:30:24,320
Speaker 5: is the only ones that produce these orbs. But this

579
00:30:24,360 --> 00:30:27,040
Speaker 5: will change, so you will have, like if you're fast

580
00:30:27,040 --> 00:30:28,800
Speaker 5: forward two three years in the future, you and I

581
00:30:28,840 --> 00:30:31,440
Speaker 5: have many companies that produce their implementation of the ORB

582
00:30:31,560 --> 00:30:34,840
Speaker 5: following the same standards that connect to the work on protocol.

583
00:30:35,680 --> 00:30:38,160
Speaker 4: Right, so you basically.

584
00:30:37,880 --> 00:30:40,560
Speaker 1: Apple could produce an or theoremically.

585
00:30:40,160 --> 00:30:43,280
Speaker 5: Apple, Microsoft, like whatever, they could produce their own ORB,

586
00:30:43,280 --> 00:30:45,760
Speaker 5: follow their own standards, make it, design it the way

587
00:30:45,760 --> 00:30:50,040
Speaker 5: they want, but it will connect to the protocol, and

588
00:30:50,960 --> 00:30:56,400
Speaker 5: that will allow for just a very foundational protocol on

589
00:30:56,440 --> 00:30:59,760
Speaker 5: the Internet that everyone can use that doesn't break. And

590
00:31:00,280 --> 00:31:06,760
Speaker 5: the token distributes the governance and also the incentive makingsmus

591
00:31:06,760 --> 00:31:08,160
Speaker 5: all around the same goal.

592
00:31:24,480 --> 00:31:27,440
Speaker 1: So I mean you talked about the importance of like, okay,

593
00:31:27,440 --> 00:31:29,880
Speaker 1: if this is going to be this foundational infrastructure that

594
00:31:30,200 --> 00:31:34,440
Speaker 1: everyone uses in some way to verify themselves or whatever,

595
00:31:35,040 --> 00:31:37,960
Speaker 1: and also is the sort of rails perhaps for some

596
00:31:38,200 --> 00:31:41,480
Speaker 1: universal basic income that we all need because you know,

597
00:31:41,520 --> 00:31:43,160
Speaker 1: AI is going to put half of us out of

598
00:31:43,240 --> 00:31:46,280
Speaker 1: jobs in theory, like okay, it's better be like not

599
00:31:46,440 --> 00:31:49,880
Speaker 1: just controlled no offense by Alex Blanya and Sam Altman,

600
00:31:49,960 --> 00:31:51,920
Speaker 1: like that'd be kind of weird. But how do we

601
00:31:51,960 --> 00:31:53,720
Speaker 1: know this not the case? And I want to go

602
00:31:53,800 --> 00:31:56,600
Speaker 1: back to a question I asked specifically about like in

603
00:31:56,680 --> 00:31:59,680
Speaker 1: the past, like have you had a IRIS database.

604
00:32:01,000 --> 00:32:04,280
Speaker 5: So how it works is and that's not it's not

605
00:32:04,320 --> 00:32:07,800
Speaker 5: a secret. Basically, when you when you verify with with

606
00:32:07,920 --> 00:32:11,640
Speaker 5: roll coin, go to the flow. So let's say here

607
00:32:11,640 --> 00:32:12,760
Speaker 5: what roll coin, you're excited about it?

608
00:32:12,840 --> 00:32:15,280
Speaker 4: Whatever, you want to sign up? You download an app.

609
00:32:15,680 --> 00:32:17,560
Speaker 5: Right now, there's only one app that lets you do that,

610
00:32:17,600 --> 00:32:20,479
Speaker 5: but soon that also will change. So right now that

611
00:32:20,520 --> 00:32:23,920
Speaker 5: it is called world app. It's a non constudial wallet.

612
00:32:24,600 --> 00:32:28,680
Speaker 5: You click verify now, a map pops up, hopefully we're

613
00:32:28,720 --> 00:32:32,200
Speaker 5: near you. You go to an orb and then you

614
00:32:32,280 --> 00:32:35,840
Speaker 5: show kr open. Thirty seconds later you receive your world

615
00:32:35,920 --> 00:32:38,719
Speaker 5: eddy and you also receive world coin on a rolling basis.

616
00:32:38,800 --> 00:32:41,480
Speaker 5: Every week receive like a small part.

617
00:32:41,760 --> 00:32:47,240
Speaker 4: Of world coin. And in that kind of flow.

618
00:32:47,600 --> 00:32:49,880
Speaker 5: Because we're in beta, we actually did not launch yet

619
00:32:50,120 --> 00:32:52,479
Speaker 5: right The launch is something that's going to happen recentaly soon.

