1 00:00:03,199 --> 00:00:21,275 Doctor Strayhorn, I presume. I'll call you that one time, and then it's David. It's David. Yes. John, it is, fantastic to be here. Thank you. Appreciate you joining me. I'm excited about your podcast. Thanks so much. Excited to have you. I've been looking forward to this conversation, and, we're here in the lovely Bitcoin Park, studio where we both, spend a lot of time. Yep. 2 00:00:21,575 --> 00:00:22,075 And 3 00:00:22,695 --> 00:00:24,155 as you and I have chatted, David, 4 00:00:24,935 --> 00:00:27,275 what I'd love to do is to start 5 00:00:28,055 --> 00:00:28,555 with 6 00:00:30,369 --> 00:00:31,189 a primer 7 00:00:31,570 --> 00:00:32,070 Okay. 8 00:00:32,450 --> 00:00:34,070 On trust from a neurologist's 9 00:00:34,449 --> 00:00:37,190 perspective. I I say that tongue in cheek, but you'll 10 00:00:39,090 --> 00:00:43,350 do a better job than me. I will include, of course, your your esteemed biography, 11 00:00:44,364 --> 00:00:46,144 in in the show notes. But 12 00:00:46,605 --> 00:00:48,464 as someone with a background 13 00:00:49,565 --> 00:00:58,545 academically in electro engineering Mhmm. As a board certified neurologist who's practicing Mhmm. And as someone who has spent now the better part of ten years 14 00:00:59,130 --> 00:01:02,750 studying web of trust and implementing different approaches, 15 00:01:03,210 --> 00:01:04,430 let's begin with 16 00:01:05,770 --> 00:01:09,070 where does trust live in the brain? Where does trust live in the mind? 17 00:01:09,530 --> 00:01:11,870 Well, in in true cyberpunk, 18 00:01:14,045 --> 00:01:15,425 the spirit of a cyberpunk 19 00:01:16,285 --> 00:01:19,425 mind is decentralized, and it doesn't live in any one place. 20 00:01:19,965 --> 00:01:23,585 And so first, I guess, I'd say the answer to your question is not something 21 00:01:24,125 --> 00:01:27,345 that's ever come up, you know, in med school or neurology residency. 22 00:01:29,030 --> 00:01:32,490 It's something that I think about a lot in the clinic when I, 23 00:01:33,750 --> 00:01:42,009 I a lot of what I do is getting a history from patients. And so epilepsy is what I focus on. So a lot of the information is just what people tell me. 24 00:01:42,310 --> 00:01:42,810 Sometimes 25 00:01:43,385 --> 00:01:49,005 the information is accurate. Sometimes it's not because just of the nature of seizures, you can't remember stuff. 26 00:01:50,905 --> 00:01:54,605 And so I spend a lot of time just thinking about how the mind works and 27 00:01:54,985 --> 00:01:57,165 how trustworthy is our self reporting. 28 00:01:58,080 --> 00:01:58,820 So clinically, 29 00:01:59,200 --> 00:02:02,900 I'm certainly not alone in dealing with that question. 30 00:02:04,800 --> 00:02:05,300 But 31 00:02:05,840 --> 00:02:11,655 some of what I would say, I'd you know, I've been obsessed probably with for the last decade 32 00:02:12,114 --> 00:02:14,454 in the on the quest of web of trust 33 00:02:15,235 --> 00:02:18,114 and how do we obtain trusted information. How do we 34 00:02:18,834 --> 00:02:20,614 to me, freedom technology 35 00:02:20,915 --> 00:02:22,454 is about how do we agree 36 00:02:24,010 --> 00:02:27,950 on any whatever on a different set of rules. Rough consensus and running code. 37 00:02:28,569 --> 00:02:30,990 How do we how do we generate consensus 38 00:02:31,370 --> 00:02:35,310 to solve whatever problem? And there are various different problems I'm sure we're we'll talk about. 39 00:02:37,415 --> 00:02:45,595 But so I'm I, a lot of times, have my engineering hat on trying to build tools to solve problems for real people. But then sometimes I'll put on 40 00:02:46,135 --> 00:02:49,435 my scientist hat and I'll think, You know what? I think the brain 41 00:02:50,055 --> 00:02:51,834 has the exact same problem. 42 00:02:53,470 --> 00:02:56,530 We people, just in real life, face the same problem 43 00:02:56,989 --> 00:02:57,890 as we 44 00:02:58,190 --> 00:03:02,610 freedom technologists solve, which is how we arrive at So just something as simple as language. 45 00:03:03,310 --> 00:03:05,890 Who decides to call, This is you. 46 00:03:06,430 --> 00:03:08,209 Nobody. There's no one in charge. 47 00:03:09,625 --> 00:03:11,165 And and I speculate. 48 00:03:11,545 --> 00:03:18,845 So I've been, you know, with my engineering hat. I'm like, here are the tools I think that are gonna solve. How do we build a platform with no one in charge? 49 00:03:20,825 --> 00:03:22,445 I have lots of thoughts on that. 50 00:03:22,850 --> 00:03:30,069 And and as I'm, like, building the the big picture and the little picture, I'm like, you know what? Maybe the brain does the same 51 00:03:30,770 --> 00:03:34,150 strategy. So if you ask where does where is 52 00:03:34,530 --> 00:03:35,670 trust in the brain, 53 00:03:35,975 --> 00:03:39,915 there is no place. Like, if you ask me which neuron is in charge, 54 00:03:40,935 --> 00:03:42,635 there is none. And, 55 00:03:43,015 --> 00:03:46,075 you know, it's not the pineal gland is not in charge. There's there's, 56 00:03:47,015 --> 00:03:49,195 Is there a and I I asked that question 57 00:03:49,655 --> 00:03:50,155 knowing, 58 00:03:51,050 --> 00:03:54,510 as as little as I know from a neurological standpoint that it's not simple. But 59 00:03:54,810 --> 00:03:56,270 but to sort of tease out, 60 00:03:57,050 --> 00:03:58,270 if not a place, 61 00:03:58,810 --> 00:03:59,310 certainly, 62 00:03:59,610 --> 00:04:00,110 then 63 00:04:01,130 --> 00:04:05,310 is there an understood process? How do we process trust? How do we 64 00:04:06,095 --> 00:04:07,475 derive or arrive 65 00:04:08,095 --> 00:04:14,835 at at what or who we trust? Yeah. That I actually don't know. There's definitely no no spot, and, 66 00:04:15,215 --> 00:04:19,715 you know, there probably are people who think about that, you know, who, you know, probably psychologists, 67 00:04:20,015 --> 00:04:20,515 psychiatrists. 68 00:04:20,960 --> 00:04:22,180 So there's no clear blueprint 69 00:04:22,560 --> 00:04:26,180 to take from your field of of medicine and science 70 00:04:26,880 --> 00:04:34,240 to apply to web of trust as a protocol clear blueprint that I know of. The analogies that I come up with are things that 71 00:04:34,895 --> 00:04:40,275 yeah. I don't know if any of that. Primitives? I mean, what what do you What are the primitives of trust? When you are mapping 72 00:04:41,215 --> 00:04:45,555 trust onto a protocol, onto software, onto code Yeah. 73 00:04:45,855 --> 00:04:46,915 Are there clear 74 00:04:47,690 --> 00:04:48,190 buckets 75 00:04:48,570 --> 00:04:58,750 Types of trust. Or or primitives that you think about that you need to map? So when I try to do that, I come to realize that there is an there's no cap on the different categories of trust. 76 00:04:59,210 --> 00:05:05,985 The word itself can make conversations about this very difficult because I might say I trust you to do something, 77 00:05:06,845 --> 00:05:11,825 or I might say I trust that something is true. It's a completely different kind. Contextually 78 00:05:12,445 --> 00:05:15,105 The word is being used in two different very different ways. 79 00:05:15,840 --> 00:05:22,020 And there's kind of even a continuum. I trust that this is true because I trust you to be truthful 80 00:05:22,480 --> 00:05:25,620 on that topic. Right. I think trust is contextual. 81 00:05:26,000 --> 00:05:28,000 And and, yeah, like and I don't 82 00:05:28,720 --> 00:05:30,900 as far as what are the different categories, 83 00:05:31,444 --> 00:05:34,025 that is something that's not for the developer 84 00:05:34,485 --> 00:05:41,785 to decide. We need tools so that your community can tell you help you to decide what the categories are. It's emergent. 85 00:05:42,164 --> 00:05:42,985 Yeah. Now 86 00:05:43,460 --> 00:05:44,360 talking about the 87 00:05:44,900 --> 00:05:45,720 brain, you know, 88 00:05:46,020 --> 00:05:48,440 how does the brain I think of the brain as decentralized. 89 00:05:49,140 --> 00:05:54,200 If you do a corpus chaliceotomy, which is sometimes done, you cut you separate the left and the right hemispheres, 90 00:05:55,220 --> 00:05:58,520 then for so which is sometimes done for a very 91 00:05:58,825 --> 00:06:00,285 severe refractory epilepsy. 92 00:06:01,225 --> 00:06:02,445 Then you 93 00:06:02,985 --> 00:06:17,720 and some you there are patients who where this is studied. There is very interesting because they don't communicate with each other. You can you can send it put information into the one hemisphere, and the other hemisphere doesn't know it. You can say, 94 00:06:18,259 --> 00:06:24,759 draw I want you to draw something, and the what you draw off your right hand and what you draw off your left hand are not gonna be the same, 95 00:06:25,875 --> 00:06:26,375 because 96 00:06:26,755 --> 00:06:30,615 probably it's your left hemisphere that understands the words, draw something. 97 00:06:30,995 --> 00:06:34,055 But I maybe I'll say copy the image that you see. 98 00:06:34,515 --> 00:06:39,190 So you took an images. And it's almost like there are two people. And yet, 99 00:06:39,570 --> 00:06:42,550 patients like this can go to the grocery store. They can 100 00:06:43,090 --> 00:06:48,870 they can not 100% normal, but can function like one person. How is it that you have 101 00:06:51,965 --> 00:06:56,305 this thing, this brain, you know, that stores information and processes information 102 00:06:56,845 --> 00:06:58,944 where there isn't anything in charge 103 00:07:00,284 --> 00:07:02,384 seem like one unit? And 104 00:07:03,289 --> 00:07:04,669 so as Freedom Technologists, 105 00:07:04,970 --> 00:07:07,710 we want tools so that in peep so that 106 00:07:08,169 --> 00:07:08,669 people, 107 00:07:09,930 --> 00:07:10,430 users, 108 00:07:10,970 --> 00:07:11,710 we can, 109 00:07:12,410 --> 00:07:18,349 in some ways, act as a unit to come to agreement in a way that doesn't require a leader, a king. 110 00:07:19,355 --> 00:07:21,215 So it is not so much that you're 111 00:07:22,235 --> 00:07:22,735 emulating 112 00:07:23,675 --> 00:07:25,695 or need to emulate an emergent 113 00:07:26,155 --> 00:07:27,535 sort of feature or function, 114 00:07:28,155 --> 00:07:30,015 of the brain so much as 115 00:07:32,540 --> 00:07:34,480 the technology can be informed, 116 00:07:35,260 --> 00:07:40,160 I guess, by some of the limitations. I mean, I think of Dunbar's number. I I don't know if there are other 117 00:07:41,180 --> 00:07:44,160 factors or features like that that would 118 00:07:44,845 --> 00:07:51,585 maybe put boundaries around something like web of trust? And and, actually, let's pause, and if you would, give us a primer 119 00:07:51,965 --> 00:07:55,345 on web of trust itself. Web of trust itself. So the, 120 00:07:57,110 --> 00:07:59,370 so the way the way I think of freedom technology 121 00:08:00,070 --> 00:08:04,490 in general, the problem is how do we arrive at consensus on solving different problems? 122 00:08:04,790 --> 00:08:08,570 So Bitcoin, the problem is how do we arrive at consensus on money? 123 00:08:10,150 --> 00:08:10,650 Nostr, 124 00:08:11,225 --> 00:08:14,205 how do we arrive at consensus on building platforms? 125 00:08:16,905 --> 00:08:17,405 And 126 00:08:17,785 --> 00:08:27,919 web of trust. So I I think the goal of what we're trying to build I I go big picture, you know, nuts and bolts, what are we trying to build, is to build what I want 127 00:08:28,460 --> 00:08:30,479 users to have that I'm trying to build 128 00:08:30,940 --> 00:08:32,400 is a store of information 129 00:08:33,260 --> 00:08:34,880 that you and your community 130 00:08:35,500 --> 00:08:36,720 can identify 131 00:08:38,885 --> 00:08:46,185 who is the most trustworthy to help you curate what's not only what's in the store of information, but also how it's structured. And so I can effectively 132 00:08:47,285 --> 00:08:48,265 I can outsource, 133 00:08:49,525 --> 00:08:57,810 in a managed fashion Mhmm. Who should I trust about what in what context? Yes. And the context, your web of trust, 134 00:08:58,350 --> 00:09:04,770 is gonna help you manage even what the contexts are. So you have a data store, and then you have your web of trust, which I call a grapevine. 