Decentralized Trust Graph WG GDC Meetup - 2026/08/31

Hey, hey folks. [clears throat] Um, as is not unusual for these uh remote locations, we're still getting connected here. Um, welcome everyone. We'll get going here shortly. Um, we're just figuring out how to get uh connected into the the screen here at Spacian Nations in uh Geneva. So, hang there for a little bit and we'll be back with you in just a minute. Are you guys hearing me? Is anyone hearing? Just a thumbs up from someone. Okay, you got at least at least basic. Excellent. How's it going? Do you hear me? Good to see you. >> Hey, German. excited. >> Awesome. All right. Oh, and Andy's there. Excellent. Andy, I know you're up late and uh we want to get to you here fairly quickly. We're still just trying to figure out uh uh we want to get you up on the big screen here. So, give it just a minute. We'll debug this. Um get your coffee. All right, folks. Looks like we have um it's great. >> Um >> yes, you can join in and that way we don't have to switch back and for I mean you can just present directly if you if you want to. [clears throat] >> All right. Um, we have uh what about seven of us here in the room and and more folks actually on online or as many as seven and seven. Um, I'm going to give me a second. I'm going to bring up share my screen and we'll get going. Uh, have to find the right thing here. Here we go. And this slideshow. Just go that I see it's redoing that. Okay. Share a screen. I'm going to turn up the uh sound here too. Okay. We're going to start here and share and go to full screen. There we go. Good stuff. This on over here. Get that far side. [clears throat] Okay. Is this uh is this showing u screen showing for the folks online? Just a thumbs up from somebody. Okay. Thanks, Gary. Excellent. All right. Welcome everybody to our our first person meetup, our one-y year anniversary meetup. Who was actually here in the last one? Three of us. Okay, that's right. Okay, some more folks are on their way on trains. Um so, uh we'll just, you know, they'll get here when they get here. Um okay, let's click here. Um all right, so [clears throat] it's going to be two parts. Uh we know folks coming in from all kinds of time zones. So whatever whatever you know time you have to come and go. Um we're [clears throat] going to start just quick introductions around so you know who's in the room, who's online. And uh I'm going to do just a short intro to the to the um verifiable trust infrastructure stack and then Andy since he's up nice and late is gonna uh talk to us about Yeah. the uh opportunity for New Zealand. Ah okay. >> It's too late. No, no, you can go ahead. We're going to we're going to get a uh Is that an owl? Is that what it is? >> What is >> Yeah, we're going to get an owl set up here for uh uh for the voice. Fantastic. Thank you. Um and then uh and then we're going to turn it over. So, uh Glenn and Jeff are going to go deep into um uh VTI Stack, OpenVTC, and we're going to do a hands-on interactive demo. Um, is this only for folks in the room or the can the folks online also participate? I'm looking over >> I think Jeff's having network issues. >> Do you want to connect to uh >> it can be connected to all >> to me? There's a USB here and can we get to that? >> Yes. >> Let's see if that works though. I know. I put it by the window. >> Okay. And then on Zoom or Teams, you have to connect it to all. >> Ah, that's right. Okay. Give me a second. We're going to try and connect in. Okay. We're going to go to the audio. >> And yes, >> that's awesome. >> All right. >> That's >> Is it What am I looking for? Is it It might not be seeing it. Uh, hang on. What I can do is sometimes it doesn't see it through this thing. I'm going to try Oops. Might need You want to pull that just a little bit further for >> No, it's >> Ah, yeah. Yeah, just Yeah, right there. Perfect. All right. Now, I'm going to pull out this and let's see if it sees it. If I put this in here. All right. H You see, now it says allow accessory to connect and I think it's switched over already. Um, leading owl three. >> All right. Okay, guys. [clears throat] We've switched over to a leading owl. Can you hear us? Audio. >> Uh, uh, I'm not seeing any. >> Uh, oh, you Steven's hands thumbs up. >> You got to switch the audio output as well to this one. I'm moving. >> Oh. audio output. Close it. All right. One second. That's right. Um Oh, it's on. Yeah. No. Can someone say something if you're online just to make sure we can hear you? >> Nice haircut. Nice haircut. >> All right. Got it. Okay. Thank you very much. >> Appreciate it. All right. Uh Okay. We've been No, no. We're gonna go back here. All right. So, All right. So anyway, and uh we're going to go over we're going to do a slide interactive VTI demo and uh see uh how frustrated we can get uh uh Ron and Jeff. Uh they've been working on squashing bugs all morning, even during lunch. So uh we'll see. Anyway, at that point, then we'll uh uh take a coffee break and um and then we'll do part two um where we'll talk about first person project, go to market, what's what we're looking at in the next 90 days because an awful lot happening this fall. And then lastly, uh prep for GDC over the next three days. Um just highlighting the sessions we're all part of. Okay, that's what we got stacked up here. And now we're going to do introductions. And uh um is is that owl actually showing us as well? Are we seeing the the room view? >> You could activate the video as well. Yes, it has video. It has 360 degrees video cap. >> Okay, let's do that. Hang on a sec. Cool. >> Uh cameras right there. >> That it will compose the screen. >> It's going to figure us out. Got highlighted right there. >> Oh, wow. It has multi focus. So, whoever is speaking is is usually put on. >> So, guess what that means? We're starting with Andre. >> Yeah. >> All right. Nice and short. Just, you know, basically enough background, but last sentence. Favorite film. >> Oh, Andre Andre Kudra CIO of Vizato has been in decentralized energy for more than 10 years now. I've done a bunch of stuff with Drummond also at Sabra and Trust IP and so on. And yeah, looking forward to catch up on on everything here. I have not been there for for for a while but uh looking forward to see what's going on. Uh favorite film uh obviously Interstellar just rewatched it so this is the all time best ever. So >> good shot Jack. >> Hi. Um am I showing up? >> Let's see if it switches to you. >> There you go. >> No. Okay, I'll keep talking. Hi, my name is Jeff Durk. I've been working with Drummond and the FPP gang for the last year plus now year since January or June last year and doing a lot of the uh work uh just uh moving things forward on testing and helping Glenn out on that front and we've got some uh other things related to Kubernetes infrastructure to help host some of the BTA and things like that. So >> and the credentials working group >> and the credentials working group. Thanks. Yeah, there you go. >> Favorite film? Uh, that's a good question. I'll come back to you on that one. I'm not sure. >> Uh, Glen Gore, Affinity. Uh, yeah, Handbomb building lots of things around verifiable trust, decentralized trust graphs, and favorite film would be the best action film of all time, Aliens. >> Aliens. Okay. >> Excellent. [clears throat] >> All right, let's go down to that end of the table there. Okay. Uh my name is Verstein but I come from another galaxy which is the domain name ecosystem >> which of course is strongly you know related but people behave as if you know the separate places and uh I'm desperately trying to establish relationships you know and you know understanding and bridges to the identity ecosystem. There's one project that stands out. In our case, we work for the sports federations for dosport and internet top level remain run by the sports federations. And one of the objectives is to be able to give identity solutions with the federation goals for athletes and coaches and and so on. And those are things that are very dangerous if there are strings attached that shouldn't be attached. Um so um we're very much interested in kind of learning about more ways and possibly also to try to roll this out with a couple of pilot projects associated with um sports federations. Of course for them there is a need also to find um best current practices to do something that they have to do by themselves anyway because every sports federation is actually an identity provider by nature. Oh, excellent. >> All right. >> Okay. Excellent. Um, we we we're friends at DNS. >> We are going to coexist and and hopefully make a beautiful marriage. >> Yeah. [clears throat] >> Anna. >> Yeah. Hello everyone. I'm also from Affinity. Um, I'm Britney in Aid Affinity and under the leadership of Adam Glenn. I'm leading the the team which is building the trust infrastructure and my team was the one who was building who did build the edge page event application. Hey everybody. Uh, Adam L, CTO for Affinity. I report to Anna actually. Um, um, I'm an uh, AI slash identity retrocomputing nerd and my favorite film is Contact with Jody Foster. >> Contact Goody. >> What was your favorite film, Emma? >> Oh, yeah. >> Oh, you didn't say. I didn't say but um I have different favorite films but the last one uh just to hint on it I don't wear >> Prada >> never we very good >> yes you okay >> all right I'm going to finish the last one in the room I think most folks know me but I am going to share because it's really hard on a favorite film but I think the best constructed film is Inception um it's just earlier [clears throat] on the other hand if you haven't seen you you Odysius I do recommend okay so we're going to go to the folks online let's bring up uh let's see well let's see I'm going to go to the uh this view big view okay Andy you're the first one up there so go for it >> hi everybody Andy Higgs. I'm uh Drummond's tour guide in in the in New Zealand. Uh uh you're all very welcome down here. Uh happy to play that role. Um I'm only 10 years into my self-s sovereign identity journey, so I'm I'm a newbie. Um and the last film, I don't know about favorite, but the last film I saw was pretty good. Recommend it. Um it's a new one, The Invite. It's a very clever film set in San Francisco about human connection, relationship, and trust. So, you'll have to check that out. Penelopey Cruz said in it, so you'll like her. >> Oh, yeah. Excellent. >> Yeah. >> All right. >> There you go. Uh Jerry. >> Hey. Hello everyone. Giddy here. Uh dialing in from Bangalore. Uh I can also be a tour guide for Bangalore. You can either walk or ride at the same speed. Uh been with Affinity for a while. Uh started with distributed computing and then now uh really excited about bringing the control asymmetry back uh to the internet where we all can exchange with more trust and reliability. Uh favorite movie Shung Redemption is something I keep going back to every year. It's like a yearly ritual. It's all about shared experiences and I think that's what we're here to make it better. into the classic and next I see Stephen Curran. He might just be monitoring us but if not um >> I'm doing well. Um Victoria, British Columbia where it's very early in the morning. Um I'm working on uh various W3C standards, working on DIDs, working on DID webb and um recognized entities um spec there at at W3C and man that it's an early morning question for that. But 1979, I walked out of a theater in Toronto after the first day that Apocalypse Now opened and was never walked through a city so stunned after watching a movie. So, I'll go with that one. >> Oh, yeah. That was >> love smile night in the morning. >> Yeah. Excellent. Thanks, Stephen. Uh, Mr. Craft, how are you? >> Good. Um, I'll get the movie part out of the way first. Hi, Wilcross uh with Hams. Uh, but the the movie my top five, I would have to say Bladeunner, the original one, not the second one. Um, but I can't specify exactly uh I have to think more about which very very favorite one. Um, we are members of the IRA association handies very recently through our friend Mr. Daryl O'Donnell U hand is comes from the creative assertions working group COG. We are a DOI foundation registration agency have been since 2023 and we provide we're the only uh digital object identifier foundation registration agency uh working in human identity and also licensed human likeness identity. So, we also cover um the instances of licensed digital replicas of humans and fictional characters. And we uh are glad to be here, Drummond. Thank you. We had a great conversation you and I did a couple years ago at or a year ago at the uh content authenticity summit here in New York City. >> Absolutely. And um thank you for all the work you're doing with with cloud. Um super. Uh let's see. I think the last one I see on the list is Vincent. Uh, >> I'm not sure what time zone he's in. LA. >> Oh, LA. Okay. So, it is early. >> Hi. >> Um, yeah. I'm in LA right now. Yes. Uh, I work closely with Jeff and we are building this uh VTA farm platform and for everyone else who want to create their VTA and VTI stack easily. That's why we are working um very hard to make other people to easy to use. [clears throat] >> Fantastic. >> Excellent. >> Thank you. >> All right. Did I miss anybody? I'm I'm just moving through folks on screen. I think we got everyone. Um and we will have a few more folks uh arriving as their trains pull into Geneva. So, we'll call them out when we get here. Um all right. I [clears throat] want to get to the well I want want to get to Andy here too. So we're very quickly we'll just I'm checking this just echoing. Um quick intro to uh what we're now calling VTI stack. Um you >> yeah this >> Oh okay. Got it. Um [clears throat] so you for anyone who's knew this and almost all of us are are fairly steeped in this. Uh these are the um five different orgs that have been involved in in germinating this. Um just so anyone who who you know intersects with Daniellea Barbosa here the the executive director of the decentralized uh I mean LF decentralized trust um in the context of Linux Foundation she doesn't like us to call it the first person project because everyone thinks it's a project at the Linux Foundation and it's not there are projects at the Linux Foundation the ones you see on the screen here plus OpenVTC is a project LFD central trust trust lapse but first person is not a Linux foundation project we just use that term as the umbrella over all this stuff I will talk about it is a digital trust cooperative is building a network but the term project is just uh it's just an umbrella term we've been using since the outset and Daniel said well it's okay if you keep using it just to make sure people understand it's not Linux foundation project by itself um as a result of that project and the collaboration among those groups We did start a working group at trustedia last September decentralized trustcraft working group. We'll be talking about that quite a bit. It is a joint working group of trust and diff. So um um anyone joining was part of either one of those um part of that um and just just so everyone you know knows what we're going to talk about here today now being called the VTI stack our public trust infrastructure um is based on the trust stack and that's been under the way for seven years we're going on our seventh year six [clears throat] months Andre was one of those folks who was there from the start Um uh [clears throat] anyway, Stephen early on Stephen helped us figure out the four layer stack um diagram there um [clears throat] that I think everyone's familiar with. So, and I think folks are also familiar the whole key to the four layer stack is the spanning layer. um the TCP stack that's the IP layer and with trust RP it's the trust spanning layer which gave rise to the trust spanning protocol PSP which you're going to see I don't actually know if that's going to be what you're seeing live today or whether it's home doesn't matter right >> doesn't matter >> that's it then we'll go into that okay so the last thing I'm going to show is a slide I've been using um starting in uh uh last March when I had to a talk uh at uh the quiet open AI, excuse me, quiet um personal AI uh conference in March and I needed to summarize all of what we're doing for an audience. I had 20 minutes for the whole talk. So I I drew this picture. I said this this is a way to think about this. So we start with keys. everything about it including PGP web address is all based on cryptographic keys. Okay. Um and if we care about quantum proof you can even address that. It's all based on on the mathematic keys. And next up is trust we call verifiable identifiers because that includes both dids and uh autonomic identifiers aids which is what carrier uses which can be expressed as a did like the web s but doesn't have to be. That's why we use that and then of course that leads to verifiable credentials. Um now the reason I build it this way is I draw a line and point out those are the core elements of what referred to as self-s sovereign identity and they wanted me to bring down yes this is this is this is what we've been building. This is what they are currently implementing down there. Andy will talk more about that uh digital identity New Zealand and what it is established digital wallets as the sort of core tool is going to be necessary for this new decentralized identity solutions. Now what I explain is that got us halfway there. Okay, these are essential tools. We have to have it. But it doesn't actually build out interoperable digital trust infrastructure especially at scale and especially with the network effects we need. And so what we've been building now decentralized trust app working group in the first person project is the next layer up and it starts with verifiable trust agents and this is what you're going to see um Glenn and and Jeff demonstrate in in depth [clears throat] with these um and this by the way does not have to be an AI agent. It can be an AI agent and we and and one node in the graph can definitely AI agents but every node in decentralized trust graph has a verifable trust agent and they all communicate and do work by trust task which Glenn's going to talk a bunch more about those now composed into verifable trust communities any group any size any place any reason and those can compose communities can have communities and oh interest color in the way it's coming across the industry and at full scale those communities imposed into verifable trust networks and that first person project it goal first person network first person cooperative is always been a verifable trust network of networks and so the point I make is that is covered up there in that corner because again I know I'll put this up for a second. Move that over verifiable trust infrastructure. Now you look at the whole stack and the big the big conceptual um shift is that now it's not just about digital logs. It's about digital agents of all kinds including AI agents. I'm looking at Adam and Anna because Glenn keeps telling me that's your big and you're going to show me some beautiful things here in the next couple days. Ah, so um that's >> Yes, we are. >> There you go. That is the uh the intro that I now use for this. And uh Andy can tell you more about In fact, next up is Andy because it's I don't know it's 1:30 in the morning for you now, Andy. Is that right? >> It's 20 to midnight, but it's it's fine. I've been on European time today. Not really. Yeah. No, I'm I'm good. I'm good. >> Good. Okay. Well, >> with you in spirit, Drummond. >> Yes. [snorts] And given that time, I would put you down in this uh corner of the screen. We're seeing up here. Um, uh, I I wanted folks to be able to meet Andy and vice versa because he did host a fantastic 10 days that I had done in in um, New Zealand in first two weeks of of August and it totally opened my eyes as to what the opportunity is in New Zealand. But no one can talk to this better than you, Andy. And uh, I previewed just a handful. I picked out some of the slides uh that that uh so do you want to just give folks uh again the short overview one question I do timing >> good yeah good >> okay but Andy's got >> okay all right Andy go ahead and I'll just uh I'll just uh you know bring up slides as as you want me to um I'm gonna start with this one is uh just shot that he shared in the newsletter um of the keynote that uh um uh that Dr. Keratiana, what's his last name? Um >> Tyuru. >> Yes. And maybe tell tell a little bit about why we did this joint keynote, Andy. >> Yeah, sure. Drummond. Hey, and thank you so much for this opportunity. Uh there's so much to to share, but I'll try and keep it keep it brief. Uh I love this photo. It's not just because of the green queen and the match of with the backdrop with the the drapes, but uh I just think Drummond it just sums up his trope. He like a kid in a candy store. He's enjoying the conversation with um his new um fast friend Dr. Karatana who's a kind of leading academic down in New Zealand on sovereign AI, ethical AI. Um and you know this is their first time together and they they really did did gel. It was a joint keynote because we were trying to make the actually Karatana was one of the biggest um not was more of a he called us out the previous year for the hoie talking about Marti and Marti um T Marti values the world of Mari their values which are very much universal values to to be Mari is to be human and at the core of that is um mana mana and mana aritanga which is kind of re respect for humans and um you know respect for human dignity and your your humanness and uh we were talking about these values but we didn't have that many well any MAI on stage apart from those uh the MC and those doing the the blessings. So this year we wanted to put Karatiana at the heart of the conversation and it really uh went well and I think the penny dropped for Drummond uh cuz we've always sort of said hey identity is just going to be like payments and um there's going to be a few trusted brands out there and we'll just decide what brand we trust. But um that's never really felt like the right answer. And I think this is Drummond realizing that the right answer is um that there's no um network that's more important than another. We're all sort of sovereign within our communities and you know we're all of equal standing. And I think if if we get that right, I think it's um it's going to go a long way. And I just want to call out the the the sport um example that came through before because you know it's like a club. Clubs are at the heart of this and sport is really the heart of the community here in New Zealand and we love our sport. And interestingly the sports fed federations down here, the major ones like like rugby, they they've managed they've got this data sharing agreement where they've retained the rights of the data to the to the players and they still haven't figured out what to do with it. So, I think that's a really good um early use case um because, you know, they're all trying to figure out how to um you know, stay afloat in this this new era where you're not allowed um well, you haven't been allowed t tobacco sponsorship for a while, but no alcohol sponsorship and no gambling sponsorship. And so, you're going to have to