620
00:32:52,520 --> 00:32:54,040
Speaker 5: It's also why we talk right now, why it's a

621
00:32:54,120 --> 00:32:57,960
Speaker 5: very important moment in time for us. There you had

622
00:32:57,960 --> 00:33:00,640
Speaker 5: two options, opt in or opt out. Opt out is

623
00:33:00,680 --> 00:33:04,320
Speaker 5: basically we don't have any data of you. Opt in

624
00:33:04,400 --> 00:33:09,440
Speaker 5: means we basically have custody of your data. For the

625
00:33:09,520 --> 00:33:12,320
Speaker 5: reason that the neural networks that are on this device

626
00:33:12,400 --> 00:33:14,960
Speaker 5: right now are still in development, and so if you

627
00:33:15,440 --> 00:33:19,760
Speaker 5: opt out, it could happen that you have to reverify

628
00:33:19,800 --> 00:33:22,560
Speaker 5: in a year or so, because we basically just updated

629
00:33:22,560 --> 00:33:25,560
Speaker 5: a model as we as we as we went. But

630
00:33:26,040 --> 00:33:29,360
Speaker 5: where this is going is very clearly by and I

631
00:33:29,400 --> 00:33:32,000
Speaker 5: think actually launch, it will be completely opt out.

632
00:33:32,880 --> 00:33:33,480
Speaker 4: This is just.

633
00:33:35,360 --> 00:33:38,560
Speaker 5: That the nature of being in beta and developing technology

634
00:33:38,560 --> 00:33:38,880
Speaker 5: as we go.

635
00:33:39,880 --> 00:33:41,719
Speaker 2: So I take the point that you're in beta at

636
00:33:41,720 --> 00:33:45,760
Speaker 2: the moment. But you know, the arc of recent technological

637
00:33:45,800 --> 00:33:48,719
Speaker 2: innovation seems to be that people often come up with

638
00:33:48,760 --> 00:33:51,560
Speaker 2: something with the best of intentions or in order to

639
00:33:51,680 --> 00:33:56,320
Speaker 2: solve an identified problem, and then it often gets used

640
00:33:56,480 --> 00:34:01,240
Speaker 2: in the worst possible way. So what's like the terrible

641
00:34:01,640 --> 00:34:05,280
Speaker 2: use case arc of this technology? Because one thing that

642
00:34:05,320 --> 00:34:08,040
Speaker 2: springs to mind, and you know, we could talk about

643
00:34:08,120 --> 00:34:11,920
Speaker 2: robot invasions and like terminator style AI takeovers, but like

644
00:34:12,040 --> 00:34:15,920
Speaker 2: one more realistic thing that would spring to mind is, Okay,

645
00:34:16,200 --> 00:34:18,799
Speaker 2: you set something like this up and then as you

646
00:34:18,880 --> 00:34:21,520
Speaker 2: walk through a shopping mall, you know, there's some other

647
00:34:21,600 --> 00:34:25,359
Speaker 2: technology that's scanning your eyeballs, identifying who you are and

648
00:34:25,440 --> 00:34:29,600
Speaker 2: maybe pitching targeted ads or something like that. Like how

649
00:34:29,640 --> 00:34:35,160
Speaker 2: do you separate the identity verification from companies using that

650
00:34:35,760 --> 00:34:37,319
Speaker 2: for other purposes?

651
00:34:38,080 --> 00:34:42,839
Speaker 4: Yep. So we actually have a quite.

652
00:34:44,160 --> 00:34:47,319
Speaker 5: In dem for extreme in the company right now in

653
00:34:47,320 --> 00:34:49,960
Speaker 5: Tools Rehumanity, and we will publish this as we go

654
00:34:50,800 --> 00:34:54,400
Speaker 5: that just like we tried to come of the worst,

655
00:34:54,800 --> 00:34:59,360
Speaker 5: weirdest scenarios of like whatever China's attacking the network with

656
00:34:59,400 --> 00:35:01,640
Speaker 5: billions of the like what could actually happen?

657
00:35:01,680 --> 00:35:02,719
Speaker 4: What could go wrong? Right?

658
00:35:03,160 --> 00:35:07,799
Speaker 5: So the thing is given seer knowledge proofs separate your

659
00:35:07,840 --> 00:35:13,200
Speaker 5: walid from the kind of the uniqueness check the only

660
00:35:13,280 --> 00:35:16,680
Speaker 5: thing that could happen. And to be clear, in some

661
00:35:16,760 --> 00:35:19,040
Speaker 5: cases that actually, like if you think ten years in

662
00:35:19,080 --> 00:35:21,000
Speaker 5: the future and you think, like rollkindness is very powerful

663
00:35:21,040 --> 00:35:26,480
Speaker 5: technology that could actually be bad is people could say, Okay,

664
00:35:26,560 --> 00:35:29,640
Speaker 5: that individual is verified with roll coin, yes or no, right,

665
00:35:29,719 --> 00:35:32,240
Speaker 5: and that might like you could definitely think of scenarios

666
00:35:32,280 --> 00:35:36,040
Speaker 5: of like, okay, China is banning roll coin because it's

667
00:35:36,200 --> 00:35:39,399
Speaker 5: literally the anti thesis of what they're building with their

668
00:35:39,440 --> 00:35:40,040
Speaker 5: in anity.

