Decentralized Trust Graph Working Group Weekly Meeting - APAC Time Zone - 2026/08/27

Mhm. >> Well, there's my India friend. Hello YouTube. How you doing Sagar Shine? Hi. >> have been having fun since morning. How has your day been? >> So, I guess it's What time is it for you right now? >> Half past 6:00 in the morning. >> So, you How long have you been up? >> Around 4 hours at this point. >> Oh my god. You got You got 2:00 in the morning? You got up at 2:00 in the morning? >> Yeah. >> Is that Is that normal or we just have a sleeping tonight? >> it's not normal. I I went to sleep yesterday with something in here left incomplete. I woke up figuring out that I had a solution for it. Turns out I did not. So, now I'm in a jam. >> Ah, okay. Okay. Okay. Totally understood. Um hey, I'll I'll >> It's a It's a fun thing, actually. So, you know that I have been trying to build out this risk and odds prevention toolkit, right? That I have on the gun. >> Yes. >> So, I thought how about trying to extend it to try and see if I can wire up an entire end-to-end flow and see if if a composite task happens. Can I measure if there is a privacy risks that emerge from that composite flow? So, yesterday evening I had a design in my mind, implemented it, didn't work. Then it was incomplete, went to sleep, woke up in the morning, thought I had a brilliant idea that that would work. And I've been trying since morning to try and get it sorted. It is the nearly and probably something will come out there today. >> Ooh. All right. Now you have me really intrigued. Um >> Yeah. >> Uh you know, all I know is what whatever you were doing when I was in New Zealand running your tool, you got Glenn to recognize a bunch of things, and that resulted in his uh you know, a bunch of the issues that he's been bringing forward. So, your >> Yeah, yeah. >> Your tool is already working, dude. >> It is. >> That's That's super cool. You know Nikki's going to come back and sort of go, "Whoa, what happened?" >> Nikki's going to come back and go bonkers because at this point my fork of the upstream repo is close to 300 commits ahead of whatever that is. So, >> See, you're another one of these Oh, genius using agents to do stuff that it's just the game is changing. It is changing so much. Wow. Awesome. Well, uh it's pretty clear again. I didn't expect necessarily anyone else from the New Zealand crowd to show up for this meeting after we talked to Andy last week. Um and uh and he is I'm going to talk to him tomorrow night. Uh just one more check-in before GDC, but we are going to line up folks to come to this uh the one in 2 weeks. So, uh >> Yeah, let let me just given that I have been so excited about what I'm doing. Let me just share uh one thing with you. >> you like. That's super interesting. >> Yeah, yeah. What the heck is this thing doing? Green border, zoom up this. I mean, take I just want to share screen, my friend. I don't want all this rubbish to happen. >> [laughter] >> Zoom's always adding in more stuff. >> Ah, I see. Now I I realize what they have done. They had it got me into the settings of them the actual thing. So, you should you should see a GitHub >> Yep. >> rendered page, right? >> Yep. >> So, this this is what I've been trying to work on. It's I I fancifully call it a privacy implementation profile. It's just a module that tests the implications of in terms of privacy. So, the workflow that workflow that I have at this point is that I have this monitoring tool that monitors whatever is changing in the entire DTG portfolio, right? So, that's the portfolio monitor. It kind of gives me an assessment in terms of what has changed so that I can do I really need to trigger risk assessment and health prevention evaluation. If does that if that actually happens and it triggers out that there is an element of quantum of privacy related topic that it should be investigated. It hands it over to this module. And which in turn will actually create a conclusion and throw up that information back to the tool. So, that's basically what I've been trying to do. And let me see if I can show you one of the issues that that I have had. >> Sure. >> Yeah, let me uh notepad. Uh get I'm getting it from a different console. Yeah. So, uh I was doing Yes. So, it's basically doing a given that there's a new trust task framework that they are trying to do across uh specification test. The kind of goes up and gives us a way in terms of like how you can throw up a This effectively creates a situation where it can make all these conclusions. conclusions and then eventually you go up to a place where it says that there is a uh reassessment that will throw up this privacy uh more trigger to do. So, what is not happening correctly at this point is that the privacy trigger is not throwing inferences back to the harms tool kit. >> Ah, okay. Okay, back to harms tool kit. >> Yeah, in a way in a way that makes it human understandable rather than throwing a whole bunch of YAML into into an issue. Because I'm pretty sure if I throw a bunch of YAML to an issue, nobody's going to have the patience to read and figure out what it is going to be. Anyways, I'm going to stop. >> Right. I see. I see. Wow. >> And what will happen is that once you have the VTI spec out, I'm going