135 00:09:05,230 --> 00:09:08,915 But these two things, what So you have one problem 136 00:09:09,375 --> 00:09:14,435 that you are now breaking apart into two problems. What is the nature of the store of information? 137 00:09:14,895 --> 00:09:18,915 What is the nature of the web of trust? How do you build each one of these things, 138 00:09:19,790 --> 00:09:29,569 and how do you build them so that they work together? So the design of each one needs to be guided by how does it fit with the other. And the and the the web of trust, 139 00:09:31,069 --> 00:09:33,889 certainly, I think I think the name itself is quite helpful. 140 00:09:34,665 --> 00:09:38,285 It is the the web of interconnectedness 141 00:09:38,825 --> 00:09:39,885 to individuals, 142 00:09:41,545 --> 00:09:50,200 for or on which I have certain levels of trust. Now the data store, maybe unpack that a bit. Is this the are these the subjects, the topics, the, 143 00:09:51,060 --> 00:09:51,560 taxonomy? 144 00:09:52,100 --> 00:09:59,960 So I'll give you an example of what I'm building right now. The the next iteration of what I'm building, which I'm not sure what I'm gonna call it. I think I'm gonna call it Brainstorm. 145 00:10:00,615 --> 00:10:15,755 Brainstorm is a personal web of trust relay for Nostr. Okay. This data store is a stir fry relay, which is one of the arguably the the most performant of the open source of of the Nostr relay. And broadly, we can say this is a 146 00:10:16,080 --> 00:10:17,860 self hosted or self managed, 147 00:10:19,120 --> 00:10:21,780 component that would otherwise live in Meta's 148 00:10:22,240 --> 00:10:24,180 data center, but that we have chosen 149 00:10:24,720 --> 00:10:47,860 to operate ourselves so we have more. So so ideally, this would be a database that you know, I'd love to to see you running Brainstorm on your phone. Mhmm. The Relay is on your phone or it's on your laptop. You don't have to. There could you could outsource the service, but the ability would be this is a database that you that you have full control over, everything about it, where it is, what it does. 150 00:10:48,480 --> 00:10:52,180 And, you know, the the web of trust algorithms, you have final say. 151 00:10:52,720 --> 00:10:52,959 They 152 00:10:53,839 --> 00:10:56,579 and that would be curated by people you trust 153 00:10:57,384 --> 00:10:58,605 and people they trust. 154 00:10:59,144 --> 00:11:12,285 So what is web of trust? Actually, I like to call it webs of trust. And I think Odell has probably said I think I've heard him say the same thing because there are different types of trust. Right. There's different ways to calculate trust. There's different 155 00:11:13,430 --> 00:11:13,930 purposes, 156 00:11:14,550 --> 00:11:18,650 you know, to you know, things that we're gonna do with the things we calculate, 157 00:11:19,190 --> 00:11:31,225 which makes it sound like a intractable problem, but it's not. I don't believe it's a intractable problem. It does it does feel, as we've talked before, that it could it could it could fractal outward infinitely. And frequently, this is what happens. 158 00:11:31,525 --> 00:11:35,305 People think, oh, here's a problem. Web of stress can solve it. 159 00:11:35,685 --> 00:11:37,865 You know, like If all you have is a hammer. 160 00:11:38,485 --> 00:11:41,490 Well, actually, no. We don't have a hammer. That's 161 00:11:42,370 --> 00:11:47,910 is kind of search of it. There are just so many you know, I wanna know what are the most trustworthy 162 00:11:48,529 --> 00:11:49,990 eCash mints to use. 163 00:11:50,370 --> 00:11:53,830 And, well, I know some I want the people who I follow 164 00:11:54,335 --> 00:12:04,835 to tell me who are the most trustworthy eCash Mints. So we're gonna build this put that in code. We're gonna build it so you can do that. That's an example of a problem you would use. Or if we if we zoom out perhaps, 165 00:12:06,335 --> 00:12:06,750 which 166 00:12:07,150 --> 00:12:14,770 publication, which website, which news organization should I look to for a particular set of opinions about a particular set of subjects, 167 00:12:15,150 --> 00:12:16,370 which health care institution, 168 00:12:18,110 --> 00:12:19,810 to something I presume more trivial, 169 00:12:20,270 --> 00:12:20,770 which 170 00:12:21,835 --> 00:12:43,680 bicycle helmet should I buy? Right. Exactly. There is I mean, there there are so many different problems. With different clearly different stakes. With different stakes. And what what often happens is that a group of developers get together and users and, like, oh, alright. Well, let's let's make it where you can do a rating, one to five or zero to five. And then, oh, let's make it just thumbs up, thumbs down. 171 00:12:44,060 --> 00:12:54,295 And the and then, you know, there you get to be so many more so many details that people can't agree on all the details, and it falls apart. Cannot be reached. It gets too complicated, 172 00:12:54,755 --> 00:13:00,855 too fast, and it falls apart. And this is a this is a I've seen this happen over and over and over and over and over again. 173 00:13:01,714 --> 00:13:05,050 It can so how do we deal with the complexity? What 174 00:13:05,589 --> 00:13:10,410 I believe that we need to deal with complexity is your web of trust, and this is what Grapevine 175 00:13:11,510 --> 00:13:15,290 is going to do. GrapeRank, that's the algorithm that it uses. It's extensible. 176 00:13:15,750 --> 00:13:18,330 Right. So the web, you you're gonna have 177 00:13:18,685 --> 00:13:23,265 as simple very simple web, you know, tools that just do weighted averages, basically. 178 00:13:24,125 --> 00:13:25,905 It will be extensible, meaning 179 00:13:26,285 --> 00:13:30,305 you can have create a someone who you trust, hopefully, 180 00:13:30,779 --> 00:13:33,920 or somebody who knows someone who can create a new algorithm 181 00:13:34,779 --> 00:13:35,279 and 182 00:13:35,740 --> 00:13:38,959 put that algorithm your database can have a table of algorithms. 183 00:13:39,339 --> 00:13:50,635 And then you can use this new algorithm to calculate a new type of stress score. I can do that myself, but I'm probably most users can be like, I I love it for other people to help me to do it. 184 00:13:51,255 --> 00:14:01,259 So when it's so so question, you know, what kind of what's the nature of the database? What's the nature of the web of trust? So just trying to build it. Is the database gonna be a relational database? 185 00:14:01,720 --> 00:14:04,620 I actually think graph database is the way to go, but 186 00:14:05,319 --> 00:14:08,620 I don't think there's just like there's not one hashing algorithm 187 00:14:09,079 --> 00:14:10,300 that will work to 188 00:14:10,845 --> 00:14:11,505 to implement 189 00:14:12,045 --> 00:14:17,824 Nakamoto consensus. Right. Maybe maybe shot you know, what we what we have is the best, but we could argue. 190 00:14:18,445 --> 00:14:19,985 And it is there in in 191 00:14:21,165 --> 00:14:22,704 for those perhaps less technical, 192 00:14:23,880 --> 00:14:37,180 you know, today, everyone's got a documents folder on their laptop. Yeah. I've got my iCloud Drive on my phone. Yeah. It is a given that I have a personal store of data. Yeah. In this case, documents. Mhmm. And so 193 00:14:37,745 --> 00:14:42,805 if we go forward along the path that I think you envision and you've been working on for for a decade plus, 194 00:14:43,425 --> 00:14:48,084 we simply have another just to zoom out, we have another data store. We have another 195 00:14:50,740 --> 00:14:51,880 component of technology, 196 00:14:52,340 --> 00:14:52,840 laptop, 197 00:14:53,780 --> 00:14:55,640 phone running in the cloud under our control 198 00:14:56,660 --> 00:14:57,640 that manages 199 00:14:58,820 --> 00:15:03,080 the subject matter, the data store Mhmm. The context, and then 200 00:15:03,620 --> 00:15:04,920 the ultimately 201 00:15:05,915 --> 00:15:06,415 presentation 202 00:15:07,435 --> 00:15:09,615 of, Dave, you should, David, you should 203 00:15:09,995 --> 00:15:16,334 presumably look to this person on this subject. You should look to this person on this. Or you can just query it. Right? And say, 204 00:15:16,714 --> 00:15:20,579 I want the best Thai food in DC. Yeah. Who should I message? Right. 205 00:15:20,880 --> 00:15:25,060 And so so we fast forward and and and if this is successful, 206 00:15:26,000 --> 00:15:30,660 then we have this just running. I mean Mhmm. You know, we could use various, I think, terms. 207 00:15:31,040 --> 00:15:31,700 I was, 208 00:15:32,240 --> 00:15:34,500 another conversation interview that I had yesterday 209 00:15:35,175 --> 00:15:36,235 was with, 210 00:15:38,775 --> 00:15:42,155 Marks, who's cofounder CEO of Open Secret. We were talking 211 00:15:42,695 --> 00:15:43,195 about 212 00:15:44,135 --> 00:15:45,275 artificial intelligence 213 00:15:45,655 --> 00:15:46,475 and the 214 00:15:47,170 --> 00:15:51,110 cyclical nature of, you know, what is old is new again. And in the nineties, 215 00:15:51,650 --> 00:15:56,630 intelligent agents now and now we have agents again. Right? And maybe they'll work this time. But my point to all that being, 216 00:15:57,090 --> 00:16:05,785 we could probably generalize this to an agent of sorts Mhmm. That is just running in the background somewhere on our behalves Yeah. Doing the research, 217 00:16:06,245 --> 00:16:08,185 strengthening the scores, finding, 218 00:16:08,805 --> 00:16:12,709 you know, the experts and and those in our web of trust who 219 00:16:13,569 --> 00:16:18,470 the the system would bubble up Yeah. Suggestions on. But I go through all that to say, 220 00:16:19,970 --> 00:16:22,870 let's talk about where we are today 221 00:16:23,329 --> 00:16:24,949 in terms of how we 222 00:16:26,125 --> 00:16:27,185 approximate trust 223 00:16:28,285 --> 00:16:32,944 and and the bridge that we have to cross to get to web of trust. So so 224 00:16:33,404 --> 00:16:34,225 what are the 225 00:16:34,845 --> 00:16:38,704 current state of affairs Mhmm. That are going to be altered 226 00:16:39,069 --> 00:16:46,050 and improved hopefully by web of trust. So scenarios perhaps. And, again, I raised, you know, shopping for a bike helmet. Yeah. Finding 227 00:16:46,430 --> 00:16:48,290 a physician for for a particular 228 00:16:49,310 --> 00:16:50,930 problem I need to solve. Yeah. 229 00:16:51,264 --> 00:17:09,360 So so we have lots of different solutions for lots of different types of problems. So what what we do is we go to Yelp Right. Or we go to Amazon, and and Amazon tells us this is the best product, and you trust that or you don't. Yeah. And are the how many of the reviews are doctored and botched and Right. Which is everyone knows that's a big problem. 230 00:17:09,980 --> 00:17:31,934 Amazon spends tons of money trying to get rid of the the BotFox. Firefox is an extension, you know, that they do a third party. Mhmm. They'll they'll normalize the Amazon score. Yeah. And often, it's pretty shocking. You know? Yeah. A five star actually ends up being a c minus by way of Mozilla's sort of extension as in Yeah. And so so we use these tools because they're better than no tool. Right. But 231 00:17:32,330 --> 00:17:36,990 we have nothing better. And so what our goal in is in is Freedom Technology 232 00:17:37,290 --> 00:17:38,510 Builders is to build 233 00:17:38,970 --> 00:17:40,030 the next best thing. 