WEBVTT - The Eyeball-Scanning Plan Designed to Save Us From AI 0:00:10.200 --> 0:00:14.280 Hello, and welcome to another episode of the Odd Lots Podcast. 0:00:14.360 --> 0:00:16.080 I'm Joe Wisenthal. 0:00:15.600 --> 0:00:16.639 And I'm Tracy Alloway. 0:00:17.120 --> 0:00:20.000 Tracy, do you remember a few weeks ago, or maybe 0:00:20.000 --> 0:00:22.040 a couple of months ago now, someone made a like 0:00:22.079 --> 0:00:25.840 a fake AI version of the Odd Lots podcast and 0:00:25.880 --> 0:00:27.760 they like replicated our voice. 0:00:28.000 --> 0:00:30.520 Yes, I think they used chat gpt to do it. 0:00:30.600 --> 0:00:33.960 And they basically asked chat gpt to create a script 0:00:34.000 --> 0:00:37.320 for odd Loots and that they had someone replicate another app 0:00:37.400 --> 0:00:38.400 or something replicated. 0:00:38.479 --> 0:00:40.120 It wasn't a someone like I mean, it was like 0:00:40.200 --> 0:00:43.680 somewhat like other like AI app. It was like, yeah, 0:00:43.840 --> 0:00:45.760 it was pretty it was pretty good. It was like good. 0:00:45.840 --> 0:00:48.200 I mean it clearly like was not us. 0:00:48.479 --> 0:00:50.159 How much did it worry you, Joe? 0:00:50.840 --> 0:00:53.319 I think I'm more of an AI domer worrier than 0:00:53.360 --> 0:00:53.639 you are. 0:00:53.960 --> 0:00:54.520 Interesting. 0:00:54.880 --> 0:00:57.840 I mean, I have concerns, don't get me wrong, But 0:00:57.880 --> 0:00:59.520 what do you think is the worst case scenario? Is 0:00:59.520 --> 0:01:00.880 it robot taking over the world. 0:01:01.000 --> 0:01:01.120 No. 0:01:01.200 --> 0:01:04.080 I think that like in a couple of years, the 0:01:04.160 --> 0:01:06.360 AI will do a really good job of making the 0:01:06.360 --> 0:01:09.840 Odd Lots podcast. And people say, I don't really need 0:01:09.880 --> 0:01:11.960 to listen to Joe and Tracy anymore. The AI is 0:01:12.120 --> 0:01:14.920 I don't know. I'm like, you know, I but I 0:01:14.959 --> 0:01:16.400 feel like maybe that's all a parandoise. 0:01:16.520 --> 0:01:19.240 No, I feel like there is reason for concern. I hope. 0:01:19.360 --> 0:01:21.319 Maybe I'm just being an optimist here. I hope we 0:01:21.360 --> 0:01:24.200 are a few years away from the odd Blots replicat. 0:01:24.240 --> 0:01:25.720 I know what, I want to do this for like 0:01:25.760 --> 0:01:27.480 another thirty years, and I don't want to just have 0:01:27.600 --> 0:01:29.480 like three or four years or a few years. 0:01:29.480 --> 0:01:32.080 That's the optimistic scenario. Is it doing this in thirty 0:01:32.160 --> 0:01:32.560 years time? 0:01:32.640 --> 0:01:33.040 Okay? 0:01:33.240 --> 0:01:33.600 All right? 0:01:34.360 --> 0:01:37.720 AI. Lots of people expressing some worries about. 0:01:37.520 --> 0:01:38.920 It, a million different kind of words. 0:01:39.000 --> 0:01:42.199 Yeah, and we see all these, you know, good ways 0:01:42.280 --> 0:01:45.040 that the technology is being used, but also insidious ways. 0:01:45.160 --> 0:01:47.760 As you pointed out, someone made an Audlots episode. I 0:01:47.800 --> 0:01:51.720 think that was in a in a very fun spirited way, 0:01:51.800 --> 0:01:54.520 a nice way. But there are people who use this 0:01:54.560 --> 0:01:57.480 technology for more nefarious things. Yeah. 0:01:57.480 --> 0:01:59.880 I think there was like a sixty minutes recently where 0:01:59.880 --> 0:02:03.640 like someone was able to like impersonate someone's voice and say, hey, 0:02:04.320 --> 0:02:07.360 you know, you know, you could imagine this scam where 0:02:07.400 --> 0:02:09.520 someone's like kids, say hey, I need money or something 0:02:09.560 --> 0:02:13.079 like that, and they call and it sounds just like them, 0:02:13.160 --> 0:02:14.760 especially in a couple of years when the tech gets 0:02:14.800 --> 0:02:17.640 better and it becomes a bit of a crisis in 0:02:17.680 --> 0:02:19.760 the sense it's like, do I know that a person 0:02:19.840 --> 0:02:21.959 is a person when they talk to me on a zoom, 0:02:21.960 --> 0:02:23.480 when they talk to me on a podcast, when they 0:02:23.480 --> 0:02:24.880 talk to me on a phone call, et cetera. 