Chat_120 - How P2P is changing the Internet [Pears]

[00:00:00] Speaker A: There are so many reasons to be interested in peer to peer. You can be interested in it for resilience because there are no middlemen. So it's going to work if any of your devices are online, but you can also just be interested in it from this angle. The idea that if you have full control of your data and the code that you're using, that you're running on that data, then anybody can fork it and modify it and tweak it and create a whole. That's where the Cambrian explosion comes in. This whole proliferation of apps that give users very different experiences. [00:00:42] Speaker B: What is up, guys? Welcome back to Bitcoin Audible. I am Guy Swan, the guy who has read more about Bitcoin than anybody else you know. This episode is brought to you by Fold and the Fold debit card. Let me put that right there, right open front. [00:00:59] Speaker C: Look at it. [00:00:59] Speaker B: The full debit card gets you sats back on literally everything in your life. And you can also push anything from your card into bitcoin with a 0% fee. If you have premium like I do and you can go in reverse, you can send Bitcoin and you can push it to your card. It is the service for being on a Bitcoin standard is essentially banking that gives you sats back on everything that you do. Plus gift cards right now for the holidays that are killing it in sat back right now, you can literally buy like one gift card that gives you more than the Fold Premium. Like a couple of months of fold Premium just on like a $500 Airbnb gift card right now. It is so easy to make way more in BTC and sats back every single month than the very low cost premium fee. And actually right now you can get 20,000 sats, which is over $20 just for signing up and using my referral code. Check it out, it'll be right down in there in the show notes. We've got an awesome chat today, one that I have been waiting to see finished and published for a while. [00:02:09] Speaker C: The audio is a little bit off. [00:02:10] Speaker B: So we're trying to fix it. It gets a little bit quiet at times. Hard to hear Math and David sometimes, but the conversation is so good. So this was during the Lugano Plan B conference. We managed to set aside a time and hang out at their new location down there that wasn't quite open yet, which is also a really, really cool place. And we just had a fantastic discussion about everything peer to peer about the Pears ecosystem, about Keat, and it's the first time I believe In, I think in this section because we did a Q and A as well afterward. But I think this is the first time that I've really talked about Pear Drive. But we will have an episode on that very, very soon as well. So stay tuned if you've been interested or kind of following my musings on that, because I am, I am so. Oh my God, the relief that we're getting to. We're like right there in seeing this in action, like really in action. So stay tuned on that front. But this is a perfect precursor and we go into so much of the history, the why and all of the different things that we can build and are being built right now in the pairs ecosystem, in peer to peer tech and how bitcoin is related and why. Why our future is a whole lot brighter than a lot of people give it credit for. I think. I think we are watching all of the foundations be built for a much, much cooler Internet in the not too distant future. So without further ado, let's go ahead and get into our Fireside chat. This is the P2P fireside with Mafentosh, David and Andrew from the Pears team. [00:04:02] Speaker C: All right, first off, how are you guys doing? How are we doing? [00:04:06] Speaker D: I think we're doing. [00:04:06] Speaker C: How we doing? [00:04:07] Speaker D: As good as you can do after two days of serious conference. [00:04:10] Speaker A: Yeah, hang it in there. [00:04:11] Speaker C: Awesome. Awesome. I guess, I guess we'll do introductions. I guess we'll do introductions. That feels right. Okay, Matthias, of course you've been on the show numerous times now, but actually, why don't we go down the line just for any new audience. Matias, introduce yourself. Why am I bugging you all the time? And why do I have you here? [00:04:33] Speaker D: Nothing. You can say the word. Yeah, I'm Matthias. I'm the CEO at Hole Punch. So I'm the. I co founded the company with Andrew some years ago. [00:04:47] Speaker C: Wait, you're a co founder? Did not even know that. Did not know that. Okay, cool. [00:04:51] Speaker D: David has been around the same time, obviously. Paolo also started a small team now. We're tons of people. But yeah, we make peer to peer apps and the best one is Keith and we work on that every day and we're trying to make that the best freaking chat app that exists. I'm very, very privileged to be working with these guys because we're making cool stuff. [00:05:18] Speaker C: David, how did you get connected to this exactly? [00:05:21] Speaker E: Oh, I was. I've been in contact with Matthias for quite a while. Before that we worked together and I actually met him at a conference in Oslo, where he was talking about Merkle Dags. And I was like, what is a Merkle Dag? [00:05:35] Speaker D: Don't say this here. Don't say this in one forever. [00:05:39] Speaker E: It has been going on for some time. And that's where, I mean, I was sort of interested in stuff at that point as well. But, like, you know, having that, like, really good, like, conversation at that point sort of like made us like each other and want to talk more and stuff. And we ended up doing some work. [00:05:55] Speaker C: So y'all get along? [00:05:56] Speaker E: Yeah. [00:05:56] Speaker D: Okay. [00:05:57] Speaker E: Yeah. [00:05:58] Speaker C: Hit or miss. [00:05:58] Speaker E: But yeah, I mean, sometimes, you know, remember about that? Bugs being triggering. But yeah, yeah. So that's. So I got a text from Matthias one day saying, why don't you do it at the minute? And I happened to be Tony Fried. And I was like, yes. Finally. [00:06:18] Speaker C: Finally he called me. That's awesome. [00:06:23] Speaker E: So it was, yeah, it's really cool to get in on this, like, at the very beginning and help define it and, you know, break down the paradigms of how this can go and how this can work. And at the time, obviously there was nothing out there like this. And so we were just like, we believe that this is something that works based on knowledge and express all these different things. We didn't know. You know what I mean? But it was just that combining of like, fuck, yeah, we can do this. You know what I mean? That really, like, for me, helped me, like, just go, like, it's hard, straight line with it, you know? [00:07:03] Speaker C: Dude, that's crazy too, because it's funny how much that like expands and like, kind of builds on itself. Because, like, after you guys built Hole Punch and then pair runtime, like, it was my using of Keat and all of this stuff that I was like, holy shit. Because, you know, in my experience, I'm not. I'm not a developer. I've just been kind of like a pro user for. For 20 odd years for peer to peer stuff. And they've all just been like, they're great from the context of, like, how comfortable I am with a computer, but they're awful. Like, I would never give it to my wife or my friends, you know, like that sort of thing. Um, and it wasn't really. It wasn't until Keat really that I was like, holy shit. Like, no, we could act. We can actually build peer to peer. Like, I've always thought it ought to feel like, you know, that's exactly how we feel about it. [00:08:01] Speaker E: Like from being like, I was a 90s kid, you know, like, with the IE 5.5 and all that kind of stuff. Yeah. Trying to make things talk together and your mental model is like, oh, I just make this computer talk to that computer. And it all feels point to point. I think even today some people still think, oh, when I'm direct messaging someone, it's going from one device to another. And obviously in almost all cases direct. That isn't the case unless it's