reinvent and it's being harder to balance the book. So, I think there's definitely something there. But this was our um we called it our north star but it really should be I got called out it should be our southern cross the the star um constellation in the in in [laughter] the southern hemisphere is southern cross. So um the uh toa no um mahu tonga mahu tonga but yeah it's decentralized trust is our is our southern cross and um yeah we sort of set out there's really three options and I think all options will be uh adopted to varying extents. I mean the smartphone has the the huge advantage and um Oracle and their wisdom have become the most open decentralized company we we we all know of and they're very very dominant in New Zealand across uh banks and um other corporations but um she just did a post on LinkedIn about the the mini mini Mac mini um trend for and even um you know Palanteer one of my good buddies was at Palanteer uh for many years and I give him a hard time but even Alex has started calling out data sovereignty. So um it's the is the issue of our age. So I really think um the time for the decentralized trust graph has come and for all the reasons we've got here you know community ownership privacy by design all the rest of it dignity really at the heart that idea of mana manaka um and may maybe that's the sort of trust model which I think has definitely earned its place and you know to me it's not about brands anymore it's about you know implementing that at scale and New Zealand's always been a great test market because it's 5 a half million people and two islands but there's nothing really more to it than that I mean we do have a trust framework which is live uh so that enables this um decentralized trust model um but we're still pre-implementation because the self-determination values are so strong here no one really wants a big brother ID everyone so even the word digital ID is is is you know I get I at um death threats as I'm sure some of yeah anyone who doesn't understand what we're trying to do doesn't isn't a supporter and I've just been at a crypto conference in Queenstown and they're like look we don't want digital identity and I was trying to explain where you guys have got to with decentralized trust and they said well if that's true maybe but you know we really don't like this idea at all [laughter] so yeah we've got to do a better job of explaining it um do you want to go to the >> you use this slide to summarize the >> Yeah. Well, it was funny because there there are a lot of a being Queenstown. Queenstown's actually in New Zealand, but the Australians um think it's part of Australia, which I I like to joke with them about because um it's like a principality. It's like Daros and it's a beautiful place and it's like too good for to be New Zealand. So, it's actually part of Australia. About 65% of the expenditure there is Australian. So, we let them think that because we want their their tourist dollars, but um so someone said at the conference, "Look guys, you just got to you've just got to um you're never going to be a price maker. You're going to be a price taker." And just just just copy us. And um and we'll show you how to do aic identity. And and I was doing the the talk on aentic identity. And I said, "Look, I think we can be a prize maker actually." and we we we're going to work with the guys at um First Person Cooperative and and you know hopefully you know we can be early and um we can we can set our own commercial terms for this and this was a slide um it was really the bridge you know we we need some catalytic funding but the 30 million number was just a number really just to say look if you're serious this is sort of what it's going to start to get going but the potential and this was more of an opportunity across Ross, you know, all of the the um extractive um capitalist kind of models or not models, more um platforms, you know, from banks to big tech to everything else that's that's taking value out. And uh I mean, I think even Alex Karp has said, look, if if you guys don't take control of your data, the world's going to be run by two and a half thousand people in California, you know, which is not a good scenario for the world. So um you know once people realize that and we had um Dr. Premod Varmmer from India the chief architect of the India stack talking to us once they understand the power of the unbundling from the platforms and you know this opportunity to put in place this new infrastructure you know 30 billion is 10% of our GDP McKenzie already said it was digital identity on its own was worth 3% but when you take that didn't allow for agentic systems and um you know uh open marketplaces um once you bring those into the thinking um by the end of the conference and of our conference the identity conference that Drummond was at people were saying yeah we can get to 30 we can see there's a $30 billion opportunity here so um it's not so challenged anymore that that's just a really high level business case and I think that's it's actually important for Drummond and and your mission because you know you've got to paint the picture of this is a big opportunity that's broader than just you know streamlining workflows this is about you the value of our of our data and creating new types of open um marketplaces where people's ideas are very valued and and and Dr. to promote really spoke to the um the small business owner, the entrepreneur, the startups and talked about the tens of thousands of startups in India building on the their sovereign stack and you know we don't really have a fintech industry in New Zealand because you know the data is not open you know even though they've been pushing open banking for a long time it's it's still hard going and you know they're forced to do screen scraping and everything else. Um the the other thing that's helping us is every single proprietary system they've tried to implement at government level has been a complete flop. So it's becoming criminal. The level of um you know tech investment that is going nowhere here. So you know we're we're now pushing this as hard as we can despite um the resistance. Uh and this the next slide Drummond. Okay. So this was the um this is our current our current reference architecture which is um something we've been working on with with government with it was the DIA the department of internal affairs now the government digital delivery agency GDDA which hasn't delivered anything yet so I'm starting to call them out for their failure to deliver but really they just haven't moved with the global you know developments and identity which um all you folk have been so um strong on. So um this was kind this is kind of where we're at with our um uh reference architecture. But if you move to the next slide, Drummond, this was um one of our leading architects here in New Zealand who's actually also an accredititor under the trust framework said this reference architecture needs to evolve to this stack which is really the um verifiable trust infrastructure. you know, the the decentraliz trust graphbased stack. Uh let's call it the sovereign stack cuz that's what we call it when you're in New Zealand, Drummond. and uh you know we've added verifiable trust agents in there and um yeah relational identity um domain specific governments uh membership that I think that's such a an important point because you know to be accredited under the trust framework you really need um well you need to prove the trust and the longevity which if you work with the right organizations who are already doing this who are already trusted I think that's where you where you can scale Um and then of course >> and then yeah this was another one that sir he he had done a whole plan around decentralized identify for the health sector working with health New Zealand but now he's added all the other ones cuz now he can see how if we get the foundation right we climb the right mountain because I keep saying we've got to make sure we're climbing the right mountain and you know is it M dos or is it you [clears throat] know jotss or is it something else um and of course it's probably all of them but you know you need to get the foundation right around the key. Um um well the the yeah the diagram that Drummond just saw which really landed well for people here um and um and then as you move up the stack you've got this opportunity and where it really um comes alive is is across the use cases and so we're working with Trust Alliance New Zealand which is the agri use case and they're very keen to see it scale. Education's starting to move, although they're quite conservative and educational institutions. Um, you know, the banks are starting to realize they because of the scale of fraud, they have to do something. So, yeah, the timing is all lining up. We've actually got an election in November. So, you know, nothing's going to happen this side of Christmas, but, you know, the new year, I think, if we we want to be really ready really to accelerate adoption into the new year. So, if we can get some pilots done this year and across some of these use cases, I think we'll be in good shape um for 2027. Um is that is that enough drummers? >> We just had a couple more folks uh join uh we just finished. This is uh Andy Higs who's a the executive director of Digital Identity New Zealand. He was a host of the trip that I took down there. Um and I just he's he's just been giving us an overview of sort of the opportunity down there. Uh we can share the slides later. Um but I just want to say anyone anyone have any questions uh for for uh Andy uh before we go live on it being 1 in the morning hearing. >> All right. It's I think Andy, thank you so much for taking the time to uh share all that. Um I don't know how much longer you want to stay on. Uh we're going to go. >> No, I'll stay I'll stay for a while. Well, I'm really keen to see the demo if we can. Hopefully, that was useful. I really thank you for the opportunity. I wish I could be there. So, it was uh it was really good to >> Oh, well, I mean, what Andy really wants all of us going down there for the uh the next two week. So, uh that that's the his his secret plan. >> Well, no. Yeah. Well, we need all the help we can get because um you know although there are a lot of open people who are you know can see this is the right answer, there's also a lot of incumbents and a lot of uh a lot of resistance which is another reason why I think people think New Zealand's a good opportunity because it's small enough to um yeah to to and and there's kind of two degrees of separation down here, one degree to two degrees. So everyone knows everyone and uh you know we think that if you can get the um you know some momentum you could actually implement this at scale across multiple industries. >> Absolutely. Excellent. Thank you very much Andy. Really appreciate all that. Um next well first we'll quickly let folks that just joined uh do we we we've done one round of introductions. We we'll wait to break and you can meet everyone else. But at least the new folks that just came in. Uh we're using this owl here. So speak to the owl and just uh uh say hello. Um we're also asking everyone to name their favorite film. >> We'll start with you, Darl. >> Daryl, executive director of IRA. Um Armageddon. >> Armageddon. >> Read into that what you write. >> Exactly. Sometimes you need cheese, [laughter] >> right? >> I'm I'm brazed. I'm a little dehydrated. Uh I think I think for this context I'm the executive director of GIF. Um and I invented something called verifiable community and I'm surprisingly gladiator. >> Gladiator. >> I think it's just an inc like it's the first time I understood what male friendship is about [laughter] because I'm I'm not male but like there's something really deep about what male friendship is about that comes out in that film. Excellent. By the way, there is water on the table right behind you over there. >> Yeah, I got some very overpriced electrolytes if anybody else. >> Excellent. Daniel. >> Oh, I'm Daniel. Hi everyone. I'm Danny Glint. Uh I run a small consultancy in AIS. Actually, that's what we do right now. And I'm also part of BIF running there. So, and then what was the last thing I'm expecting? >> Uh your favorite film. >> Oh, favorite film? I don't know, but probably a recent one. The Invite. I don't know if you've heard that. >> It's a new one. It's quite a cool film. I recommend it. >> What's it called? >> The Invite. It's about the invitation for a dinner that goes sideways. >> Yes, it's one one of mentioned that earlier. >> That was my one that that's already been taken. Sorry about that. >> [laughter] >> Uh yeah, Mitchell Travis, I'm a privacy and decentralized AI researcher and my favorite film probably The Fellowship of Ling. >> That one New Zealand film there. >> Yeah, well there actually when I was in New Zealand I saw someone dressed up as Gandalf walking around Christ Church. So >> that's pretty normal. >> That's [laughter] true. have a regular day probably. >> All right, I'm just quickly going to uh for folks that are come later. So, we've done introductions, just did a brief intro to the public trust in stack and then Andy uh just gave us his his overview of New Zealand, the opportunity there. Um and uh and now we're going to uh the balance until we get to our uh coffee break is going to be these two gentlemen here. Um uh showing uh the I'm going to stop sharing and turn it over to them. uh just this is our our deep dive interactive demo of open VTC in in preparation for maybe we'll explain a little bit about that um a month from now almost exactly to the day right now I guess a month and couple days we're going to be in Prague at the Linux uh plumbers conference that's their annual conference of the uh kernel team showing them this as uh the pending replacement for the current PGP weather trust. So, no pressure for these guys. Um I will go to uh not I guess I I'm the one to go. >> There you go. All right. >> We are just little the window. There's one on the left. All the way. >> That one. Yeah, exactly. >> There goes. >> There goes New Zealand. Uh, sorry, we won't find any Australian, New Zealand jokes. Uh, as you go through this, welcome everyone. Uh, I'm Glenn for those that just joined. If you haven't met before, uh, this is going to be as interactive as you want. Um, please interrupt at any point in time. If you have questions, something doesn't land, uh if you don't believe what I'm saying, and you should not believe what I'm saying at any point during this session, just jump in. It's part of this is actually some of the practice and pre-work for the plumbers conference that Drummond was talking about. And there's going to be a whole bunch of super brains in the room, but they've never seen anything to do with decentralized identity before. Uh there's a lot of stuff not in this deck though as well. Um so just because Mitchell's in the room, there's a whole bunch on ZKP that we just don't touch on because until it solidifies a bit, it's just confusing [clears throat] and distracting to everyone where you may hear something that may not uh go as you go through it. Um all right, let's jump into this. So you know the whole premise of VTI stack is that every entity and when we use the word entity entity can be a human, an AI agent, a building, a transaction, uh a business, they all actually need a digital twin that represents themselves. And this digital twin is effectively your shadow self. Jung use a different word but uh in terms of just how you represent yourself in an online world. Um and the thing is you actually have digital twins today. You just don't know about them and you don't have control of them, right? Facebook has one for you, Google has one for you. Apple has one for you. Amazon has one for you. You probably have about 400 of them out there and they're all fragmented. And this is part of the problem. If you think about the sum of the world's information, it's heavily fragmented and you can't get a single holistic view of that for yourself. But that is also true for communities. And again, you know, building on what Andy was saying in New Zealand, in Australia, First Nation tribes have exactly the same challenge, which is they don't own their own information. you know, it's being given to the government and universities who then monetize that data uh on behalf of the First Nations people. They don't get to monetize it themselves as they go through it. So, this is really what this stack evolves uh solves. You'll see there's two themes mixed into this. I will try not to talk too much on the AI side of things, but this started from helping people, but really the reality is we are all heading towards a future where you're going to be working side by side with AI whether you like it or not. So you might as well jump in with both feet and actually give proper structure, authority, consent, approval to AI from the start instead of us repeating the mistakes of kind of the content platforms of today and just giving it away. Um there's really kind of two things that we talk about. One is there's identifiers that you control. That's when we talk about a graph that's the node. And then there's the relationships that you have and you form with people. Relationship with a company, relationships with other people, relationships with a device, relationships with websites. They all get represented by edges on the graph as you go through. So I'm a member of a community, I know this person, etc. Uh the digital twin is not a profile. I just want to make this really clear because it gets um confusing sometimes. your digital twin. It's the graph of the relationships that you form over time and what those relationships start collecting because you start collecting a whole bunch of credentials and meta data as you go through this. And you know these relationships must be portable. They travel with the parties. They don't travel with the platform that you're on as you go through. They have to be reciprocal. I would love to say that I know I don't know who's someone cool right now. Uh doesn't matter. I know Drummond. But if Drummond doesn't say something back saying I know Glenn, it's not worth anything, right? I could know whoever Peter Jackson, you know, unless there's a reciprocal reaction there. It's just a one-way statement. Uh these things have to be revocable. I know Peter Jackson. He knows me. We then got in an argument, had a falling out, and now he wants to revoke that relationship. It has to be composable. So, one of the things that you have to do is take all of these credentials and be able to construct the graph from the credentials, not the other way around. Uh, a lot of platforms will manage the graph as a database and then construct credentials from the graph. And the problem with that is someone's owning the graph and they're claiming ownership on behalf of a network as you go through it. This is the kind of other way around as you go through this. Um, yeah, whatever. We're going to go through some of this. Uh, I will say, and I know some of you have been in this industry for a while, so here's I do say some provocative things sometimes. I think the original philosophy of SSI, it was great. It did its job, but it's also failed. Um, it did not drive the change that people wanted. But as Newton would say, we all stand on the shoulders of giants. This really is kind of kind of ter like it's SSI 2.0. It's standing on the shoulders of what SSI1 did. Uh so it's just a better implementation, a more pragmatic implementation and really the focus is stitching a lot of the discrete components because part of the failing of SSI you had people solving DIDs but not talking to the people that were working on verifiable credentials, not working with people who were doing messaging, not working with people who were trying to create you know the first registries of this. So you ended up with all these wall gardens and you know what really to me brought this together was actually your first person project white paper Drummond uh which really just brought a lot of this together and the concept of the verifiable trust agent was just a no-brainer of it's one of the missing pieces of the architecture as you go through this and my background is architecture so you know if you don't get the design right uh good luck building code to it or anything else as you go through it. So what we're building is the VTA, your digital twin, the VTC, the trust community, which is uh managing the policies, the governance, the collection of people, and then we're building the ecosystem around this as well. But we also have AI agents working in there. And it's crazy what we can do with some of this technology already. And we're just at the very start of it. uh none of what you're going to see existed in January of this year. So this has all been built in the last six months. And to kind of just give you an idea of how driven the working groups and the teams are on this, there's, you know, many different DTG working groups on this. There's uh DID groups that have uh tapped in to help with some of this. But, you know, you're talking over 4,000 commits, 24 repositories created. Um, huge millions of lines of code already as you go through this. Uh, it's just crazy complex. Um, very cool as you go through it. But what we're seeing is now that a lot of the foundation is built, building on top of this is much easier as we go through it. Um, while we're in Europe, I think it's always good just to give a contrast. People sometimes ask, does VTI replace things like EUDI, IDAS? No, it doesn't. They're kind of two different approaches that coexist. There is and always will be a model where governments want to have kind of state driven identity, state driven credentials. Uh, they will publish that. Um, they have certificate authorities. It's accredited. There's a whole bunch of things and for good reason, right? A driver's license allows you to drive a car on the road and can be internationally recognized. Your passport gives you the right to uh cross borders and that is universally, you know, kind of recognized by people as you go through it. The wallet consumes trust. It's a presentation of a government credential. It's actually not a vouch. The government's not saying you're a good person. They're not saying you're a good driver. They're not saying you're going to be a good tourist as you go across those boundaries. Uh that's the limit on VTI. It's driven by the community. And by the way, you can be a community of one at its smallest. You have many devices. You uh have relationships with those devices. Sounds so weird when you say it that way sometimes. Um there's kind of trust data which is queried live. And so a big piece of these are things like the trust registries. And the trust registry is what keeps things honest. I've given you a credential. I can't take the credential away from you, but what I can do is update a trust registry and let it know that it's revoked or there's a something has changed on that as you go through. Uh the anchors are the DIDs themselves and the keys in those DIDs. Membership is based on the governance and what it means to be able to join a community as you go through this. And the community itself is vouching