0:02:25.080 --> 0:02:27.760 Well, even on Twitter on a much more basic level, 0:02:27.800 --> 0:02:29.760 you have all these bots, and I don't know about you, 0:02:29.880 --> 0:02:32.200 but there seems to be a new Tracy Alloway with 0:02:32.320 --> 0:02:36.600 the l as like an I and fooling everyone almost 0:02:36.680 --> 0:02:38.520 like every few weeks, and then you have to go 0:02:38.560 --> 0:02:41.800 through the whole bureaucracy of reporting to Twitter and proving 0:02:41.880 --> 0:02:44.720 that you are in fact the original Tracy alloyway. 0:02:44.880 --> 0:02:47.359 And not to get into much of a tangent, but 0:02:47.440 --> 0:02:49.600 wasn't Elon going to solve this because he said like, oh, 0:02:49.600 --> 0:02:52.200 I'm buying Twitter to solve the bots problem, and I don't, 0:02:52.480 --> 0:02:54.120 as you say, it's not been solved. 0:02:54.160 --> 0:02:55.600 I think that's a whole other episode. 0:02:55.800 --> 0:03:00.320 But even with like really rudimentary like human you know, 0:03:00.560 --> 0:03:03.520 fake traces and fake Joe's like, that's the most rudimentary thing, 0:03:03.600 --> 0:03:06.000 just like taking our avatar and like changing like one 0:03:06.000 --> 0:03:07.840 of the l's and trade in L a way to 0:03:07.880 --> 0:03:10.760 like an I or whatever or one like. That's just 0:03:10.800 --> 0:03:11.480 gonna get worse. 0:03:13.000 --> 0:03:16.280 You know what's a really good business model creating a 0:03:16.400 --> 0:03:19.800 problem and then creating the solution to it. 0:03:21.240 --> 0:03:23.880 Yes, that's exactly right. So you have all So we 0:03:23.919 --> 0:03:27.440 have all of these ais, and they're going to like 0:03:27.480 --> 0:03:29.120 replicate us, They're going to look like us, or they're 0:03:29.120 --> 0:03:32.040 going to talk like us. But the good news, in 0:03:32.080 --> 0:03:34.280 a sense good is that others are working out the 0:03:34.360 --> 0:03:37.400 problem of how do you prove that you are who 0:03:37.440 --> 0:03:41.400 you are in a world in which AIS can replicate 0:03:41.400 --> 0:03:43.480 our voices and our faces and our expression. 0:03:43.640 --> 0:03:46.320 Technology can save us from technology? 0:03:46.520 --> 0:03:48.440 Is the thrust of this, right exactly. So we are 0:03:48.480 --> 0:03:52.800 going to be speaking to someone working on the technology 0:03:52.840 --> 0:03:57.280 designed to address this concern in part. And as you say, 0:03:57.360 --> 0:03:59.400 some of the people working on it are some of 0:03:59.440 --> 0:04:02.920 the same peopleeople behind some of the most cutting edge 0:04:03.000 --> 0:04:05.880 AI research that's sort of like blowing everyone's minds in 0:04:05.960 --> 0:04:07.680 both good and scary ways these days. 0:04:07.720 --> 0:04:09.520 So I'm just going to go ahead and lay out 0:04:09.560 --> 0:04:12.440 my priors here, which are that I find this sort 0:04:12.440 --> 0:04:16.920 of instinctively dystopian the topic of this podcast what we're 0:04:16.960 --> 0:04:20.200 going to be discussing. This technology creeps me out a 0:04:20.240 --> 0:04:23.080 little bit. But that said, I don't know that much 0:04:23.080 --> 0:04:25.039 about it. I don't know much about how it works, 0:04:25.040 --> 0:04:28.039 and so I am very interested to hear the bowl 0:04:28.120 --> 0:04:30.040 case on eyeball scanning. 0:04:30.560 --> 0:04:32.440 So you just jump bright. We're going to be talking 0:04:32.680 --> 0:04:36.080 eyeball scanning. We are going to be speaking with Alex Blania. 