669
00:35:39,800 --> 00:35:43,200
Speaker 4: System, because it's privacy preserving that.

670
00:35:43,280 --> 00:35:46,279
Speaker 5: You might actually have it might be bad for you

671
00:35:46,360 --> 00:35:48,759
Speaker 5: to be verified with roll coin or whatever. Right, So

672
00:35:48,800 --> 00:35:50,640
Speaker 5: there's like there's some set scenarios here. But I think

673
00:35:50,680 --> 00:35:53,160
Speaker 5: that's that is pretty much the worst case scenario you're

674
00:35:53,160 --> 00:35:57,399
Speaker 5: looking at, because otherwise the only like, let's go even

675
00:35:57,400 --> 00:35:59,840
Speaker 5: to the far extreme, which to be very clear, is

676
00:35:59,880 --> 00:36:00,880
Speaker 5: not what is the case.

677
00:36:01,400 --> 00:36:05,000
Speaker 4: Let's say there would be like a huge database.

678
00:36:04,600 --> 00:36:08,360
Speaker 5: Of of of biometric images of all users.

679
00:36:09,120 --> 00:36:12,440
Speaker 4: It's not connected to your actual account, so.

680
00:36:12,840 --> 00:36:17,440
Speaker 5: You you still have the most there's like multiple players,

681
00:36:17,880 --> 00:36:19,120
Speaker 5: layers of privacy in between.

682
00:36:19,160 --> 00:36:21,160
Speaker 4: But did make sure that nothing really bad is happening there?

683
00:36:22,520 --> 00:36:27,279
Speaker 1: Well, I'm gonna get into a slightly macabre arc of

684
00:36:27,320 --> 00:36:30,759
Speaker 1: this or maybe sci fi angle, but since we're just

685
00:36:30,800 --> 00:36:34,320
Speaker 1: talking about bad things happening, I'm just gonna go there,

686
00:36:34,800 --> 00:36:39,800
Speaker 1: which is eyeball theft? Someone cutting out, uh if a

687
00:36:39,960 --> 00:36:44,320
Speaker 1: really important person who uses world coin to verify themselves

688
00:36:44,360 --> 00:36:49,920
Speaker 1: on some network, a criminal attacker physically maiming someone like

689
00:36:50,800 --> 00:36:54,000
Speaker 1: I imagine these these these sort of scenarios are digging up

690
00:36:54,480 --> 00:36:57,759
Speaker 1: the dead for their eyeballs. Can these must come up

691
00:36:57,800 --> 00:37:01,120
Speaker 1: in your conversations about really extreme scene. Can you talk

692
00:37:01,160 --> 00:37:03,319
Speaker 1: about some of these sort of like weird out there

693
00:37:03,360 --> 00:37:04,400
Speaker 1: physical risks?

694
00:37:05,280 --> 00:37:07,319
Speaker 5: Yeah, they all of these sign ups would not work

695
00:37:07,400 --> 00:37:12,200
Speaker 5: because the art would realize that you are that person

696
00:37:12,320 --> 00:37:16,160
Speaker 5: is these eyes are not alive and are not connected

697
00:37:16,200 --> 00:37:17,960
Speaker 5: to an actual This is the.

698
00:37:18,239 --> 00:37:22,440
Speaker 2: Moment on all thoughts when we talk about degenerating eyeball tissue.

699
00:37:23,120 --> 00:37:24,680
Speaker 1: But I don't know.

700
00:37:24,600 --> 00:37:26,719
Speaker 4: Anything part of discussion.

701
00:37:26,800 --> 00:37:29,000
Speaker 1: But no, no, no, I mean I am sure. I'm make

702
00:37:29,000 --> 00:37:31,240
Speaker 1: sure really like this is like what is a liveness

703
00:37:31,320 --> 00:37:33,040
Speaker 1: check like? And I guess it goes back to your

704
00:37:33,120 --> 00:37:36,440
Speaker 1: question of like why eyeballs not thumbprints or why not

705
00:37:36,480 --> 00:37:39,160
Speaker 1: fingerprints or face ID or your gate or something like

706
00:37:39,239 --> 00:37:42,719
Speaker 1: talk to us about that verification. Why the eyeball can't

707
00:37:42,760 --> 00:37:45,520
Speaker 1: be replicated? Why wouldn't work if not attached to a

708
00:37:45,560 --> 00:37:48,440
Speaker 1: live person? Like these must come up? This would this?

709
00:37:48,480 --> 00:37:49,399
Speaker 1: What would freak me out?

710
00:37:50,719 --> 00:37:50,959
Speaker 4: Sure?