built on pair. It kind of gives. [00:08:30] Speaker D: Andrew had a really good talk yesterday. I don't know if you saw, but just about this problem like this unlearning all the things we learned. I went to university and I had to take some computer science. It was all about all this complexity we actually added because we can do that. [00:08:46] Speaker C: Yeah. [00:08:47] Speaker D: And it's just like canonical, like this is the stuff you have to learn. And then you come out of it and you start doing peer to peer. You're like, oh, I don't need any of it. [00:08:53] Speaker C: Now tell me a bunch of stuff that I can't, I shouldn't even use. [00:08:57] Speaker D: It's interesting because then when you talk to developers, they're, they're, it's, it's very, it's imprinted in your system. [00:09:03] Speaker C: Yeah. Right. [00:09:03] Speaker D: So you have to unlearn a lot of stuff. [00:09:05] Speaker C: The number of people that I talk to who just kind of don't believe it, they're like, oh, well, no, that doesn't work like that. It's just not, it's just not going to work. [00:09:13] Speaker A: You know, just because you're trained. [00:09:14] Speaker C: You know, when you, when you study. [00:09:15] Speaker A: Computer science and you start building things in the world, you're trained to think that there's all this complexity out there that's just required. It's just inherently essential in building anything. I should introduce myself also because I'm Andrew. [00:09:25] Speaker C: Oh yeah. [00:09:29] Speaker A: But yeah, so I'm Andrew. I've been doing this, I guess for about six years now. I'd say maybe seven years. And I remember first meeting with T.S. because I was building kind of a very traditional web service that I wanted to peer to purify because I was having so much trouble operating it. And I was kind of doing all the ops for it. And this was in a research lab that I think Matthias came through and that's when we first met. But the whole idea behind this was you have scientists that wanted to share research data with each other and they would upload it to this big cluster, this Kubernetes cluster that I was running and it was such a pain to do and you know, and I was just Thinking like all these scientists have the analyses on their computer, they have the data on their computer, they have the code and the people they want to share it with are right next to them. But they can't do it. And so they have to rely on me, who's a thousand miles away and it's 4am and so, you know, I'm just, I'm the middleman and all of this knowledge sharing that just should be so much more seamless. And so I remember in that time trying to find alternatives and. And I found some of Matthias modules and started playing with them and they were really, really cool and you could just see a lot of potential in them and where they could go. [00:10:34] Speaker C: Was it hole punch at that time already? [00:10:37] Speaker D: We've been building the building blocks for open source for probably 10 years. [00:10:42] Speaker C: The hypercore stuff is like 12. Tell me how old I am. [00:10:48] Speaker D: But two different phases. Because I actually. And I was just talking to one of our guys just like. Because the origin story I tell once in a while for myself was I was just a young, in my early 20s, probably not against what we've gone on for a while. [00:11:06] Speaker C: Just no numbers. [00:11:08] Speaker D: But I was just sitting on my couch wanting to watch BitTorrent stuff on my TV and I couldn't get it to work. I'll just write a bite. And then there's a path from there to the making modular software, yada yada, and to discovering, coming up with tons of feedback, making it 10x better, yada yada. And it's just been a path from that to that. But it's like, it's very organic actually. And then obviously peer to peer, peer to peer, very open source, just people meeting up at conferences. And then at some point we're like, okay, we actually. Because I think you actually talked about it really nicely saying is this for nerds or is this something that we want to do for everybody? Yeah, because we're very good at making it for nerds. That's why these guys discover. That's why they all stay right. Because it's like a virus inhibitory comes in. You're like, I just want to do more and more with this. But we were like, actually we need to make this good enough and friendly enough so we can have it run everywhere so that you can understand it without reading a paper that we can make apps that people actually don't even care that they're peer to peer. Like the only reason we actually say that key to speedy peer is because it's cool marketing, but you shouldn't Care. Right. It should just be like a really good privacy app. [00:12:16] Speaker C: It just works and it just works. [00:12:17] Speaker D: It costs nothing. So it's like no trade offs, etc. And that was the thing that was actually the next phase of our cinematic universe. In that phase where we're like, we got to remain and we're like, let's get serious. And we formed the company and started working with Paolo in that way. I've been working with Paolo for a while but you know, get someone involved just to be. [00:12:39] Speaker A: But that, that phase is actually like a relatively recent one. Absolutely cool. Lunch is like three and a half years old at this stage. But this, this, you know, this whole story has been going on for 10. [00:12:47] Speaker C: Yeah, 10 plus years, long time before it. [00:12:50] Speaker D: And actually I think that's. [00:12:51] Speaker C: You can track it on GitHub. [00:12:53] Speaker E: Well, that's the thing because it all comes from the open source community. I was aware of Matthias before that conference because I used his modules, you know, and I like making open source modules and Andrew likes making open source modules and that spirit of collaboration and that spirit of like, you know, mutual cooperation for like mutual benefit, which is kind of antithetical to a lot of like highly competitive, like aggressive sort of strategies in the marketplace. Like that's missing a lot like an open source, the open source sort of revolution from the 90s onwards, maybe a bit before was like, was like fundamentally it made a whole huge shift like from onto, you know, lots of services. Let us know. Right. And the next phase of that like P2P and open source like this, you know what I mean? Like you need both to make it all work. [00:13:44] Speaker C: Yeah, yeah. Now there's, there's something really interesting and I've talked about this on the show quite a bit and I've actually wanted to do a talk about it because I don't think people quite get that the understanding of exactly how economic incentives like align cooperation or competition and how it is that. I actually think bitcoin is a critical part to this story, especially when it comes to funding because most of the time you don't get. I think the reason it stays in the developer community and in the cli, you know, interfaces for nerds, basically is because it's so unbelievably hard to fund a product that isn't competitive with some, that isn't fighting for the network. But what's interesting is that, and not to turn this into a fiat money discussion, but when you have that constant inflation and the expansion of debt, what you actually do is you create a competition for who can get the resources that we have fast enough before the prices go up. So, like, you get the new money and you're like, I have to buy these houses while they're still 250,000 because they're going to be 320,000. So you, you suck up as many resources as you can and you're in a competitive environment with other people. However, when you have sound money, if you make your house better, you make the value of my money go up. So you're actually in a more cooperative position. Which is why, like, talking about, like, standing on the shoulders of giants and like, moving this thing up layers is that every time you guys built something and then you found the modules and then you got together with him, like, I