for its members and kind of publishes the roster if it wants. So a community can be public public something like the Linux kernel for example but it can also be private and if it's private it may not be recognized by any other community. Uh it may be private but only be want to be recognized by a limited number of communities. That's their right and they can do this. And a question that comes up a lot is, well, what happens if someone creates an a hate group community? Um, how do you stop that? The simple answer is I can't stop it. If people want to create a hate group of whatever that is, go for it. But what will happen is that hate group would be very unlikely to be recognized by any other community or group that doesn't subscribe to the same philosophy. So the idea being is that you kind of starve out the relationships or the graph of responsibility and this is again where things like accredited trust networks will come into play which is you know a trust network for health care is unlikely to recognize a group that might be I don't know antivax for example uh if the government has a vaccination status would be a good example of how that works. Um, so they don't replace each other. They I think they work together uh as they go through this. It's really VTN's or the verifiable trust networks and the verifiable trust communities. They're very much a federation or a fabric of what they choose to do. By the way, there's not a single VTN. First person project can define a VTN that they will vouch for what it means to be a member of the first person network. But I can go create a competitive network and see if I can curate a better list with a difference governance policy as we go through this. And this is where they start to stack up. And a community may choose to be part of the FPN and the GGM. Register that name now as you go through it. Um, so there's a lot of flexibility in this, but that's also where some of the complexity uh comes from as you go through it in terms of, you know, kind of a a trust stack. I just realized people at the back of the room can't see that. If you join Zoom, you can see it on your screen easier. Uh but the biggest thing about it is at the lowest level is the foundations the DIDs the keys trust registries witnessing infrastructure kind of you don't really need to see that on a daily basis then there's the trust spanning layer on top of that which is things such as the asynchronous messaging systems that it use relays delivery and we'll talk a little bit uh about that then there is something new called trust tasks so before we started down this path, you know, didcom was a messaging transport that had built in it not just a messaging layer but also the start of a protocol layer uh which worked well but the problem is it was didn't work outside the world and you start using TSP tsp uh infamously went we are not a protocol layer wheelie but uh neutral so trust task is effectively now a cryptographic way of between two parties saying, "Adam, I want you to do a task. Here's the def definition of that task. Everything that is needed to complete that task is within it." And then Adam gives a cryptographic response back saying, "I either uh completed that task and here are the side effects of the task or I've rejected the task or you know there's a next step you need to do." Uh very very simple. There's 350 trust tasks defined already. uh they're extremely strong. One of the things in the decentralized trust group is they have a risk and harm prevention uh kind of research team and doing systematic reviews. It's being focused very much on these trust tasks and looking at you know how they work. You then take those trust tasks and create trust ceremonies out of them. So trust ceremony is if I need to stitch together more than one task to do something like joining a community well I need to do some KYC I might need to check a few other things that is a ceremony we'll talk about that credentials consent and then at the top of this are the the applications open VTC which is the CLI tool that Linux uh kernel developers will use there's agent memory MCP bridges uh AI tooling whatever it is that you want and this kind of loosely follows the trust over IP model as we go through it. Um and we kind of talk through that. So one of the big things we did was there was a workshop in January February uh some of you around the table and online were there and we realized that while there's a lot of standards and specifications there was very little that was working across it and people who had been trying to build across the standards you just run straight away into incompatibility issues and so we decided that what we needed to do was build the code at the same time the specifications are being drafted and worked on. And in hindsight, yes, it was the right thing to do, but at the time is a little controversial. Uh people were not fully convinced on it. Um but I would say that what has happened from as we build is it's not that the build is ahead of the specs. It's not that the specs are ahead of the build. They're kind of doing this all the time and moving around. But what I love is that we're getting evidence of where the specs are really solid and absolutely helping clarify things, but also where they just simply don't. Uh I'm not going to go into the details, but if you're on any of the working groups, you'll see there are issues galore being talked about, PRs happening left, right, center, and that's good. That's what you want to see in a vibrant community. uh some of those communities were pretty dead before this occurring. They' kind of been frozen in time uh as you go through this. So let's see have a look here. Uh this is kind of some of the examples of what you can do uh based on this. So one of the concepts of your VTA is the VTA holds your keys and it holds your credentials and can hold kind of meta data. when that VTA is working on your behalf, it's not handing out the keys. Um, you can sign, you treat it like a signing oracle. You can kind of hand it information and it will generate the signatures for you, hand them back. Um, it can enroll and revoke devices that you connect to the VTA and therefore to other aspects. Um, there's a lot of really fun things in there that you can do with the VTA. the what it actually holds is a whole bunch of information. So, your VTA is holding all of your key material uh and signing keys and you have control whether they leave the VTA or as in they don't leave the VTA or they're actually going to be given to certain applications and services as a proxy as you go through. So, I'll show you how that works with Open VTC when we get a demo. Uh it's a credential store. So it does all the things, issues, holds, presents, and revokes your credentials. Uh everything you expect there. It's a secrets vault. So again, we want to kind of shift left and work with existing website. So it can be a password manager for you if you want. Uh it has agent memory. So you can actually tie your AI agent to your VTA VTA and use it as a durable secure store for your AI interactions. uh which is really really powerful when you take that capability with a VTC you can create AI community memory which is really powerful so this one of the things that we want to unlock whether it's before Linux uh plumbers event or afterwards but imagine the DTG working group had an AI shared community memory that had all the specs latest decisions being made everything in one place so that when you join that and you say I'm going to use an AI agent to help just distill and explain things. That AI agent instead of trying to figure it out on its own is now talking to the DTG shared or memory as you go through this. Uh, a lot of people ask, how do we move so fast on the build if you're kind of looking at the build site? That's part of the secret right there is everything is going into my VTA memory for myself and my agents work off it. Uh, it has application state. So, one of the fun things, if you've ever uh configured an application, then you lose a device and you think, "Oh crap, I didn't back up the application. I've lost that aspect." The VTA can do that for you. Uh it has approval policies. So, a big thing you'll see on the VTA and VTC is everything is policy driven. And when we say say policy driven, it's open policy agents, OPA. And one of the things we've done is taken the policy management and pushed that into the approval chains. So for the first time you can have intelligence as to depending on who is asking for an approval and what the trust task they need approval for I can handle that approval flow differently. So again, this comes from the expectation that in the future, you don't want to wake up in the morning in an AI world and say, "Good morning, Glenn. You have 123 things to approve." Right? That is what is going to happen. It's going to be GDPR all over again. Approve all cookies. Just click, click, click, click, click. Right? So what I can do is I can say actually my VTA, you are allowed to approve some things that are low risk on my behalf. I may plug in an AI agent to deal with things that are of medium risk. And then it's only certain critical things that meet the policy criteria I set that comes to me asking for a human intervention. The idea being if I only get asked once a day, once every two days, I'm going to pay a lot more attention to it than the 100 plus approvals a day if we kind of do this. And again, this is really powerful. This also works by the way in the VTC. So really good example here are things like a family unit. As a parent, do you want to approve things on behalf of your children? You can start programming that. For example, do you want to approve things on behalf of elderly parents that might be at risk of getting scammed? You see an unusual financial transaction. You can program the approval to be, hey, all of the kids, the adult children, if three out of the four approve it, then it's okay. So you can really do a lot of fun things with this consent management, device agents, uh identity and operations with this as you go through. So uh lots and lots of things you can do with it. Uh one of the questions that comes up a lot is well what about in the VTA? Do I just have one identity? No, you do not. You have uh trust context. Trust context you can break up however you want. This is just one example. I can have my personal IDs for Apple, Google, Facebook. Each of those is a context. What that means it's like a logical partition. They can't see each other. And I can start doing policy saying anything in personal always keep separate from work. Apple can't see Google. Something at the personal level can see Apple, Google, Facebook. Uh you can really do whatever you want there. This makes it really easy by the way. If you sign up to like a new site or a new application, you just create a context and if a 24 hours later you don't like it, just blow the context away. Everything gets shredded with that trust context when it goes away as you go through it. Uh this also means by the way you can have multiple VTAs. So you don't have to have one VTA doing everything. You can have a mesh of VTAs and there's actually a fleet management tool uh that helps you manage that. And so you can start spreading out the trust context across a fleet of VTAs uh if you want. Um just be warned, it gets very very complex very very very quickly as Jeff and I will attest uh on some of this. You'll see a bit of uh a multiVTA demo as you go through it. Uh the verifiable trust community. This one's really simple. Uh a community owns its own policies. We will have out ofthe-box policies that allows a community uh to get started, but they get to choose what it means to join. Must have proof you're an adult, must be in Switzerland, must be a member of DI. Sure. Whatever it is that you want, you get to set. Uh that is set at a policy level. Then there's a policy engine that runs all that for you. Uh as you go through the trust ceremony, there are side effects created. And those side effects are things such as you got accepted and here is proof of that by a verifiable membership credential, a verifiable endorsement credential. Uh the VRC's the relationship credentials you create within that community and people outside that community if they want to ask you know Glenn are you really working on the Linux kernel project and if so are you a contributor or are you maintainer or are you a core level? Well that's where the trust registries come in. You can ask the trust registries that every VTC has for that type of information and they get to choose what goes in those trust registries, who can talk to the trust registries, whether they're truly public or you must have a credential saying you're recognized by the trust network. Uh so a lot of this uh is all here. All of this runs on DIDs. If you I just find it amazing that even in our own industry, people still don't understand really what a DID is. So again they really do three things and you must understand these three things are the entire backbone everything in BTI works on first of all it is the unique identifier they are globally unique second they are what holds your public key infrastructure proof of ownership is sign something with the key in the if you can't sign it you don't have ownership of the DID as simple as that the third one is what are the service endpoints how do Do you reach me? How do you talk to me? Anna, do I talk to you using Ditcom or REST or uh TSP? Do you have a trust registry? Are you a community? Are you an individual? All of those details are the service endpoints in your DoD document. Uh as you go through it, uh the ones we support by default, you can kind of see there the usual suspects. I will say for us did webv is unashameably the star of the show. Um it's designed to be you know highly highly durable. Uh it's verifiable history because and you'll see this there are already that we are using in this stack that have got hundreds of changes to them as the keys are getting rolled service endpoints change etc. And webv just works beautifully. chef's kiss. Uh, and it's extremely performant compared to other did methods that try to do this uh, as you go through it. So, I think Stephen, you're on the call. Thank you. Uh, great protocol. Uh, as you go through it, um, things are ced with a DID. It's okay. Um, you can have this kind of immutable methods like DID key. If you want to change the key, it's an entire new DID with a new identifier. or it can be mutable like webvh uh and others that you go out there. I will say that u if you want to work with a vti stack there's kind of a few core protocols you must support. If you can talk did webvh, you use curve 25519 as your cryptography algorithms and you can talk didcom or tsp, you can fully interact with everything in the vti stack. That's the basic three things you need. If you don't have those three things, it's just another walled garden and you can't uh talk to it. Everything else is built on those fundamentals uh as you go through it. um in terms of then uh how transport works. So there's your one VTA for simplicity, there's the trust tasks which really kind of define how this works and I'll send some links out for this. Uh but how you transport the information across, we actually don't care. uh it can be trust spanning protocol it can be didcom v2 it can be rest there's lib ptp coming uh these are just different transport mechanisms that work and this is being used right now when you see the demos you'll see divid tsp and rest all being used interchangeably the system does not care we do have a weighted matrix which is if you have a did document that says I talk rest didcom and tsp it will negotiate in order of TSP first. If that doesn't work, fall back to DIPCOM. If DIDCOM doesn't work, fall back to rest. And that's again based on the opinion from a security and performance perspective, we think they go in that order. uh in terms of just modern aspects of encryption. Um, one of the important things with any messaging system is well, how do I talk to you if you are on a air gap network, if you're behind that gateways, if you're moving around? Um, the great thing about your VTA, if you don't use REST, which we would not recommend for a VTA, then you can move your VTA anywhere in the world and there's no URL that you have to keep updating and repointing via DNS to where it goes. You've got these messaging relays or mediators that allow you to abstract where you are. And this is what allows you to punch holes through firewalls, through NAT gateways. It can obuscate where you exactly are. There's some pretty clever things you can do. Again, I come from a bit of an interception background, so you can do things like hide behind mediators in different countries to create constant timing attacks or even based on latency. I don't know where you are. And you can do some really fun routing such as mediator to mediator to mediator. So for example, if um Daryl and I want to send messages to each other, but I might have Jeff in the middle, Daryl really never knows what my mediator I'm using. Jeff will know what mediator I'm using, and Jeff will pass that to Daryl, but I won't know actually what mediator might be between Jeff and Daryl as you go through this. So this is where you get a whole bunch of security coming in. If you've ever built an async messaging bus that has multihop routing like this, uh read between the lines. It's super complex, hard to get working. I wouldn't say we've got it perfect yet, but uh it's going really well. So, what does it look like on the wire? Uh it's pretty efficient. Um one of the things, again, you'll see this in the demo. You may be thinking, "Ah, problem with anything decentralized, it's fat, heavy, and slow." You're not wrong. Um but what we've been doing is driving this to an efficiency where for example the authentication flow we have is the fastest authent authentication flow on the internet right now. It's faster than anything else. Anything open ID oorthth even simple username password it's as fast or faster. It's crazy what it does and it's because of uh some of the things it's doing. Uh basically at the inner is your trust document. It is also signed. You then wrap that in the messaging layer, a sealed to the recipient. There's then routing envelopes that go around it. Again, we don't care whether it's limp P2P, DIDCOM, TSP, and then you have a transport agnostic, whether that goes over a web socket, HTTPS, HTTP3, quick, uh you can print it out on a piece of paper for all I care, it doesn't matter. And again, this is really important. And you know, Adam, uh, our CTO, you're saying is, you know, if you're going to do it, do it two or more times because it just makes you have an interface and abstraction point for it. And so this is some of that design thinking coming into this. Um, what this means from an implementation point of view, if you use the trust task library, which is all code gem, so you get that for free. It also has the bindings which means you can plug and play and move from DICOM to TSP without really any code changes whatsoever. And this is something that before we had this library was almost impossible to do uh from people as you go through. In terms of what this looks like in reality and why we are big fans of TSP, I was a big fan of Dipcom. I love Dipcom. still has a soft spot for in my heart. But the problem that DcomCom has is as you start to route it, if you do lazy routing, you're losing some of your privacy. I can see everyone in between Drummond and I can see that Glenn and Drummond are trying to send a message to each other. So the way you would protect that is I would nest the envelopes. I would say I'm going to send a message to Drummond, but I'm going to send it via Adam. Adam's going to send it by Anna. And every time you do that, the message grows 1.6 to 1.8 times per hop. Or a, you know, kind of 100 kilobyte message quickly gets 20 to 40 megabytes in just 10 hops, which is not that many out there. So what TSP does is TSP brings a lot of what DIDCOM does. Uh but that 10 hops of 100 kilobytes only grows to about 120 to 130 kilobytes which again if you're dealing with mobile devices you know I live in Southeast Asia uh networks are not always the best in developing countries people on cheap devices you don't want to have uh very heavy protocols or be sending around megabytes of traffic as you go through it. So this has been really really good learning. When we started this, we didn't know that we would be using TSP until we ran straight into the DICCOM uh size library. Yeah. >> What about postquantum computing crypto primitives? >> Fully fully supported. So for >> this will bloat all your uh device coding there. >> Not not by too much actually. Okay. >> Not by too much. So if you look at like MLDDSA 44, the keys get bigger, but the actual message being transferred over the glass, it's still the same size. >> Okay, cool. >> Yeah. >> Uh the speed at cryp encrypting and decrypting slows down the problem because this will have a latency effect, >> but it's not too bad. Um again, depends which PQC you're choosing to use, but yeah. Um all right, trust tasks. So I'm not going to go through this in detail but basically everything is a URI 350 specs uh five transport bindings. Basically what happens is um you have the ceremonies and you can effectively build whatever you want out of them. Everything in VTI that you'll see is all running over a trust task of some form. uh and the ceremonies that you go through this you can also do some fun things and I'll give you a demo of this in a second is one of the problems let's talk about DID management if I have a web VHD ID it has pre-rotation key setup for example which means you can only legitimately rotate to the next log entry if you have proof that you can have ownership of the pre-rotation key and you sign the next log entry with one of those keys. But if I'm hosting that DID on a did hosting service, that means I'm going to give the keys to them if they're going to be able to edit that DID or make a change to it. So what we've done is created uh delegated trust task execution which allows you to effectively have sensitive data published onto a third party but never share the keys or the ability for that person to make changes to your information. This started off from the did management from a UX uh aspect but interestingly has really become powerful in things like healthcare industry. You could imagine being able to do things like share a medical record but no one else can edit it for example and if they do want to it would be a delegation to a doctor or a nurse to make changes to your electronic medical records as you go through it. Um it's really cool. Let's have a look here. What else we going I'm just going through this. we get into demo time. Um there's a lot of learnings that we've gone through this. Uh there's a browser plugin. There's, you know, we've one of the things we wanted to test was could you have Chrome talk natively didcom and TSP from a plug-in without using any REST. Uh good news is you can. Um, and so when you see the browser plugin, it's all using native messaging protocols uh from it, which is really, really cool. Good luck getting that through the Chrome store. Having that