711
00:37:51,040 --> 00:37:54,520
Speaker 5: So, so the device has multiple senses in front, there's

712
00:37:54,520 --> 00:37:56,759
Speaker 5: like a lot of computer on this device. There's like

713
00:37:56,800 --> 00:38:01,000
Speaker 5: a separate GPU and this seven year old firks that

714
00:38:01,320 --> 00:38:07,640
Speaker 5: basically at all time run and what a large part

715
00:38:07,680 --> 00:38:09,919
Speaker 5: of these networks is doing is just to make sure

716
00:38:09,960 --> 00:38:14,040
Speaker 5: that whatever we see is an actual human being alive,

717
00:38:14,160 --> 00:38:17,560
Speaker 5: no fraud attack, no kind of weird cases. And then

718
00:38:17,640 --> 00:38:20,439
Speaker 5: there is in front, there's multiple sensors that make sure

719
00:38:20,480 --> 00:38:22,920
Speaker 5: that that's the case. And you could even go as

720
00:38:22,960 --> 00:38:25,680
Speaker 5: far as like an AI trying to attack the system,

721
00:38:25,760 --> 00:38:28,080
Speaker 5: because that can also happen in the future, right, so

722
00:38:28,160 --> 00:38:31,719
Speaker 5: all of that should not happen. And so what the

723
00:38:31,840 --> 00:38:34,880
Speaker 5: ORB does It images multi spectral, So what that means

724
00:38:34,920 --> 00:38:39,520
Speaker 5: is across multiple wavelengths and the electroc neetic spectrum all

725
00:38:39,560 --> 00:38:45,239
Speaker 5: at the same time. And this is just like you

726
00:38:45,239 --> 00:38:49,160
Speaker 5: can think of very sophisticated optical table attacks like I

727
00:38:49,239 --> 00:38:51,160
Speaker 5: did a lot of these physics when I was younger.

728
00:38:52,280 --> 00:38:57,279
Speaker 5: You can think of these things, but it's extremely expensive

729
00:38:58,120 --> 00:39:00,480
Speaker 5: and I think even that will not be possible the future.

730
00:39:00,520 --> 00:39:02,920
Speaker 5: So it's just really really hard to fool this device.

731
00:39:04,840 --> 00:39:08,200
Speaker 5: I mean, of course, to be clear, the infrastructure is there,

732
00:39:08,200 --> 00:39:10,440
Speaker 5: like the new networks are still we were still building up,

733
00:39:10,480 --> 00:39:12,880
Speaker 5: so right now we are actually aware of a couple

734
00:39:12,920 --> 00:39:15,960
Speaker 5: of tacks that can still happen, and that will still

735
00:39:16,000 --> 00:39:17,319
Speaker 5: be the case for the next year. Like all of

736
00:39:17,320 --> 00:39:20,560
Speaker 5: this technology is not perfect yet, so we'll still take time,

737
00:39:20,640 --> 00:39:22,960
Speaker 5: but at least the technology is there to make sure

738
00:39:22,960 --> 00:39:24,719
Speaker 5: that none of these attacks will ever go through.

739
00:39:26,080 --> 00:39:30,560
Speaker 2: So setting aside the worst case scenarios maybe a slightly

740
00:39:30,960 --> 00:39:36,120
Speaker 2: easier question, but what is the path to adoption that

741
00:39:36,160 --> 00:39:39,319
Speaker 2: you see here, because like, there are a lot of

742
00:39:39,360 --> 00:39:42,000
Speaker 2: things it feels like to overcome. One is the creepiness

743
00:39:42,040 --> 00:39:45,239
Speaker 2: factor of having your eyeball scan. But secondly, you know,

744
00:39:45,239 --> 00:39:49,200
Speaker 2: if you're touting this as a way to identify people

745
00:39:49,280 --> 00:39:52,080
Speaker 2: who might be in need of some sort of official

746
00:39:52,120 --> 00:39:55,760
Speaker 2: identification in countries that don't necessarily have those systems yet

747
00:39:56,000 --> 00:39:59,319
Speaker 2: or have subpar systems, I imagine like there has to

748
00:39:59,360 --> 00:40:01,960
Speaker 2: be a reason and why they would want to have

749
00:40:02,000 --> 00:40:05,799
Speaker 2: this done. You know, in most places, in almost all

750
00:40:05,840 --> 00:40:08,320
Speaker 2: places in the world, no one is getting universal basic

751
00:40:08,400 --> 00:40:11,839
Speaker 2: income yet, So what is the path to adoption and

752
00:40:11,880 --> 00:40:15,239
Speaker 2: what are the the incentives for people to do this?

753
00:40:15,960 --> 00:40:16,280
Speaker 1: Right?

754
00:40:16,640 --> 00:40:20,600
Speaker 5: This is kind of the the everything question, Like, of

755
00:40:20,640 --> 00:40:23,080
Speaker 5: course if I would have the perfect.

756
00:40:22,760 --> 00:40:26,480
Speaker 4: Answer to this, I would tweet it out.

757
00:40:26,600 --> 00:40:33,560
Speaker 6: But that's a good answer every question we've ever Keep going,

758
00:40:34,040 --> 00:40:39,359
Speaker 6: So we're worried about being replaced. We don't need that anyway, Sorry,

759
00:40:39,440 --> 00:40:40,000
Speaker 6: keep gone.