0:04:36.400 --> 0:04:40.159 He is the co founder of world Coin, an entity 0:04:40.200 --> 0:04:43.680 that was set up also by Sam Altman, who everybody 0:04:43.800 --> 0:04:46.680 a name that everybody knows now because he's the co 0:04:46.800 --> 0:04:49.920 founder or the founder of open ai, which powers chad 0:04:49.960 --> 0:04:53.560 GPT and other tools. Other co founder Max nor Frenstern. 0:04:53.800 --> 0:04:56.719 We're going to be talking about how an eyeball scanner 0:04:57.160 --> 0:05:00.760 can allow us to theoretically prove our human in this 0:05:00.920 --> 0:05:04.240 future AI world. So Alex, thank you so much for 0:05:04.600 --> 0:05:05.880 coming on odd Lots. 0:05:06.400 --> 0:05:08.080 Thank you for having me. I'm excited to talk. 0:05:08.880 --> 0:05:12.640 So I'm trying to think, I mean, like the threat, 0:05:13.240 --> 0:05:15.520 how real is it that you know, in a few 0:05:15.600 --> 0:05:21.120 years do you think not whether AI will be able 0:05:21.120 --> 0:05:24.640 to create a perfect replica of the podcast, but that 0:05:25.400 --> 0:05:28.880 let's say we're having this conversation and I might really 0:05:28.920 --> 0:05:32.880 not know that I am talking to the real Alex 0:05:32.960 --> 0:05:37.719 Blania versus an AI that could perfectly replicate your voice 0:05:37.720 --> 0:05:40.400 and replicate your face. Like is that two years away? 0:05:40.440 --> 0:05:43.080 Is that three years away? Like? How soon is that coming? 0:05:44.640 --> 0:05:49.479 Well? I think partially we are already there. Right the. 0:05:51.080 --> 0:05:53.520 Technology is not completely open source yet, so you don't 0:05:53.560 --> 0:05:58.159 have widely deployed, deployed systems that can't do that, but 0:05:58.279 --> 0:05:59.800 rather you have right now you have like a few 0:06:00.080 --> 0:06:02.960 teas that actually can do things like that or would 0:06:03.000 --> 0:06:05.279 be able to do things like that. But this is 0:06:05.320 --> 0:06:08.279 going to change because of course the progress is increasing. 0:06:08.360 --> 0:06:12.440 So you had it started with Dolly and you had 0:06:12.480 --> 0:06:15.440 stable diffusion and many other things that on the image 0:06:15.440 --> 0:06:20.200 generation side became increasingly impressive. And I think are ai 0:06:20.360 --> 0:06:21.880 R is one of the coolest things of the of 0:06:21.920 --> 0:06:25.360 the last couple of the last year, to be honest. 0:06:26.080 --> 0:06:28.360 So that's on the image generation side of things, and 0:06:28.400 --> 0:06:30.080 then of course you have chat GIPT on the text 0:06:30.080 --> 0:06:32.320 generation side of things. So I think on Twitter, as 0:06:32.360 --> 0:06:36.920 you discussed in the introduction, you are I think already 0:06:36.960 --> 0:06:39.720 there that it could happen that you just chat with 0:06:39.760 --> 0:06:41.920 a neural network and you have absolute no idea that 0:06:41.960 --> 0:06:44.200 you chat with a new network and you might argue 0:06:44.240 --> 0:06:47.880 about odd lots for like an hour and don't realize that. 0:06:48.120 --> 0:06:51.520 I'm just thinking, like half the people I argue with 0:06:51.560 --> 0:06:52.839 on Twitter probably are. 0:06:53.040 --> 0:06:54.280 Bots neural networks. 0:06:54.480 --> 0:06:58.039 Yeah, well, maybe just to step back in time a 0:06:58.040 --> 0:07:00.280 little bit, I mean, world coin has been around for 0:07:00.400 --> 0:07:03.760 a few years now, and to some degree, as Joe mentioned, 0:07:03.839 --> 0:07:07.599 you guys kind of anticipated this threat. What was the 0:07:07.760 --> 0:07:12.240 sort of aha moment that made this all come together, Like, 0:07:12.480 --> 0:07:16.240 how did this actually develop into a company, and what 0:07:16.480 --> 0:07:20.200 was the use case thesis when you were just setting. 