to I'm going to kind of wire it back into this VIP thing that I've been doing. So, it allows me to that entire spectrum of the behavior of the part to be codified and be able to turn it into a series of tests. >> Oh, wow. That cuz it's like cuz yeah, this can be a really amazing tool. >> Yeah, hopefully. I mean, that as long as long as this thing is useful. I And the the good part I realized is that good it it did find a whole bunch of useful issues in code that Glen had pushed. So, it was nice. >> Well, that's that's what I'm saying is is he, you know, you definitely uh in you know, in a couple of his posts he was addressing, "Hey, this the tool found this. We need to uh address it." A lot of stuff came out of that. And I was like, "Wait a minute. What is this tool? This is really cool." >> Yeah. Yeah. You see, this is what this is what happens when Glen pushes that mad volume of code and I try and keep my head above water on it. >> [laughter] >> I'm I'm I'm drowning in his code push at this point in time. It's just completely insane. >> I know. And he's he's pushing out to uh the next What let's see. It was the What was the C No, the I forgot the name of the release that started with C. Um >> I press this I press this and yeah. >> Yeah. And now he's preparing the D1. Uh I can't remember the name. He he he mentioned to me. He said he's he's working on the cuz he wants to have a um next Monday afternoon in in um uh Geneva we're going to have a uh uh by the way, I think uh I can run a Zoom on that you know, if if any for anyone who wants to join remotely. I I I just finally reserved the room and of course got Wi-Fi and a screen, so we could run a Zoom for any of that you want to. It's going to be from 1:00 Let me see what this shows up. It's from uh Zurich time uh Central European time it's going to be from uh 1500 to Wait a second. 1300 to 1700, so >> Yeah, that's not a bad time for me at all. I think Zurich is around 4 hours off or something, so I >> Yeah, yeah, yeah. That's that ends up being from 4:30 to uh uh to to to 10:30. >> Yeah. >> No, no, no, to 8:30 your time. Yeah, exactly. Okay, good. So >> If you are able to run a Zoom, please share the link on the signal and I'll open. >> I I will I will definitely do that. I think we should uh Okay, so >> Before we kind of close this one given that we are this is such a heavily attended instance of the meeting. >> Yeah. >> It struck me yesterday while we were discussing a lot of the topics is that for those of us who follow this on a day-to-day basis and and and and are constantly poking around and and all. Even for us sometimes it becomes challenging to have the entire picture in our heads. >> Yes. >> I think we need to be a bit more kind to those who attend our calls on a weekly basis and and try and present to them something that is sufficiently high-level for them to then drill down on their own rather than try to give them the firehose of stuff because yesterday was a firehose. >> Yeah. Yeah, it was. And I knew it was going to be and I was I was worried about that um and I I saw your your note in uh in Discord and I was like you're right. Um, but I'm wondering should do you do you think we should uh >> My suggestion is >> uh that on a Tuesday uh evening or something you feed into your AI agent uh a list of things that they should create as meeting highlights to post or share. And then just or two slides to add and just add add those slides and be done. Because that allows people sufficient high-level uh understanding of what is happening. Like yesterday's uh that the house with no one where no one has any names. It's a very interesting and very involved topic, but it is not a topic that will ever get discussed thoroughly with everyone interested being able to find a time to speak on a call, right? So, we'll have to tell everyone that if you want to speak and you think it is interesting, we'll spin up a separate call. Somebody will lead it and it can go on, but here are the highlights. Here is what you should note for for the project. And the rest of it is just uh technical stuff that's going to happen. I I mean I don't want to be unkind to folks who attended, but I I think yesterday was a good example of like, okay, we are overwhelmed by too much. >> Yeah, I you're absolutely right. I I knew it was. So, um so, are you uh I just want to make sure I understand your suggestion. You're you're you're suggesting that the uh the task force reports get simplified to a set of uh slides that that can get that information across so we don't have as much there and then uh what are you suggesting then? >> Yeah, I'm I'm saying that you remember in the early days of Trust Over IP when you were running meetings used to actually have a slide set where it where it was very high level stuff that says this happened this happened this happened. And here's a link go go find and take your own time to figure this out. I think we need to go back to that thing just to ensure that because these are well attended calls that I think we had around 35 or 40 folks. Yes. >> I know. There were a lot of folks. That's I see what you're saying. So so what we could do is Yeah, it would make it Yeah, it would it would make it easier for someone new or someone who's not involved in