760
00:40:41,120 --> 00:40:45,160
Speaker 5: Okay, So there's like there there's a couple different arcs

761
00:40:45,200 --> 00:40:47,719
Speaker 5: to the response to this question, right. So one is

762
00:40:48,400 --> 00:40:50,879
Speaker 5: when we launch, there's going to be a token token

763
00:40:50,920 --> 00:40:54,120
Speaker 5: with a capsule apply that will play a very important

764
00:40:54,160 --> 00:40:58,279
Speaker 5: role in that kind of the whole evolving nature of

765
00:40:58,360 --> 00:41:03,080
Speaker 5: this project. And when you sign up, you basically receive

766
00:41:03,600 --> 00:41:05,680
Speaker 5: on a running basis, you receive ownership in a network

767
00:41:05,719 --> 00:41:08,920
Speaker 5: through the token, And I think that's exciting for a

768
00:41:08,960 --> 00:41:09,520
Speaker 5: lot of reasons.

769
00:41:09,600 --> 00:41:10,960
Speaker 4: Like one, it's a direct incentive.

770
00:41:10,960 --> 00:41:13,719
Speaker 5: It's a direct like you literally get money basically that

771
00:41:13,800 --> 00:41:16,239
Speaker 5: you could you could sell, or you could do anything with.

772
00:41:17,200 --> 00:41:21,279
Speaker 5: So it's already like a very small form of It's

773
00:41:21,280 --> 00:41:23,520
Speaker 5: not basic because there's not enough money, but it's a

774
00:41:23,680 --> 00:41:26,360
Speaker 5: small form of UBI in some sense, I think that

775
00:41:26,400 --> 00:41:28,239
Speaker 5: will motivate a lot of people. And we already see

776
00:41:28,280 --> 00:41:31,160
Speaker 5: it's like moving a lot of people to like we

777
00:41:31,200 --> 00:41:34,680
Speaker 5: have the crazy Like literally right now is the craziest

778
00:41:34,680 --> 00:41:38,080
Speaker 5: time for me running this project because it's blowing up,

779
00:41:38,160 --> 00:41:42,440
Speaker 5: like all the dashboards completely go vertical, and we have

780
00:41:42,520 --> 00:41:46,360
Speaker 5: like people flying from Japan to Portugal to sign up

781
00:41:46,400 --> 00:41:50,600
Speaker 5: and whatever, like it's pretty really, it's pretty crazy. So

782
00:41:50,680 --> 00:41:56,280
Speaker 5: that certainly does something that's one too is of course integrations.

783
00:41:57,120 --> 00:42:01,719
Speaker 5: So making sure that kind of large technology companies or

784
00:42:02,719 --> 00:42:08,200
Speaker 5: services actually integrate with world ID. That's gonna I mean,

785
00:42:08,200 --> 00:42:09,839
Speaker 5: it's gonna take a while. Because you have a chicken

786
00:42:09,920 --> 00:42:12,080
Speaker 5: egg problem. You need a lot of users for that

787
00:42:12,160 --> 00:42:14,920
Speaker 5: to be interesting too, products and.

788
00:42:14,719 --> 00:42:16,319
Speaker 4: So and so forth. There just like a little bit

789
00:42:16,360 --> 00:42:18,880
Speaker 4: of a balancing act. I'm sure it's going to happen.

790
00:42:20,440 --> 00:42:22,560
Speaker 5: And once that is the case, once you have like

791
00:42:22,640 --> 00:42:28,040
Speaker 5: major integrations, then kind of an actual flywheel is starting

792
00:42:28,160 --> 00:42:30,480
Speaker 5: because as a user you will be able to use

793
00:42:30,480 --> 00:42:32,160
Speaker 5: a lot of services that you're just not able to

794
00:42:32,239 --> 00:42:34,480
Speaker 5: use without it. And so that's kind of that's the

795
00:42:34,600 --> 00:42:35,920
Speaker 5: that's the short of it. And I have a lot

796
00:42:36,200 --> 00:42:40,359
Speaker 5: more specific specific answers are much more localized as well.

797
00:42:40,440 --> 00:42:43,839
Speaker 5: So for example, we are in Bonos Aires and so

798
00:42:43,840 --> 00:42:49,400
Speaker 5: we will work on local integrations in Argentina. So it

799
00:42:49,440 --> 00:42:52,719
Speaker 5: will not start with Twitter or Meta. It will start

800
00:42:52,760 --> 00:42:58,520
Speaker 5: with like very localized applications. But yeah, that's where it goes.

801
00:43:00,160 --> 00:43:03,000
Speaker 1: Since this is odd lots, I have a really quick question.

802
00:43:03,120 --> 00:43:05,400
Speaker 1: You know, you built this I'm going to pick it

803
00:43:05,480 --> 00:43:07,000
Speaker 1: up again because it's really kind of fun the whole

804
00:43:07,239 --> 00:43:09,279
Speaker 1: or you built this piece of hardware, Yeah, this ORB.