0:07:19.960 --> 0:07:24.440 Out, Well, we honestly did not. We didn't start off 0:07:24.480 --> 0:07:28.800 to solve a threat, all right, that is not brief 0:07:28.840 --> 0:07:32.080 history in time. Sure, back then, I did research. 0:07:32.720 --> 0:07:34.600 So we started the company a little bit over three 0:07:34.680 --> 0:07:35.200 years ago. 0:07:35.800 --> 0:07:40.160 And I was doing research in deep learning applied to 0:07:40.240 --> 0:07:44.520 quantum physics. So I basically predicted quantum system of neural 0:07:44.600 --> 0:07:47.440 networks and that's I think it's a very promising field 0:07:47.440 --> 0:07:50.920 in physics. And Sam was already working on the idea 0:07:51.640 --> 0:07:54.120 overall kind together with Max, so I think they've already 0:07:54.160 --> 0:07:55.960 been on it for like six months. But of course 0:07:56.280 --> 0:07:58.480 Sam had a full time job at opening Ie. Max 0:07:58.520 --> 0:08:00.480 had a full time job back then, so it was 0:08:00.520 --> 0:08:03.880 all like in an early idea face. Then I got 0:08:03.880 --> 0:08:09.240 an email from from Max back then to Poop with 0:08:09.320 --> 0:08:12.800 like an early paper describing the vision of world cooin 0:08:13.680 --> 0:08:15.520 and that it should come by and talk. And then 0:08:15.960 --> 0:08:18.520 at a couple of interree years with Max uh one 0:08:18.560 --> 0:08:20.160 with Sam, we spent a lot of time together and 0:08:20.560 --> 0:08:23.440 actually decided to. 0:08:22.560 --> 0:08:23.280 Proceed, but. 0:08:25.200 --> 0:08:27.440 To go back to the initial proposition, and I think 0:08:27.440 --> 0:08:30.640 that is very important. It's not about solving a threat, 0:08:30.680 --> 0:08:34.839 but rather how I mean, Sam was already breaking in 0:08:34.880 --> 0:08:37.240 open EIE and this was three years ago, so back 0:08:37.280 --> 0:08:41.520 then it was not yet an accepted topic that AI 0:08:41.679 --> 0:08:43.240 is going to happen and it's going to be as 0:08:43.240 --> 0:08:44.280 powerful as today. 0:08:44.320 --> 0:08:45.280 It was like a very. 0:08:46.040 --> 0:08:50.120 Niche topic I think, although open of course was already 0:08:50.200 --> 0:08:53.760 quite famous entity. But of course Sam was motivated by 0:08:53.760 --> 0:08:56.160 the idea of AI and what it's going to do 0:08:56.200 --> 0:08:57.320 to the world and it's going to be this very 0:08:57.320 --> 0:09:01.839 powerful technology. And so the initial proposition of Rocline was 0:09:01.880 --> 0:09:06.600 really what is the societal infrastructure that we can build 0:09:06.600 --> 0:09:10.280 that makes sure that AI is for the better of 0:09:10.400 --> 0:09:13.559 all of society versus a few and so there's a 0:09:13.679 --> 0:09:15.839 there's a many things actually that come together here, like 0:09:15.880 --> 0:09:19.480 proof of personhood is what you just described, is the 0:09:19.520 --> 0:09:22.040 ability to prove that you're actually human being on the Internet, 0:09:22.040 --> 0:09:25.679 which is basically just a much more capable identity system. 0:09:25.720 --> 0:09:27.360 But it goes as far as. 0:09:28.880 --> 0:09:32.200 We think eventually UBI will become a thing, and it 0:09:32.240 --> 0:09:34.240 actually is like a big chance if you have like 0:09:34.240 --> 0:09:36.640 a very very powerful technology and it's not about kind 0:09:36.640 --> 0:09:39.560 of political redistribution, but you have a very centralizing technology 0:09:40.200 --> 0:09:42.720 that you want to share the upside with the world with. 0:09:43.760 --> 0:09:46.800 And there was another motivation, kind of Rokin can actually 0:09:46.800 --> 0:09:48.920 be the infrastructure to do that, because you have a 0:09:49.000 --> 0:09:51.760 very capable identity system, you have an economic system that 0:09:51.800 --> 0:09:55.960 connects everyone. So there were like a couple of very 0:09:55.960 --> 0:09:59.360 ambitious reasons why we think something like that. So basically 0:09:59.360 --> 0:10:05.880 creating and bootstrapping the largest identity and financial network that 0:10:06.040 --> 0:10:12.160 is actually open and privacy preserving versus relying on centralized 0:10:12.679 --> 0:10:15.559 government systems that might not work in large parts of 0:10:15.600 --> 0:10:19.760 the world is a very important preposition for the coming decades. 