would never have even, like, imagined or attempted to build something like the hypercore stack and Pair Runtime. But after you guys built Pair Runtime, I had a day where I was trying to move 40 gigs from my MacBook because about once a month I run out of space doing the media stuff and everything. Because, you know, one, one interview is like, oh, yeah, sure, 40 gigs, why not? And I was trying to move files from my MacBook to my Linux. They were three and a half feet apart. They were sitting right next to each other. I had an SMB server running on my Linux. I have NextCloud. I have 10 supposed solutions to this problem. And I spent an hour trying to get it to work, in which I was in the middle of a recording and it said, you're out of hard drive space. Please delete something before you can start recording again. And I couldn't do it. I ended up opening up a new hard drive and chucking it into a bay, plugging it into my MacBook, dumping, and then moving it over to my Linux to get it off. And the only thing that had worked through all of that time to move any individual files and that I've been using constantly was keyed. Was keyed. I just couldn't do 40 gigs of a folder of 800 files. And I, I and then I had been messing with par runtime, and it just struck me. I was like, no, no. Like, I have Bitcoin. I can do this. Like, I, I can have someone build this thing, and this problem can actually be solved in a way that makes sense. And that puts just these two devices. Isn't that the way the Internet was always supposed to work? [00:17:09] Speaker E: Really? And that's what Pair Runtime is supposed to inspect Inspire. Yeah, you know, it's the nerds building it. But like it's, it's actually multi. Multiple layers of nerds until it gets into, into like just normal general people. And that's like so Keats about reaching the, the mass market. [00:17:27] Speaker D: But it's hard for us actually even as. Because we're building it. We had a case actually Andrew, like two weeks ago or a week ago, Randall was in the Meaton Files. And we, we both live in the same city. And it was really fast because he was just yarding for Keith. And it was just the speed was insane. I was like, right to Ender, like, this is really fast. And end was like, dude, look at. [00:17:47] Speaker C: These bits per second. [00:17:49] Speaker D: Because it's like, you know, because we know where the body is. Yeah, it's really fast. And Ender was like, no, no, no, you don't understand this. It's because we live near each other. And then that made me. Andrew was kind of like saying, no, it's because we're near each other. I'm like thinking that's because we're still trained to think that that shouldn't matter. It's actually. [00:18:05] Speaker E: Right. [00:18:05] Speaker D: It's actually not like a hack that we're near each other. It's just because we are near each other. And that's that as fast. Because from Nehru, it's just a consequence of peer to peer technology. It was still, you know, it would still be as fast as it could if we were further apart. But it's so cool that matters. [00:18:18] Speaker C: Yeah. [00:18:18] Speaker D: Because that takes that thing away where. [00:18:20] Speaker C: Like, you know, it brings geography back into like it's a. It's a. It's more natural. It's a connection to the real world rather than a middleman. Well, it's like kind of like it's limited. [00:18:29] Speaker A: It's only bounded by your network capacity. It's only bounded by your hardware. Which I think is what's so cool about it. Like there's no other artificial constraints that have to be imposed by somebody else. [00:18:37] Speaker C: It's raw. Like it's, it's like the actual Internet connecting device to device, which is just epic. [00:18:43] Speaker D: Because I was thinking a lot about it in your talk. You did a plan B, which I really like by the way, because you were talking about how the current economic system is really hard for us to think about because we've been living in it so long and it's like we're shaped by it. And I think, I think it's the same way about centralized technology. It's really, really hard to not think centralized, even for us as like people who know yeah, because we're social. [00:19:06] Speaker C: It's just the frame. It's just the frame of reference that has always been when you interact with things. Yeah. [00:19:12] Speaker D: And then, and then the only other reference we actually have with PHP technology is actually piracy because it's like, for me at least it was bitter back in the day. [00:19:20] Speaker C: Legal. Legal for piracy, of course. [00:19:26] Speaker D: And then that actually trained my brain then to be like, bad. [00:19:29] Speaker C: Yeah, bad. [00:19:30] Speaker D: And that's actually what I hear with a lot of people. Me and Andrew, when we were just making modules, we got our funding from grants and like we're pretty successful as you can be with that. But we did that a lot. People were just like, okay, so it's piracy. No, we're actually trying to frickin. [00:19:50] Speaker C: So you're doing illegal things. [00:19:51] Speaker D: Yeah, yeah. [00:19:51] Speaker C: What kind of illegal things are you doing? [00:19:54] Speaker E: Change that expectation. [00:19:55] Speaker D: It's pretty hard, right? [00:19:56] Speaker C: Because it's like, no, we're just building the Internet without a middleman taking a rent. [00:20:01] Speaker D: And the next thing they're saying is, how do I make my money back? No, we're making infrastructure and then we can make things on top. [00:20:07] Speaker C: Yeah. [00:20:08] Speaker D: And bitcoin, the environment there was the only place which reaction by people who had that mindset and that changed my. Because I was not shaped by the bitcoin community that way because I came up for peer to do. But I changed my opinion about the bitcoin community a lot because there was actually people who had money. People in open source community don't have money. [00:20:26] Speaker C: Yeah. [00:20:28] Speaker D: And we're willing to put them because. [00:20:29] Speaker C: They'Re in open source. [00:20:31] Speaker E: Hopefully we can change that. [00:20:32] Speaker D: Hopefully we can change that for sure. But like, but there was like a, there was like a shared sense of like implied radicalism I guess. That was very interesting for me. [00:20:42] Speaker C: Yeah, yeah. There's a, there's an element of like this is one of the really powerful things. And it's funny, I remember conversations and like 2012 and 2013 where we were talking about this because you know, you're watching like one of the first bubbles happen and you're thinking about what happens when like you get a bunch of paranoid crypto anarchists and open source people who have literally designed their own money and then that money is successful. Like bitcoin is literally, it's, I mean obviously it's literally the open source money, but it's also the money of open source. And when you can actually capitalize on it and you think about it in the context that sound money actually puts people in a cooperative stance. It like I have no reason to think of anybody who is building something in the bitcoin space as truly a competitor because they are making everything better. And I think that's going to be such a floor to kind of lift up the open source development at the exact same. I mean, it has. It's literally done. It's funny you say that because it's. [00:21:47] Speaker D: Exactly the same thing. Basically, from day one, we started the company about. Like it's not a serum game. [00:21:52] Speaker C: Yeah. [00:21:53] Speaker D: Like it's all about just making the market bigger. [00:21:55] Speaker C: Yeah. [00:21:55] Speaker D: And have this. The English leaders call it Cambrian Cam. But like, you know, just have tons. You know, we need tons of apps. Yeah, we need tons of apps to compete with each other. So actually the best app comes out rather than it's just like. No, it's the centralized one you're logged into. But that's very interesting for me. [00:22:19] Speaker C: Yeah. [00:22:19] Speaker