805
00:43:09,800 --> 00:43:12,640
Speaker 1: Do you have any supply interesting supply chain issues in

806
00:43:12,719 --> 00:43:16,000
Speaker 1: terms of the optics, so the semiconductors or I mean

807
00:43:16,040 --> 00:43:18,160
Speaker 1: it's a and how much does a right now cost

808
00:43:18,160 --> 00:43:18,799
Speaker 1: to build an ORB?

809
00:43:21,800 --> 00:43:26,280
Speaker 5: So we had especially during COVID, we had crazy supply

810
00:43:26,440 --> 00:43:30,640
Speaker 5: chain crunches. So it's like generally, like Elon talked a

811
00:43:30,680 --> 00:43:33,640
Speaker 5: lot about this in the earlier days, but I definitely

812
00:43:33,680 --> 00:43:36,160
Speaker 5: also ran into this. It's like you have a perfect

813
00:43:36,280 --> 00:43:39,960
Speaker 5: running prototype and then it just feels trivial to produce

814
00:43:40,000 --> 00:43:44,960
Speaker 5: this thing, but it's just not. It was absolutely terrific,

815
00:43:45,160 --> 00:43:46,839
Speaker 5: Like we had this, We had this prototype, and then

816
00:43:46,840 --> 00:43:49,320
Speaker 5: building up a production line and really being able to

817
00:43:49,360 --> 00:43:51,640
Speaker 5: produce like thousands of them. It was a completely different

818
00:43:51,680 --> 00:43:53,719
Speaker 5: challenge that it almost took like a year or two

819
00:43:53,960 --> 00:43:57,600
Speaker 5: to complete. Doctors locked us in and then within COVID

820
00:43:57,600 --> 00:44:02,480
Speaker 5: we had crazy supply chain issues, but it's now mostly gone.

821
00:44:02,680 --> 00:44:05,719
Speaker 5: The kind of the industry has completely recovered.

822
00:44:06,000 --> 00:44:08,279
Speaker 2: What was the hardest component to get.

823
00:44:11,960 --> 00:44:15,880
Speaker 4: Oh, when it was it was death by a thousand cuts.

824
00:44:16,239 --> 00:44:24,040
Speaker 5: It was everything from the computing unit to random capacitors,

825
00:44:24,400 --> 00:44:29,239
Speaker 5: weird ones that we use. We have like that the

826
00:44:29,320 --> 00:44:33,200
Speaker 5: actual lens right now is single source supplier, which is

827
00:44:33,239 --> 00:44:37,280
Speaker 5: something that obviously has to change. But since we customly

828
00:44:37,280 --> 00:44:39,120
Speaker 5: had to build this lens, there's only one company in

829
00:44:39,160 --> 00:44:41,920
Speaker 5: the world that builds that lens, so that that was

830
00:44:42,160 --> 00:44:43,640
Speaker 5: a really tricky one to kind of ramp it up

831
00:44:43,680 --> 00:44:46,879
Speaker 5: over time. How much does it cost right now? Right

832
00:44:46,920 --> 00:44:52,040
Speaker 5: now is still We're still pretty expensive. But honestly, it

833
00:44:52,080 --> 00:44:55,920
Speaker 5: doesn't really matter because so right now it's like four thousand,

834
00:44:56,280 --> 00:44:58,480
Speaker 5: four to five thousand dollars depending on the exact patch,

835
00:44:59,080 --> 00:45:01,440
Speaker 5: it's coming down to be around one thousand, five hundred

836
00:45:01,440 --> 00:45:04,120
Speaker 5: dollars at mass manufacturing, and I think in a couple

837
00:45:04,160 --> 00:45:06,600
Speaker 5: of years it's going to be below five hundred dollars.

838
00:45:06,760 --> 00:45:09,319
Speaker 1: We got to put the THOT to put the ORB

839
00:45:09,400 --> 00:45:12,000
Speaker 1: into the FEDS inflation basket. We're getting so much better

840
00:45:12,040 --> 00:45:14,520
Speaker 1: at producing eyeball scanning orbs.

841
00:45:15,040 --> 00:45:19,520
Speaker 2: Yes, yes, healthcare and medical care massively inflated, but we

842
00:45:19,560 --> 00:45:22,919
Speaker 2: can all scan our eyeballs cheaper and cheaper excellent.

843
00:45:23,320 --> 00:45:26,440
Speaker 1: Alex Bonya, this was such a fascinating conversation. It's like,

844
00:45:26,480 --> 00:45:28,640
Speaker 1: I mean, there's like a million more things I could

845
00:45:28,719 --> 00:45:30,279
Speaker 1: want to ask, and maybe we'll have you back in

846
00:45:30,280 --> 00:45:32,799
Speaker 1: a year and then in two years when we're all

847
00:45:32,920 --> 00:45:35,279
Speaker 1: on UBI because we've been put out of work by

848
00:45:35,360 --> 00:45:38,319
Speaker 1: AI and so, you know, whatever it is. But this

849
00:45:38,440 --> 00:45:40,680
Speaker 1: was this was far and away our most sort of

850
00:45:40,719 --> 00:45:45,360
Speaker 1: like futuristic sci fi ish conversation we've ever had not

851
00:45:45,480 --> 00:45:47,800
Speaker 1: something I would have ever contemplated in the past. So

852
00:45:47,920 --> 00:45:50,400
Speaker 1: thank you so much for coming on outbox amazing.