debate with them right now as you go through. Um, one of the things we also did is we are targeting the Linux kernel and therefore by extension Linux foundation. Um, one of the things we do is at Affinity, we really think about scale a lot. happens when you've got Adam and myself ex AWS um tech talent. Uh and so one of the things we modeled is what would all of the Linux foundation look like? You know, you got over 256 communities made up of six big networks, 8,300 entities, 18,000 credentials. This is in reality what uh these things kind of look like as you go through it. And so this is it here live. You can go in and kind of see what is happening. And you can actually see the credentials here live as you go through. So these are real DTG credentials. This is a real graph made up from a lot of these relationships at scale. And what we realized when we did this is just to model the Linux foundation the credentials here is about 600 megabytes of credential information. So it's big. Um there's opportunity to reduce that down. Uh you know number of edges is extremely high. You can all do types of fun things. I can send this out. You can kind of have a look at it uh if you want as you go through it. But interestingly, even at 600 meg, creating the graph from that and working with the graph is super quick, super fast. Um, as you go through it, it's just the signing things. This is where PQC will by the way. Um, having the much bigger signatures will blow that out to a couple of gigs as you go through it. Um, if you are on the DTG working groups, you'll see there's quite a lot of tension between Drummond, myself, Jeff, and others, Brendan, Alberto. The reason why we're very very pedantic on keeping the DTG credentials as small as possible, as generic as possible, and kind of trying to overload them a little bit so that they can do as much as they possibly can. It's to keep this size down. And again, it's something you don't understand until you build it at scale and try to interact with it uh as we go through it. I don't think any of us thought a Linux graph would be 600 meg in size. If you start thinking about this going out to, for example, all AI agents joining a network, there's going to be a lot more than 8,000 nodes in that network. Um, China with a billion plus AI agents. That's a damn big graph as we go through. And so these are what we are building and designing for uh kind of as you go through it. Um, you can kind of do routing, you can do everything here. So what can you do today? Um you can define one of three communities. So you can define a public community. So this would be the uh Linux kernel would be the exemplar. Um you know everything you see is able to be seen by everybody else. I will note that while there is kind of KYD know your developer that doesn't mean that publicly I have to be known as GL. So, I can still be anonymous to people out of the community. It's just part of my vetting. I just had to show that I'm Glenn. I'm an Australian citizen. I'm not from a um sanctioned country. Uh I know people who vouched for me, but I can be Donald Duck within that community to others. Uh there's no important that is a governance policy decision at a community level. The second use case is what we call the house where nobody uses names. to get into the house, you have to be known, you have to pass some checks. But once you're inside the house, moving from room to room, well, if you're in the house, I know you've already been vetted. It's okay. You don't need to do anything else as you go through this. And so one of the examples here that was talked last week on the prem school I think Drummond this came from one of your meetings in uh New Zealand was what about things like women's shelters for example extremely high degree of privacy but they also need to have proof of government identity so they can get assistance and the right you know kind of health support mental health programs etc. There's a lot of support material just making sure that is being managed effectively as well. So how do you do these two things? I need to be known. and I need to get access to the government programs but I want to be absolutely you know no one else knows that I'm getting this assistance or help as you go through. So that's that uh you know kind of the pair-wise aspect. There are you know onetoone relationships and then there's a community um aspect and the model here is actually think about an enterprise company they inside the company actually may be public but it's private to outside but even inside the company they may have relationships with their ecosystem that are private depending on who the ecosystem player is. So this is where you get a lot of kind of mixing together of some aspects of being public, some aspects of being pairwise and you're using those in a hybrid model as you go through it. Uh so that exists today. U by the way you as one person you can actually be a member of three communities with three different styles here and you can choose whether you want to use your same identity or whe you want to have three different identities as you go through this. Uh it's your choice. Second thing is the join ceremony exists. So today the way that works is you have openvtc which is the developer tool. Uh I will give you fair warning. It's the Linux kernel team. They live in the terminal. They love text braced user interfaces. It is not beautiful. It's not sexy. Uh that's what they want. That's what they get. Uh you've got VTC which does have a web interface and then you've got your VTA as part of. So those three tools all work together. Open VTC asks VTC for what's the join manifest? Drummond, what do you need for me to join your community? That's the join manifest. I then go and gather the material. Drummond said, I want to know I'm an adult. I uh am a member of LFDT, for example. I then submit that back. It'll present that back to Drummond's VTC. uh where he then does a whole bunch of verification and validation checks and then he decides based on that material whether he's going to accept me into that community or not and let's say he rejects me then I just I may not get anything or I may get a rejection notice or he's accepted me and then I get a verifiable membership credential and an endorsement credential saying what's my role within that community. So it's quite simple uh as you go through it. And then there's the recognition. If I want to be recognized outside of that in a way again this is where the trust registry query protocol comes in where a community can start saying hey Adam do you recognize me uh because I'm saying I'm coming from Drummond's community and Adam might say I have trust because Drummond's community is actually part of Grace's greater community of communities as part of diff for example. So you can start building up uh the hierarchies as you go through this. So that exists there today. Um the other one I'm not going to show this today. Uh but this is really what we're building for the Linux kernel is if you know you're developer then apply that to the world of git. So when you do a git commit which is committing your changes to the code best practice is you should sign that git commit and if I can sign it then it's proof that it was uh developed by glam submitted by glam committed by glenn and then if I have my you know use GitHub here as an example if I have GitHub actions that get run on every single time I do a commit or when I do a pull request the GitHub action can now talk to the open-source community and say, "Hey, is Glenn even allowed to do a PR? Can Glenn merge his own PRs?" Uh, does Glenn need to get Mitch to approve that? And once Mitch has approved three or eight or 10 of these above Glenn, we're going to elevate him so that he doesn't have to go through a uh vouch from another maintainer as you go through it. So, all of this is like really quite fun. Some of our own team are using this already inside Affinity uh for code development on the platform side as you go through. It's really really cool. What's nice is we hacked git to get this to work. So for the geeks in the room uh git only allows uh GPG signing or SSH signing. But in an SSH sign, you can put what the key identifier is. and it does take a div identifier and SSH supports ED25519. So we've hijacked the back of SSH and interestingly when you do that it shows up with the green verified tick on GitHub. >> Did you check it in as G10? [laughter] >> Did you check that in with G10? >> No, I checked that in as GG. Uh no, not the ZX hack. Uh for those that not know what Daryl's inside joke is there. Um the other thing you can do is with your trust agent you have this concept of linking multiple devices to your VTA. So I can have my laptop, my phone, my tablet, desktop, whatever I want and I can choose depending on the trust context what gets shared across these things. One of the most powerful uses of this is actually when you start using it for agent memory. So I can store something and say I don't know Drummond's favorite food is beer and uh is that a food group? I think it is. Um and then when I ask that on any of my devices, it's going to come back through it. So those of you in the AI world, if you know things like ME zero or Enthropic has just been adding this themselves into their own tools, this is a way of having agent memory that is agnostic to any of the AI tooling you have and you can plug into it via MCP. There's also an MCP interface for your VTA. Be careful with it. Um, but you can ask your agent to manage your VTA. So depending what I got this credentiing, do I want to do something? So you can start doing if this then that type triggering as you go through it. Uh, there's plugins for this. There's an MCP server agent. There's a claude code plugin directly that you can use. Super cool. Uh again, just be careful if you give it right access. Um default is just have it have read only as you go through it. Again, you can have your AI agent plug into the approval policies so that depending on what it does, it will ping you and say, "Hey, your agent's trying to access something. Do you give permission to it?" Yes or no? All right, enough talking. Uh I'm going to apologize in advance. Jeff, is your computer working? >> You're back online. So Jeff's having a lot of things. So maybe I'll just give you a quick little demo and then Jeff's got a hosted version of this. Uh let's see. Everyone pray to the demo gods. Uh and see what happens as we go through this. So um so first thing I'll show you is this is open VTC. So I'm going to tell you this is live. Um, nothing here is pre-recorded. It's probably going to crash at some point to prove that it's real. So, this is a blank uh profile. It's basically asking me I can recover from backup. We're going to do a new profile setup. And so, the first thing it's going to ask me is, can you give me your VTA's DID? And so over here we have the personal network manager tool which I can already see there's my uh web VHD ID here for this. Look at that. There's an error down there as we go through this. So all I need to connect is I'm basically using my public infrastructure. So I'm telling my client here who doesn't know about my VTA here's what it can do. It's going to resolve that DID and get the service endpoints and basically negotiate how to talk to it. Now what it's saying to me is okay you need to because this is a new setup I want you to create a trust context we're going to call it open VTC and can you give access to this administration DID that we just created itself which is just an ephemeral key and this is going to expire in 1 hour so sure no problem we're going to run that over Yeah. Okay. So, this is the P&M. So, the PNM is just the interface to your VTA. Remember the VTA is not running on my my my machine. My VTA is actually running in Ireland at the moment. Um, so it's created the context. You can see, you know, there's some tech info in terms of the way the context and keys work. They are derable, which means I don't have to back everything up in the VTA on every single transaction. I can kind of derive it after the fact as we go through this. So, what I've done now is I've effectively OpenVTC has figured out how to talk to the VTA. It has created a ephemeral key on the client side and it said, can you create an ACL entry on the VTA so I can connect to it? So we say yes, you can. And now it's negotiating access and it's just rolled the keys. So you can see here it just rolled the admin key. It went from the ephemeral key to a DID key that the VTA actually just sent back to itself. Uh as this went through this, it's authenticated via TSP, so it could have fold back to didcom or rest. Uh as you go through this, um everything's good. Uh now it just goes through the rest of again because this is the Linux kernel. Um particularly you know people like Lenus and GKH they will have some mandates on using hardware tokens but we're just going to skip that today. We're not going to require an unlock code because we live on the edge. Yes I understand that's dangerous. Uh that's it. I'm in. I'm connected and now I can start interacting with things. So I can go down to my VTA service. We can see that uh I'm connected. I don't have any persona DIDIDs. I have no invitation credentials. I don't have anything. I've got all my logs. So, what we're going to do is we're going to join a community now. And so, similar again, the join flow is it's going to ask me for a um so this is what a VTC looks like. So, this is variable trust community and here is my community did. So I want to join this community. I put that in. And so what it's doing here is I've just created from the VTA myself. I've just created a brand new DID for this persona. So I'm not joining as my VTA. I have then asked the VTC what do you need to join? I don't have a verifiable invitation credential. So effectively it has uh joined a queue and I'll show you this in a second. And now what it is is I'm in a wait state. I've basically said here's my persona did. It's called sorry dessert. It makes up names on the fly. Uh and it's waiting for a response. So we can actually see some of this pretty fun. So if we go over here, this is a DID hosting platform. I'll come back to this and we can should be able to see sorry dessert is here. So this is the did that was just updated. We can see there's no traffic to it. Here it is. It's a web VH. I didn't have to worry about setting up keys. I didn't set up pre-rotation. We can see here, you know, it's a portable DID. Um, it does have the key management uh pre-rotation on. It's not deactivated. It doesn't have any watches or witnesses. Great. This is working perfectly. If I come over to the VTC and log into this, I can see here everything is looking good. My trust registries will arrive. This is the trust ceremony. So I need to basically work through this for a join flow. And what's happened is I've effectively hit the problem where I don't have a valid invitation. I do not have a VIC from the VTC. So what's happened is I've gone into a moderator queue. So this is an example where the policy starts to work. So if I go over here to join requests here, I can see sorry dessert is in pending and I haven't presented any credentials because this ceremony just for simplicity is not asking for anything. So if I go let's just make sure this is going to work. I'm going to click on approve and I've approved that and over here we're active. So I didn't have to there's been no transfer of secrets, passwords, anything whatsoever as I go through this. You can see here that I have my membership credential and I also have a RO credential and I can see these by going over here and I go to membership and you can see here here is the actual decentralized trust graph credential. Uh what is this one? This is membership and it's got 29 days left and so it will auto renew and revoke things as you go through this. So again did I have to know anything about the VCs, the DTGS? I just joined a decentralized trust graph and didn't even know it. And I can prove that because here is the trust graph uh starting to come in as you go through this. So there is sorry dessert and if we click on that we can see that it is a full connection. Uh it's a VMC pair and that sorry dessert is a member of data. Forget the UX for a second. Uh some of this stuff is coming hot off the brush. >> That trust wrap. So for this one, this would be the VTC administrators own. So this is a only within the community at this point. >> And then if it's a network of networks, >> then the community would choose to export parts of that to outside of themselves via the trust registry. >> Exporter, expose, sorry, expose. >> Does anybody need to see the trust graph? >> Because Adam might be running open-source community. I might be running an open source community. And I want to recognize that if you've been an active member in Adam's community and he says you are a good developer, you get instant access to my community. So to do that, I want to see his trust grow. >> Why would I need the graph? I just need the verifiable credential from the person who says I belong. >> That also allows me to start seeing the graph. So I don't have to see the whole graph. Just seeing the credential allows me to start seeing part of the graph. >> Why would you let me see part of the graph? All I need. It's your choice if you want to. Some communities will want to expose the graph. You have the ability to either expose all of the graph, parts of the graph, or none of the graph. >> When I'm an individual showing my credential, how do I know what I'm exposing about my friends? >> Because the VTC part of the policy of what it does should be describing what what is it doing with your information as you go through this? I >> just really don't understand why I would want to expose part of my graph to another community. I don't understand the purpose of that. >> Like if I say this is a trusted member, they're a trusted member. Why would I need to expose any of my graph? I don't understand. >> But that is that is exposing part of your graph. That is a by definition of DTG credential. It's just you're exposing one node of your graph. >> One node fine, but I'm not exposing. >> But that's a graph. A single node is still a graph. So this is a semantics. You don't have to expose the entire graph in its entirety. But if we start to share 10 members from one community to another, you've now got 10 nodes. That is the side of a graph. And they may start to create VRC's amongst themselves. Now you've got a new graph starting to form independently across two communities. >> Like like you're calling something a graph when I'm kind of like, okay, it's a list of members of a community. >> Not a list. It's a it's a graph. It says knows link to notes. >> Yeah, that's why it works. >> A list is a graph. >> But why do I have to know who knows who? >> You you don't. It's if you want to. That say I would just say >> I'm just saying like if I knew everybody who was in this room. >> Yeah. >> I still didn't don't need to know who issued who a relationship. Like I might have more of a direct relationship with five of you, but not a direct relationship with any of you. But I'm still a member of the group. Why would it matter who I was related to? in his only way was really relevant for giving the access. >> So, so it's not endorsed by that guy. So, you need to say, "Well, I'm endorsed by that guy. So, that's that's why you should let me in." >> Yeah. It's not shared proactively unless the community chooses for transparency. They are going to show all relationships within their community. >> So, the assumption is that these vouchers are endorsement that you belong to a group. >> Again, it depends on what the community wants to do. So for Linux kernel, >> it is very public that if you're a member of the Linux kernel, that is something they publicly expose today via the web of trust with PGP today. >> And and that's and the way that they do it is is that if enough members say you're a member, then you're a member. >> Yeah. It's it's the their own >> that and you can see the link which members brought in other members, how far away from Linus they are. It's all it's all graphed out. I can totally understand why within the network you need that information like who vouched for this bro, right? Like that totally is clear to me, but it's totally unclear to me why outside of like if I know like why outside anybody would care who vouch for that bro. >> Just to make it clear that's under your control and the community's control >> always. There's no defaults for setting in the trust graph. So >> I'll turn it on. >> Yeah. Yeah. Don't turn it off. So whatever the community says this is what we want to expose or not, they control that. >> And if there's only three people who are authorized to give membership, let's say there's three people and they have to give membership. The trust graph will always show just links to them and not links to each other among like the trust. >> Are they giving membership in the community? >> No, membership comes from these three. There's a membership community. The member there's a membership board of three people. I'm just calling it that. Those three people have to vote you in and out and that's it. The community doesn't have to vouch for you. It's those three people. >> Yeah. So in that >> membership board allows you in in that graph. Would it just be like a downward graph? >> Yes. So well so different. So again some of this to be fair Grace we're still figuring out. So these are good questions. But in your example there, the three authorized members, they effectively be giving some form of a vouching credential saying I I check Jeff and Jeff's real. And then when the three of the approved um administrators do that, they actually send that back to the VTC. And the VTC says when I collect three of those from people who match the policy, the VTC is the one that is creating the membership credential, not the three maintainers. And so those three maintainer credentials, again, you could choose to never expose to anyone. That's just internal to the VTC governance policy. All the graph would show is that the VTC has decided that it is a member of that community, but you may never know who the three were that actually vouched for them. >> I see. So just like and the network draft would just look like a everybody is coming from this command. >> Correct. So this and this is the difference between a vouch or an approval mechanism or like an endor. So an endorsement credential is not an edge or a note for example. >> I understand >> it's not it's not an edge or a note. It's a it's like >> like I can endorse you. It doesn't mean much because I don't know you very well but I could endorse you. >> That's right. >> Right. So any form of organizational structure like a a corporation that has you know a hierarchical corporation and just design it that way and and then it would look the graph would end up looking like a hierarchical org chart, >> right? Um and yeah, whatever membership rules they want to do that the the the whole idea of the pure vouching is something that's that's the Linux Foundation PGP web trust rules. >> That's clear to me. I understand that. >> Yeah. And that's inferred like if I know that to