234 00:17:40,490 --> 00:17:40,990 So, 235 00:17:41,290 --> 00:17:43,630 yeah, then then all of those services. 236 00:17:44,090 --> 00:17:45,630 Yeah, to answer sometimes 237 00:17:45,985 --> 00:17:54,325 questions that are just subjective, what's the best coffee shop Right. But sometimes very important. You know? Who's the best doctor in some specialty? 238 00:17:55,664 --> 00:17:59,690 What's actually happening in the world? Do you trust ABC News? Do you trust this 239 00:17:59,990 --> 00:18:01,929 news organization or that news organization? 240 00:18:03,030 --> 00:18:04,570 Those are the only ways 241 00:18:05,149 --> 00:18:05,690 you know, 242 00:18:06,230 --> 00:18:08,410 right now, we have only basically centralized 243 00:18:08,870 --> 00:18:10,330 entities to go through, 244 00:18:10,630 --> 00:18:13,929 but we often don't know how are they filtering the information, 245 00:18:14,765 --> 00:18:18,625 and can we trust them. Right. And You and I talked earlier there. 246 00:18:19,804 --> 00:18:21,424 There's a piece I read this morning. 247 00:18:21,804 --> 00:18:22,465 I think, 248 00:18:22,765 --> 00:18:25,664 I'll find it linked to it. It's the free press, organization. 249 00:18:27,030 --> 00:18:33,050 They referenced a subsegment of Gen z who went through lockdowns Mhmm. And who have come out the other side 250 00:18:33,750 --> 00:18:40,090 trusting nothing Yeah. Effectively. And, you know, this is my generalization. But what they referenced and and it was an interesting term 251 00:18:40,470 --> 00:18:40,970 that 252 00:18:41,705 --> 00:18:42,525 these individuals 253 00:18:43,145 --> 00:18:46,605 who went through formative years in lockdown had only the Internet 254 00:18:46,905 --> 00:18:48,125 as their as their, 255 00:18:48,825 --> 00:18:49,325 companion 256 00:18:50,505 --> 00:19:03,160 would now rather, as they say, magpie Mhmm. And source this information broadly Yeah. And find ways to reach consensus among their peers Yep. As to what is real or true or valuable Yes. And have wholly discarded 257 00:19:03,540 --> 00:19:07,960 traditional institutions. Right. And and tools that do exactly that are what 258 00:19:08,420 --> 00:19:13,025 I want to what I'm attempting to build Right. And that we will build. 259 00:19:13,485 --> 00:19:15,665 So, you know, I didn't grow up through the lockdowns, 260 00:19:16,205 --> 00:19:23,265 but I also don't trust our institutions nearly as much as five, ten, fifteen years ago. Absolutely. So I think they have good reason 261 00:19:24,080 --> 00:19:24,820 for that. 262 00:19:25,280 --> 00:19:32,420 Skepticism is is is well founded. Yeah. And and and so with this as the next major sort of hurdle 263 00:19:32,960 --> 00:19:33,700 in how 264 00:19:34,640 --> 00:19:35,460 we arrive 265 00:19:35,840 --> 00:19:36,340 at 266 00:19:38,715 --> 00:19:39,695 who who 267 00:19:40,075 --> 00:19:45,055 or what we trust about what. Mhmm. What are the major hurdles to get there? What are the 268 00:19:47,515 --> 00:19:50,255 difficult to intractable problems that you're 269 00:19:51,400 --> 00:20:03,660 knocking your knuckles on now and trying to solve or others are working to solve? What what what's in the way of getting getting there? So, really, for me, it's just a matter of enough time to build it because I'm not you know, 270 00:20:04,120 --> 00:20:12,515 being a developer was not my career. Right. And it's not what I dreamed about. I wasn't eight years old, you know, and and dreaming about maybe building computers 271 00:20:13,055 --> 00:20:16,195 or writing assembly code or anything like that. So I'm, 272 00:20:17,295 --> 00:20:18,675 that's the slowest thing. 273 00:20:19,390 --> 00:20:21,230 I mean And are you are you 274 00:20:21,950 --> 00:20:24,770 certainly, I appreciate that. Are there 275 00:20:26,429 --> 00:20:32,610 we talked about PGP early nineties, you know, as perhaps one of the first implementations or approaches 276 00:20:32,990 --> 00:20:34,290 to a web of trust. 277 00:20:36,174 --> 00:20:37,154 Are the domains, 278 00:20:37,934 --> 00:20:46,034 be it the science, the computer science, the algorithms, are they well established? I mean, so, you know, to what degree can you stand on the shoulder of giants, or to what degree must you 279 00:20:46,414 --> 00:20:48,835 invent a lot of this from scratch? I think it's, 280 00:20:50,360 --> 00:20:52,700 I think it's a whole new thing. Like, imagine 281 00:20:53,080 --> 00:20:55,980 if you were trying to build Bitcoin, if you were Satoshi, 282 00:20:56,360 --> 00:21:01,095 and you asked the same question. You know? What would he say? He would have said 283 00:21:01,655 --> 00:21:04,875 difficulty adjustment. I just thought of that yesterday in the shower. 284 00:21:05,495 --> 00:21:13,674 Yeah. He did have eCash. He did have Right. So definitely Chami and eCash, etcetera. Yeah. We did. You know, there's that timeline showing all the component. 285 00:21:14,660 --> 00:21:20,760 Models. Pieces to the puzzle. Yeah. And I think that, you know, Bitcoin is one of those if you have nine out of the 10 pieces of the puzzle, 286 00:21:21,299 --> 00:21:26,039 it doesn't work 90% of the way. Still some critical breakthroughs there. It doesn't work. Yeah. It's 287 00:21:26,419 --> 00:21:28,985 the last piece and and then putting it all together. 288 00:21:29,685 --> 00:21:30,585 And and, 289 00:21:31,605 --> 00:21:32,985 avoiding the temptation 290 00:21:33,445 --> 00:21:38,825 to do ad bells and whistles Mhmm. Because this is the nature of freedom technology is that, 291 00:21:39,925 --> 00:21:44,230 in order to you have to anybody can walk away from it. You know, if 292 00:21:45,170 --> 00:21:50,630 if Satoshi had made Bitcoin, the consensus rules more complex than they were. If he had added more, 293 00:21:52,050 --> 00:21:52,550 parameters 294 00:21:53,090 --> 00:21:57,510 to make it so that he can add more, he or she or whoever, make more, 295 00:21:58,130 --> 00:21:58,630 features. 296 00:21:59,175 --> 00:22:04,235 That way, every additional parameter is one more thing that people can disagree about. Or stumble across. 297 00:22:04,695 --> 00:22:05,195 Stumble, 298 00:22:05,655 --> 00:22:08,635 one thing. Just fight over. Fight about. I mean, one parameter, 299 00:22:09,095 --> 00:22:10,395 the block size 300 00:22:10,775 --> 00:22:11,675 almost completely 301 00:22:12,070 --> 00:22:15,770 you know, destroyed. We had we had block size wars about that. 302 00:22:16,550 --> 00:22:29,445 Every arbitrary parameter that you add is a reason for people to say, well, I would have used a different value for that because of this and that and the other. And then your community never forms. Right. And it and the entire endeavor fails. 303 00:22:30,065 --> 00:22:30,565 So 304 00:22:31,424 --> 00:22:36,725 in, you know, in that case, I think it was not only getting all the pieces of the puzzle together 305 00:22:37,050 --> 00:22:42,430 and figuring out how they fit together, but avoiding the temptation to do anything more 306 00:22:42,850 --> 00:22:43,350 and 307 00:22:43,770 --> 00:22:45,710 find and figuring out what that means. 308 00:22:46,490 --> 00:22:49,550 And so in web of trust. Well, I'm trying 309 00:22:49,945 --> 00:22:50,685 to learn 310 00:22:51,385 --> 00:22:52,445 from that example. 311 00:22:52,985 --> 00:22:54,285 I want the simplest 312 00:22:56,345 --> 00:23:00,365 product that will be a store of information, a web of trust, extensible 313 00:23:01,145 --> 00:23:06,809 so that because if it's extensible And I wanted to make it as simple and bare bones as possible, 314 00:23:08,390 --> 00:23:10,809 knowing that if it's extensible, then 315 00:23:11,510 --> 00:23:13,450 you can start to add more trust algorithms. 316 00:23:13,750 --> 00:23:19,245 Let's say you start out with just a very simple database, a couple of tables, table of users, table of events. 317 00:23:19,545 --> 00:23:23,885 Your web of trust is going to be able to hopefully keep out the spam. That's all it does. 318 00:23:24,425 --> 00:23:29,965 But it has the ability to make you know, your web of trust can say, you know what? You need another table of trust algorithms. 319 00:23:30,345 --> 00:23:31,885 You need another table of 320 00:23:32,770 --> 00:23:33,490 e cash, 321 00:23:33,810 --> 00:23:36,470 mints, another table of this, that, and the other, and 322 00:23:37,250 --> 00:23:37,990 it can 323 00:23:38,450 --> 00:23:38,950 become 324 00:23:39,410 --> 00:23:40,230 as sophisticated 325 00:23:40,690 --> 00:23:43,190 as it needs as as you want it to be. 326 00:23:44,130 --> 00:23:54,924 But you have to this is what I I'm trying to you know, the the discipline is to avoid the temptation to add all the bells and whistles. It's don't add those things that that need to be done because 327 00:23:55,225 --> 00:23:57,164 your community has to do them. 328 00:23:57,640 --> 00:24:07,580 Because I could say, here's the best way to categorize trust, but then, you know, you put two developers in a room on how to categorize trust, you're gonna get five different opinions. 329 00:24:08,200 --> 00:24:09,880 So your community has to 330 00:24:10,755 --> 00:24:17,495 that's the only way to make it make it happen. And is that I mean, you and I spend a lot of time on using and talking about Nostr. 331 00:24:19,315 --> 00:24:22,055 There's a lot of development going on within 332 00:24:22,890 --> 00:24:24,030 or on Noster. 333 00:24:26,330 --> 00:24:29,070 Is is it I I I I don't necessarily 334 00:24:29,450 --> 00:24:29,950 have 335 00:24:30,570 --> 00:24:34,830 enough insight into other domains to know if web of trust work is developing 336 00:24:36,005 --> 00:24:38,024 well Mhmm. Or rapidly 337 00:24:38,404 --> 00:24:39,865 elsewhere. Mhmm. So 338 00:24:40,325 --> 00:24:50,585 is is it the hot spot? Is Nostra the hot spot for web of trust? Nostra is definitely the hot spot for web of trust. So before I discovered Nostra, I was working on IPFS, a different technology base. 339 00:24:51,090 --> 00:24:51,590 And, 340 00:24:52,049 --> 00:25:00,630 but I when I there people are definitely excited about Nostra and optimistic about it and building tools that are actually working. And what's your take on why? 341 00:25:01,010 --> 00:25:02,230 Why Nostra? Why 342 00:25:02,770 --> 00:25:06,390 are people excited? Why is So I I think that the reason that 343 00:25:06,735 --> 00:25:10,195 Nostr is the place to be is because Fiat Jaf 344 00:25:10,895 --> 00:25:16,434 made Nostr as simple as possible. Exactly what I was just talking about. Bitcoin 345 00:25:17,535 --> 00:25:20,950 was complex enough to do the thing and no more complicated. 346 00:25:21,970 --> 00:25:27,590 And then all the complication is on layer two, layer three, layer four. This is what Fiat Joffe did. Nostra 347 00:25:28,290 --> 00:25:28,790 is 348 00:25:29,570 --> 00:25:32,870 as simple bare bones as it can possibly be to 349 00:25:33,575 --> 00:25:35,675 implement the vision, and then the complexity 350 00:25:36,135 --> 00:25:40,555 is added on top of those. So I think this is why Nostra is working. 351 00:25:41,575 --> 00:25:45,195 And and and this I mean, this is why I'm so excited about it. And 352 00:25:47,800 --> 00:25:52,220 And and and is there a clear sense in your mind of why it lends itself 353 00:25:53,160 --> 00:25:55,500 to solving the web of trust problem? 354 00:25:55,960 --> 00:25:56,460 Well, 355 00:25:59,000 --> 00:26:03,180 you know, because it's not a cathedral. It might Mhmm. Maybe just what I'm I 356 00:26:03,855 --> 00:26:06,755 was just trying to make a book. The Cathedral and the Bazaar. 