0:10:20.160 --> 0:10:23.800 So let's just jump right to it. If you're listening 0:10:23.920 --> 0:10:26.360 people who are listening to this podcast, you might want 0:10:26.400 --> 0:10:29.040 to actually watch the video version. We're recording this on 0:10:29.200 --> 0:10:32.880 video as well. People just the audio butt here on 0:10:32.960 --> 0:10:35.719 the set. I'm just gonna I was gonna wait a 0:10:35.760 --> 0:10:38.640 little while before unveiling it, but this is I'm holding 0:10:38.640 --> 0:10:38.960 in my. 0:10:38.880 --> 0:10:41.199 Hand like we need extra music, a big. 0:10:41.160 --> 0:10:44.440 Orb It feels about six pounds. I don't know how 0:10:44.760 --> 0:10:47.360 heavy is this, Alex, I'm sure you know the exact weight. 0:10:48.360 --> 0:10:50.640 It's a little bit over one kilogram. 0:10:51.120 --> 0:10:53.400 Oh okay, so I was wrong on So that's like 0:10:53.480 --> 0:10:58.440 just on like about two and a half pounds, Okay, okay, 0:10:58.520 --> 0:11:00.840 that's fine, about two and have found and it's an 0:11:00.880 --> 0:11:03.400 eyeball scanner, and so here it is, and it's not 0:11:03.440 --> 0:11:06.240 working right now, or we're not scanning right now, so 0:11:06.280 --> 0:11:09.160 I can like tilt this in our face. It's not scanning. 0:11:09.360 --> 0:11:12.040 But let's like talk about what this piece of hardware 0:11:12.160 --> 0:11:13.920 that I'm actually holding it is that I'm going to 0:11:13.960 --> 0:11:16.800 set it on the table in front of me. There's 0:11:16.800 --> 0:11:18.120 an eyeball scanner. What does it do? 0:11:19.240 --> 0:11:23.000 So the show to answer is one it will start 0:11:23.040 --> 0:11:25.200 with shortened than the longer answer. The short answer is 0:11:25.240 --> 0:11:29.199 that it first ensures that whatever is interacting with the device, 0:11:29.240 --> 0:11:32.400 is an actual human being, So okay, not a display, 0:11:33.320 --> 0:11:36.480 not a photograph of an eyeball, not a photograph or 0:11:37.080 --> 0:11:40.160 like more sophisticated optical attacks, because there's like many. 0:11:39.960 --> 0:11:41.040 Things that you can do here. 0:11:41.720 --> 0:11:44.000 So that's the first step checking that whatever we see 0:11:44.000 --> 0:11:47.040 is an actual human being. And then the second piece 0:11:47.800 --> 0:11:52.120 is it takes an image of my eye, calculates a 0:11:52.240 --> 0:11:57.040 unique embedding out of this picture. Everything happens on a device, 0:11:57.920 --> 0:12:01.720 and that embedding then signed by the device, and that's 0:12:01.840 --> 0:12:05.040 then the only thing that actually leaves the device and 0:12:06.320 --> 0:12:10.240 gets compared against all other users. And if that uniqueness 0:12:10.320 --> 0:12:14.480 check is successful, then me that I just verified with 0:12:14.480 --> 0:12:17.720 an ORB I then can later do a zero knowledge 0:12:17.760 --> 0:12:21.880 proof that I am included in that set. And if 0:12:21.920 --> 0:12:23.880 you're not familiar with your knowledge proves, what it allows 0:12:23.920 --> 0:12:27.160 you to do is it lets you prove something without 0:12:27.200 --> 0:12:32.520 actually revealing the underlying information. So what that enables us 0:12:32.600 --> 0:12:36.560 to do is it builds a self like essentially a 0:12:36.600 --> 0:12:41.040 self