be a member you have to be approved by the committee. If I show you a membership card, I know the committee approved you. That's right. So I'm not that network doesn't expose something I don't already know. I know you got approved by the committee if you're holding a membership card. So those relationships are already built in. >> Correct. Correct. And this is where one of the first use cases of ZKP is exactly for that is the vouching aspect. So that even at the admin level, you may not know who actually vouched for them. >> Right. It might just show the committee and the committee might be voted in every year might be different if I'm an old member, new member. >> Yeah. Clear. >> So just the good questions discussion. The thing I'd say is I'm looking at this from a platform perspective. The platform doesn't restrict what you can or can't do. >> I understand. >> Uh how you set up your BTC is very much, you know, how you set it up Grace is probably radically different to how Lenus wants to run his community. Uh sure. Same tool though. um as you go through if I could add a couple point I think what you're sharing is that you have a perspective within a graph >> rather than you're sharing the whole graph every time because this also gets into the PPC thing like I sit within this graph and I have a perspective because I have someone has been invited which is why it's often one to one and that's the only thing that gets shared but if you had multiple perspectives within this room then you would actually see us all starting to connect and your view of me within the graph will grow if that makes sense. So it's kind of like yeah you can look at a map but then behind the map there's a territory and you can't see the territory but you can see the map. And then the second point I wanted to make is that I don't actually think and I've made this point a couple times. I don't think we're always going to be computing the whole graph all the time. I think we should stay away from that as much as possible. I think it's we have our own version of the graph. You compute your own version of the graph. That's your perspective and you can choose what >> you would like to share and that's [clears throat] going to save us on the again the keys and signatures as we go. >> Yeah. Well, in some ways, like you could say something. I mean, it's kind of a silly way to say it, but in some ways, even though I know all of you, the only node that vouched for me was Drummond because he invited me to this. And so, even though I can see you, there aren't and there's nothing on the network graph about our connection to each other. They're only >> as it relates to this room, I only have one voucher in this room. >> And that's why when you think about the graph, there's three dimensions to it. There's the node, there's the edge, and then there's annotations on both the edges and the nodes. So things like a vouch, an endorsement, um, invitation credential, they are annotations that allow you to have extra information that is tied to either the edge or the node. And to Mitch's point, that is often very much from your perspective that you are looking at the graph from, not from a holistic like the whole graph probably wouldn't have all of that information in it. I don't see why there's a reason why you would want to unless you're an extreme public. >> Yeah. I mean, you know, sometimes people have fun giving each other pull offs and stuff, but other than that, we generally don't. We're taking selfies together, but generally most of the people I know, I don't have a selfie with them. Yeah, but this is just true to this again. I'll send this link around, but this is where you start getting into >> you get into different graph [clears throat] theory and you know it's very hard to come up with a system a single system and framework that has to deal with these radically different extremes. You've got one extreme that is everything should be known and the other extreme nothing should be known. And we're trying to basically can we have one implementation that deals with both of those plus the in between anything in between the two extremes. >> Yeah, totally makes sense. >> Yeah. Uh all right. So we can do something else. Uh so we're going to create a new persona. So remember I joined this DID this VTC. I'm known as sorry dessert. Um so now we can do something. I'm going to be known as we're going to join as Drummond. Um >> so I've just created a new persona DID. So this is really as kind of as simple as it gets and it's an orphan Drummond. You have no community. >> You have no community. >> So this turned into a bullying session very quickly. >> Uh so now what we can do is I can do something interesting though because Drummond and I met and we're just going to do a little hand wavy thing here. And what am I invitations? Um, so Drummond and I met and Drummond said, "I want to join your community." And I said, "That's great, Drummond. What's your did?" And he's created this uh he's going to be soon virtual. Oh, that's like looking around the corner. You're going to be soon AI. Uh, soon virtual. So, what I'm going to do is I'm going to create a VIC uh verifiable invitation credential to him. And, you know, I can set what it's going to be. I can say what role uh here is. So, I'll just say you're a member. And I've now created an invitation for Drummond. And you know, if we had this, you could do a QR snapshot of it. But we're going to copy and paste this. And so Drummer now, and I send that, I could email that invitation credential back to Drummond. I'm Drummond over here. So I can go here and I can say, I want to join another community. I'm actually going to join the same community, but with a different persona. And so I'm going to join. So I could see here I could well I can't join as sorry dessert because I've already there can't have it more once but I'm going to join as Drummond. And uh yes it links me across everything. Uh I also have uh an invitation credential that's sitting in my clipboard. So I'm just going to say uh pasted that in. And so you can see the invitation just got pasted in. Now, what's going to happen here is when we join, fingers crossed, when we join this VTC, because Drummond has a valid invitation, he's going to not go into the pending queue. He's going to go straight into accept uh as we go through. So, we can kind of just double check all this uh members been logged out. Uh there is no member here. There's kind of sorry and bamboo. uh you're going to be soon virtual. There's no join request for you. We have the invitation credential here. Um okay. So if we click over here, this is now submitting. It's presented the aspect and there he is active in the BTC. So it's literally as simple as that. >> Acceptance >> as you go through that. So Grant, you now have friends with him. >> You're friends with yourself. uh uh as you go through and again the VTC is done. All right. Some other fun things you can do with this. So you may have noticed when I was logging in, did anyone notice how I'm logging in uh >> I using a self-issued open ID provider that is coming from my VTA. So I have a little browser plugin running over here. I can have multiple VTAs. So, this is using the GDC uh demo one, but I could use my US-based one. And I can do things like load what are the websites that I'm allowed to join as I go through this. [clears throat] And so, you know, if I sign in with a pass key, we all know what that experience looks like. I do this, it thinks about it, it lets me in. If I just do this, I'm in. And that's logging in by the VTA. There is no my login credentials for my VTA are not in my browser and this is talking TSP and DIDCOM from the browser plugin directly to the relying party with the VTA's approx. So we're here in Geneva. This is running in Ireland. Uh we did one earlier with Jeff and he was his VTA in Frankfurt. It's no difference in terms of speed uh as you go through this. So that's fun. Um >> there's no cookie. >> What? Sorry, no cookie. >> Nope. Nope. It's like really quite clever. Uh so then you can do some fun things like >> if we go to our plugin, you can see here it's all designed to help educate people. Um but I can do things like show me my network. So I can see here's my trust agent. It's connected to this messaging relay. This is my wallet, my browser plugin. I have a separate approval identity that's actually running that'll do step up approvers. And then I have these authorization authentication profiles that allows me to sign into these different websites as I go through this. >> So do you have any other authentication means for the wallet itself? >> Yeah, you can uh Yep. >> But because if I mean if you just click obviously you're pre-authenticated to the wallet and and presenting whatever keys you need but uh you obviously want to have some ring fencing of the wallet itself, right? >> Yeah. So the wallet if you if I do this. >> So first of all everything is encrypted on I can't the wallet's blocked now. So uh I have to unlock it. So >> okay so use the the the biometrics on the device for >> you question uh do you have any provisions for having you mentioned hardware keys for for key stores. So you could have HSM and other >> hardware criteria certified stuff. Okay. Yeah, we just don't do it because it's just doing it all the time. But yes, um so you can see here this kind of describes to people, you know, how it's working, >> approvals work, uh testing this stuff is really hard. So we have a lot of kind of like livveness checks, etc. But you can do something really cool. So this is the DIDs uh that have been created. We can see Drummond's soon virtual. So remember this DID is sitting inside the VTA. This is a third party did hosting service that is not connected to my VTA as in it doesn't have any of the keys. So we're going to edit Drummond's document and uh >> frozen berries. >> You want to make frozen berries? >> Yeah. Sorry. when we were in health sync even translated a menu and it said >> frozen nose bed please. So there's some fun things here. Again, if you're working with WebVH, and again, it's a web VH is very powerful, but it's difficult sometimes to think about I've got to edit the DID and then I've got to create new signature sign. I've got to create new pre-approved keys. There's a lot happening around it. You can see we're on version one, but I can just say publish with my agent. Uh it's not going to work. This is not in the right space. this one. We'll do this one. So, I just didn't give access to it, frozen berries, and there you go. The VTA just updated it and it's now on version three and frozen berries in it. It's as simple as that. Uh, all the keys have been updated, changed. You can do whatever you want. Um, we can do things such as there's an agent name here. So, those of you who haven't seen an agent name before, these are really cool. You can go into a browser and straight away resolve to my DID as I go through this. And that should also mean that if I go back to agent open BTC and go to my, you'll notice here the VTA is known by its agent name, not by DoD. So this is some of the UX stuff that's starting to come through as you go through. So you can resolve a DID one of two ways. You can either resolve it by its full name there or as simple as that as you go through it. And they both reserve uh they're both protected uh from each other because the also known as goes through this. So there's a lot of kind of powerful oops uh capability that goes into this as you go through it. I'll stop there because that's been a lot. But yeah, um again, first target is Linux kernel. So a lot of this is very skewed towards their use case. I'm kind of skipping over it. There's a whole bunch of things um we skipped uh as we go through that credential. >> Uh so you did exchange credential. You just didn't see it. The thing reason I'm not going to show you a VRC is because then I've got to set up multiple relationships and we can do it. So Jeff, I don't know if you want to show hosting capability if it's working. >> It's Yeah. Well, my computer uh yes, we'll get the just for protocol. Are we having coffee break at some point? Because sorry the idea is once you're done with the demo, we we could do it now. Do you want to do it now before? >> Uh, I mean, let's let me just give a quick just to show what's built really fast. Um, >> the coffee break, there's a coffee place just like a block. >> So, we're [clears throat] going to just migrate over there for whoever want. >> Where's the sauna from last year? >> There's a gym at the end of the hallway. [laughter] >> You really want to sweat? >> It's pretty hot out there. So, I'm glad this time we have a more temperate. >> I'll I'll make this really quick. >> Okay. >> Just I think everybody wants a break, but it's good that you can wrap with this. >> So, as Glenn's been building all this stuff, the question that came to my mind and Vincent's mind who's been working on this and doing the bulk of the work that you're going to see here, the question's been, you know, how do we make this easier? Because setting up a VTA, you had to stand up a server and then you had to get a lot of ducks in a row to do it. and we learned to do that. So what he's done is he's gone uh Vincent and has built a um Kubernetes deployment of the VTA uh stack. So there's a couple bits to it. We have it all deploying on top of Kubernetes. So for instance, uh we've been using Hner as our main um as our main source of uh is where we're storing stuff. So we've got different stacks here. So we have a staging stack. Notice there's three of them. So it's a high availability cluster. We have what we call our production stack. And then we have a rancher which is the Kubernetes orchestration also has high availability. So and you'll notice they're in different places. We have some stuff in Nuremberg. We put the staging in Helsinki. So if we needed to do a demo and things crashed in Germany, we could hopefully be running in Helsinki at least. Um what we've done here, let me just show you this real. Yeah. So all this is controlled through rancher which is a basically it's by soux it's a way that works quite easily we look here what we've got I'll show you the production one um I need to sign yeah I don't have that one Yeah. Oh yeah. Sorry about this. Okay. Now it's this guy. What we have is it. So we some basic statistics about what we're running. There's a couple admin accounts. Remember what she's doing here. Not any users. you've just been playing around with that Vincent and you made mainly sessions. So, we've got what we call the platform stack. So, this is the full VTI stack that Glenn has been demoing. So, you've got the mediator, you've got the VTA, you've got the DID host, and finally, you've got the VTC. So, we can stand that up. You can actually drill down into there, see all the DIDs, you can make changes to it. You can add an administrator to the VTA for it. Got all the private keys there and anything that you would need. Um, and then each user goes and sets up what's their own GTA VTA. So like I've set up one here, which you know then uh that allows you to do other things. I've connected also for instance browser plugin and got that working. Been playing around with that today. Uh that's all being driven through the VTA. So what will happen uh is uh in in the stack itself when a regular user goes in like here I'll go sign in my regular user. Uh, creating a VTA. You basically go here. You say you want to create a VTA. You give it a name. I'll just see GT 2 because I've already got one called GT. You select the image, the latest. We'll be using the latest image from this morning. And then you go ahead and create session. What this does now in the background is it's basically spinning up a Kubernetes pod that's creating the VTA. It does a few things. And then within a few seconds you basically get a DID key that you link to the VTA to the PNM that you're running locally on your computer. And then so you'll see I'll just show you that. So it pops up. So here it gives you a gives you the like that. So what I would basically do is go into my terminal um and I would see PM setup. Let me call this GT2 since that's where I named it. I'll paste the DID there. It gives me this DID key right here, which is basically the master key that I'm using on my local machine on my PNM to talk to the BTA and manipulate it, do everything I need to do. So, if I get back here, paste that key in, provision the agent, and in a few seconds that'll be done. Now, at that point, I've got a fully functioning VTA up and running. I can then you do like Glenn was doing run open VTC start uh creating bids, start joining communities, everything like that. That's it in a nutshell just to show some of the other stuff here. So like as you can see the VTA farm it's we we're using firstperson.dev is the domain for it right now. So we've got hosting here. I'm going to do the the fancy scop login. Boom. I'm in. And so we've got that all stood up. We've got the DTC community as well. Same thing also under firstp person.dev dev. It's all there numbers. Well, me I've been speaking this this morning, the dashboard, everything that Glenn was showing you. This is all running on Kubernetes. Now, uh to set this up, we made a couple opinions. We thought we're going to use uh virtual private servers. We chose Hatsner as the provider and we're using Cloudflare for DNS. Those are the only sort of opinionated things we've done. uh we're using uh Terraform or Open Tofu, the open source fork of it to orchestrate the building of all these Kubernetes servers that you see that are running the cluster. Um thing that's nice about Open Tofu, it's just it's all declarative. You're not in there at the command line doing it. You're just telling it what to do. You plug in a couple API keys uh to build this whole thing that you're seeing here. takes I don't know Vincent if you if you're still there what's what's your record like 30 minutes what how long does it take to >> 30 minutes >> the whole process yeah take like 30 minutes is simple enough >> a lot of that's just waiting for the computers to do their thing waiting for the cluster to get to a stable state before it adds the next node in actual stuff that you're doing at the keyboard probably less than five minutes to get it all up and running >> so that's what we've done with that um you know we're eventually going to make the code public open source once we've made some more improvements to it and whatnot. But uh yeah, that's our little side project to this whole endeavor, at least for the last couple months. >> Does the hosted VTA have a TE? >> Not yet. >> Is that on the road map? >> It could be. Yeah, could be. >> Yeah. Yeah. The trusted execution environment is definitely something we're interested in. Yeah, we're we're going for the low hanging fruit first. So at least the low hanging fruit. >> That's awesome. >> The the cost of the cluster was roughly. >> Yeah, that that cluster is a little bit beefier. I I mean you can build a cluster that's Vincent's in the process of doing some load testing right now, but you can have a cluster high availability cluster running for about 30 $35 US per month. Yeah. And that could comfortably handle couple hundred BTAS. >> Couple hundred. >> Yeah, probably. Yeah. Okay. >> We're gonna load test it and see, you know, is that real reasonable? But as far as memory constraints of the cluster itself, 150, 200, something like that, >> right? >> So you're talking 30 cents per VTA per month. >> Yeah. Just so again that we didn't talk about it, but the efficiency of the VTA and the media is very high. It's very very very high. >> So if I was running a sovereign node, I could run this VCA 10,000 times. >> One one thought that I've got is is setting up uh is this a trust task in of itself? >> Could be. >> That's actually that's a good one. Yeah. >> Right. Like if it's half an hour of runtime, that's a lot of attention a person is putting on a computer. It's pretty high like personhood veracity. >> So that's 30 minutes to set up an entire cluster for the whole thing on its own. If you want like a VTA with BTC, like say you want just create a new community, Mitch's community, that's >> how long minutes >> or VTA plus BTC. >> Yeah, probably. Yeah, the question is about the mediator and what you're going to use for that. That's why we just put it all together. >> But uh actually, I should the 30 minutes includes setting up the >> the head. >> Yeah. Well, the control plane and everything like that. If you if we were to spin up another instance of uh this stack now, it would probably be like five 10 minutes. >> Okay. [clears throat] >> Yeah, >> that's so pretty. >> Yeah, it's pretty good. Yeah, >> very cool. It's all real. It's all real and it's an amazing amount of information and I can tell we folks need some coffee. So, um first thanks. That was amazing amount of stuff. These guys have been cranking on this for, you know, months now. Um, and uh, it's good. By the way, Andy finally um, signed off, you know, in middle of the night for anybody said, told me wants to copy your document and he said there's a bunch of people down there that they saw this, they would be okay. >> Happy send it to everyone. >> Yeah. >> Yeah, please. >> Excellent. Um, all right. So, for the folks online, we're going to take um our our coffee break now, which means we're gonna some of us going to walk over to a nearby coffee shop and we're not doing a inerson coffee service going on something better than that. So, uh we'll be back um likely, you know, let's give ourselves 30 minutes um and then we'll come back and we'll talk about state of the uh uh first person project and go to market. will also talk about um proper GDC. So that will be the final portion. We're we're we're only here through fine. So >> um let's let's take a half an hour for coffee. I'm going to stop the recording now. We'll have a second recording for the >> I'm guessing the reason we didn't really do a ceremony is because only three of us have Phoenix. >> Uh no, no, you did do a ceremony. You just didn't see it. So the ceremony happened when you joined the community. It just happened. It's automatically applied from the policy. That's the beauty of it. There's no need to be involved on that one. So when I created the VIC for Drummond, that was the ceremony going through the ceremony. >> I can do it again. >> You mean a big You mean everyone joining the VTC? >> Vouching for each other. >> Uh >> I mean Jed vouched for me. I think we have some >> hoping to get like I only have two VCs. >> Yeah, we didn't want to talk about why we [laughter] >> I was so excited when I got my first bid. Like you just have to understand. [laughter] >> What you got a What laptop do you have, Grace? Is it an Apple? >> No, Linux. >> I have an Auntu. >> Abuntu. You've got a >> I've got a I've got in a bunch of >> you how to turn off the recording. >> Problem is right now you got to compile open BTC and it >> nice work. >> Yeah, >> you should join the open BTC tool chat. Speaking of waking up to 120 messages, >> coffee. >> I'll leave my keyboard. Yeah. >> I looked it up. I brought so many walk together. If you don't try again yesterday by your Well, I'm a cat person, but I can say that. >> Yeah, I'm only [clears throat] >> um sorry. enjoy. >> Uh, I I got messed up getting here. I didn't have I lost a lot of time, so I'm probably just going to catch up on stuff. [clears throat] coffee drinker. Uh, >> I'm probably adequately caffeinated to sleep tonight. Yeah. >> Yes, I understand. >> Yeah. >> Who's Scott? >> Scott, nice to meet you. >> Cool. Scott Jones. >> Stephanie. >> Cool. I'm trying to see if I remember you from from a Zoom call. >> Probably not. This is a very new space for me. >> Cool. relatively new for me, too. Kind of engaged in these communities starting a year ago. >> Oh, cool. Um, I've been working in the space probably like closer to three years. >> Cool. >> Uh, but like this was very that was very technical for me. I know you came in a little late. Um, and I understand for the most part, you know, but there's, you know, the technical sense, there's a lot of terms and things being thrown out. And >> and what is your what is your specialty? What are you doing? >> I guess I'm for I see myself as a human rights practitioner. Uh, I work for a law adjacent uh nonprofit. It's a coalition. It's a >> say law adjacent nonprofit. Yeah, because it's like a technical it's like law Jason >> it's more legal related. Um how can I explain it in simple terms? Um I I'm part of the Gleonet Alliance uh founded by Richard Wit who is the ex Google attorney. Um and what we want to do is well that's all great what was just being talked about then from the technical sense. Uh but we really believe like cool that's awesome but there's no legal duty for that business or that entity or the company or the tech stack and protection against like the users or what we call them agents then where's the recourse where you know like her question was like okay why do I have to do this like why I don't know if you were here but she kind of had this question she said like why do I have to expose my graph to like somebody else like why she kept like you know like then they kind of kept insisting you know well you don't have to it's like the community leads how they want to expose that you can you yourself can say you don't need to expose it if you don't want to um and as she's talking about that I'm thinking well the work that we're trying to do is build up something that we call an fiduciary so in the sense of a business or a digital product or like a service, anything whether that from like a civic standpoint, a government standpoint, a nonprofit standpoint, like any type of realm in the digital space that has this sort of ownership and power over others in the sense like users, not even those users. um they can take on this digital intermediary fiduciary like uh role so that they are protecting um the people that they're working with and that in turn would uh continue to like bring more authentic relationship between consumer and business so that there's more trust eventually. Uh so but it goes back to to law. So we're trying to push for this, you know, all this is founded on common law. Like when we think about professions where there's high trust like a doctor, you wouldn't trust some person with a scalpel to take something out of your brain. to trust someone that has a certificate or a law like school, you know, the professional associations that like can make sure they can take away this, you know, scalpel from the doctor. They're doing terrible things and they're not qualified for that, right? Um, well, why isn't that something that's happening on our digital services realm? Um, and so of course it it gets highly complex because there's everything wild now. Like it's out there and um we're really standing in the space of like nobody's talking about us like a regular human like me. Who's empowering me every day? God >> in the lane center. That that was my little pitch, but I probably botched it. >> I'm uh I'm coming from a different angle, but trying to solve a similar problem. I work on privacy preserving biometrics. >> Privacy preserving biometrics. >> The idea that uh I I've built this technology that allows you to attest you're a human. You're unique to a community or to a transaction. You demographically are who you say you are. >> You built it. >> My company. >> Oh, yeah. >> I'm a leader. >> Okay. Yeah. >> And the idea that uh >> uh it's an anonymous system and once you go through a launch, you can then be reauthenticated. It's >> the idea of trying to bring humans into the loop where kind of to your point whether I'm going to a website trying to join the first person project as an open source developer whatever I can use this mind or similar technology to say yeah I'm a real human I am unique but you don't know exactly who I am. >> Yeah that's awesome. Where was the the core like why why did you want to build it? Uh we got into it through a relationship with meta actually. So they were looking for a face verification model. We entered their RFI in one. >> They've been using a hypers scale about 12,000 times a second globally. >> What is RFI? >> Requests for information. >> Oh yeah, they're terrible. >> Yeah. [laughter] So we entered theirs and we provided a model. They won our competition and they deployed us and they've been using us for account Oh my god. Account unlocks. >> Okay. Oh. Oh wow. That's crazy. You know, I was locked out of my Facebook account because I'm I'm a US citizen. >> How did you for years? >> Oh, I was I live in Spain now and I've been there for eight years this month actually and I lost my Facebook account for 4 years because I kept like I travel between and I didn't have a phone number and I would try to contact them if possible until they must have updated something. Maybe it was three years >> sued a lot and so three years ago >> something happened because I recovered it >> very end of 2024 they rolled us out as a alternative path to that >> wow stopped it for me >> and you you would have submitted a video selfie >> no I didn't I >> I went back to the US >> ID or something >> no like I think it was still like it was strange like just a combination of my like my old phone. I probably got access to like a really old email like the very first email that I ever used it with >> and then I think I connected my Instagram and through a phone number the connection because now they have purchased Instagram and like the WhatsApp and maybe it was like the connection with what's up the phone number unsure but finally got access which which I thought like this is terrible terrible terrible I don't understand how like why is this so bad but anyway I don't care about Um what I care when you're talking about your technology like I think it's super cool like um when we think about biometrics like I studied migration a lot like I I grew up in Tucson so that's the quarter town um and for me like moving abroad going to Spain when I was doing my human rights program I was like really focused on like how with everything with the African population saying what's going on and you know we got really into the studies of you know just I guess a barrier example before go was designated, you know, like the biometrics at the grocery store, you know, for me that was like so bad. This is like terrible that like that like >> what >> is that how they pay? >> That's how they were Yeah, that was like their trans their right now was their transaction to like get goods out of like a shelter like grocery in a tent which is like horrifying even begin to consider it. That is the extreme. >> That is so horrible. >> Extreme version of non privacy preserving measures. >> Yes. Yes. Um anyway, you should you should form part of our circle in Gleonet, which um I'd love to, you know, I think my boss is coming in a little while. >> I like that. Where are you? Uh, I'm in North Carolina. The company's based in Europe. Headquarters is in London and our tech teams in Hungary. >> Nice. >> CTO's in Barcelona. >> Oh, that's so cool. Maybe I didn't meet him in one day. Um, how long has this been around for now? >> 20 years. >> 20 years. You started it? >> No, they got they met at Oxford, the co-founders. I joined four years ago. >> Wow. Okay. Whoa, that's a long time. >> Yeah. And I always joke we've been waiting for this AI renaissance to catch up. >> A like now is the moment. >> Yeah. They got started in the context of advertising with proprietary models for measuring attention and emotion. >> Okay. >> And doing ad testing studies. >> Okay. Totally different. >> Yeah. And then I joined to lead a new platform business. And when I joined we were still focused on attention and emotion but then identities. >> A and that's like your bigger >> that's where I saw the opportunities. Yeah. to disrupt this surveillance state, this dystopian thing that's taking shape right now. >> But like I'm not knowing, right? But um Meta's whole Rayban thing with the glasses and wouldn't like from an out way far outsider. It seems like Meta doesn't care. >> We actually So we actually won their competition around the Ray-B bands. So, our models were >> they actually put them on rebrands earlier this year secretly. >> Whoa. >> Without telling us, but um the plan was to debut it at their developer event in September, but they really messed up the PR. >> Oh, September. >> They messed up the PR on it on purpose. So, in Q1, they leaked a very dystopian version of what the solution would do. They said it would dox anyone. Just walk around and be like, "Oh, >> they were the ones that leaked that." Of course, >> that's what they do. They see what happens like, "Oh, that's how bad it is." has been 6 months no one remembered but this time it kept ramping up and that's very thought we're not going to do that with the plan is what they patented in their plan is actually >> still doing it >> no what they is actually quite friendly it only recognizes your friends and family doesn't know strangers >> oh >> and it would help people with like disabilities sort of interesting compelling use cases I don't know why this is actually for good and a counter to But how are they going to fight that narrative? >> They have to do better. >> I feel like that's so far farfetched for where we are. >> That's why we're trying to help them. >> That's nice. It's very optimistic. Uh I do think the public narrative is like we hate you like terrible. >> Just get a good president like 17 billion. >> Yeah. 17 billion. Yeah. That's like the first step towards them being like tobacco companies. >> Yeah. Yeah. That was pretty >> led the way. >> More to come. >> Yeah. Uh why do we have to wait so long? You know what? Like why do we all have to wait so long, right? >> All wait for cigarettes. >> That's true. That is totally true. >> Yeah. >> 100 years or something like they were saying it was good for your baby. >> Yeah. Like or [clears throat] the smoking on planes like you look back at movies. Yeah. Yeah. Context does matter. Absolutely. So what is so your local product here? >> Uh those those terms work with you know technically being here. >> What did you do before this? >> I'm like 17 years into product strategy and innovation leadership >> attemp. Wow. What was your first job coming out of like school when you started this? very boring analyst >> like >> economic consulting and higher education and social policy. >> Wow. >> Nerdy stuff. >> Me too. But normally from the social and behavioral sciences. Yeah. I've been in tech now for like I guess seven years since I since I moved to Barcelona. Um but I wanted to be an attorney. Um somehow I I met Richard at a Ricecon event. Do you know what Ricecon is? Um, Reichcon is one of the largest forums at the intersection of technology and human rights. And so a lot of the it's put together by this nonprofit called AIS now. Um, and a lot of like the good players go to this conference to sort of connect. Of course, the the other players are also there. They usually sponsor like a Meta or Amazon. Um, but the thing that you said about Meta like patenting like the nicer version. >> Yeah, you can look that up. >> Yeah, >> people have done follow on stories saying, "Oh, well, they just patented this. This isn't as scary as we thought." But it still sucks. >> Well, mainstream media, right? They love to play the narrative that like get more clipped. Yeah. >> If it bleeds, it leaves. Yes. Did um like travel on Spanish and just Spanish in Spain for very long at all? >> I'm Mexican. I'm Mexican American. So speaking Spanish is [clears throat] >> Spanish is different from Latin American Spanish. And then of course because I live in Satalan and I kind of understand it more so now after you know picking it up but like if you're talking fast like I'm not going to really understand >> and even just saying when you say >> yeah no I haven't picked that up >> a little bit scared people >> the jet >> [laughter] >> Scott prices. Nice to meet you. >> Made sure. >> Great presentation. >> Thank you. So, how long have you been here from Singapore? >> Yes, sir. What's that? And you're like, okay. >> No. Okay. My brain was like all I was trying to fall asleep. >> Yeah, some damn Red Bull. >> You're really good at holding it together. How would we know? >> I would be a mess. >> I am a mess and I did not do what you did. >> I'm a mess. I'm a mess inside. >> Yeah, but you you are you like German? Are you kind of all around the world all the time? >> I am. Yeah. >> Yeah. Probably more. Like last year 197. >> Wow. I I have other hats. >> Yeah. So, you're like essentially the pilot. You get to sit way up there. Imagine your rewards. >> They look That's true. Once on Christmas Eve, I got stuck somewhere like the airport like you in like airport shutting down flights. Christmas and kids sometimes my dad come with me just come and went to the store and said there's a flight taking off and got USC we kicked other people off we have USC I'm like >> that's that's where that's the status where you get upgrades and there status where I've got They will guarantee you a flight a seat no matter what. They will kick someone else off which is a shitty thing to do. >> Badass to have that. >> Yeah. Yeah. >> Where are you? You arrived this morning. >> I I came to London yesterday. >> Oh wow. Where can you >> North Carolina? Okay. So east coast. >> Yeah. And I'm doing a cool hack I got into in January when I came from offside at Oxford. There's an hourly hotel at East. >> Ah. So I show up at air hotel and I stay room for three hours. >> Makes you feel >> Adam. [clears throat] Zpp. >> Oh, awesome. He reports to >> Oh, cool. >> Reports to Adam. Dennis has Yeah, he's >> he's a deep guy. >> Yeah. Yeah. Hey, by the way, I was a nuclear physicist and no one's surprised. Yeah. Yeah. Yeah. Yeah. He's got his own cyclopron at home and we keep waiting for the day where we see on the news and be going, "Are you surprised?" Like, >> no. >> Yeah. I'm surprised. >> He's very excited about what he's doing as well. He's one of those guys that >> if he's doing something he doesn't think is like really hard or um really worthwhile, it's hard to get him motivated to be honest. But with this, he's just like, "This is what I want to do. This is what I want to do." And yeah, >> he's very handsome. Like, I want demos. I want things to show people. Awesome. Yeah. That's pretty much where we're at is zero knowledge protoc. >> You didn't by chance pick up my chocolate bar, did you? >> I don't know where I put it. >> You picked it up. There's nothing. >> Yeah, I thought I did. You put it in you stashed it everything. >> Yeah. Okay. >> Did you eat it by accident? >> I don't think so. >> Makes sense. Scott first opening that >> I'm still very much drinking from the fire hose trying to have things come into focus. >> It's just more if it's like thumbs up then let's just go. But if it's like no no no there's a misunderstanding of the use case and how it can be better to know that now >> then this will become like the north star >> and then like heart is over that as well everyone gets a line to that one that What's the next crazy thing you're adding to agent whatever? >> Uh, no, I'm working on Genie. So, how you doing? >> I've um it's connecting the two of them together. So, uh I can now here is agent stream. You'll see there are 23 services. If I type in Genie, there's nothing but I can go over to Trust Genie and I'm logged on. I've got organizations account building and everything. >> So I want to enable LM safety. So I do this and >> that's cool. >> In a second it will be done. Okay. So it's done because there's nothing to spin up. >> Yeah. Yeah. >> And then I go >> when you say there's nothing to spin up >> because the instance is already running. It's a surface. It's a onetoone mapping. >> This is multi-tenant. >> Yeah. >> You've created multi-tenanted host. >> That's that's what I'm Yeah, I got you. >> Cuz it'll cost little bit. Yeah. Yeah. Yeah. >> 200 users. >> Absolutely. >> So that's going to give us the >> Okay, that's pretty cool. >> Yeah. >> So the bundles are really just surface features. We'll curate them and call it a bundle. And the names can be completely separate because no one's ever going to see it. Interesting. >> So we can say I want an OPA policy. >> How do you do the isolation between them or just >> different systems? What this the trust? know that if I you know if I go into trust union as well and spin up something >> um because every ID that's generated has a pattern which contains your account ID your project ID whatever and it's checked >> this is actually you hit on exactly the problem because the back end was never built for that so I need to >> make sure that it's not possible >> it's in the VTI stack the way I solve that is why I just religiously only use TSV or DICOM which is I accidentally send the wrong message somewhere M where it doesn't matter. You can't do anything with it. Yeah. >> Unless you got the key in which case >> but as an administrator if you log into the back end of a gateway whatever you have full like the default you have full control of drag things and press save. Well it's all fixed. >> That's what you're doing there as well. Um yeah yeah yeah it is only like two days in so it's still a bit brittle but >> what I think >> it's okay to be pragmatic and just say if you don't like them >> then you upset >> yeah most of my family it's really you But I guess it gets cold in North Carolina. >> We have four seasons for sure. Winter's Winter is hilarious. Somebody who's been in the Midwest winter is really like North Carolina. >> North Carolina will get dusted. >> Yeah. Um >> but because they don't have infrastructure, they shut everything down. >> Yeah. But then if you do catch people trying to drive through it, they have no idea. >> Yeah, >> it's crazy. >> And the first time I experienced that, I thought I was in a different dimension. Like I was steady and doing all the things. Everyone's literally spinning out. I said, [clears throat] "What's going on?" >> Yeah. >> Take your foot off the gas. [laughter] >> Yeah. Yeah. Yeah. Come from the desert, so not humidity. Awful. It's really humid in >> Barcelona from the dry heat. Yeah, you can definitely don't have good >> is why I think for both of us we've got to do a bit of a time out and just watch videos and showcasing >> the original >> there's so much gold and like like that whole thing I just actually get it properly structured with layer when you start layering that now actually >> working. So you can use trusting but you don't have to do the case of trust like this. >> Hello. Good to meet you in person for >> Yes. Adam dog. Hello. Good to meet you. >> Good to meet you too in person. >> Excellent. >> You're the ones on the train, was it? >> Yes. >> Thank you. It's not just presentation. of one of the forces behind >> one of hope it's more ready people talking about something experience that he just thinks about. >> Yeah. Yeah. I don't know. Yes, sir. >> Yes, I was. Yeah. This is Very you're like fraud and the way that they had the problem to verify me is super lame. It's like a interview. >> Is that verified? I think it verified that ID me or something and like a person the person has you hold your ID. >> Yeah. Videos >> sketchy, right? Guys, you saw World Hunger and Peace while we were gone >> twice. >> Thank you. Did you take notes? >> Actually, not World Hunger because I bought myself the chocolate bar and then I lost it. I don't know where it went. [laughter] So, quite the opposite. the same person. Why are you deciding who that really is? Yeah, >> you can get your computer on that if you open a web. >> I think I see if I open a web browser and click. >> Yeah, you you pick with the spaces Wi-Fi. Pick the spaces network, >> open a web browser, and then it won't load the page you ask for. There'll be a little click here top left. You click that, and then it opens up. You just face to That's what I did earlier. How do you say weird Linux people? >> You know, I actually don't know how you would say that. [laughter] >> It's not you. It's me. Okay. >> You mean you talk to somebody and they have Yes. Okay. Now I understand. Okay. >> But we need Yeah. So you can connection that we have a strong for tonight. So I have recommend some of them working on this. We could try to create a framework so that each mistakes. >> Yeah, it's interesting. >> Gary loves cricket so he'll be right into the sport. >> I was at a business in Palm Beach, Florida >> years ago. Accountant guy. He's actually rich dad. You ever heard of poor dad? >> American book. This this is a guy was written about this guy's the funniest account I've ever met in my life. >> But he says you know the language of business is account language but he went and read >> a cricket play to the American football crowd >> and he said they might play understand cricket four,000 put them up. This is going through when people in the audience are dying laughing. >> People you follow they speak the language you know when he goes to Australia he does the opposite football installed >> this thing that would be in my browser. or Carly tell us it won't work. >> Have you got the same problem you have? Do >> you want to see something? >> But what is the thing that I click on? >> It'll tell me like in my notification, right? >> It'll say you have to log in. there connected >> open go to my bundles >> and have been just a surface here. You're very >> judge and jury different bundle. The actual content doesn't matter. Like I'm starting up section. >> Yeah. >> I mean I just tethered it to my phone. >> Yeah. >> All right. Let's do that. >> So she attacks her brother. >> See this screen. >> All right. >> Yep. So that's what >> I say for complicated. >> So we get to decide what is the simplest possible thing >> and then it just gets turned into a circle. >> Oh. restaurant and the gateway everything >> there is no automatic connection if you take that >> that will be all right at the moment it's proxy but that'll be a fleet >> but you can start up again. >> So, this is what I'm working on now is making sure >> anyone who's got the Zoom up, if you can just keep an eye on the chat questions that are we still have a couple folks. So, we have uh >> we have this place till 5 and we want to uh you know cover a few more things in our in our last um uh you know 45 50 minutes but also leave time for hopefully folks got time you know chat but um uh then the