357 00:26:07,695 --> 00:26:09,535 Oh. If that's a rep yeah. So it's, 358 00:26:11,695 --> 00:26:30,130 I'm suddenly drawing a blank on the author, but it it it it presents the two approaches Uh-huh. To build the cathedral versus an open bazaar. Yeah. And Noster is the bazaar for sure. Right. And, you know, blue sky would be the cathedral. Right. So the fact that Noster is the bazaar, it's as simple as possible. It allows people to experiment. Right. And 359 00:26:30,775 --> 00:26:35,995 and I Yeah. And people are experimenting with so many different things. The portability of the social graph. 360 00:26:37,255 --> 00:26:39,355 Yes. You have all these different 361 00:26:39,655 --> 00:26:41,195 apps with all these 362 00:26:41,655 --> 00:26:42,155 different 363 00:26:43,830 --> 00:26:44,330 purposes. 364 00:26:44,630 --> 00:26:49,450 But imagine if you could port your identity and your social graph between 365 00:26:49,830 --> 00:26:51,850 Twitter and Facebook and Instagram. 366 00:26:52,470 --> 00:27:03,905 Nostra can do that. Right. And different people can so that lends itself to the web of trust because of the portability. Right. So yeah. And so and and and, again, for those who may be coming up to speed or new, 367 00:27:05,005 --> 00:27:06,865 the social graph is 368 00:27:07,804 --> 00:27:17,990 a crux or key to a function in web of trust? Yeah. So and so the social graph for Nostr is usually people think of follows. Mhmm. You know, I follow you, 369 00:27:18,450 --> 00:27:25,910 and you you know, I follow 500 people. You follow 500, whatever, a thousand people. And and the people who have implemented web of trust, 370 00:27:26,495 --> 00:27:26,995 are 371 00:27:27,615 --> 00:27:31,795 doing basically, that's the raw data Right. Follows. Those are the strongest signals 372 00:27:32,175 --> 00:27:33,315 we have or they 373 00:27:33,775 --> 00:27:36,275 have determined that we have? Those are, 374 00:27:37,615 --> 00:27:40,975 I think that we are we are accustomed to using follows data. That's what legacy social media uses. And, it's, 375 00:27:41,790 --> 00:27:42,290 social 376 00:27:42,990 --> 00:27:44,850 media uses. And, 377 00:27:45,790 --> 00:27:52,450 it's you know, the data is available, and it means something. You know, if I follow you, I've got skin in the game. I've chosen I've supposedly 378 00:27:52,990 --> 00:27:58,145 followed you. It's not an empty gesture. I'm actually gonna see your content on my feed, so 379 00:27:58,605 --> 00:28:00,385 skin in the game. It means something. 380 00:28:00,924 --> 00:28:13,970 And and this is a conversation that I that I see happening in Nostra. I've seen nowhere else is what should be the best raw data to use to calculate how much you trust somebody in a context, certainly in a given context. 381 00:28:15,390 --> 00:28:18,130 The obvious critique of using follows 382 00:28:18,430 --> 00:28:18,930 as 383 00:28:19,390 --> 00:28:31,475 a trust indicator is that I might follow you. It doesn't mean I trust you. You know? I might I might wanna see your content, but I'm not gonna trust your advice, financial advice Right. Or this or that or the other. And it certainly doesn't 384 00:28:32,415 --> 00:28:34,735 lend itself to contextual trust. You know? 385 00:28:35,455 --> 00:28:37,555 I can't say I followed you for this reason. 386 00:28:38,180 --> 00:28:39,080 And so 387 00:28:39,700 --> 00:28:42,440 so many people in Nostra, we we've had conversations. 388 00:28:42,820 --> 00:28:54,855 Follow doesn't mean I trust you. Mhmm. So where do we go next? Since I'm interested in what you have to say. I'm interested in what you have to say, but I but I but if I want you know, trust needs to be context I I wanna say I trust you 389 00:28:55,315 --> 00:29:06,294 for this context. Because I follow you or rather that I follow you because you post great memes Mhmm. Does not mean Does not mean I mean I wanna hear your opinions about I wanna know events. Who are the best 390 00:29:06,620 --> 00:29:08,160 people to tell me about 391 00:29:08,700 --> 00:29:13,360 the overlap of Bitcoin and monetary policy. I want my I want 392 00:29:14,620 --> 00:29:24,705 my web of stress to tell me that Lynn Alden is the best, even if I don't already know. She's two, three, four, five hops away from me. This is a topic I wanna know about. It's a needle in a haystack. Can you give me 393 00:29:25,085 --> 00:29:26,465 this needle in this haystack? 394 00:29:27,005 --> 00:29:27,505 And 395 00:29:27,805 --> 00:29:30,065 how do we do follows doesn't do that. So 396 00:29:33,085 --> 00:29:36,145 a year ago, I would have said what we need are explicit 397 00:29:36,820 --> 00:29:39,160 trust attestations where I can just say 398 00:29:39,620 --> 00:29:40,360 I trust, 399 00:29:40,900 --> 00:29:44,120 you know, Lynn. Mhmm. You know? And I'm very confident 400 00:29:44,660 --> 00:29:56,325 in this because I've, you know, read a lot of stuff. Or maybe I'm not confident because this is based on one tweet that she did. Sure. But, you know, 99 of a hundred, you know, this is why I trust her in this context. 401 00:29:57,184 --> 00:30:04,085 And then and and a lot of people will say we need explicit contextual trust attestations. That's better than 402 00:30:04,385 --> 00:30:07,125 follows because you don't really know what that means. 403 00:30:07,700 --> 00:30:16,840 And the obvious objection to that is that it would make for an awful user experience. Yeah. We're always in. We're just not gonna do it. Yeah. Yeah. For whatever reason, you know, 404 00:30:17,140 --> 00:30:22,520 anyone who's good at product is just gonna tell you no one's gonna do it. Bad UX. It's just not gonna work. 405 00:30:23,105 --> 00:30:24,805 For whatever thread that needle? 406 00:30:25,425 --> 00:30:25,925 So 407 00:30:26,705 --> 00:30:28,245 I believe that 408 00:30:28,865 --> 00:30:33,925 what we need to do is use all of the data that is available to us. 409 00:30:34,785 --> 00:30:35,525 It follows 410 00:30:35,825 --> 00:30:37,285 whatever if What other signals 411 00:30:38,309 --> 00:30:41,370 are there that you see are useful? So just as an example. 412 00:30:41,750 --> 00:30:45,210 The in in the personal WebTrust Relay that I'm building right now, 413 00:30:46,230 --> 00:30:48,090 it calculates a whitelist 414 00:30:48,630 --> 00:30:49,529 and a blacklist. 415 00:30:49,830 --> 00:30:50,570 And so 416 00:30:51,895 --> 00:30:55,435 that's based off of follows, but it's also based on mutes and reports. Right. 417 00:30:56,295 --> 00:30:59,515 And that's to make do a little better job of 418 00:30:59,975 --> 00:31:01,835 keeping the high quality content 419 00:31:02,295 --> 00:31:05,355 but getting rid of the spam and the bots. Right. 420 00:31:05,940 --> 00:31:17,080 So follows. That's and then then reports and mutes. But there's lots of other signals. You what if z what if you don't like Volos, and what if we'd rather do Zaps Mhmm. Or replies? Right. Or, 421 00:31:17,539 --> 00:31:19,159 But it's some sort of engagement metric. 422 00:31:19,485 --> 00:31:22,785 Right. Exact maybe something I've never thought of as a developer. 423 00:31:23,325 --> 00:31:24,625 And so the solution 424 00:31:25,405 --> 00:31:29,425 this thought process is what led me to design how the grapevine's gonna work. 425 00:31:29,805 --> 00:31:30,684 It will 426 00:31:31,340 --> 00:31:34,960 one of the most important steps in the grapevine is something called influence, 427 00:31:36,140 --> 00:31:37,360 or I'm sorry, interpretation. 428 00:31:37,820 --> 00:31:38,220 And, 429 00:31:40,300 --> 00:31:42,080 so imagine if someone you trust 430 00:31:42,380 --> 00:31:42,880 writes 431 00:31:43,180 --> 00:31:47,575 a script that will take whatever piece of data you think is useful, 432 00:31:48,435 --> 00:31:51,495 you know, replies or or zaps or something. 433 00:31:52,355 --> 00:31:53,415 And that 434 00:31:54,035 --> 00:31:57,735 takes the data and puts it into a format that is ready to be 435 00:31:58,035 --> 00:32:00,870 utilized by the grape rank calculation 436 00:32:01,170 --> 00:32:02,130 engine Mhmm. 437 00:32:02,610 --> 00:32:03,110 Then, 438 00:32:06,530 --> 00:32:15,845 then you can calculate a new kind of trust score using new using whatever data is possible. And so or whatever data exists, whatever data you want. And how by the way, David, how, 439 00:32:16,245 --> 00:32:20,184 it might make sense to take a moment and talk about grape ranks 440 00:32:21,924 --> 00:32:30,024 or PageRank, Google's PageRank's influence on GrapeRank. So Yeah. Give us give us a brief understanding of PageRank that many of us 441 00:32:30,460 --> 00:32:30,960 have, 442 00:32:31,500 --> 00:32:36,080 in effect, trusted for our web searches for fifteen plus years now. Yeah. 443 00:32:37,100 --> 00:32:39,120 How does it work at a high level, 444 00:32:39,500 --> 00:32:41,520 and how has that now influenced 445 00:32:41,820 --> 00:32:45,455 your idea for GrapeRank? I so I think that's a great question because, 446 00:32:47,115 --> 00:32:52,735 Nostra developers probably should have a greater appreciation for PageRank. It changed the world. In 447 00:32:53,275 --> 00:32:54,415 1998, 448 00:32:54,555 --> 00:33:00,950 it launched Google. Mhmm. Google is a trillion dollar company Yep. Because of this algorithm, which is a very simple algorithm, 449 00:33:01,649 --> 00:33:02,149 but 450 00:33:03,169 --> 00:33:06,710 got rid of the spam. I mean, how How many percentage of the population could explain it? I can't. 451 00:33:07,250 --> 00:33:09,269 I couldn't have until I started, really, 452 00:33:09,570 --> 00:33:12,870 thinking about this. All I knew I remember in 1998 453 00:33:13,315 --> 00:33:27,335 when keyword search if you did a key a search for whatever keyword, nine out of 10 results are just junk. Sure. Then PageRank comes, and nine out of 10 are what you're looking for. It was magic. All I knew is this was magic. Right. Get rid of the speed. Got the job done. The outcomes were great. 454 00:33:27,809 --> 00:33:33,190 Despite the fact that it's pretty simple. So the basic idea is, so instead of follows, it's 455 00:33:33,649 --> 00:33:35,590 calculating a score for every URL, 456 00:33:36,450 --> 00:33:42,230 and it's based on a hyperlink. If my website links to your website, it is an indication you're probably not spam. 457 00:33:42,825 --> 00:33:46,845 And so it's not a trust score. It's an important score. Right. It is not 458 00:33:47,145 --> 00:33:56,284 explicit. It's not an endorsement, but it is It's just a thing that's useful. It's a signal. And so if I link to your website, then it boosts your score. 459 00:33:56,779 --> 00:34:00,240 And if the if my score is higher, then it has 460 00:34:00,620 --> 00:34:03,840 a bigger boost of your score. Right. That's it. And so 461 00:34:04,220 --> 00:34:06,880 that's pretty simple. But it was ridiculously 462 00:34:07,340 --> 00:34:09,359 good at getting rid of spam. 463 00:34:09,975 --> 00:34:13,355 Not contextual, but it didn't have to be. So job one is, 464 00:34:14,535 --> 00:34:25,319 deflect the barbarians at the gate. Right. Hold back the spam, hold back the junk, and I'll you know, we could I'm sure have an interesting conversation about where we are today in that regard. And, you know, I use Kaji 465 00:34:26,339 --> 00:34:34,440 Yeah. As a result of not no longer, you know, wanting to see all the paid ads and the sponsored links, etcetera. But as you say, in the beginning, 466 00:34:34,740 --> 00:34:42,935 it was novel, and it was incredibly effective. Incredibly effective. Yes. And and and what we are finding in Nostr is just using follows 467 00:34:43,395 --> 00:34:46,855 is incredibly effective at getting rid of the spam. And so the 468 00:34:48,115 --> 00:34:54,380 the bridge from page rank to page rank to great rank So the problem that so I'm gonna aim it kind of 469 00:34:54,680 --> 00:34:56,220 in reference to page rank. 470 00:34:56,520 --> 00:34:57,260 Page rank, 471 00:34:58,200 --> 00:35:03,260 just calculates one number, not contextual. Mhmm. Grape rank is designed to be contextual 472 00:35:03,800 --> 00:35:06,859 and to be able to use data from any source. 