853
00:45:50,440 --> 00:45:53,319
Speaker 4: I take that as a compliment. Thanks for absolutely question.

854
00:45:53,719 --> 00:46:08,560
Speaker 3: Yeah, that was really interesting, Tracy.

855
00:46:08,600 --> 00:46:09,880
Speaker 1: Are you are you going to get scanned?

856
00:46:11,600 --> 00:46:11,920
Speaker 3: You know what?

857
00:46:11,960 --> 00:46:13,960
Speaker 2: I think I'm gonna wait a little bit until there's

858
00:46:13,960 --> 00:46:17,520
Speaker 2: some visible upside for me, which which is kind of

859
00:46:17,560 --> 00:46:19,799
Speaker 2: that last question that I asked Alex, which you know,

860
00:46:19,920 --> 00:46:25,280
Speaker 2: I am reasonably I understand the use case of having

861
00:46:25,400 --> 00:46:29,520
Speaker 2: some sort of like wallet or portable identification that you

862
00:46:29,560 --> 00:46:33,920
Speaker 2: can take with you across the internet and like various outlets.

863
00:46:33,960 --> 00:46:35,399
Speaker 2: That kind of makes sense to me, and I think

864
00:46:35,400 --> 00:46:39,840
Speaker 2: we've spoken about that use case before. However, at this

865
00:46:39,960 --> 00:46:43,520
Speaker 2: point in time, I don't see a lot of upside.

866
00:46:43,760 --> 00:46:46,000
Speaker 2: Uh you know, maybe it would be easier to file

867
00:46:46,120 --> 00:46:49,840
Speaker 2: some claims against Twitter for like yeah, bought copycats, but

868
00:46:50,200 --> 00:46:50,839
Speaker 2: it's not much.

869
00:46:51,000 --> 00:46:53,600
Speaker 1: Yeah, but and to your point though, like or and

870
00:46:53,640 --> 00:46:57,840
Speaker 1: really or sorry to Alex's point, like, so it's on Twitter,

871
00:46:58,040 --> 00:47:00,279
Speaker 1: like do they want to integrate the world coin? Yeah,

872
00:47:00,480 --> 00:47:03,720
Speaker 1: which I don't actually get the impression that, Like I'm

873
00:47:03,840 --> 00:47:06,319
Speaker 1: it doesn't seem like any of the big networks like

874
00:47:06,360 --> 00:47:08,920
Speaker 1: are anywhere, like that's on the roadmap of anyone, right.

875
00:47:09,000 --> 00:47:12,320
Speaker 2: Well, this is the other thing I'm thinking about, is, uh,

876
00:47:12,360 --> 00:47:15,319
Speaker 2: you know, part of the world coin use case is

877
00:47:15,360 --> 00:47:19,840
Speaker 2: this decentralization aspect of it. But what we've seen time

878
00:47:19,880 --> 00:47:24,319
Speaker 2: and time again throughout history is that the world tends

879
00:47:24,440 --> 00:47:28,759
Speaker 2: towards middlemen and verification of some sort. So I can

880
00:47:28,800 --> 00:47:31,800
Speaker 2: imagine a future where maybe you do get your eyeball scanned,

881
00:47:32,040 --> 00:47:35,960
Speaker 2: but maybe it's not by world coin or decentralized entity.

882
00:47:36,160 --> 00:47:40,000
Speaker 2: Maybe it's Twitter itself, and they this is even more discian,

883
00:47:40,120 --> 00:47:42,640
Speaker 2: you know, like Twitter keeps its own database of your eyeballs.

884
00:47:42,640 --> 00:47:44,440
Speaker 1: Well that's the thing, like I used clear, you know,

885
00:47:44,560 --> 00:47:46,600
Speaker 1: like I already like scanned my ball at the airport,

886
00:47:46,680 --> 00:47:49,560
Speaker 1: and it's like one hundred dollars. That's a centralized database.