custodial identity network where the data control is with 0:12:41.080 --> 0:12:44.080 the user and there's no centralized entity that actually has 0:12:44.200 --> 0:12:47.000 the information of who you are, where you're from. All 0:12:47.000 --> 0:12:50.199 of these things but rather this person is a unique 0:12:50.280 --> 0:12:55.160 human being and has verified revocan before, while the user 0:12:55.520 --> 0:12:58.320 could decide to attach more information to that. So that's 0:12:58.360 --> 0:13:00.600 kind of the short of it. I think we should 0:13:00.640 --> 0:13:02.560 talk much more about the why and for a second, 0:13:02.559 --> 0:13:03.960 because that's not trivial. 0:13:04.000 --> 0:13:05.079 But yeah, that's it. 0:13:05.520 --> 0:13:08.319 Yeah, maybe just to help us get to the why, 0:13:08.440 --> 0:13:11.480 you could talk about what would be a practical use 0:13:11.559 --> 0:13:14.920 case for this technology. Because if I think about something 0:13:15.040 --> 0:13:18.600 like you know, you mentioned UBI universal Basic income. If 0:13:18.640 --> 0:13:21.880 I think a government starts some sort of UBI program, 0:13:21.920 --> 0:13:25.600 then I would assume that they identify people by something 0:13:25.679 --> 0:13:28.760 like a social Security number or whatever the equivalent might 0:13:28.800 --> 0:13:32.160 be in that country. So what does this do that 0:13:32.240 --> 0:13:35.120 those traditional systems of identity. 0:13:34.679 --> 0:13:36.760 Do not well? 0:13:36.800 --> 0:13:40.640 So when you start with traditional identity, the really hard 0:13:40.679 --> 0:13:44.679 thing for governments is actually making sure everyone only has 0:13:44.720 --> 0:13:48.960 one password, and that is with an identity is called deduplication. 0:13:49.640 --> 0:13:53.160 It's a really hard problem because all the systems, all 0:13:53.160 --> 0:13:55.800 the technologies you use usually in your life, let's say 0:13:55.840 --> 0:13:59.360 your iPhone Phase ID. What that actually does is just 0:13:59.520 --> 0:14:04.080 it re authenticates you, so it realizes I'm the same 0:14:04.120 --> 0:14:06.240 person again logging into that phone. And that's a one 0:14:06.240 --> 0:14:09.079 to one comparison, and that is fairly easy to do. 0:14:10.160 --> 0:14:12.200 What is really hard to do is making sure okay, 0:14:12.320 --> 0:14:15.360 Alex did not sign up compared to a billion other people. 0:14:16.280 --> 0:14:20.360 And so that's why many governments actually have biometric systems 0:14:20.360 --> 0:14:24.320 in place, because otherwise also their Social Security number system 0:14:24.320 --> 0:14:27.880 would completely break. But that's not rolled out globally. So 0:14:27.960 --> 0:14:31.920 in fact, a good mental model is about fifty percent 0:14:31.920 --> 0:14:35.560 of the global population actually have a digitally verifiable identity. 0:14:36.120 --> 0:14:37.920 But that's a huge number of people. 0:14:37.680 --> 0:14:40.400 That just don't have it right, And so I think 0:14:40.600 --> 0:14:45.840 United States is pretty good, or at least decent. It's 0:14:45.880 --> 0:14:50.120 actually not a really good at any solution. Europe in 0:14:50.240 --> 0:14:56.880 parts is pretty capable. The strongest is India, although there's 0:14:57.000 --> 0:15:00.240 definitely privacy concerns here because they don't use you knowledge 0:15:00.280 --> 0:15:01.120 proofs and things like that. 0:15:01.160 --> 0:15:02.800 But for example, India. 0:15:03.000 --> 0:15:06.080 They have a program called at HOR that also used 0:15:06.880 --> 0:15:09.680 ours recognition to kind of roll out to the global 0:15:09.720 --> 0:15:12.160 population and it really accelerated the whole economy like crazy. 