473 00:35:07,595 --> 00:35:08,494 That's the main 474 00:35:09,194 --> 00:35:12,494 difference. Well well, I guess Rather than just an inbound link, 475 00:35:13,515 --> 00:35:14,015 to 476 00:35:14,474 --> 00:35:23,295 ascribe a certain PageRank score, you have the potential for multiple You can use signals. Follows. Or you can use like, I've already built, you can use 477 00:35:23,730 --> 00:35:25,110 reports and mutes, 478 00:35:25,650 --> 00:35:27,190 or you could use zaps. 479 00:35:27,570 --> 00:35:28,710 You could use whatever. 480 00:35:29,330 --> 00:35:36,070 You could interpret any piece of information that you see and that you want that you trust, that you think is gonna be useful 481 00:35:36,615 --> 00:35:39,675 and create a different score for a different category. 482 00:35:40,615 --> 00:35:43,515 And does this look like I mean, if we if we lay out, 483 00:35:44,215 --> 00:35:45,275 a bit of an ecosystem 484 00:35:45,895 --> 00:35:47,435 or layers from 485 00:35:49,655 --> 00:35:51,115 grape rank as an 486 00:35:51,780 --> 00:35:52,840 algorithm, as 487 00:35:53,460 --> 00:35:54,120 a machine 488 00:35:54,660 --> 00:35:56,440 to the end user who 489 00:35:57,620 --> 00:36:13,664 will probably never care deeply Mhmm. To understand all this. Exactly. Yeah. Are there are there obvious layers? So we've got an engine Mhmm. Grape rank. We have a it affects a web of trust. Mhmm. We have plugins. Mhmm. We have providers. I mean, in any sense, I know we're we're a little early, 490 00:36:14,125 --> 00:36:15,585 but how does it stratify, 491 00:36:16,765 --> 00:36:17,984 in in your view? 492 00:36:18,765 --> 00:36:20,785 Or how rather do we get from 493 00:36:21,339 --> 00:36:23,839 this web of trust engine, my words, not yours, 494 00:36:24,140 --> 00:36:25,680 to an end user 495 00:36:26,380 --> 00:36:35,119 who finds these delightful moments when they pick up their phone. You know, Apple intelligence doesn't look like it's going to be it, at least not for a while. 496 00:36:35,705 --> 00:36:39,645 And I've got recommendations on that great Thai restaurant. I've got recommendations 497 00:36:40,025 --> 00:36:47,325 on that specialist. I've got recommendations on that bicycle helmet that I'm Yeah. I'm looking for. So so my product road map, 498 00:36:47,990 --> 00:36:51,050 I've been in basically in r and d mode for a long time, but I 499 00:36:51,510 --> 00:36:54,250 my next iteration, I think, is gonna be something that's actually 500 00:36:54,790 --> 00:36:57,530 useful for actual users. And it's going to 501 00:36:57,990 --> 00:37:11,675 all it's gonna do at first is is personal level trust relay for and so you will be able to get it running, and you'll be able to connect to your relay. You can tell your friends, here's a relay. It does a good job of getting your spam. Right. And it has 502 00:37:13,655 --> 00:37:15,994 it it it processes as follows, mutes, and reports. 503 00:37:16,340 --> 00:37:18,280 So And it does the job. I mean, I I 504 00:37:18,580 --> 00:37:29,935 love the jobs to be done approach. And the job it will the job I'll hire this to do To get rid of spam. Just to get rid of spam. That's it. To for starters. Right. You're gonna you're gonna get it running, click a button, and that's it. It'll work. 505 00:37:30,235 --> 00:37:31,055 So then 506 00:37:31,755 --> 00:37:39,135 I what I want to to showcase is the ability of the grapevine and grape rank to curate specific lists. 507 00:37:39,435 --> 00:37:41,855 So let's say you want a list 508 00:37:42,315 --> 00:37:42,815 of 509 00:37:44,369 --> 00:37:46,150 of different Nostra applications. 510 00:37:46,690 --> 00:37:48,390 You wanna know what's the best 511 00:37:48,769 --> 00:38:00,405 or or just what are my options. What are the what are the different ecommerce apps? I don't know. What are they? So there's a NIP called NIP 89, which generates data that can be useful for that purpose. 512 00:38:01,105 --> 00:38:05,765 And it's not perfect, but the whole point of the grapevine is it doesn't have to be perfect. 513 00:38:06,224 --> 00:38:06,724 So 514 00:38:07,025 --> 00:38:11,765 one of my, you know, I'm going to one of my iterations is probably gonna use 515 00:38:12,220 --> 00:38:14,240 nip 89 data, which exists. 516 00:38:14,700 --> 00:38:19,600 And you're gonna be able to go to a page, and you're gonna be able to see here's a list of apps 517 00:38:19,900 --> 00:38:20,880 curated by 518 00:38:21,180 --> 00:38:22,000 your grapevine. 519 00:38:22,700 --> 00:38:24,320 And it will be a 520 00:38:24,645 --> 00:38:25,145 list. 521 00:38:25,445 --> 00:38:28,505 It'll be I mean, there are lots of people who have 522 00:38:28,885 --> 00:38:29,385 different 523 00:38:29,845 --> 00:38:42,440 you can go to different websites and find, here's a list of all the different nostra apps, but someone's always in charge. Yeah. And it's that's a grab bag. And it's a grab bag. And maybe this one is great, but what if this person gets bored and stops maintaining it? Right. 524 00:38:42,900 --> 00:38:43,400 What 525 00:38:44,099 --> 00:38:44,599 this, 526 00:38:45,460 --> 00:38:46,099 what brainstorm 527 00:38:46,579 --> 00:38:49,559 if I if that's what I decide to call it, I think I may call it brainstorm, 528 00:38:49,859 --> 00:38:51,160 is going to do is 529 00:38:51,615 --> 00:38:52,595 it's going to be 530 00:38:52,975 --> 00:38:55,235 a list that is that is maintained 531 00:38:55,615 --> 00:39:03,875 not by one person, but by your entire network. Right. And maybe this person does a great job today. Maybe these three people contributed to it to it tomorrow. 532 00:39:04,480 --> 00:39:09,060 No one there's gonna be no point of failure. This will be a list that is very 533 00:39:09,440 --> 00:39:09,940 contextual 534 00:39:11,760 --> 00:39:12,980 and curated by 535 00:39:13,600 --> 00:39:22,464 your community. And so There's something I want surfaced to me. At least that's the way I think of it. It's something that you want. The ability to what's my next read? 536 00:39:22,845 --> 00:39:31,185 You know, I can go to Goodreads. I can ask Yeah. On Noster or whatever social media, but, you know, I I would love the, the ability to 537 00:39:32,020 --> 00:39:33,640 have among my 538 00:39:34,180 --> 00:39:36,760 friends' followers' web of trust Mhmm. That 539 00:39:37,220 --> 00:39:41,000 next book that I should read in a given category, fiction, nonfiction, 540 00:39:41,620 --> 00:39:49,055 whatever it is. I mean Yeah. Is that is that a suitable application for this? So as long as there's data out there that has signal. Right. 541 00:39:49,435 --> 00:40:00,815 That's that's the it. If there is data that has signal. And and and how how structured are unstructured? I mean, I think about It doesn't matter the structure. Okay. If you can if the fact that somebody posted about a great read, 542 00:40:01,740 --> 00:40:07,360 presumably, that can get ingested. That can inform somehow this So yeah. So if, 543 00:40:09,020 --> 00:40:10,880 if somebody could write an algorithm 544 00:40:11,500 --> 00:40:14,560 that will ingest the data and know what it means, 545 00:40:15,075 --> 00:40:29,255 Then so someone would have to write a script. But you don't have to be the developer yourself. Just someone Yeah. There's a there's a marketplace of plug ins, extensions. Call them what we will. And and it will be something that interprets data, that that spits out effectively 546 00:40:29,635 --> 00:40:30,135 ratings, 547 00:40:30,990 --> 00:40:32,690 interpreted ratings of books. 548 00:40:33,390 --> 00:40:34,130 And then 549 00:40:34,510 --> 00:40:54,474 once you have that, the great rank engine takes care of the rest, and now you've got ratings that is decentralized. It is curated in a by your web of trust. I've got my own personal tomato ometer, my Right. Yeah. Phone calls through. And you don't have to wonder, well, okay. Well, did the people who run this website take a lot of money in order to promote this product? Right. 550 00:40:54,775 --> 00:40:59,275 Because if someone gets into your community who's doing that, then, 551 00:40:59,930 --> 00:41:12,830 see, eventually, what's gonna happen is that you're gonna route around all the bad bad bad apples, and the problems will be fixed before you even know about them. And I'll tell you one product I really want to see built in the not too distant future is a is Wikipedia. 552 00:41:14,565 --> 00:41:17,805 The, Nostra pedia is Yes. And there are 553 00:41:18,244 --> 00:41:21,625 you know, we have in Nostr. We have decentralized wikis. 554 00:41:22,244 --> 00:41:28,325 Question is, how do you know which authors to trust in which context? So And and and the the prop that solves 555 00:41:29,760 --> 00:41:30,740 tell me a bit more 556 00:41:31,360 --> 00:41:32,180 about Wikipedia. 557 00:41:33,760 --> 00:41:34,260 So 558 00:41:34,640 --> 00:41:36,260 so the problem with Wikipedia 559 00:41:36,720 --> 00:41:40,580 is that there, you know, there are many people who feel that 560 00:41:40,880 --> 00:41:46,275 it is politically biased on certain topics. Mhmm. I think it's great on Sure. Many topics, 561 00:41:46,575 --> 00:41:47,694 but, you know, there are 562 00:41:48,575 --> 00:41:59,850 it is ultimately centralized. The history of the baleen whale Yeah. I can probably feel good about that. My son is my son is very into Ah, okay. You know, wait a sec. Baleen is the French word for whale. I forget. I'm embarrassing myself. 563 00:42:00,150 --> 00:42:06,410 But the point is there are certain subject matters that That you don't that you wonder. Even in America, we can't politicize, 564 00:42:07,270 --> 00:42:18,115 and there's probably little question as to whether it's somebody's camped out on that page Exactly. And making edits versus those things which are Right. So we want something that cannot be taken over 565 00:42:18,655 --> 00:42:29,490 by any centralized entity. Right. And so maybe all of it is very trustworthy, but do you know that? Can you feel confident in it? So I would like to see an alternative to Wikipedia, 566 00:42:30,589 --> 00:42:32,690 which I believe is gonna be built on Nostr, 567 00:42:33,069 --> 00:42:34,290 and where 568 00:42:34,670 --> 00:42:36,369 you can see, okay, 569 00:42:36,750 --> 00:42:38,849 five people wrote an article on this topic, 570 00:42:39,355 --> 00:42:48,175 and my community trusts this person. And I you can actually if you want to, you can dig in and see why. Like, how is this number calculated? Do I trust 571 00:42:48,875 --> 00:42:53,055 the metric? Or maybe you don't wanna dig into the details. You'll just know that 572 00:42:55,140 --> 00:42:58,359 that that it was, you know, I trust people who trust people, 573 00:42:58,660 --> 00:43:09,585 or I chose the people who choose the people who chose the authors of this topic right here. They chose different authors for that topic. Maybe and so all there has to be a signal. So maybe the signal is that 574 00:43:10,065 --> 00:43:13,845 people who have a high trust score in your network, they gave thumbs up 575 00:43:15,585 --> 00:43:16,244 to these 576 00:43:16,545 --> 00:43:19,045 three articles written by this author 577 00:43:19,345 --> 00:43:19,845 in 578 00:43:20,224 --> 00:43:22,645 the category that and, you know, the author 579 00:43:23,040 --> 00:43:25,540 might say, here's an article I wrote. Here's the category. 