A: I think what's so cool about this also is that there's so many reasons to be interested in peer to peer. You can be interested in it for resilience because there are no middlemen. So it's going to work if any of your devices are online. But you can also just be interested in it from this angle. The idea that if you have full control of your data and the code that you're using, that you're running on that data, then anybody can fork it and modify it and tweak it and create a whole. Where the Cambrian explosion comes in, this whole proliferation of apps that give users, you know, very different experiences depending on what they want. And so it's kind of like it can create this very competitive but also creative environment. Creative you can't get. [00:22:52] Speaker C: Yeah, yeah. [00:22:52] Speaker A: It's kind of like the dream of API mashups that flink was never really realized because, you know, people want full, you know, to seize control of their own APIs. But in the peer to peer world, you can't really do that because it's permission. [00:23:04] Speaker C: It was permissioned. Trying to be permissionless, but it necessarily couldn't be. But peer to peer can actually be that. Like this. This can be. [00:23:13] Speaker D: And it also has. Andrew actually talked about this in his talk. [00:23:18] Speaker C: Which stage was it on the utxo. Okay. [00:23:21] Speaker D: Full House was really cool. But he's talking about this thing about data because that's one of the challenges in PewDiePie. But it's also one of the features, I think, where you know the systems because at the end of the day it's just data and code. And that's a challenge because all the code is always out there, all the old versions, but also means it always works. You can have an old version and if it worked at that point, it's going to work again now within my reason. So you can make things and you can push them, put them out in the public and you don't really need to maintain it in the same way. There'll still be box, the box will still be there, but it'll still work when you dust it off like five years later. And I had this experience because continuously as an open source developer, I, I found one of my old modules which I'm using every day. The other thing was like seven years since I still worked because it was software, it's still working. The only thing that was broken on that software was that the readme that was displaying was using a centralized service to display a picture. And that picture was just completely gone, just gone. And it made the whole thing just. [00:24:24] Speaker C: Lost to the graveyard of the Internet rendered terribly. [00:24:27] Speaker D: And I was just thinking like all this work that went into this and the reason why this looks really broken. [00:24:32] Speaker C: Is because of the one centralized thing. [00:24:33] Speaker D: The one centralized thing and everything else because it's just code and data kind of work. So I think that's kind of mentality. [00:24:41] Speaker C: It's interesting. [00:24:42] Speaker A: I also like this idea of being able to dust things off later and just use them. It's kind of like there's such a huge gap when you use something like Microsoft Word or Paint or something. You have local software, it's for you, you save it to a file, it's on your disk. If you have a program that can read that format later, it's just going to work. But it's so funny just how much more complicated that gets. As soon as you want a network to Microsoft Word, as soon as you want to have any kind of collaborative experience, there's suddenly it just, just blows up. There's no reason why, you know, you couldn't have some kind of social app, for example, that just has all the same properties of Microsoft Word. You know, it's just something that you have on your computer. All your social data is just on your computer. And you dust it off at some point in the future and it just works. It's like that's. I think that's what peer to peer is kind of trying to bring us to. [00:25:22] Speaker C: Yeah, you know, I had a. In fact, I actually, I got to talk about this just after your talk on stage, but I had this like crazy. I've had a couple of aha moments with the pair runtime stuff. One of them was when we were. I was working with the two developers that we've been, you know, trying to solve the. Oh, shit, I can move files between my computer problem. And we were looking at stuff, doing our Monday meetings or whatnot and doing a rundown of stuff. And then as we were looking through it, he realized there was an error or something like, messed up with. And he's like, wait, wait, wait, let me, let me do this real quick. And he fixed it and then immediately sent me the key and I ran it the fixed version, like instantly. And that was like my first, like, holy shit. Like, I'm running this off his computer, like with nothing, you know? You know, like this just, it just happened, you know. And then the other one was when I was doing this by myself because like, like I said, like, so much of the bane of my Internet existence have just been trying to get my two devices in the same place to. [00:26:28] Speaker E: Talk to each other. [00:26:29] Speaker C: It's so ridiculous. It's like, oh, log into your Apple ID and. But I was building a. Building. I was using an LLM and asking it to build an app for me on Pair runtime and using the. Just the normal desktop application framework. And I'm building a little drag and drop thing for files because I have a bunch of little single purpose apps that I've built to like convert a gif, change the format so it's ready to upload to YouTube, drag and drop, transcribe it, give me subtitle file so I can put it back in like. And I've done all of this with just AI, just like basic LLMs. [00:27:12] Speaker E: Well, very exciting to me in itself, like people building on Pair using LLMs. [00:27:16] Speaker C: Exactly, exactly. [00:27:18] Speaker E: And build on it. [00:27:19] Speaker C: Yeah. And I'm trying to put. I'm trying to make a little framework for it where I can have a little like, just kind of like a sidebar app, drag and drop files or folders or anything and then run all of my separate things on those files. Which. It's a level of design that I'm not used to. So I've had to reiterate and do a bunch of things. And you know, LLMs are only so good at it. But it was funny that I finally got the first function to work. I got my current file object, everything. Like the main framework was finally working and I just was. My base function was like, copy file. It's like, just, can I do this with the file? And I copied it and sent it to a new location on my Mac and I was like, oh, it's working, it's working. My stupid app that doesn't do anything is working. And so I immediately was like pair stage or in a pair seed or whatever and got the key. And then I was like, is this going to work? And I jumped over to my Linux machine, I said, pair run and put it in and then bloop pop right up on my Linux machine. I drag something from my desktop and I hit copy and then bloop brung brings up the thing and I saved it to a new folder. And I was like, holy, this works. This is so cool, man. So cool. I love it. But yeah, it's been, it's been a trip. [00:28:35] Speaker D: I love the thing about Pewds pewdiepie because it's always like that. It's always like it's one command. [00:28:39] Speaker C: Yeah. [00:28:40] Speaker D: And we should see endless face when the run back because he, he knows all the layers that's going through, like. [00:28:47] Speaker C: All the errors that can happen. He's like, it didn't air. It didn't. It didn't. Again. [00:28:53] Speaker D: It used to be because we've been working on infecting that basically 10 years, right. Like obviously it's like iterations and because we solve one problem, we solve the next problem. We solve next. Like five years ago that would have worked 50% of the time we open. Like that's freaking amazing. [00:29:08] Speaker C: Yeah, it's amazing. [00:29:10] Speaker D: And after we got really serious and looked into