the day-to-day to go, "Okay, this is what's going on. Here's exactly where I can go follow up on this, follow up on this, follow up on this." >> And and to I mean 6 years back you needed to sit down and hand write and today it's like a 5-minute AI LLM job. >> That's a great That's a great I you know every single person I'm talking to Sai Krishna in this space that's you know, that's that's deep into it that has has a stake in things. They're all using tools not I think you are an example, Glenn's an example, Jeff Turk. They're deep deep deep into those tools. But everyone's taking and I'm still at a very early stage. I haven't even gotten a subscription yet. So I I I mean Carly is another one. She's pretty deep into it. >> Yeah. I know the two of you were discussing how it was so I think that's a great idea. We need to uh I think that's I think that the editors call a week from not not next week. Oh, that's right. I got to send that the note just canceling meetings next week. But the week after that editors call, I think we should bring this up and see if everyone agrees and we can come up with a uh you know, uh Yeah, just a formula or routine we can use that will be the minimum time for everyone that gets everything out there and still allows folks to go, "Okay, and this is, you know, this is the level that we should go into with any one of these things on on a call like that." Um yeah, that's a that's a good good >> the the other thing that might help is uh so when newer teams like Andy Steves and others come in, they have an access to at least a very high-level slide deck to say, "Okay, here's the level set. >> Yes. >> And I go forward from here rather than sitting down and listening to 3 hours of let me get you up to speed." >> That And yeah, that that um I saw that come up several in the chat today and it's come up several times. That's a really good idea. Um I do I I was frustrated because I was working starting mid-July on a a new white paper in a in a very um uh very accessible format. In fact, I'll put uh I'll put in uh I'll put it in Signal so you have it there. I'm going to find Signal chat. >> I'm going to put a link to this um You you might even recall this, the Sovereign white paper. Uh I don't I don't know uh >> Yeah, yeah. >> Do you remember the Sovereign white paper? >> Yes. >> Uh I'm looking Where do I have that? Um do do do do do do I wonder Oh, I know. I got to put this in front of it. No, that doesn't I swear I've had it up that way. Anyway, I'll find it. Um uh I don't think it's the protocol white paper. It's a different one. Anyway, that format, you might remember, it had it had a >> Yeah. >> one you know, it was almost like slides and you know, it was like one headline and three subheads and diagrams on each page. It was just really, really accessible. I have drafted the headline and subheads and a little bit of the body text and a couple of diagrams for a paper like that. It's not It's going to be much shorter than the first-person paper, much more accessible. That will be another tool um I I I I need to finish the body text and then and then have folks, you know, review and then we need to lay it out. Which is probably something, you know, I think Margie is interested in doing it cuz she's very um you know, graphics oriented, but at the same time it might just be an AI agent thing. Um I don't know. Have you used a Have you tried using AI agent to lay out a nice-looking paper? >> You You can't do that. I have seen people do it because now uh folks like uh ACM allow their templates to be used in an AI agent to like ACM has a very template standardized template in which papers need to be submitted, right? So, they use They actually allow uh AI agents to build the template if they're using LaTeX or something. >> Mhm. >> And uh Yeah, I mean it's not unusual to do. A lot of companies are doing their standardized documentation using AI agent templates. So, this is quite This should be quite easy to do. >> Yeah, yeah. That's That's what I was thinking. Um anyway, that would be another I I I I love your idea of just a you know, uh sort of starting point slide deck um that anyone coming in we could just like right the start of each meeting and go, "Hey, if you're if you're new and and but we could also have sort of current state uh you know, update the update slide deck that we're able to use for the meeting. I think that could make a lot of sense. Really, that's a good idea. I will bring it up on the editor's call and I'll also discuss it we'll we'll uh you know, next week in the in the Geneva meeting. And I think actually what I'll do is I'll get the Geneva meeting on the calendar just so that you know, folks you know, who want to come it's it's there's no you know, particular it's it will be somewhat centered the the the actual testing actually you know what even that we could do the first part of the meeting where Glenn and and and Jeff want to test the open VTC stuff, but you don't have to be in the room to do that. So that's a good idea. Excellent. All right. I'm going to I will go get these things underway and uh then I will end up talking to you next well, it'll be Monday my time Tuesday your time. Awesome dude. Take care and get get some get some sleep. All right. Bye-bye.