580 00:43:25,920 --> 00:43:33,460 Lots of people you trust gave it thumbs up. Here's another author in the same category, and people you trust said thumbs down. So 581 00:43:33,840 --> 00:43:44,335 so a score gets generated from this data. You don't have to know exactly how it's done, but you just know that But it's auditable if I care to. If you care to, it's auditable, and you're not wondering, 582 00:43:44,715 --> 00:43:45,615 did the 583 00:43:46,155 --> 00:43:48,095 this particular large entity 584 00:43:48,555 --> 00:43:55,210 influence it for reasons that I don't may or may not know what they are. Now on the one hand, I see that, and I think in my mind, I'm thinking 585 00:43:57,109 --> 00:43:58,490 almost like a diff, 586 00:43:59,030 --> 00:44:02,570 a git diff. You know, I'm I'm going to see a Wikipedia page 587 00:44:03,990 --> 00:44:04,490 reflect 588 00:44:05,445 --> 00:44:07,625 my web of trust. I'm going to see edits 589 00:44:08,005 --> 00:44:16,505 that are closest to or most enforced by my web of trust. Yeah. And that to me is interesting. Now on the flip side, I think, you know, does this just feed 590 00:44:17,540 --> 00:44:35,105 the already fractured nature of what is fact, what is truth? We all get to see our own little version of it. Yeah. So It's a fantastic question, which comes up a lot, and I think about it a lot. Are we gonna make better echo chambers? Right. Or, you know, more powerful, more echoey echo chambers? And 591 00:44:35,484 --> 00:44:40,065 and I don't want that to happen. And so I do think about this. And and, 592 00:44:41,325 --> 00:44:47,500 I think that if you want to build a better echo chamber, you will be able to. Sure. I cannot stop you from doing that. 593 00:44:48,440 --> 00:44:48,940 So 594 00:44:49,640 --> 00:44:52,859 so that's but, you know, my answer to that would be 595 00:44:53,160 --> 00:44:55,660 we don't we already have echo chambers. 596 00:44:56,040 --> 00:44:59,660 If you want an echo chamber, we've got them. You don't you don't my tools, 597 00:45:00,435 --> 00:45:03,655 you don't need them. The web of trust, you don't need those. 598 00:45:04,915 --> 00:45:07,095 But what if you wanna get away from 599 00:45:07,475 --> 00:45:15,510 the echo chambers? I don't want to know the political answer to this question. I wanna know the actual truth. Right. So As measured by 600 00:45:15,970 --> 00:45:17,589 So I want five neuroscientists 601 00:45:17,970 --> 00:45:19,510 in my web of trust who 602 00:45:19,890 --> 00:45:22,869 have, at the very least, the background, the, 603 00:45:24,049 --> 00:45:24,549 credentials. 604 00:45:25,010 --> 00:45:29,025 Yeah. You know, I can sort of maybe flip that switch and show me the most 605 00:45:30,205 --> 00:45:30,705 normalized 606 00:45:31,085 --> 00:45:31,585 factually 607 00:45:32,205 --> 00:45:42,225 Yeah. Show me. Sound Here's a question I have, and I want to know the answers by people who are board certified in this. Right. And I want it to be not filtered by any large news organization, 608 00:45:42,780 --> 00:45:43,440 not filtered by 609 00:45:44,060 --> 00:45:44,560 a, 610 00:45:45,660 --> 00:45:47,820 by a social media company Right. 611 00:45:48,300 --> 00:46:01,025 Whose motives I may or may not know. Yes. Advertorials and all these other things. Yeah. So say what you will about credentialism as some Right. Maybe you trust that. Maybe you don't. That's part of the thing that you have control over the algorithm 612 00:46:01,404 --> 00:46:03,184 that you don't have. So 613 00:46:04,924 --> 00:46:09,840 Yeah. No. No. At no point, and I think it's, you know, certainly a a a firmly held belief that I held hold, 614 00:46:10,320 --> 00:46:11,060 is that 615 00:46:11,520 --> 00:46:12,660 if you want to 616 00:46:15,520 --> 00:46:29,795 if you want to engage in the digital equivalent of locking yourself away in your basement Mhmm. Not my job to stop you. Okay. Right? But what I can't what we well, our job is to give you tools to walk out. Right. Is to leave, to exit this 617 00:46:30,255 --> 00:46:30,734 this, 618 00:46:31,615 --> 00:46:34,755 jail cell Yeah. Yeah. That that don't really exist 619 00:46:35,055 --> 00:46:35,760 right now. 620 00:46:36,240 --> 00:46:37,620 What would stop someone 621 00:46:38,400 --> 00:46:39,140 at Meta 622 00:46:39,600 --> 00:46:46,260 Mhmm. From having an and maybe they are, from having an epiphany Mhmm. And deciding to do this? What what is 623 00:46:47,120 --> 00:46:48,500 inherent to the incumbent 624 00:46:48,960 --> 00:46:49,460 Mhmm. 625 00:46:51,875 --> 00:46:54,055 Social media companies, I think, most relevantly. 626 00:46:54,835 --> 00:46:56,775 Maybe you'll have a different opinion on that. 627 00:46:57,395 --> 00:47:00,535 What's in the way of them embracing web of trust? 628 00:47:01,395 --> 00:47:09,600 I think that if they wanted to find the solution, they would have found it, but it is not in their financial interest to implement things this way. They they 629 00:47:09,920 --> 00:47:14,100 ultimately, they make money by monetizing data. Sure. And so 630 00:47:14,400 --> 00:47:16,180 they they wanna 631 00:47:16,720 --> 00:47:19,540 gather that data and then control it and monetize it. 632 00:47:20,465 --> 00:47:23,205 And some would argue they also monetize outrage. 633 00:47:23,665 --> 00:47:26,325 Exactly. This is one of the things that's very pathological. 634 00:47:26,945 --> 00:47:29,845 The the more your eyes are glued to the screen, 635 00:47:30,305 --> 00:47:31,445 then the more 636 00:47:31,825 --> 00:48:00,135 money they make because they have more users who are on there for more time. They can get more advertising dollars. Right. And so their AI algorithms, they don't even it's not that they're trying to generate outrage. I don't think they have to be doing that. But AI just discovers, hey, if I show you this thing that outrages you, you stay on longer. I'm going to do more of it. Correlation is not causation and vice versa, but AI isn't trying. It just does what works. And what works is we want to make more money. And 637 00:48:02,320 --> 00:48:02,820 so 638 00:48:03,520 --> 00:48:06,180 they don't really have an incentive to fix this problem. 639 00:48:08,000 --> 00:48:10,020 What I want to see is 640 00:48:10,480 --> 00:48:11,380 a world where 641 00:48:11,680 --> 00:48:12,420 your data 642 00:48:12,960 --> 00:48:15,540 is not monetized by these large entities. 643 00:48:16,655 --> 00:48:24,755 You have the option of giving it away if you want or you monetize it yourself. So for example, let's say you are an excellent movie reviewer, 644 00:48:25,055 --> 00:48:31,380 and my web of trust says you you're two, three, four hops away from me, but everyone recognizes, you know, you're Roger Ebert. 645 00:48:32,180 --> 00:48:35,319 And and I might be willing to pay you some Satoshis. 646 00:48:35,700 --> 00:48:36,920 Right. And 647 00:48:37,380 --> 00:48:37,880 so 648 00:48:38,500 --> 00:48:39,400 you monetize 649 00:48:40,660 --> 00:48:41,160 your 650 00:48:41,859 --> 00:48:44,599 what you know, not the value that you create. 651 00:48:45,059 --> 00:48:48,494 It's not Twitter that's monetizing it. So these tools 652 00:48:49,275 --> 00:48:52,655 have the ability to do that, but why would they build them? Yep. 653 00:48:53,435 --> 00:48:58,575 We have to this is this is why we have to build these tools, and they can be done. I know. 654 00:48:59,355 --> 00:49:03,215 And does that present to those incumbents? I mean, you know, it's why 655 00:49:03,890 --> 00:49:11,910 the numbers on Nostra are relatively small, and and I will continue to use it as an example throughout these discussions. But I think it is 656 00:49:12,450 --> 00:49:12,950 the 657 00:49:14,850 --> 00:49:15,350 counterexample 658 00:49:15,730 --> 00:49:16,630 to these centralized 659 00:49:17,184 --> 00:49:19,444 social media giants. Mhmm. 660 00:49:20,305 --> 00:49:21,525 Do they need 661 00:49:22,464 --> 00:49:33,410 to break in order for Web of Trust to succeed? I mean, what's your take on I don't think they success? I don't think they have to break, and I don't think that they're gonna see, you know, Oster as a threat. No. Just like 662 00:49:33,710 --> 00:49:38,690 the, you know, Federal Reserve didn't see Bitcoin as a threat, still doesn't. 663 00:49:38,990 --> 00:49:43,390 You know, we all thought it would, but that was probably very naive. Right. And there's 664 00:49:44,855 --> 00:49:48,555 no. Those don't have to break. All we have to do is is build, 665 00:49:50,375 --> 00:49:53,355 options, build alternatives that don't exist, and people 666 00:49:53,734 --> 00:50:01,140 and we and and that allow you to do things that you cannot do anywhere else. And and that's a great point I was just going to ask. I mean, what 667 00:50:04,080 --> 00:50:08,340 what is the what do you think are the intrinsic motivations or the incentives 668 00:50:09,360 --> 00:50:09,860 that 669 00:50:11,295 --> 00:50:14,035 not billions, but some significant number of individuals 670 00:50:14,335 --> 00:50:16,515 need to have Mhmm. To 671 00:50:16,815 --> 00:50:24,994 embrace something like this. Assuming we get over the UX hurdles, assuming it becomes consumer friendly Mhmm. You know, assuming in my mind at least that there's 672 00:50:25,770 --> 00:50:45,924 an agent, I'll just use that term very loosely Uh-huh. That employs this web of trust engine to get jobs done for users. Uh-huh. What do my motivations need to be in order to take a to take a risk, take a leap, and try this? What what gets me interested as a I think the same reasons that people use, you know, Facebook for the first time, 673 00:50:46,385 --> 00:50:50,724 their friends use it, and they tell them, hey. I I'm I'm using this new platform, 674 00:50:51,505 --> 00:50:52,964 and I love it because 675 00:50:53,424 --> 00:50:53,910 it 676 00:50:54,230 --> 00:50:59,609 does this thing that I no one else does, or it does it better than anyone else. I think, ultimately, it's going to, 677 00:51:00,630 --> 00:51:10,835 web of trust on top of Nostr is gonna just simply do a better job at finding the content you want. And and people are going to say that it does that. They're not gonna be ideological. 678 00:51:11,375 --> 00:51:13,235 Right. We we will never 679 00:51:13,615 --> 00:51:19,555 Nostra's not gonna win because people are purple pilled. Right. Just like Bitcoin's not gonna win because people are ideologically 680 00:51:19,935 --> 00:51:29,890 love it. Yeah. People Bitcoin is going to win because people ideologically hate it, but use it anyway Right. Because It's a superior store of value. It's better payments. It's whatever. 681 00:51:30,350 --> 00:51:31,070 They're gonna be like, 682 00:51:31,630 --> 00:51:38,785 I'm not using it, but I'm Yeah. But, I I'll use it over watching my my USD melt away. Or yeah. They they'll just 683 00:51:39,425 --> 00:51:43,925 even not that. They're just gonna be like, well, because the person down the street takes it. Or Network effects. 684 00:51:44,385 --> 00:51:50,820 Or or what's that? Network effects? Or yeah. And I mean, I guess that cuts both ways. Right? I mean, so, you know, we see we hear many critics, 685 00:51:51,520 --> 00:51:52,900 who might otherwise be, 686 00:51:54,480 --> 00:51:59,060 I think, prone toward Nostra as an example, dismiss it Yeah. 687 00:51:59,440 --> 00:52:00,900 Because ex Twitter 688 00:52:01,365 --> 00:52:04,185 has that has that tremendous network effect. Well, 689 00:52:04,965 --> 00:52:15,850 Do we have to do we have to get that? Do we have to get do we have to clear that hurdle? Well, I mean, that is a hurdle, but we clear it when we can offer things that no one else can offer. Right. Right. That's that's it, 690 00:52:16,390 --> 00:52:20,890 and that will Do you have a and this is this is, perhaps a tricky question, but do you have 691 00:52:22,230 --> 00:52:26,570 an image in your mind, David, of what that moment looks like when someone says, 692 00:52:27,665 --> 00:52:34,085 I'll step away from x. I'll step away from Facebook. Or rather, you know, not either or, but I'll spend some of my time Mhmm. 693 00:52:34,705 --> 00:52:37,445 Using product x, product whatever, 694 00:52:38,464 --> 00:52:40,885 that employs web of trust. Like, what is that 695 00:52:41,190 --> 00:52:43,210 magic moment look like do you think? 