spent a ton of resources on like we got now actually, it's actually 100% also clear, you know, nothing. And that's just amazing to me because that's like where we can get these unlocks where we can actually have. Because it sounds so complicated. This is our problem because we mega nerves. We like to make it sound complicated. [00:29:31] Speaker A: We. [00:29:32] Speaker D: People always ask me, I'm like, let me tell you about dhts and stuff. [00:29:38] Speaker C: You want to explain how it works? You know, everybody's like, well I can't use that because you just explained 80 things. [00:29:44] Speaker D: Then they come back and they're like, oh, I read about them and it's like, actually you don't need to know any of them. It's like a freaking stage at one up or like, you know, it's really, really simple. Do the things that's like a continuous path also also just like to. To make it approachable. [00:30:00] Speaker A: Yeah. But I think in many ways that's also the story of like a hole punch because you know, this is a long journey and like we've been doing this before but in the early pre hole punch Days, it was always 50%, 50%, 60%. I think hole punch is just a story of how do we go to 100%. And that's a huge amount of work to get those last percentage points. And that's just really, really tough because you can get to 98%. But some things are going to be problematic on some networks, but not others. And we need to find ways of debugging those weird situations. And we're getting a lot better at that over time. But those last few percents are really, really tough. [00:30:33] Speaker C: Yeah. And it's a. [00:30:34] Speaker E: It's a. [00:30:35] Speaker C: It's a limit, right? Like, it's. It never actually. Never actually gets an obstacle. [00:30:41] Speaker D: But then that also creates these situations where I actually been. I think on the last year, four flights I took, I got free wi fi because the network, the pewter engine just breaks. Dude. [00:30:50] Speaker C: I've read, I've run in that a couple of times where it's like, you're not allowed to send images or videos. And I bring up Keaton, there's an image. I'm like. [00:31:01] Speaker D: It'S like. It's like me. It's like me going, I'm on the plane, right? I'm like, actually, I want to stay connected. So I go. And I'm more buying citizen. I go, actually to pay while I'm paying keep. And I'm like. I'm like, oh, I guess I paid already or something. [00:31:17] Speaker C: It's like, does he have my hotel room? [00:31:21] Speaker D: And then I was like, oh, okay. But it's just like, you know, it's fine. [00:31:24] Speaker C: Like, they just found a route, not a code. [00:31:26] Speaker D: Like, life finds a way, kid finds a way. You know, it's like probably rounding. I'm sure the planet is going off. [00:31:32] Speaker C: Insert Jeff Goldblum. Keith finds a way. [00:31:36] Speaker E: That's one of the things that I don't like when people don't believe it. [00:31:40] Speaker C: Right. [00:31:40] Speaker E: It's because people don't have that model of like what we're named after, the hole punch. [00:31:46] Speaker D: Yeah. [00:31:46] Speaker E: Like, they don't. Like, all their lives, all our lives, we've been like, blocked out from being able to use things without, you know, being gatekeeped and all that kind of stuff. [00:31:57] Speaker C: Right. [00:31:57] Speaker E: So that's what we're used to. And the idea that you can just now with this tech, just go point to point and it punches through anything is just like, is mind blowing. It's hard to believe, but it's real. [00:32:09] Speaker B: Yeah. [00:32:10] Speaker D: Actually, it makes it also. Sometimes it makes our life really hard. Because when you make PHP normal, it's all about defense and depth. Kind of like a bitcoin everybody's anniversary. So we have all these defensive, like all this didn't work because somebody's probably attacking it. So does this all around. So we have tons of boxes where I would never see them because the network just a gaps. And then at some point we're like wait, how this code never ran and never worked. It's like, oh, it just took the alternate around and then you fix it, it becomes much faster and stuff. So it's actually really hard because you're just. The software is finding itself but at. [00:32:42] Speaker E: The same time it's super resilient. [00:32:43] Speaker D: It can't, it doesn't go down makes it that way. AI is very scary to me because that's where like we're not going to. [00:32:48] Speaker C: Be able to shut it off. [00:32:53] Speaker A: In. [00:32:53] Speaker C: Terms of the Cambrian exposure, the Skynet. [00:32:58] Speaker A: Getting ahead of ourselves. [00:33:00] Speaker E: But there is, there is, there is a piece of like, in terms of like not just applications but not necessarily services, but there's the idea of like an ecosystem of P2P tech including AI, including different things. And that's where it can get really interesting in terms of reciprocity between pairs and like, you know, I need some storage stuff, but can I have some AI stuff? And like that's, that's a whole new level of human interaction as well. That's completely just between pairs. [00:33:31] Speaker C: Yeah, maybe that's actually. Oh yeah. [00:33:34] Speaker A: I was just going to add that like you know the one cool property of any peer to peer network is that peers help other peers. So the more peers you have, the better it gets. So you know, any, any of these two person point to point things, it's really, really cool. But the more peers you can add to that, the better things can become. You know, we, we recently, you know, the last year we got IDs out in Keats. You can have cross device identities. But that also gives us this really cool social backbone for building out lots of other features like social mirroring and you know, selecting a set of trusted peers that you might want to be, you know, you might want to use to enhance your experience. And it's kind of like, you know, these are things that you can do now or we can start investigating what we want to build there. But yeah, it's that any, anything that we can do to get more peers is really good. [00:34:13] Speaker E: I love how it incentivizes good actors whereas other systems can incentivize bad actors. Do you know what I mean? It sort of incentivizes, let's work together. It incentivizes that, that behavior. And I think we should, I think just on a philosophical level we need. [00:34:28] Speaker D: To see more of that. [00:34:29] Speaker C: Yeah, yeah. I think it's also like super critical in. Because it's extremely hard to overstate exactly how much if you, if you tweak like the balance or the incentives at a very low level, how much that actually continues to manifest up, up the layers without even people understanding that it's happening. Like that your behavior just changes, how you react to it changes. [00:34:57] Speaker E: And, and that's how we shift the frame. We're talking about being in that centralized frame all the time. The, there's a primitive right at the bottom of all of this, which is the Kademlia. The thing, you know, going all of this right at the bottom and then that, that, that sort of each level of the stack has changed the way we code. Like we code. That's. That seems to have affected our frame and we don't. And the tools that we use and the tools that we make and how we work with each other. It's very decoupled but like very, very effective. And putting all of that together, it's just like I think that that is how we move forward as a species. [00:35:35] Speaker D: I think it's actually, I think totally, totally agree. And I also think to a degree this is what you're saying because we've had this discussion internally a lot where we've written tons of open source code for PewDiePie and everything and it's running everywhere and it's running in Google, it's running in Microsoft. It's has, if you go to the download size, that's billions and billions of downloads. Right. We never made any money off just open source. We just did it fun. And I often had this