0:15:12.280 --> 0:15:14.600 So they had at HOR and then they build UPI 0:15:14.720 --> 0:15:17.000 on Top, which is a payment rail, and so whenever 0:15:17.040 --> 0:15:18.960 you talk to people that are in that kind of 0:15:19.560 --> 0:15:22.480 area of technology, everyone always brings up India as the 0:15:22.520 --> 0:15:26.480 most prominent example of like what identity plus a strong 0:15:26.520 --> 0:15:29.600 financial rail can actually do to accelerate an economy. 0:15:30.320 --> 0:15:32.160 So, and this is what roll coin is doing. Basically, 0:15:32.240 --> 0:15:32.800 roll coin. 0:15:32.760 --> 0:15:38.400 Is is able to create a d duplication set, so 0:15:38.720 --> 0:15:40.960 it provides me with the ability to prove that I'm 0:15:41.000 --> 0:15:44.640 actually unique against everyone else. But then also there's a 0:15:44.640 --> 0:15:47.160 world ID, which is an identity protocol on top of that. 0:15:48.040 --> 0:15:49.840 And let's start with the basics. 0:15:49.920 --> 0:15:51.640 So the first thing I could do is I could 0:15:51.800 --> 0:15:54.840 log into Twitter and I could say I don't already 0:15:54.840 --> 0:15:58.640 have a Twitter account without actually really my name, And 0:15:58.680 --> 0:16:00.920 then a few things a little bit forward. You could 0:16:00.920 --> 0:16:04.040 even cryptographically sign every tweet you make, so that you 0:16:04.480 --> 0:16:06.440 kind of lay a trace of like, okay, this is 0:16:06.480 --> 0:16:10.200 actually Alex versus anyone else. And I think this is 0:16:10.240 --> 0:16:14.280 really where this is going to go, right, because this 0:16:14.720 --> 0:16:19.920 progress around AI is definitely an exponential So we will 0:16:19.920 --> 0:16:21.960 get to a place where you will help You will 0:16:21.960 --> 0:16:24.240 have to authenticate pretty much everything you do on the internet. 0:16:25.080 --> 0:16:25.880 So that's the basic. 0:16:25.960 --> 0:16:28.920 But then with rold eddy, which is a protocol that 0:16:29.040 --> 0:16:33.880 lets you add verifiable credentials and many other things, with 0:16:33.920 --> 0:16:36.080 an identity space, you can actually bootstop a whole identity 0:16:36.120 --> 0:16:36.760 system on top. 0:16:37.480 --> 0:16:39.960 So this was like a quick grant. But no, it's good. 0:16:40.520 --> 0:16:44.040 So it's not that it's proving that you are, you know, 0:16:44.120 --> 0:16:46.720 a Joe Wisenthal or a Tracy Alloway. It's proving that 0:16:46.760 --> 0:16:51.040 there is only one of you without necessarily the ID 0:16:51.240 --> 0:16:52.800 attached or the name attached. 0:16:53.720 --> 0:16:55.960 That is exactly so that that is the I. 0:16:56.520 --> 0:16:59.600 That is the whole point because then but then what 0:16:59.680 --> 0:17:04.600 we're id lets you do so if you're not into crypto, 0:17:04.840 --> 0:17:08.720 but this whole idea of self custody is very important, 0:17:08.880 --> 0:17:10.160 right of course, I'm pretty sure. 0:17:10.080 --> 0:17:10.800 It came up before. 0:17:12.560 --> 0:17:15.679 What that lets you do is you basically you verify 0:17:15.720 --> 0:17:19.679 for well, can you receive your world Eddy? And then 0:17:19.720 --> 0:17:22.000 it might be that for some applications you actually want 0:17:22.000 --> 0:17:26.040 to prove that you're over eighteen or you're Joe viisatal. 0:17:26.800 --> 0:17:30.119 Then the world Eddie protocol, which is literally just like 0:17:30.119 --> 0:17:33.000 an open source protocol and every developer, every company could 0:17:33.040 --> 0:17:36.120 implement with that, You as a user reader can then 0:17:36.160 --> 0:17:39.760 decide to also attach additional information to your world Eddie, 0:17:39.800 --> 0:17:42.239 so that is my name or that is my can 0:17:42.320 --> 0:17:45.400 we see information and then you can decide to reveel 0:17:45.480 --> 0:17:47.840 that to some applications with zon knowledge proofs as well, 0:17:48.240 --> 0:17:51.360 but no one other than you actually has that information. 