idea is anyone who does want to go to dinner um uh Andre is researching restaurant options being a, you know, a sort of local close to local. >> Close to local, you mean close to local here or you mean close to local food as it gets? >> Uh, >> someone who actually Yes. Actually, I guess Martine is the other almost close to local. I mean >> local [clears throat] >> two two people are better equipped than me doing their stuff >> because of their generalizations. >> Please do feel free to recommend a restaurant for anyone who wants to then uh water down to dining after. >> Well, do you do do people want to stay on this side of the lake? >> Well, I wanted to go get pick up my bag. All right, we have an hour to talk about cons substantive stuff. Don't worry about that. I'm cool with that. >> All I was saying was anyone who does want to check into that while we're doing this. Okay. >> But yes, we have actually a little less than our let's say 50 minutes. So, um I have some slides that sort of summarize what the state of play is right now. Um and and a few topics I want to bring up just for all you know for folks that are interested in decentralized trust graph verifiable trust infrastructure and what we're doing with the first person project and the cooperative um so we can do that to sort of start up a conversation but I do want it to be a conversation because this is a project and you all are contributors or um um you know directors of how that's going to go forward. So, um, [clears throat] some of this is just, you know, to make, you know, I know a few of you are new to what we're doing here, so I just want to make sure you understand where we are. So, we really kicked off sort of publicly at a first person project um, meetup a year ago and now that we have a full room, how many folks were actually at that meeting? >> Okay. You gave a talk there. Exactly. And Martino was there. All right. So there like six of us here in the room and some of you were online. So in any case, uh it was rather hot. >> I was here. Yeah. Yeah. >> Yeah. It was it was very hot, >> but it was still uh it it was great and it gave us a a good uh good momentum going into GDC. We had no idea what the reception would be for the few sessions [clears throat] we had planned there. And it was I thought it was great. There were a lot more interest than than I expected. That's one of the reasons we're we're back. So, >> beyond interest because like this because here I am at GDC again like beyond interest how many of the people who are interested ended up following up beyond just saying that's cool >> like in terms of results. Okay. So, we have some people who joined the group. >> Yes. Well, last year >> it's one of the reasons I have this slide up. >> Okay. We went from only having a actually at that point we actually trusted we could approve this new working group but it wasn't starting up till September and we hadn't even uh engaged diff to approve it being a joint working group. Now we have I'd say there's roughly 30 folks attending once a week the the main meeting. We have three dedicated um uh task force meetings that are also meeting once a week for an hour. And all seven of these uh task forces are like >> moving. Yeah. >> You still haven't approached DI about a joint working group though >> about a >> a joint working group >> for [clears throat] >> I don't know. I would know if you'd approached Dick, right? >> Oh, no. That happened last last September. It is a >> Oh, the Oh, the joint work group. We've already had Okay. >> Yeah. That no that happened is before trust approved it I think in June. >> I was speaking >> but we said we weren't going to start out till September and then we started September. Okay. >> Yeah. And so yeah no it's a joint working [clears throat] that's why we have members from both def and trust people. >> Awesome. >> So anyway oh go ahead. May May I do one thing? The background is blurry and this hurts the video. >> So if we don't do that >> because then there >> Thank you. This is why we have technical experts for this group. >> Good catch. Uh so yeah so this is again why I put this slide together to be I mean there's just huge and I'm going to say this so I've been doing um based work [clears throat] on this standards in this area on this trajectory for 34 years I've never been part of a standards effort that's moving as fast as this one um or has has as many and part of it is this that we adopted this idea of code first with uh with the development of these standards and uh and you saw you know I'm I'm pointing at you know the folks that have been leading that's been really awesome that's that's [clears throat] why next slide um yeah no actually so just on deliverables from those groups okay first two now have working draft specs and when I say working draft I mean full uh you know trust RP standard template that we developed finally after after you know six years of working on that stuff and um these are these are pretty robust working drafts already. I mean they're a long ways from okay being final ready you know for 1.0 We own all that, but we're also just ordering up um uh the the repos for ZKP verifiable data structure agent names and thanks to Daniel and um Nikki who's not here. She's still on her sbatical and [clears throat] Margaret Marguy Noatne who will be here for GDC but she's been hiking in the dolomite since she gets here in Geneva tonight. Um we have a a a full you know leadership and a lot of interest in the human trust experience which [laughter] >> it is and I I will say one thing I don't know I didn't see Seren on there that last risk and harness prevention he's built I don't know what you call an agent harness for >> yeah a risk and harm prevention AI tool that looks complex multi-repository software stacks and then what are vectors of attack or misuse that they could be used for. >> Exactly. He's applying what what risk assessment hard prevention which was our first task force came up with in terms of analysis and he he's built that and it's already resulted in improvements to these uh for specs. So anyway, it's just it's moving and it's in this next year it's going to be I I expect we're going to have filled out all of those and we're going to have infrastructure up and running even before these specs are all final. >> [clears throat] >> So uh another story there's one person in the industry will who will go unnamed not presently in this room who has assaulted me for the best part of my career. Where's the code? Where's the code? You're always talking these good ideas and sometimes even good standards, but where's the code? And my favorite thing about this project is not only what you saw here, the open VTC, and again, I'm just blown away how fast that's moving, but in the last couple months, there's a second trust stack based on lib P2P being driven by um Erica Duni, I believe that's how you pronounce her last name, and Shannon Bailey. Um they're based in California. They're both also Wikipedia alums. Uh, but they wanted to to to bend to the edge. And we have another member of that team now called the Trust Stackers sitting in the back of the room. Uh, Martina, is there anything you want to say about the uh this this virginly new uh uh uh trust trusters.com is now the site? >> Yeah, the site uh you can uh check it out. this uh trust stackers.com. Um there's a mailing list you can join. Trusters trust stackers one word. There's a mailing list you can join. Um there's also a signal group I can share. It's not up there but I can uh share the QR code afterwards for us who like to join. And uh the next big goal will be IIW uh where I intend to have something to present your thoughts maybe even in the decentralized transcribe working room like two three weeks before and then have some sessions that go into the details. Um, so yeah, it's been becoming more and more. We're like, I don't know, seven months behind the other stack. But the goal definitely is to have it interoperable. Um, so that we can also, you know, benefit from each other and make this entire thing bigger and it's accessible for everyone. I love the fact that now they're going to be these two stacks. It basically um forces the the question of interoperability uh trust suite which we need anyway. You know, we you don't have all infrastructure without it. I actually think it's one of the things that might be foundation has leaned into and done pretty well for for quite a while. Um but we need it if this infrastructure is going to get to where it's going to go. Um, a reminder actually to both teams on that, make sure you get into the um, IW demo queue at the start because it fills up every time and we missed out being able to do demos last time because we weren't in the queue. I'll try to remember that. Um, but yeah, no, that'll be that'll [clears throat] be awesome. Um, by the way, uh, Martina was not able to make the call, but, um, I had a call with, uh, Erica and and, u, uh, Shannon sort of prepping for this and and and, uh, discussing various [clears throat] things about their plans for, uh, the P2P trust stack and, uh, sort of a show of how determined they are uh, how much firepower they bring to the table uh, is the Burning Man um uh vehicle that uh Shannon was the primary architect for and and their team built together and I guess it they were one of the very early uh burners uh and this vehicle eventually was donated to Burning Man and is still still does the rounds there. So, and that is a real fiveheaded fire breathing dragon. So, if you don't think these two women [laughter] are serious, this is in I can't remember some some article about the burning man. It's one of the icons like >> um Yeah. Okay. So, they're serious burners. >> They bring the fire. >> Um we were talking at at coffee. It's sort of like okay what's the actual human trust experience going to be of using this uh infrastructure and uh even though they can't be here today um this key ring is a being driven by um Alberto Leon and Brendan Miller uh who are they were for the Harvard applied social media lab they got refunded and now it's the Harvard applied technology lab um it's now associated with the like uh technology and business school whatever. Anyway, they still have a major sponsorship at Harvard Harvard really likes what's happening here and wants that to grow and expand and uh so what currently key ring you uh Alberto is giving a talk on it. I I'll I'll list that um later on. [clears throat] and uh and both of them are coming to the um uh to the the um Linux the plumber summit and OSS um Linux open source some Europe in in prog in in first week of >> I just say a shout out for Brandon and Alberto for DTG every issue now that we're raising from a build perspective they independently co-build in another language stack and verify >> so implement aspects. Yeah, that's being really helpful of just clarifying that what we see from one perspective matches another build. >> Oh, totally. >> I also want to give them a lot of credit. They did the the first >> uh graph of the credential spec. >> They've been staying right on top of it. The editors teams are really leaning into this stuff. Uh it's setting a really high bar for all of us and >> very easy to work with. >> Yeah. Yeah. It's a great uh great team and they they are getting personally quite invested in where this is going. So uh it's a really good sign [clears throat] and uh there's another interesting potential mobile app um called Union. You can read sort of this is a set of bullet points on the GitHub of um Phil Broedale who was the founder of Second Life. And I'm not here saying, "Oh, he's ready to pile in and, you know, totally deliver a uh uh, you know, DTG enabled app." He, uh, spent half a day with Erica and Shannon. Uh, this was, I think, in early June going over what his requirements were, and they had been working on the P2P trust act to see if they can supply what he what he needs to be able to do all this with EPG. Um, but he's pretty serious about this app and he's got a pretty good track record of uh uh one thing [clears throat] that I learned talking to him is there is still um uh in Second Life 20 years later a billion dollars a year in the economy uh um of that. Yeah. So, the metaverse, you know, a certain Z guy never produced, but Second Life is still producing. Uh, yeah. Okay. So, there's that. And then, um, people ask about, well, what are what are the other apps? I got had some great dialogue in New Zealand when people were saying, well, what about this, what about that? And I just wanted to list that we for quite a while we talked about um uh agent agent pages uh and using VTAs as simple contact agents. Really it's a pretty simple function but it's your policies now. It's entirely independent. There's no LinkedIn or any other mediator in between. I can hardly wait to get that to market. The irony is that our best propagation vehicle for uh for folks getting on this could end up being LinkedIn, right? As post folks announced, hey, in addition to my LinkedIn page, here's my agent page, right? Um I've been waiting for like 25 of those 34 years for what I personally have always believed is a killer app for this stuff, priority messaging, right? And just how many of us are so delused with messages? My ex-wife. Every time >> every time you text her, this is really going to work. >> Which one was it even? [laughter] >> It's somebody sent me an email. I didn't know it. I didn't do anything. And I'm like I go to her inbox and I'm like, "No wonder." Well, it was important. like [laughter] yes, but email doesn't know it's important, right? She needs she's just a perfect example. You know, how many folks and how many businesses would like to have a private channel with you where they know it's not fishing and if it's priority, they're going to get it. >> I see you. >> Welcome. >> You have to do an introduction before you can even sit down. >> Okay. Uh, Richard Wit just flew in from San Francisco and I'll be speaking Wednesday on Tuesday. >> Tuesday. Oh, no, no, Wednesday. That's right. That's right. >> Today's Monday. >> First of the two days of the two sessions. >> I know my days are off. Yours should be. >> Anyway, really pleased to meet you all. Sorry I'm late. Uh, just, you know, a bit. author of >> a little book called Relieving the Web. >> Oh, >> found the need for new governance of uh viant fiduciaries and personal AIS >> and first person cooperative be a perfect building in my mind. >> That's why in the 30s sorry to interrupt but >> no I'm just really glad you can make it. >> Yeah, I'm really happy to be here. >> By the way, trains have been late. [clears throat] People have been um you know >> Yeah. >> trickling in and it's all good. It's all good. We're just We have about a half an hour left, okay? And we're just basically going over what what's been going on with the uh the project, what's the state of play, and where's it going. So, perfect perfect timing. >> In fact, [laughter] literally, he walks in, as I'm about to say, and then there's fiduciary. >> Oh, right. Oh, sure. >> I mean, were you like waiting until I got to that point >> dramatically? Yeah. For 12 minutes outside the door. >> Yeah. Yeah. It was like, okay, okay, I'm going to wait for that. [laughter] Seriously, this is something that that folks will die. They heard us give several talks about this. That's where uh I last saw Richard in person. And uh [clears throat] I'm understanding that we've only seen half of the uh the amazing [clears throat] stuff coming from Infinity because the other half they won't show us yet because it's too top secret. >> Ask that man. >> Yeah, it's these two right here that are worth. >> We have no secrets. >> AI. No, I'm I'm It's a squirrel secret stuff. But [clears throat and cough] I will say this at the February um um summit on human agency that we had the the uh Affinity app which there was called H2 Connect um you were connected with both people and AI agents in one interface one they could be part of the dialogue and uh I'm hoping that's where this is going to go. >> Cool. The big thing Richard's gonna add after, you know, he's actually a real lawyer at the table is >> apologize. There goes any conversation. >> I was just going to say, well, this is >> this is going to get really expensive then obviously [laughter] >> what I love is the fact that he was inside at Google for 13 years, >> right? And then he broke out of the cage. He's [clears throat] >> probably You can export yourself from Google. [laughter] >> Exactly. Um [clears throat] anyway, I think there's a lot of potential there. Last one up there is the a more recent subject but one that um is getting a lot of attention because I've been working on the challenge of funding for the first person block for the first person that one of my first person project since I left Jen which is now 14 months and I gave myself six months to get that done. Um the the the thing that bought us more time I want to acknowledge was Affinity helping us out in the early part of this year. Um and that really helped us get the Seattle workshop, the the February events uh with the Linux Foundation, but we're still going down that um that uh path. And um the aha came this summer when not only was the the uh open BTC starting to look like whoa, you know, this these trust tasks were really they were multip I mean we're up again. What was that number again? >> 350 >> 350 trust tasks are now in the registry at trust.org. Okay. So it's Yeah. Doesn't that figure worry you a little bit down? >> Worry me. Yeah. [laughter] Anyway, um also when the P2P trust act work was going, I found out that um Erica had separately been working on in a separate app, a crowd a crowdfunding app that she recognized she could basically uh VTA enable and suddenly like, oh, that's an app that can be built right into VTA. Imagine that you're joining for the first time. uh a trust community, we get an invitation from someone, a person or from the community, right? And in the process of joining and getting a VTA, you're getting a wallet with hooking it up to whatever just like you do, you know, when you first get one of these wallets, you give it credit cards. You you you you enable it with means of payment. Bingo. Now it's it literally uh a click to join a crowdfunding and if this is an app then any BTC any BTN out there can do a crowdfunding. This changed our perspective on how much we should be looking at other funding sources versus bottomup funding for this stuff. So, [clears throat] so now I'm just going to briefly let you know what so what is the state of play now with first person first cooperative first where are we going with this idea and so this again just background to make sure we're all on the same page um people hear this you know oh are you going to create a cooperative we created a cooperative we we we did about nine months worth of research and we're very surprised has to find that the most progressive cooperative law in the world that at least fit our needs was not outside the US. It was in Colorado largely due to these two law firms there that worked together quite a bit. All they do is social enterprise and they actually help work with Colorado this like 12 years ago to instantiate a new thing called the limited cooperative association or an LCA. It's a hybrid of a conventional patron uh patronage cooperative and um an LLC. So you can have investor members that actually get shares just like you would in a C corp, but they're there's no unless the cooperative is designed to have an exit. Some of them do have exit to community as an option. Um they're typically just repay a multiple of their investment. So it provides a means for cooperatives to get the seed capital they need to actually you know uh grow a new business. Um but they don't participate in governance. It's pure funding. The patron members are the ones that actually uh govern the cooperative. And that was our whole goal because the idea was if [clears throat] we were going to establish a VTN although it's clearly now a btn of VTN's um for where the the the member communities and network said we're going to have common governance for person the paper that we so widely site around that. Um [clears throat] then how do you want to govern that? how do you want to govern as as a global public utility? And you know, based on the time I've spent in different governance options, it was like, you know, if it's going to work for people, then people should be the member owners. Um, so [clears throat] that's what that's what we set up. We're in the process of putting in place. We basically started with uh uh some classic, you know, placeholder bylaws as we figured this out. We're in September should be approving the first bylaws that are really underscoring you that will that will actually be operational at which point now we're ready to actually start spinning up BTTS VTN's and having members that will then >> I've got a guy that can help you. >> Yeah. Yeah. Exactly. Excuse me. Believe me, a lot of interest in that charge you $30. So if we go back to this picture and some of you who weren't here at the outset but you probably seen this before [clears throat] >> early on we had as I as I put in the the newsletter we were thinking of these btns and like okay they're going to be a couple of those like first person network they're trying to solve this broad stuff after I trip to New Zealand as for those who are here to listen to Andy uh uh Higs I would I step back and oh no no no there are going to be many VPNs right and uh I did my own research to say how actually did that work on on the uh you know with the [clears throat] internet routing and uh and discovered I did not know about autonomous systems or as I brought it up with Glenn Glenn immediately goes oh yeah internet handles autonomous systems and uh [clears throat] that I learned from the Wikipedia article there are now roughly 120,000 in the internet system. They in turn have their own >> tree out. Some of them have hundreds or thousands of lands that are part of that autonomous system. >> I didn't, you know, I'm not a protocol architect at that level. I didn't know. But that's that's I think a much more accurate picture of what could happen. So, so [clears throat] now it's much clearer. Oh, but the first person cooperative, it's not just people or VTCs. [clears throat] We want to have VTN's as members. Um, and and they're they're the ones that probably have the biggest stake in the cooperative being successful. >> Okay. >> Oops. >> One thing I could add to that drummer would be because a lot of this is now built out up to just under the VPN. Nothing. There's been no red flags as to why this architecture doesn't work and nothing has been thrown out. I just want to say that because that's a key point that none of this has actually been put together before. And so one of the things we're always pressure testing is there a better way to do this and can we actually get rid of one of if you can get rid of a layer that's the best thing you can do architecturally but actually all the abstraction points that are there are there for a very important and legitimate reason and that's being reinforced as the build goes through. So I'll just say that's an important side effect that's being tested