like. And it's very popular and we always, sometimes I think like, you know, why did we not make any money off it? And to a degree, it's because you make money on products, some good products. The reason why we make a lot of good open source software is because it's really easy. We just write software, we just put it out there. People can download it. It's easy. It doesn't cost us anything. Just do it. Making products is really hard. And it's really hard because really quickly somebody wants you to take the credit card out of your pocket. They're like, you know, if I want to make an app, hosting it, whatever cost money. So it's like, I'm not gonna just go, I'm not gonna make a Thousand of them. [00:36:39] Speaker C: Yeah, I make one. [00:36:41] Speaker D: That one I make is probably going to be bad because I'm fan like I'm not that good at writing software. I'm just good at writing a lot of software. [00:36:48] Speaker C: Yeah. [00:36:48] Speaker D: So I've written a bajillion software. One percent of it is really popular. If I could do the same with products, that would be a huge success story for me. [00:36:55] Speaker C: Yeah. [00:36:56] Speaker D: It was too hard. Peer to peer for me solves that by just. I don't know if it's incentive, but it's definitely just removing that. Disintentive. [00:37:03] Speaker C: Yeah. [00:37:03] Speaker D: Because I actually think there's a disintensive products right now is that's a good. [00:37:06] Speaker C: Way of putting it. [00:37:09] Speaker D: And that's one of the reasons why originally I was super excited about this because it's like just sit at home, sit on my couch, might be shitty, somebody might pick it up. It'll make it better. That iteration cycle. [00:37:20] Speaker A: So powerful. [00:37:21] Speaker C: Yeah. [00:37:22] Speaker D: So it's for me it's all about. [00:37:24] Speaker C: Yeah, yeah. Now this might actually be a good segue when you're talking about like reciprocity and like, you know, I'll give you some AI if you give me some storage. Because I'm curious where things are headed, where the next thousand little modules and little things that you can look at, look at the reactions. Oh my God. It's a big topic. [00:37:52] Speaker D: Yeah. I mean, no, it's a great topic, thinks I. Yeah, well, you know, I. [00:37:55] Speaker A: Mean to take a key focused angle on it, I think uh, you know we, so we, we have Pair, which is the platform for building all these crazy tools and modules and making little apps with them and or big apps with them. And we have key which runs on top of Pair and it's, you know, it's our chat app. We, we decided to focus on chat because if you actually decompose what's really hard about chat, it generalizes to, it's primitives to everything. [00:38:17] Speaker C: Like I thought about that from the very beginning. I was like, why didn't they do the low hanging fruit of like just getting files between computers. But then like I think about it, it's like if you can get the chat app to work, you can get anything to work, you know. [00:38:30] Speaker D: Exactly. [00:38:30] Speaker A: So I kind of. Excuse me, sorry. You can add that up. [00:38:36] Speaker C: Everybody stop and clear yourself. [00:38:39] Speaker A: We've had an intense few days. But so the way I think about how we built Keat, it's kind of, it's kind of like Jira Dreams of Sushi, if you've seen that movie of someone who just focuses on making kind of one thing really, really well. For 30 years, you know, we are kind of refining and refining and refining these models for chat because we know that when we can actually make them scale when we want them to, you can make something like a Twitter or a Facebook or any kind of social. [00:39:01] Speaker C: App or any kind of file sharing. Yeah, yeah, exactly. [00:39:04] Speaker A: Or file sharing. [00:39:05] Speaker C: File sharing is easy after you do. [00:39:07] Speaker A: That'S what style media sharing. All of these things you can put, you know, if you squint, you can fit them into kind of a chat oriented framework or you could just take the data structures we built for that and easily convert them to work in those situations. So the hard problems are all captured by that. So I think we're now kind of reaching this point where we think we've come across a really solid set of data structures and algorithms for doing chat incredibly well. And from there I think there's so many possible branching points of other things you could build with that. [00:39:35] Speaker E: It's a basis for smart contracts as well, right? [00:39:37] Speaker D: Yeah, you can totally run anything, like close networks. And it's actually funny because all this stuff that Andrew was mentioning, we can't solve. Basically this year we've been like hunkered down and just in like crazy optimization performance mode and we're rolling all that stuff out. The thing we're really bad at, that we're trying to get bad at is just telling people how it works because we're sitting on Amazon open source. So much stuff. [00:40:03] Speaker C: Yeah. [00:40:04] Speaker E: Wouldn't bring any vision bottleneck in a way. [00:40:06] Speaker D: Yeah. It's just hard for us because we're too ingrained in our own nerdiness. But I think also that's a huge, huge part of it for us, like, you know, making it more tortureful for developers because it's already out there. It's not already really simple. We're just mad at sometimes explaining it because we go straight to Kevinian tickets. Like that's really interesting for us because we actually, we could build Keat two or three times faster if we didn't care about anything else. Me, Andrew, had this discussion the other day because we care about solving this for the world. So every time we're like, you know, Kita is not a jet app. It's like collaborative, peer to peer app. All the data structures are generalized that way. All open source. If we just wanted to make a chat, I'm sure we could just cut a ton of corners. But we're always thinking about the general case because we wanted this to be a huge revolution and we had all those unlocks and we actually already passed that. So now we just need to like, you know, seal the deal on that and roll the last gens out with planning on doing by this year. It's really exciting. So there's going to be tons of releases this year on all the open source things for tons of performance. It's been a journey for us for sure just getting the stuff to where we want it to be and it's. [00:41:25] Speaker A: Definitely getting very exciting. [00:41:27] Speaker C: How's the database migration to RocksDB? How big of an undertaking is that? [00:41:33] Speaker D: I had said originally it was going to be two weeks and then I. [00:41:37] Speaker C: Was four months ago and then it. [00:41:40] Speaker D: Took six months because I was in charge and then somebody. I can't be in charge. And then I put one of our really smart guys in charge and he basically just finished it this week in like three months which is really, really. He started to look like me. [00:41:58] Speaker C: Is he any good? How much hair has he lost in the last two months? [00:42:01] Speaker D: I'm stressed at Logan. [00:42:03] Speaker C: Oh man, that guy's good. Look at the shine. [00:42:07] Speaker E: I never doing anything with Rockstar. [00:42:11] Speaker D: Well the hard thing for him is Carlite. You know he'll. He'll work on it and then he'll be like I find this bar but it's your code. So I don't. I think it's my fault. [00:42:20] Speaker C: It's like no, it's probably my bar. [00:42:23] Speaker D: So. So it's just. [00:42:24] Speaker C: At least he's polite. [00:42:25] Speaker D: Yeah. [00:42:26] Speaker C: At least he's nice about it. [00:42:28] Speaker D: But refinement just actually caught turned glass corner. I think I'm going to be rolling out some stuff. Yeah. In the next week. It's very. It's looking really, really, really intensely good. Andrew's been working on some crazy stuff. [00:42:41] Speaker E: Yeah. [00:42:42] Speaker A: And along with the rocks to be one we've also been really iterating on