696 00:52:43,510 --> 00:52:55,050 I I'd say I don't know. It might look very different for lots of different people. Just like why do people use Bitcoin? There's no one answer to that. Right. It's gonna be lots of different reasons for lots of different people. Yeah. 697 00:52:55,665 --> 00:53:00,805 And and so for those who are building products, who are engineers, who are heads of product, 698 00:53:02,385 --> 00:53:02,885 founders, 699 00:53:03,825 --> 00:53:06,565 what's in it for them? Why would either 700 00:53:07,740 --> 00:53:08,960 maybe it is 701 00:53:09,340 --> 00:53:13,120 a media product, maybe it's a social media product, maybe it is 702 00:53:13,660 --> 00:53:19,760 a networked application. There's some social element. Presumably, there's a there's a network effect at play. 703 00:53:20,165 --> 00:53:37,280 Mhmm. Why should they consider this? Why should they put some some engineering effort Into Nasr. Into Web of Trust. Into Web of Trust. Presumably, Web of with Nasr as maybe a a a a path to adopt Web of Trust. But Mhmm. Why should they be experimenting with Web of trust? That's 704 00:53:37,980 --> 00:53:42,160 a challenging question. I mean, ultimately, you cannot say I have a hammer. 705 00:53:42,940 --> 00:53:48,080 What can I do with it? Right. You have to say I have a problem. Sure. And how do I solve it? And so, 706 00:53:49,065 --> 00:53:51,645 you know, what is the problem that other companies 707 00:53:52,585 --> 00:53:53,405 what are the problems 708 00:53:54,265 --> 00:53:56,205 that that exist that, 709 00:53:57,545 --> 00:54:03,244 Yeah. What's deeply broken that you think is And so that they should be looking at? There's gonna be lots of examples. So one example 710 00:54:03,545 --> 00:54:04,910 is that companies 711 00:54:06,089 --> 00:54:27,925 have data. Now we've talked about how the fact that companies monetize your data, but also company data can be toxic. Mhmm. It can be hacked. Yep. There are a lot of companies who would love to not have to hold your data. I would love to not get one of those emails every week offering me for, you know, free credit monitoring Right. In return for having linked my personal health information, you know, whatever it is. So if you have, 712 00:54:28,645 --> 00:54:30,905 if you would like to somehow 713 00:54:31,285 --> 00:54:34,250 not have to keep track of all your customers' data, 714 00:54:34,790 --> 00:54:36,250 I think we're gonna end up 715 00:54:36,630 --> 00:54:40,650 offering solutions to that problem. So the personal data store, which some would call it, 716 00:54:41,510 --> 00:54:45,690 I've I've worked in that field previously where, if I if I hear you correctly, 717 00:54:46,065 --> 00:54:55,365 web of trust can operate on a personal data store, which is that I am holding Mhmm. Those data, not company a b c. Yeah. Yeah. It is the whole 718 00:54:55,744 --> 00:55:00,000 philosophy is that you can you are in charge of your data. You're not 719 00:55:01,580 --> 00:55:05,360 So if if we take and maybe this is the way to go at it. It's something I think about. 720 00:55:05,980 --> 00:55:09,120 It is a big if, but if we take as a given that 721 00:55:09,500 --> 00:55:12,060 there is a trend toward personal data ownership 722 00:55:12,380 --> 00:55:15,040 Mhmm. And that's a big if. I mean, I've I've, 723 00:55:16,275 --> 00:55:21,494 we've we've all seen, I think, many, many attempts at this. I've I've taken a shot at it myself. It is tough. 724 00:55:22,195 --> 00:55:30,855 Many people would give away everything for a free slice of pizza, so how do we get across that? But if that's the arc, if if personal data ownership is is where we're headed, 725 00:55:31,290 --> 00:55:33,710 would it be fair to say that web of trust 726 00:55:34,570 --> 00:55:35,950 is a means of, 727 00:55:37,690 --> 00:55:42,510 while I may lose access to your personal data because now you own it, I don't have to lose access 728 00:55:42,810 --> 00:55:44,595 to useful operations on it, 729 00:55:45,155 --> 00:55:47,255 and therefore, delivering value as a product. 730 00:55:48,595 --> 00:55:49,415 Say that again. 731 00:55:49,795 --> 00:55:51,655 So if in the future, 732 00:55:52,435 --> 00:55:56,375 you as Meta don't hold all my data Yeah. A billion users 733 00:55:56,799 --> 00:55:58,480 hold their data. Hold their data. Yeah. 734 00:55:59,119 --> 00:56:05,140 Doing useful things with that data Mhmm. Is the next, I think, opportunity to monetize 735 00:56:05,760 --> 00:56:22,125 or becomes the core. I can I can no longer sell your data, and so I have to find a way to do something useful to do operations on that data? So maybe companies should be thinking not not what is the problem we wanna solve, but what is the problem that our cuss our future customers wanna solve. Certainly. And 736 00:56:23,060 --> 00:56:23,880 how can 737 00:56:24,580 --> 00:56:29,720 that happen using these tools that are being built by Freedom Technologists? 738 00:56:30,260 --> 00:56:35,240 And and what what are the we talked a little earlier about some of the challenges to getting here. 739 00:56:35,845 --> 00:56:39,705 What perhaps are the next few building blocks that take us to, 740 00:56:41,685 --> 00:56:42,985 a minimum viable, 741 00:56:43,845 --> 00:56:45,865 you know, web of trust that the 742 00:56:46,405 --> 00:56:49,410 average individual, maybe early adopters, going to find useful? 743 00:56:49,890 --> 00:56:51,430 You I mean, we already have 744 00:56:52,049 --> 00:56:54,950 webs of trust that people find useful. We already have 745 00:56:55,250 --> 00:56:59,349 personal level of trust relay Mhmm. That your colleague, UTXO, 746 00:56:59,650 --> 00:57:12,625 has built. That is and and many obviously, many people credit that. I use it. You use it. Now I'm a hyper nerd, and, you know, I've got a virtual private server that I run it on, etcetera, etcetera. So Yes. But you for the normie. But but you can, 747 00:57:14,045 --> 00:57:25,940 establish one, do the work to get one running, and then It's official one. And then your friends can connect to it. Yes. And you can connect to five different personal web of trust relays, and now your feed is devoid 748 00:57:26,340 --> 00:57:40,765 is is much, much higher quality. Yeah. I mean, for me, it has all but eliminated spam. Exactly. And so I'm you know, have the ability, and I'm sufficiently motivated because who wants spam? Right. So we already have tools that are Great point. That are doing that people are using. 749 00:57:41,385 --> 00:57:43,645 And and so we just need to make them 750 00:57:44,505 --> 00:57:46,765 we need to we need to make them more sophisticated. 751 00:57:47,465 --> 00:57:51,830 Yeah. So this, for example, point and click. So maybe someone like, 752 00:57:53,030 --> 00:57:57,530 like, Relay Tools, CloudFinder's Relay Tools could make a point and click version of that. 753 00:57:58,150 --> 00:57:59,370 Or, you know, CloudFinder 754 00:57:59,670 --> 00:58:00,570 has is 755 00:58:00,950 --> 00:58:03,930 is one of three people who have built the grape rank 756 00:58:04,244 --> 00:58:04,904 calculation engine. 757 00:58:05,845 --> 00:58:09,384 He's he's just built the code just to to see how it works. 758 00:58:09,765 --> 00:58:19,470 Manny Me has done the same. He is his one of his goals is building a a library that other other developers could use to implement this. 759 00:58:19,950 --> 00:58:21,890 So making it point making it easier. 760 00:58:22,510 --> 00:58:23,010 Packaging. 761 00:58:23,310 --> 00:58:30,770 Packaging definitely is always important. It needs to be easy. It needs to have it needs to be able to do more things, 762 00:58:31,390 --> 00:58:34,315 but not too many things. Right. And so I think the extensibility 763 00:58:34,615 --> 00:58:35,355 is gonna 764 00:58:35,895 --> 00:58:36,635 be a big 765 00:58:37,174 --> 00:58:43,015 no. That doesn't really exist yet. Yep. But it can. So making it so that you can decide 766 00:58:43,575 --> 00:58:44,954 that's gonna be a big step 767 00:58:45,255 --> 00:58:48,234 where you can decide what data you wanna use, 768 00:58:48,910 --> 00:58:50,770 and then that can happen 769 00:58:51,150 --> 00:58:51,650 without 770 00:58:52,670 --> 00:58:58,770 me or Sure. Doing it. So a tool that allows the extensibility to happen. So you're you're 771 00:58:59,550 --> 00:59:00,050 obvious, 772 00:59:00,670 --> 00:59:02,610 perhaps obvious customer at this stage, 773 00:59:02,915 --> 00:59:06,214 not an end user, of course. Mhmm. But it is someone building 774 00:59:07,155 --> 00:59:08,595 Nostra apps. It's someone 775 00:59:09,075 --> 00:59:20,670 My users for when I right I right now, I've got this alpha product brainstorm. My initial user is gonna be people like you who already have a personal level of stress relay and are sophisticated enough to do that. So the bleeding edge early adopters? 776 00:59:21,690 --> 00:59:24,349 At first. Yeah. At first. And then gradually 777 00:59:24,650 --> 00:59:26,109 make it easier to use 778 00:59:26,490 --> 00:59:27,230 and gradually 779 00:59:27,609 --> 00:59:30,829 start to add features that no one else is is is 780 00:59:31,155 --> 00:59:34,375 offering. Now you reserve the right, of course, to to change these 781 00:59:34,675 --> 00:59:35,655 predictions, but 782 00:59:35,955 --> 00:59:37,895 how much time would you say, David, until 783 00:59:38,835 --> 00:59:40,295 well, so in phases, 784 00:59:41,315 --> 00:59:44,295 you've got something that the early adopters will use. 785 00:59:45,490 --> 00:59:48,390 What are we looking at? Six months? What do you think? I would 786 00:59:49,490 --> 01:00:02,434 I think there's a good chance that within a month Okay. Great. I'll have something that you can play with like I just described. Forward to it. Not all the blood bells and whistles. But but you'll be able to load it, pull it up, click a button, start my relay, 787 01:00:03,454 --> 01:00:14,035 turn on my white list, turn on the black list Mhmm. Or not. And then that'll like, just that right there, that right there actually already works in alpha, but it's not Yeah. 788 01:00:14,335 --> 01:00:16,230 So I think in a month, I could 789 01:00:16,710 --> 01:00:17,930 have that running. 790 01:00:18,390 --> 01:00:20,170 And then beyond that, when 791 01:00:20,470 --> 01:00:29,210 when does it become a software development kit? When does it become something that someone who's, I don't know, you know, a talented TypeScript developer 792 01:00:29,925 --> 01:00:32,585 understands the basics of the Nostra protocol 793 01:00:33,365 --> 01:00:34,745 can can build on? 794 01:00:35,365 --> 01:00:35,865 That's, 795 01:00:37,685 --> 01:00:45,065 that's a tough question. You know? A year out, do you think? Is it yeah. Yeah. So so my goal is to have something that users are using 796 01:00:46,119 --> 01:00:47,020 and and say, 797 01:00:47,320 --> 01:00:56,380 builds I wanna get to a point where it's something that cannot be done else otherwise. Right. And then I want users and developers to say I wanna learn more. Sure. And then I can have that conversation. 798 01:00:56,920 --> 01:00:59,340 How do we work together Right. 799 01:01:00,165 --> 01:01:02,825 To to do this. So Yeah. 800 01:01:03,285 --> 01:01:09,145 Nice. And I mean and and the the reason I ask these questions is to sort of level set for for the audience. 