887
00:47:49,600 --> 00:47:52,000
Speaker 1: I save a few minutes. It's like I've already I

888
00:47:52,080 --> 00:47:55,040
Speaker 1: could when it comes to privacy, I've already like basically

889
00:47:55,200 --> 00:47:56,960
Speaker 1: given up. I've already given it all up to save

890
00:47:57,000 --> 00:48:00,600
Speaker 1: about five minutes in the airline. So I guess I've like, yeah,

891
00:48:00,680 --> 00:48:04,680
Speaker 1: you know, I'll get scanned, why not. But I'm skeptical,

892
00:48:04,719 --> 00:48:07,080
Speaker 1: you know, actually, honestly, it's not even like the privacy

893
00:48:07,200 --> 00:48:10,920
Speaker 1: or anything like that, or like the dystopian scenarios so

894
00:48:11,040 --> 00:48:14,480
Speaker 1: much as like, I don't really want a scenario in

895
00:48:14,560 --> 00:48:17,239
Speaker 1: which like computers put us all out of jobs and

896
00:48:17,280 --> 00:48:19,239
Speaker 1: we need the UBI, and I think we didn't really

897
00:48:19,280 --> 00:48:22,560
Speaker 1: get into like this scenario in the future, Like where

898
00:48:22,560 --> 00:48:25,719
Speaker 1: will the political impulse from UBI come from. It'll come

899
00:48:25,760 --> 00:48:29,200
Speaker 1: from in an economy that's so lopsided in who reaps

900
00:48:29,239 --> 00:48:31,239
Speaker 1: the benefits of AI that we like have to have

901
00:48:31,320 --> 00:48:34,080
Speaker 1: it for political stability, which is that's like the scenario

902
00:48:34,120 --> 00:48:35,880
Speaker 1: that like freaks me out more than anything.

903
00:48:35,719 --> 00:48:38,799
Speaker 2: Right in the future where you really need world coin,

904
00:48:39,120 --> 00:48:41,680
Speaker 2: Like the most insidious thing is that you need it

905
00:48:41,719 --> 00:48:42,920
Speaker 2: because no one has jobs.

906
00:48:43,040 --> 00:48:45,479
Speaker 1: That's right, right, Like more than anything else, the scan

907
00:48:45,760 --> 00:48:48,600
Speaker 1: or whatever, it's this scenario in which, oh, I need

908
00:48:48,600 --> 00:48:50,840
Speaker 1: to get my UBI from like somewhere, and like they

909
00:48:50,840 --> 00:48:53,719
Speaker 1: attack sam Altman so that we can all Like that's

910
00:48:53,760 --> 00:48:56,160
Speaker 1: like the part more than the SCAN or the privacy

911
00:48:56,320 --> 00:48:58,440
Speaker 1: or the you know, the digging up the dead that

912
00:48:58,480 --> 00:48:59,160
Speaker 1: freaks me out.

913
00:48:59,280 --> 00:49:01,960
Speaker 2: Yeah, I'm telling you a great business model create the

914
00:49:02,000 --> 00:49:04,920
Speaker 2: problem and the solution. On that happy note, shall we

915
00:49:04,960 --> 00:49:05,279
Speaker 2: leave it there.

916
00:49:05,360 --> 00:49:06,680
Speaker 1: Let's leave it there, all right.

917
00:49:06,600 --> 00:49:09,360
Speaker 2: This has been another episode of the All Thoughts podcast.

918
00:49:09,400 --> 00:49:11,879
Speaker 2: I'm Tracy Alloway. You can follow me on Twitter at

919
00:49:11,920 --> 00:49:12,640
Speaker 2: Tracy Alloway.

920
00:49:12,760 --> 00:49:15,480
Speaker 1: And I'm Joe Wisenthal. You can follow me on Twitter

921
00:49:15,560 --> 00:49:19,280
Speaker 1: at the Stalwart. Follow our guest Alex Blanya, as he said,

922
00:49:19,360 --> 00:49:21,160
Speaker 1: if he might just tweet it out, he might just

923
00:49:21,200 --> 00:49:23,000
Speaker 1: tweet out the answer to all of these things. He's

924
00:49:23,080 --> 00:49:27,240
Speaker 1: at Alex Blana. And follow our producers on Twitter Carmen

925
00:49:27,320 --> 00:49:31,359
Speaker 1: Rodriguez at Carmen Arman and dash Ol Bennett at dashbot.

926
00:49:31,520 --> 00:49:34,080
Speaker 1: And follow all of the Bloomberg podcasts onto the handle

927
00:49:34,200 --> 00:49:38,040
Speaker 1: at podcasts and for more Oddlots content, go to Bloomberg

928
00:49:38,040 --> 00:49:40,799
Speaker 1: dot com slash Oddlots, where we have a blog, we

929
00:49:40,880 --> 00:49:43,920
Speaker 1: have a transcript and a newsletter that comes out every Friday.

930
00:49:44,000 --> 00:49:47,400
Speaker 1: And check out our discord. It's super fun people chatting

931
00:49:47,400 --> 00:49:50,160
Speaker 1: about odd loot stuff with other listeners twenty four to seven.

932
00:49:50,440 --> 00:49:53,400
Speaker 1: Go to discord dot gg slash odlogs.

933
00:49:53,520 --> 00:49:56,760
Speaker 2: Yeah, the discord is very fun. You should also stream

934
00:49:56,880 --> 00:50:01,799
Speaker 2: Bloomberg originals on Apple TV, Samsung Roku. Tune in at

935
00:50:01,880 --> 00:50:30,040
Speaker 2: ten pm