the database. The database that runs inside the actual room. Kind of the higher level database that stores all the chats and invitations and the members. And I think we came up with one of these funny things where we've gone through many iterations with it and the first one was pretty focused on. [00:42:59] Speaker C: Just like just chat. [00:43:00] Speaker A: Can we actually solve this for the chat use case? And over time it becomes more generalizable just as along with that theme. And I think we've landed on something that should work incredibly well for our cases but should also be very upgradable so that we can change it a lot in the future and it'll make it a lot Easier to prototype different kinds of social apps as well. And this is all going to come as part of this one big kind of database push. [00:43:22] Speaker D: It's really hard to roll out and make events, unfortunately. So we're just going to be. We're going to admit, like, there's going to be like, K 2.0 is the same stuff, but it's just like optimize to death. But it's really, really good. People always text me. They're like, because Keith is really fast. And then it's really slow at some things right now, to be real. And people always text me, like, I'll be cool if you can just not do all this networking stuff because it's slow and I search for rooms or something. [00:43:46] Speaker C: Something. [00:43:47] Speaker D: I'm always like, you don't realize. It's just because we're. [00:43:50] Speaker C: It's not the networking. [00:43:51] Speaker D: We're just really bad at writing a local code that searches it because we've been working on these really, really hard problems and we've been trying to scale the team. It's actually just because it's like a loop that's intensely intense to slow. So Andrew just took partial app. [00:44:03] Speaker A: So I mean, just the tldr of it, without getting too nerdy about it is instead of. So right now when you do a search, it does just really, really naive stuff. Like if you type in the search bar over your rooms, it's going to have to basically go into all those rooms. And when it does this, it has to kind of engage the networking stack and, you know, just spin up a bunch of stuff. [00:44:19] Speaker C: Oh, wow. [00:44:20] Speaker D: Okay. [00:44:20] Speaker C: Yeah. [00:44:21] Speaker A: And it's going to do this, like it's going to scan over your rims and do that. So it's really, really, really silly. And we've basically just, you know, we're getting rid of the silliness and the next one is actually, we're going to be a lot more careful about how we, you know, how we spin up those different pieces. We're going to introduce a bunch of scheduling to do that very efficiently. And it should just be lots, a lot more local. [00:44:39] Speaker D: I'm sure marketing will like it because it's one of those slides we can make. We're now a thousand times faster and sounds really good, but somebody's going to be like, wait, was that just because of really show before? [00:44:49] Speaker C: Yeah. [00:44:51] Speaker D: So that's really exciting. [00:44:52] Speaker C: Everything's relative, right? [00:44:55] Speaker D: And like Andrew said, we took on all these problems because again, this is all problems everybody's going to face. And we generalize them. We Put them into our data structures. Rolling out, also open source. Add a little bit about that in my talk. Going to be rolling out really soon. It's going to be super exciting, very exciting. And it's going to be a lot easier to build apps. [00:45:15] Speaker A: I think the dream is always thinking. I don't know if this is a great metaphor really, but Firebase was such an unlock for people when it came to building centralized apps. When that was launched, it made it trivial to spin this up without needing to do database configurations or deal with user management. Could just create this. And you had your users for free and you had all your data models basically for free. Really, really digestible. And I think that's like a really nice end goal to get to when it comes to how developers can actually. [00:45:41] Speaker C: Interact with all of our stuff. [00:45:43] Speaker A: But it's just, we haven't really had the right ideas for how to approach that problem. And I think it's one of these interesting things where the more we iterate on this kind of the closer that seems to be falling out. It's just like an actual consequence of what we're making. So this new database, for example, it's just very easy to digest and understand and work with in a safe way. A safe, upgradable way. And I think that's like, I think we're kind of slowly reaching this point where developers will be able to use this stuff really easily to make things at huge scale. [00:46:09] Speaker C: Hell yeah. [00:46:10] Speaker D: That's all you have to do. Are you going to bring us out or nurture? [00:46:17] Speaker C: Well, I really think, like, so much about producing blocks that other people can build, you know, like, it's just. And that's what's. And it's funny because, like, that's generally my relationship to it because, like, as I said, I'm not a developer, but when they kind of reach a degree of maturity some, at some point, suddenly I can build with it. Suddenly the developer that I'm working with can build with it, even though they're not super familiar with the tools. And the, the unlock is, is not linear, it's exponential, you know, and I think we're reaching. I've been feeling the momentum for a long time. Like, I've been in the key rooms for two years, you know, like, and that momentum is definitely, definitely slowly starting. I mean, the paradevelopment room just reached 500 people. And there's so many, like, little conversations like, oh, I'm doing a to do app, I'm doing a thing, I'm doing a thing. And, and we'll probably be in this situation where, you know, a thousand people build 2000 things and then three of them are just like, oh, shit. Oh shit. This is crazy. [00:47:23] Speaker D: We have that. I mean, we have a lot of that internally. I don't know if you saw how you have to talk about just what Tether is doing tillers during banks. And a lot of it is building on peer to peer. And for us at least it's been an insane journey just having other teams work with this stuff, building things completely independent of us. And then we're just trying to just freaking works. Building all the stack. We were never involved. We're not used to that being enterprise. We used to be modern everything because it used to be very rough. So that just that unlock for us is very scary, but also just insanely cool. And that's where I'm like, the next iteration of that is kind of like this true unstoppability where we are not even involved in it. Know. We don't know. I love this idea. We don't know anything that's going on. [00:48:09] Speaker C: Yeah. [00:48:10] Speaker D: I would already get this from Keith where I talked all these people at the conference and whatever to have conversations on Keith or like running their internal comps on Keith. And I'm like, I'm just always thinking, I have no idea. [00:48:21] Speaker C: Yeah. [00:48:21] Speaker D: Because I couldn't even find out if I wanted to. [00:48:23] Speaker C: Yeah. [00:48:23] Speaker D: And I love that it's such a. Such a. [00:48:27] Speaker C: That is. It's the way like it's. It's the only way to actually like align natural incentives and make conversations genuine again. That they literally naturally disperse the connections. Naturally disperse. The network naturally grows and builds. And you. You're back to what I think is like a real Internet rather than this. This series of hierarchical platforms that we all just like filter through. You know, like this was the. The birth of the Internet was that, you know, for a long time. [00:49:01] Speaker D: I also think it's one. Because I was talking to about this yesterday. I think it's, you know, there's that quote about any power corrupts kind of thing. I can't remember. But I've been thinking that's just building Keat because Keat is decentralized and the tools