801 01:01:11,285 --> 01:01:22,440 Right now, we're in the bleeding edge. Yeah. But we've got proof of concept. We've got running code. Yeah. And it sounds like though there is a very healthy debate about 802 01:01:22,740 --> 01:01:27,935 how one should implement web of trust, what are the signals, what are the data points that that that's progressing, 803 01:01:28,235 --> 01:01:31,055 call it six months to twelve months away. We've got, 804 01:01:31,995 --> 01:01:36,255 a stable foundation to build on. Yeah. And and then 805 01:01:37,115 --> 01:01:42,369 what is your crystal ball for when do we go mainstream? When does web of trust become something 806 01:01:42,670 --> 01:01:45,089 not even so much grape rank per se, but, 807 01:01:45,710 --> 01:01:54,305 when does it become something that is out in the wild and cannot be ignored? You know, the future is coming at us faster and faster. If you want time predictions, 808 01:01:54,925 --> 01:01:57,185 I just started using AI to help me code 809 01:01:57,645 --> 01:02:09,500 maybe three weeks ago, and I felt like I went from one of me to five of me. So Watch out. Watch I mean, this the we're going to be able to we've been having these kind of high level debates 810 01:02:09,800 --> 01:02:10,300 on 811 01:02:11,000 --> 01:02:13,980 at the abstract level of how to solve, what's the problem, 812 01:02:14,280 --> 01:02:15,180 what's the solution. 813 01:02:15,560 --> 01:02:16,860 And we are now, 814 01:02:17,160 --> 01:02:17,820 I think, 815 01:02:18,280 --> 01:02:26,065 transitioning into building we're starting. I mean, we've actually already done this to an extent, building actual tools. How many roughly roughly how many people would you say are involved, 816 01:02:26,545 --> 01:02:28,085 actively involved in 817 01:02:28,385 --> 01:02:30,565 the design phase of this? 818 01:02:30,865 --> 01:02:34,645 Oh, I think probably every every Nostra developer is talking about. 819 01:02:35,025 --> 01:02:35,525 And 820 01:02:35,900 --> 01:02:46,000 and, I just posted not too long ago that I so the one of the tools that Brainstorm is going to do is, you know, calculating the scores is the hard part. Once you've done that, 821 01:02:46,300 --> 01:02:48,720 putting them in whatever format makes you happy 822 01:02:49,275 --> 01:02:51,454 is is actually the easy part. Mhmm. So, 823 01:02:52,155 --> 01:02:57,694 so Vitor has a, you know, he's the he's the chief architect of Amethyst, the the premier, 824 01:02:58,395 --> 01:03:01,775 Android client. Android. Yeah. So so he has 825 01:03:02,155 --> 01:03:03,615 a format for 826 01:03:04,490 --> 01:03:07,310 for trust scores that I think he would like to use, 827 01:03:08,329 --> 01:03:10,349 for a variety of purposes on, 828 01:03:11,290 --> 01:03:12,349 you know, for himself. 829 01:03:12,730 --> 01:03:18,430 And so I've already built an alpha version that exports the scores in that format. 830 01:03:19,355 --> 01:03:19,855 FranZap, 831 01:03:20,234 --> 01:03:30,255 Papalia Absolutely. They have, the app store. App store. Very exciting. And they have a FranZap has authored a web of trust DVM, NIP. So it's another way to 832 01:03:30,635 --> 01:03:31,275 for someone 833 01:03:31,790 --> 01:03:34,690 for a product like Brainstorm, to offer personalized 834 01:03:35,070 --> 01:03:43,570 trust scores to a product to a client developer. Can you can you describe, David? I mean, I think Zapstore, I'm very bullish on. And if we zoom all the way out, 835 01:03:44,190 --> 01:03:45,490 can you describe Zapstore? 836 01:03:46,075 --> 01:03:59,855 Yeah. I may not be able to do it justice. I I I know that they are calculating page rank. Or rather so so, again, if we if we if we zoom way out, we think about jobs to be done. It is an alternative to the Google Apple App Store. Mhmm. 837 01:04:00,890 --> 01:04:03,710 And it is a web of trust applied to 838 01:04:04,330 --> 01:04:05,310 what are the 839 01:04:05,610 --> 01:04:06,110 so 840 01:04:06,490 --> 01:04:16,270 broken free of a central authority, Apple or Google, who get to determine what I should run, what I can install Yeah. On my device. Correct me if I'm wrong. This 841 01:04:16,745 --> 01:04:18,845 uses my web of trust Mhmm. 842 01:04:20,265 --> 01:04:28,125 To apply a score or or or some other measure of this code is good. You know, five other people that you trust are running 843 01:04:28,585 --> 01:04:29,325 this application 844 01:04:29,865 --> 01:04:30,365 Mhmm. 845 01:04:30,675 --> 01:04:31,175 To 846 01:04:31,779 --> 01:04:36,359 do whatever, Maps, Uber, you name it. Mhmm. So it if I understand, it's 847 01:04:36,980 --> 01:04:37,960 it's a decentralized 848 01:04:38,339 --> 01:04:53,855 app store Mhmm. Which gives me both the freedom Mhmm. To run whatever I want, should I should I choose, but also the safety and trust Mhmm. To know that I'm not out there in the wild Mhmm. Installing rogue code. Right. And obviously, a very important problem. 849 01:04:54,315 --> 01:05:08,210 Now maybe you know this I I don't I don't I'm not sure what is the raw data that they're using as a small signal. I don't. I don't. I I do know that they're they're use they're calculating page rank, and I'm glad to see that showing up in the Nostra ecosystem. Right. 850 01:05:08,670 --> 01:05:12,530 And, you know, that's that's a step more sophisticated than just follows and follows, 851 01:05:13,150 --> 01:05:14,290 or follows and follows. 852 01:05:15,725 --> 01:05:21,725 Yeah. I I'd like to I I should know this, what what they're using. Yeah. We'll we'll dig in and and get that in the show notes. 853 01:05:23,165 --> 01:05:30,940 Where do where does this take us? Where what does what does five years out look like? Do you have a I mean, you've been cranking on this for a decade. Yeah. 854 01:05:31,640 --> 01:05:35,819 What is your dream scenario? My dream scenario is that I build this to a point 855 01:05:36,200 --> 01:05:36,700 where 856 01:05:37,720 --> 01:05:44,685 it does the thing so so well that it can start growing on its own. And What does that mean? I'm building this so that I can use it. Right. 857 01:05:45,225 --> 01:05:49,165 What does that look like? It means that you'll it means that people will I think 858 01:05:49,625 --> 01:05:55,405 I I think once this gets good enough so that people can start discovering content in exactly 859 01:05:56,345 --> 01:05:56,845 the 860 01:05:57,250 --> 01:06:21,885 context that they want, I think we're gonna start seeing people migrate from the legacy social apps. And Nostra users' usership is gonna right now, it's about 20,000 act I think, per day ish. Yeah. The 30, I've heard, on a good day. Yeah. Yeah. I think I think that that number is gonna start to grow when people realize it can do these things better than any place else. I mean, does it does does one of the potential outcomes, 861 01:06:23,400 --> 01:06:31,099 I'm clearly not the demographic for TikTok, but, you know, it's it's a massive success in setting aside all the recent controversy about ownership. 862 01:06:31,400 --> 01:06:42,835 It's a black box algorithm. It's a wildly successful black box algorithm. Right. You know, do we is it feasible that we get to a point where there's content suggested to us that could be as compelling? 863 01:06:44,175 --> 01:06:44,675 So, 864 01:06:44,975 --> 01:06:47,635 you know, I've heard people talk about TikTok as addictive. 865 01:06:48,015 --> 01:06:48,515 Yes. 866 01:06:49,070 --> 01:06:49,570 And 867 01:06:49,950 --> 01:06:53,730 this you know, what we're building, the the goal is not to be addictive. No. 868 01:06:56,110 --> 01:06:59,410 It's a good question. Are people gonna leave the addictive stuff 869 01:06:59,710 --> 01:07:08,535 for an algorithm that that is that they have control over. I mean, they, you could choose the addictive algorithm if you want. Yeah. Yeah. Right. That says you can't. People may, maybe, 870 01:07:08,835 --> 01:07:11,575 I wonder if people are gonna try to emulate that. 871 01:07:12,195 --> 01:07:19,640 I think if people have the tools to build better algorithms, it's it's gonna be hard to predict. I think people here's what I want. 872 01:07:19,940 --> 01:07:23,560 And what I'm cautiously optimistic we're gonna see this is that 873 01:07:24,099 --> 01:07:27,960 people are gonna say that I never knew thought this was possible. Mhmm. This just 874 01:07:28,340 --> 01:07:39,625 is a different experience. So it's not replicating or knocking off. No. It's like asking to predict what Facebook is gonna look like before, you know, way before before even Myspace or Friendster. 875 01:07:40,405 --> 01:07:43,705 It's hard you can it's hard to predict that. But it was, 876 01:07:44,005 --> 01:07:44,665 I mean, 877 01:07:45,360 --> 01:07:46,660 astounding what people 878 01:07:48,160 --> 01:07:51,300 what can be done. You think we're ready for that? I mean, are there 879 01:07:52,080 --> 01:07:54,900 this is always the question, I suppose, with an emerging technology. 880 01:07:56,160 --> 01:08:00,340 Is it a roomful of nerds who want something to exist and that's beautiful? 881 01:08:01,045 --> 01:08:01,545 Mhmm. 882 01:08:02,805 --> 01:08:03,305 Or 883 01:08:04,565 --> 01:08:07,385 are you betting on the fact that this produces, 884 01:08:09,125 --> 01:08:11,705 by whatever measure, a better, healthier 885 01:08:13,125 --> 01:08:13,625 relationship 886 01:08:13,925 --> 01:08:14,425 to 887 01:08:15,180 --> 01:08:17,920 media technology, the things that we are 888 01:08:18,460 --> 01:08:18,960 recommended 889 01:08:19,820 --> 01:08:23,840 choose to trust and pursue? I think it will ultimately be more healthy. 890 01:08:24,620 --> 01:08:25,120 And 891 01:08:25,580 --> 01:08:27,600 because I believe in free speech, 892 01:08:28,074 --> 01:08:29,614 I believe in autonomy. 893 01:08:29,994 --> 01:08:36,415 I believe in a system where you're not controlled by people who are going who have the power to abuse that control. 894 01:08:36,875 --> 01:08:40,335 I believe that when it comes to money. I believe that when it comes to information. 895 01:08:41,389 --> 01:08:44,210 And how exactly that plays out is gonna be 896 01:08:44,750 --> 01:08:48,290 hard to predict, but I believe in the fundamental principles. 897 01:08:49,869 --> 01:08:51,329 And I'm excited to see 898 01:08:51,710 --> 01:08:53,170 how that's gonna play out. 899 01:08:53,595 --> 01:08:59,695 Perfect place to end it. Thank you, David. Thank you, Sean. Pleasure. I look forward to following the success. And, 900 01:09:00,315 --> 01:09:05,370 before we wrap, and I'll include these in the show notes, there are a number of places someone could go 901 01:09:05,770 --> 01:09:13,470 to dig into your work to dig into the code. Mhmm. One or two sort of top points to to learn more about web of trust Mhmm. Perhaps broadly, 902 01:09:14,250 --> 01:09:15,310 and more specifically, 903 01:09:15,850 --> 01:09:23,115 the work and the code that you're shipping. You know, I think probably the best places, and and I'm sure this will go in the show notes, is is just go follow me on Nostr. 904 01:09:23,495 --> 01:09:27,515 I will keep that updated. Excellent. And and I do have, you know, a GitHub 905 01:09:28,615 --> 01:09:30,715 my GitHub account, w d s four, 906 01:09:31,210 --> 01:09:33,150 that has my thoughts 907 01:09:33,450 --> 01:09:38,670 as they've been developing over years. If you wanna dig into how I think the database 908 01:09:38,970 --> 01:09:39,870 should be organized 909 01:09:40,250 --> 01:09:49,635 and how I think these two things fit together, Go and read about concept graph and grapevine. Just dig into my that GitHub account. Are you looking for contributors? 910 01:09:50,281 --> 01:09:53,882 If someone's inclined? If someone is inclined, then contact me Okay. 911 01:09:54,521 --> 01:09:59,021 Through Nostra. That's the best way. Terrific. Alright. Thanks so much, David. Thank you, Sean.