0:17:51.840 --> 0:17:55.960 So I think, God, I've I have like hundreds of questions. 0:17:56.000 --> 0:17:57.800 There are so many different ways. So you know, one 0:17:57.840 --> 0:18:01.399 basic but one idea. You keep mentioning zero knowledge proven 0:18:01.520 --> 0:18:03.399 just to sort of like the basic idea, And this 0:18:03.480 --> 0:18:06.320 is the shortest. Like you know, I go to websites, 0:18:06.359 --> 0:18:09.760 sometimes government websites, you entering your social security number. But 0:18:09.800 --> 0:18:12.560 the idea is with zero knowledge proof, it would be 0:18:13.040 --> 0:18:16.200 I prove to the government that I know my social 0:18:16.200 --> 0:18:19.480 Security number without telling the government what my social security 0:18:19.520 --> 0:18:20.000 number is. 0:18:21.080 --> 0:18:23.160 Well, I mean that that one is actually tricky because 0:18:23.520 --> 0:18:25.720 in that cause you would reveal the social security number. 0:18:25.760 --> 0:18:28.240 What is a much more what is a much better 0:18:28.359 --> 0:18:33.680 way to explain it is a website asks you are 0:18:33.680 --> 0:18:35.639 you on any kinds of sanctional list? 0:18:36.520 --> 0:18:36.720 Right? 0:18:37.200 --> 0:18:37.480 Right? 0:18:38.560 --> 0:18:42.399 Or are you over eighteen? Or are you actually a 0:18:42.440 --> 0:18:43.080 US citizen? 0:18:43.960 --> 0:18:44.240 Right? 0:18:44.359 --> 0:18:47.600 And then these things you could prove of zon knowledge 0:18:47.960 --> 0:18:49.800 And then basically what comes out of this is like 0:18:50.600 --> 0:18:54.639 approve runs on your phone that you're not included in 0:18:54.680 --> 0:18:57.280 a sanctions list. Let's say and that the only thing 0:18:57.320 --> 0:18:59.359 that comes out and ends up with the website is 0:18:59.520 --> 0:19:02.920 yes or no. But the website does not know what 0:19:03.080 --> 0:19:05.399 is the social security number you check against it. 0:19:22.160 --> 0:19:24.719 So then the other question I have, and the obvious 0:19:24.760 --> 0:19:29.160 thing is look, scanning your eyeball it's gonna make anyone anxious, 0:19:29.200 --> 0:19:31.960 and I'm like totally upward even I'm like, you know, 0:19:32.000 --> 0:19:37.040 a little bit anxious about it. How do you establish 0:19:37.680 --> 0:19:41.359 so there's no unlike say clear at the airport, which 0:19:41.359 --> 0:19:43.560 also uses an eyeball scan so that you can go 0:19:43.640 --> 0:19:48.520 buy security, there's no centralized database of our irises, And 0:19:48.560 --> 0:19:50.800 how do you sort of like establish or prove to 0:19:50.880 --> 0:19:53.960 people that there's no database Because if my understanding, there 0:19:54.000 --> 0:19:56.679 has been reporting that at least at some point in 0:19:56.760 --> 0:20:00.359 the process of the development of world Cli and as 0:20:00.359 --> 0:20:02.720 you've grown as a startup, that at one point or 0:20:02.760 --> 0:20:05.400 at least currently there was a Irish database. 0:20:06.200 --> 0:20:08.600 Okay, so there there's many things here. Yeah, we've got 0:20:08.600 --> 0:20:08.960 to start. 0:20:09.000 --> 0:20:12.919 I'm going to start with explaining how to develop comfort 0:20:12.920 --> 0:20:16.960 around the idea and why why it's necessary. So, first 0:20:16.960 --> 0:20:20.399 of all, I hated the idea of building the door, 0:20:20.560 --> 0:20:20.879 like when we. 0:20:20.880 --> 0:20:22.280 Actually started working a rockin like it. 0:20:22.359 --> 0:20:26.399 It was absolutely horrific because it's We've been four people 0:20:26.480 --> 0:20:29.760 sitting in a smart, small apartment in San Francisco and 0:20:30.040 --> 0:20:34.560 the idea of building hydred devices that take biometric data 0:20:34.600 --> 0:20:37.240