literally daily um on the spec side of this. >> Yeah. [clears throat] And the the both the the powerful thing and the risky thing here is that we're getting down the road here with again specs with code with plans for governance in this next year. We need to instantiate this thing. We need to actually get VTAs and V and VTC's up and running and the first VTN's, you know, happening out there. And [clears throat] that's going to be the big difference. If we're here a year from now, I don't know what's going to happen with GBC or whether we'll even decide that's an important milestone, but that's what's coming up. [clears throat] And so again, what that the last thing I'm going to tell you about here relative to to this is where we're landing, especially after the feedback from New Zealand and the fact that they're looking at several VTN's that could happen in New Zealand by itself. Maui want uh sovereign uh infrastructure. There's a big movement for uh Maui data sovereignty. So the reception to our message down there is just it was like open arms like this is the only thing we've seen that actually would work versus the the digital identity um options they've been looking at for the last 10 years. [clears throat] So and that's that's just one market. Um so the aa went on it's not just that we could crowdfund for firstperson network we could crowdfund multiple networks and this collaborative or cooperative crowdfunding it's just it's we're moving into a new paradigm. [clears throat] So this is a table and a doc that I've written up some of I've shared with some of you. I'm happy to share it with anyone who's interested on this cooperative crowd campaign with what we have sensed are the best options right now. I'm going to talk about this one in just a minute. We know when we're going to um the Linux Clumber summit a month from now that Jim Zan's plan from the outset had been it's not we're not just doing the kernel. He wants us to do the kernel first because they have a web of trust and if they adopt it's going to be clear sailing for they're already lining up Kubernetes to be the second community that we go to and [clears throat] then it's a Linux foundation VPN covering all their projects. They build it into um LFX, their platform that's based on GitHub. Jim's going to be here this week. We are going to talk to him. Uh turned out I'm on a panel with him. I didn't know that. But we're going to get Danielle said we're going to get some time with Jim. He's also going to be at home. So, so we're really going to cement that um as sort of our anchor candidate. [clears throat] Um there's a lot of interest here. I didn't want to list anyone there because the thing is we don't tell anyone else. They need to say, "Oh, we want a VTN. We will put it together." But um Andy [clears throat] uh Higs is already assembling the folks down there that would want to put together um at least one indigenous BTN. Go ahead. You're gonna ask something. >> Yeah. Um, my question is, okay, this is great and I'm excited about the crowdfunding. Is there a more long-term sustainability plan in terms of like paying per member, you know, like this is kind of passport as a service for, you know, identity card as a service. And so the initial crowdfunding, I can definitely see it coming from the people who want to give it a try. But then is it would it be something like okay if you're one of the initial crowdfunders you'll get this many identities for free and then I don't know are we thinking about a longer term sustainability plan not just crowdfunding. So yes, and the big shift and I'll talk about a minute is recognizing that every btn and the the the stakeholder or population that it's serving can decide for itself. Like again I was in sort of the mindset of oh first person going to figure this out as a cooperative for you know all the folks we're working with um and we'll come up with some model that works you know globally right I now I'm like no no we don't have to do that we can let every bn figure out its model what what really convinced me about that is malry approach to how the Mauy want to do this all the way down the economics that's their business Right. We should The only thing is if they say, "Yeah, we want to be recognized as issuing personal credentials." Then they decide to join FBN and that's all that's involved. There doesn't have to be any economic relationship there because if they're self- sustaining and they figure that out, fantastic. Now, that does mean First Cooperative has to have a model and I'm about to talk about what might work for us. But this idea that each VT I mean um one of the things would be uh the other thing that's come up is rather than think at the level of you know you know member states and these larger organizational units cities cities are perfect for this stuff. Um, one thing I'm I'm not talking about here today, but I would just will share in New York. I'm going in uh on on October 20th to uh be part of a a one day thing at the uh it's called Green Build. It's the green US Green Building Council has an annual conference and like 500 architects there. It's all about using AI with smart buildings and the trust component there is it's just a major it's like the missing piece. I didn't know they had a whole Linux foundation project on automation for buildings. They now as it's getting AI, you know, AI enabled, they need the trust component. An architect attended a talk I gave in in uh uh in Los Angeles in March and got right back to me the next week and said, "Could you give that talk at our next online conference?" And I've now done it twice. So >> [clears throat] >> um oh so his thing was it's not just a trust architecture where every green building is a verifiable trust community but because of the interoperability you now have a a a clear path to a smart city where the buildings can you know there are a lot of things that you got you know floods or first responder things or power conservation where they need buildings to collaborate and all all the owners together. So I I didn't I wasn't even thinking brain buildings when we started down this path and it's like fantastic. So um anyway, even in New Zealand, we started talking about what about just tackling a VPN for a city like Auckland, right? There's a couple million uh people involved there. When I went to Melbourne on the way back, they said that's the best way they felt to start over in uh in Australia because there's the nice thing about a city is people actually live in the in the city. You can have first person relationships that can be verified, you know, throughout the city. >> [clears throat] >> So anyway, I'm just I'm I'm putting these out there as so you know what the thinking is about how this stuff could actually get off the ground. And of course, what about Zurich? What about Geneva? I mean, I don't know how many folks we actually have that live here, but Well, half a million. >> Say again. >> The country of Geneva is half a million and half of that against city. >> Oh, >> yeah. So, it's not it's not many people. Okay. >> I don't think the number is as important as the fact that it exists in a trust network. Like I was thinking of a model where like a a forest >> uh which is a national park could have a verified trust and a trust pass as I went on a hike right >> and bio regions >> bio regions voice to the nature through this. common interest and and and and therefore intersections that are >> are valuable to have to be able to verify. That's >> obviously more people helps with the adoption, but yeah. >> Yeah. The River Dawn Project is trying to do something like that. Makes a lot of sense. >> What >> there's something called the River Dawn project and they're doing a project where people make their commitment to the river like I'll visit the river or I'll figure you know and this would be perfect for that. >> Yes. Andy actually Andy Higs is there's a similar thing uh in already started by by Malry in in New Zealand for conservation of their lands and he thinks that might be the the best starting point down there. [clears throat] So anyway again I'm just sharing these sort of plant these ideas. Um I'll point one last one which is this idea of uh the founders network or founders btn and uh Martina is going to recognize these diagrams because we've been trying to figure out for the cooperative how do we be decentralized right how do we practice subsidiarity as a digital cooperative because we're trying to maintain this utility but the last thing I want to do is centralize uh control of governance so how do you do that so we have these ideas of okay well how do we how do we provide a way for someone that's not in this room that's not connected to us to be able to say well I want to be part of that network and u so [clears throat] the idea of first person circles is oh okay go get a set of and the number seven is somewhat arbitrary um maybe Martine has come up with really good reasons why that's the best number we're not exactly sure but this is an example of seven but the idea is some founder gets six other folks together. They all need to know each other because the test which the VTA will we we need to basically build well [clears throat] a VTA for a VTC that will do the test everyone has >> that's already exists. So when we show the VTC there's a VTA behind it already. >> Right. Right. But then the policy would basically be show me credentials that everybody knows. Right. It's not it's not a big challenge. >> [clears throat] >> If that founder then could get an invitation from another founder, it's a way to start to build it organically. But what if you have one of those where none of the founders know any of us, but we want it to still grow organically. We want an even stronger trust route. And that's the idea of founder circles. Great. Go take your six. They have to fill out their first person circles. Now you've got 49 unique individuals. That's the idea is everyone has to get six people. Right now you've got a large enough essentially group trust anchor that if the cooperative provides way for them to meet or get introductions to other founder circles, we can start to grow organically. And [clears throat] the one thing about the founders network is to your point about sustainability. If you want to be a founder and get a founders card as a verifiable credential from the cooperative, it is an annual membership payment. Um, and so the math is we're talking like 10 bucks. And so if founders network grows enough, we'll s be sustainable by just that one VTN as a cooperative. Every other VTN can have its model and we have sustainable infrastructure. One model wide open to better ideas. Please come talk to us. If Margie was here, she would ask the same thing. So that's that's all I've got on that. Um we have 13 minutes. I I'll tell you what, just so we can go straight to the questions. just so you're aware what's happening this fall in the next 90 days. There's four more milestones. Um there's there's plenty of other stuff happening, but we've already um actually the first one is something I learned about a couple weeks ago. It's uh I don't know if Richard is saying you're you're partially responsible for this or you were just one of the key speakers. um a little bit about yeah I mean Ben Moscow is is the lead guy really pulling it together with the future of life foundation and project and a few others but um yeah it's just an attempt to build an alternative AI stack is the way they think about it and the ultimate objective is to bring this to the global AI summit happening next June uh in Geneva Switzerland. So they want to walk in the to the room and basically say all you people who are like trying to create transparency and accountability blah blah blah for the MLMs and the big guys. No, no, no. We are the we're building an alternative sort of that that's the I think this is the initial sort of convening of these 3400 people I think are coming and thanks to Richard Ben invited me and I'm like just what I need another trip across the country but I'm going so I'm definitely going to be there and thank you again for that. go uh you know again my premise is and you saw it here this morning we are building the the trust foundation that's needed for that so yeah the full requirements for producer AI well that's what that's what happens building uh you'll show us that at the next point [laughter] I'm not here >> who just decided to leave >> I apologize I have an interview in 10 minutes and I need to >> a little bit [clears throat] way to make kissing [laughter] a good excuse to go. Anyway, um actually I said four to five. Um Bron's coming up in just over a month. Um yeah, I know. Uh then as I mentioned the green build conference um and by the way that lead the lead architect on that has agreed to um come and give a talk to uh decentralized trust working group um to explain why the vision of smart buildings next internet identity workshop first week of November and then two weeks after that both ITF in LA and the SEI summit um state endorse digital identity in Utah uh at which they've already shown a strong interest in showing or [clears throat] having us give a talk on how setting credentials will work as personal credentials for BPA it's sort of obvious but anyway that's all coming Um, and then the last thing I'm going to show you is this is a quick overview of the talks. >> And I I'm certain I've missed some, but these are the ones that I know we are involved with. Now, the ones I missed I know are ones that um Martina is giving. And by the way, I' I've heard conflicting things about Christopher Allen being here or not being here. Do you know the truth? >> He's not. >> Okay. because I talked to two people yesterday who were stunned to say, "What do you mean? Of course, he's going to be here. He wasn't been on all these calls and everything." And I said, "I don't think he's going to be here." So, Martina, can you quickly just tell folks when your um talks are? >> This is somehow conflicting as well. So, I checked this morning and I saw the first one is about South identity and the next 10 years. So it will be kind of a recap whatsoD is uh what the community achieve and what we are currently doing and how we envision the next 10 years the responsibilities that come with self identity it will be on the second day which is Wednesday at 9 a.m. in the morning and the second one will be uh the least worst centralized government. Yeah, [laughter] that's the very last day >> which is together with Elizabeth um on the last day on Thursday the last session slot. >> Oh, last session. Okay. >> Unfortunately, I think that is also the one that >> No, no, no. That one's been moved. I will be able to come to that one. They put me on this one which I [clears throat] Well, anyway, I think it's the announcement about what's going to be happening with open wallet. So, it's directly opposite. I saw that I was like, I won't be able to go to that one. Anyway, so those are the two that I don't have listed. Um, these are the rest of them. I highlighted the ones that we actually submitted. Um, and uh, this is the one that this is largely going to be Richard. I'm just doing Daniel and I did a very short intro. This is the one that primarily Well, you shoot a bunch of slides from it already. And this is how we'll be doing the key ring. The other ones are just ones that we've been asked joining to speak on. So that's that's the questions or thoughts or discussion items now that we're so close to our end. >> There's too much at once. That's >> I think there's so many questions that we probably would want to stop because it's a so many things that are possible as soon as we get into this paradigm like hard to complain myself to think you know just coming from my own area of uh domain names where you have a huge problem of identifying people and it's just not resolved and people have been taken this is the new normal of not knowing >> which is disaster and then the whole thing relies on domain names and people pretend that they know what they're dealing that almost feels like a whole subject that we could yeah >> we could dive into separately which by the way is also >> fine we should connect >> and uh but but that that might even be a good topic um to bring up or or have you give a talk to our decentralized trust work I'll be very much interested in this >> okay good let's let's follow up on that Martina Yeah, >> that's questions in chat. >> In chat. Okay, good. Let's uh up over on that. >> The entity supports TRQP. Are they a VTN or where do they sit? >> Or are they stuff? >> Uh they're a VTC. >> Yeah, just found that one speaker. Yes, that that would be my answer. >> Thank you for that. Appreciate it. Uh yes. Um well, we could go into in fact will your use cases um around hand and kind of both register you need and potentially how you would use VTAs. That's another very rich subject that might be worth um >> Yeah. >> Yeah. Will's Will's Will Will Will's world would fit well. >> Exactly. >> Speaking of that, so also not on that list was Daryl's um talks there at least two that day two beta room is trust registries ranging from the the patterns that repeat because it's the same pattern every time. got a credential need to know if it's real. um through to tech through to other other that's all beta room to Wednesday sorry day two um from 9 till pretty much the end of the day it's all trust registry stuff I'm speaking I'm speaking at 9:00 a.m. But it's like the whole in that room the whole room is trust registries all day and >> how the trust registry fit in with the trust network >> the the onion diagram that Drummond shows meaning you have a credential you need a trust registry somewhere in the chain you have to answer the following who said that issuer is authoritative and the answer can't be oh they have a credential because if it is it turtles all the way down >> but isn't that the same as the trust network like to me the trust network and the trust registry [laughter] >> it's an as it's an aspect of a trust >> okay >> trust registry >> every trust network must have a trust registry if you want anybody outside >> not registry is a trust network like if you say a network of networks then it's a list of >> the trust registry >> list of organizations which is a registry is what keeps a trust network honest a trust network may not always be accurate at point in time >> what does Does it mean the trust network might not be accurate? >> So if you you could take a copy of certain credentials, but something has happened which means I've actually revoked some of your credentials. And so the trust registry's role is that the community owner, the VTC would updated its trust regist and said something that Mitch is no longer uh he's been kicked out. >> We have different ideas going to be in this registry. So that's a really good day on it. I think is how I completely it's good. We're gonna have a whole day to discuss this. This gets to the same question where the trust registry is the map but if there isn't a trust registry it's just that >> and if my perspective has changed because there's been some revocation or whatever the there needs to be a registry to be updated but if that's not important to the community >> yeah that's may be useless >> where the vacation sits right to me the network at the need to know who's in the network Yes. >> So if a community's been kicked out of the network, that would be at the verifiable trust network level. Whereas if a individual's revocation that would be at the community level. >> That's right. Yes. That's right. >> Yes. Okay. >> Exactly. >> It's it's at a very very high level. It's like DNS for trust. Uh and I know it's a it's a dangerous analogy, but it needs to be as efficient as that. That's >> I think I understand it correctly. Yeah. It makes a lot of sense and I'm I'm really looking forward to my slide, but if if I had thought of this one, if we're talking about what's happened last year, >> Darl just told me this. >> So last year here, the question we got about trust was literally what is a trust registry. The question now and Apple is Apple and France chair the task force that I'm I'm on it. >> You are a co-chair of that. >> Yeah, it's not listed. They call it Daryl's track, but yeah. So, um the 220, so you have Apple and France basically saying we need global registries to hit to understand because they're hitting, you know, there is no answer for the United States driver's licenses. AMA is a partial answer. California is not in their digital trust service. So, you have that's in one country, you don't have an answer. They need one credential in one country. one type of credential in one country, you don't even have an answer yet. So they need a global registry of registries, a decentralized network of these things and they want to understand how does it become operational and it was in need of funding because it's the term they use in the GDC context was it's crushing them right now. >> Yeah, they were pushing this ball up the hill, this stone up the hill about driver's license. Get in your wallet getting down like uh oh uh >> oh it's starting to roll down the hill fast. It's nice to see and grace you're absolutely right that picture totally depends >> on trust registry is interoperable at all which is why this >> it's why this stuff >> you know came out of out of IRA um so >> so we got to solve these problems together my um the optimistic view on it is because we have a common way of doing it because we're building this infrastructure we're actually going to show folks not only the role but how those registries work and how they can multiply Why? Um, and in the end, I guarantee you, we're going to be I mean, this is why when Darl was asked to co-chair this thing with France and Apple, I'm like, whoa, folks are starting to pay attention. Yes. So, anyway, we are uh hitting time and I just want to thank all of you for taking the time to come this afternoon and hope that you have a fantastic You are all, by the way, um, uh, what's the right word for it? um delegated >> as ambassadors um >> first person project decentralized trust BTI whatever the hell if you want to carry the word >> any and everyone that you talk to where you're going oh that person's interested feel free to put them in touch with any of us um and you know we just we'll carry this forward um John and another word of describing delegate deputized deputized. Perfect. [laughter] >> You're all deputized. I would give you a picture as soon as you have it. >> Yes. Anyway, thank you all very much for coming. What's that? >> Um um dinner. Where's dinner going to be? >> Yes. >> The important questions. >> As the huggers guide would say, that's somebody else's problem. It's a message something who actually wants and what time because I heard people want to pick up the badges >> badges becoming available what time >> 5:30 to 7 I think is badges >> 5:30 we want to aim for seven then we >> and where because this is now going to pal expo and then back here is this >> probably not back here it's the phone. >> So shall we close the session? >> Yes. >> Yeah, I think we should. Thank you everyone who's with us. We really appreciate >> and uh Alex who is interested in having dinner. So in this manner