are like very defensive. And they're so defensive that they can defend against even us. Right. We can't do anything really. And we have. Sometimes we have box where you're just like, you just wish you could into our server. Let me just look at the. [00:49:29] Speaker C: I can just fix this really easy. [00:49:31] Speaker D: Two minutes. I could just log in and Just look. Why is this hair being full? Yeah, yeah, yeah, yeah. And, and, and so if I had that power. If I had that power, I would be so tempted. [00:49:40] Speaker C: Yeah. [00:49:41] Speaker D: I would be like, I'll just go ahead and look at that message. Because we spend. We can't do it. We literally can't do it. And it's so frustrating, but it's so good. [00:49:49] Speaker C: Yeah. [00:49:49] Speaker D: And then it takes us two weeks. [00:49:51] Speaker C: Yeah. [00:49:51] Speaker D: Or something. Right. But it's like, you need to have that. Right. Because even us as developers can feel that. No, it's like, it's tempting. So the technology just needs to have that built in from the beginning. And that's so cool. [00:50:06] Speaker A: You download key from the app store. It says, this app does not collect any user data at all. There's zero there. [00:50:11] Speaker C: Love it. [00:50:11] Speaker A: That's a blessing and it's a real curse. [00:50:13] Speaker C: That's a blessing. And it's also hard as shit for us to shoot anything. That's the way it is. [00:50:20] Speaker D: But I mean, it's like, it's kind of like almost ruling a time like that. But it means we need to have really good support people. [00:50:26] Speaker C: Yeah. [00:50:26] Speaker D: Because we need to like, ask people a lot of questions and hopefully they're willing. We need to have good users. You know, that can explain. [00:50:32] Speaker C: I'm a good user. [00:50:34] Speaker D: No, but seriously, it's kind of like, you know, how did this happen? What could you remember what you saw? I think that's really interesting. [00:50:43] Speaker A: Right. [00:50:43] Speaker D: It's like, it makes, it slows it down, obviously, but it also just makes it so solid. [00:50:47] Speaker C: Yeah. I was about to say it. It slows it down with. It's a trade off of slowing it down and making it more difficult, but with the benefit of robustness, resiliency. And it's just going to last. [00:50:58] Speaker E: I think it creates more social cohesion as a result. It allows people to connect. Like anyone who's building peer to peer apps wants to connect with any app at all. [00:51:07] Speaker D: We talk so much with our users. [00:51:08] Speaker E: Yeah. [00:51:08] Speaker D: We're talking about users a lot. That's probably. I mean, everybody talks to me. Users. We're pretty talking. [00:51:14] Speaker E: Yeah. Because we're not analyzing them. Because we're not. Because we're not like putting them into, you know, big data machines to see what they think. [00:51:22] Speaker D: The level of like, sometimes where I'm like, I had one, I had one block where I was working with Sasha sitting over there. I was like, at some point, I was like, sasha, I'm getting in the car, I'm driving over to your house. We're Going to look at this now because now we need to go back this device. [00:51:37] Speaker C: Yeah. [00:51:37] Speaker D: And we can't do that a bit more because we even internally don't have those tools because the technological just won't allow us. And then we fixed it. [00:51:44] Speaker A: It was really funny when I, when I. [00:51:45] Speaker D: So I. [00:51:46] Speaker A: We were grant funded for a very long time and I moved to Copenhagen from, from the US to work more closely with Matthias. And when I was in Copenhagen I had to move apartments all the time and I moved to one apartment and I was working there and I just, I discovered that like something about the network was like causing big issues with the hole punching. And the only way to fix that was for Matthias to come out to the apartment for us to do it in this actual. [00:52:08] Speaker C: What's wrong with just. [00:52:09] Speaker A: But it was really tricky because I was, I was going to move apartments. [00:52:11] Speaker C: Again right after that. [00:52:12] Speaker A: So we had to find ways to work around that. Like maybe you know, can we, can we stay for one more day? [00:52:16] Speaker C: You're a bug hunter. You know, it's just like, let's go to the next WI fi or whatever. What brand is this? [00:52:21] Speaker D: You know, one time one of the guys texted me, he was like, you know how those hotel WI fi's are really hard. At some point I can just be the guy who goes around, stays at hotels. Totally happy. Truth. [00:52:37] Speaker C: Awesome. So the question is how can keep be used to connect local communities online using the example of like a small town. [00:52:45] Speaker D: I think that's interesting. Super interesting question. Because I think peer to peer, this is like actually one of the mega. [00:52:53] Speaker C: Strengths of peer to peer. [00:52:55] Speaker D: Like these local networks that are where people actually live together and you can almost just envision it as like, you know, we could literally just have a cell tower in the city. Right. And we could just be connecting people on a mesh and we can do peer to peer things. I think that's, that's the future we're looking for. Keep by itself actually does this like the technology will already do this. [00:53:15] Speaker C: Yeah, yeah. [00:53:16] Speaker E: Create a room for the town. [00:53:17] Speaker D: Literally just create a room for the town. And then, and then, and then maybe at some point the town invested in more infrastructure so people are connecting just like you know, on land. That's really exciting. [00:53:29] Speaker A: You can even, you know, maybe in the future we'll add things like sub rooms inside of rooms also where you have like a big community, you know, a big like cross town room. And each room, each town could be a little sub room in there if you want. Yeah, I think those. [00:53:42] Speaker D: It's actually Funny, because I think it's one of those questions that sound on sounds like fun and quaint, but it's super important. [00:53:48] Speaker C: Yeah. [00:53:48] Speaker D: Because with the way the world is going, all these like natural disasters also, you know. Right. Local networks that don't go down. Insanely important. [00:53:58] Speaker C: Yeah. [00:53:58] Speaker D: And if all those networks require us to connect to some overseas server or somewhere, it's going to fail if you can't get communication. It's a very, very quick path, not communicating to full anarchy. [00:54:09] Speaker C: Yeah. [00:54:09] Speaker E: There's still governments using Twitter to do announcements about crises which, you know, there's no resilience there really. Like Twitter goes. There was actually an incident I think in the Netherlands a year or two ago where they put out an emergency alert on Twitter, but Twitter was down. And so from there they moved away from that. But like, yeah, what we have, you. [00:54:32] Speaker C: Gotta get burned to change. Yeah. [00:54:34] Speaker D: And I think it's, I think. And I also think it's a couple, you know how, you know, city. Cities also have those. I don't know if that's in the U.S. but everywhere in Europe we have those digital loudspeakers that they can turn on. The sound goes off, something's wrong. You could literally have those things that just beam out of peer to peer signal. [00:54:47] Speaker C: Right. [00:54:48] Speaker D: Because the data doesn't matter where it comes from. Like a message. [00:54:52] Speaker C: Yeah. It's interesting because like one of the things about like how the Internet worked that always was intriguing. But then also like why, why is there this disconnect? Is that actually going back to something we talked about in the previous thing was the local or the lack of geography online. And you know, I would search something in Google and like trying to find something in my area that was like relevant to me, like where I was was always awful. And then the only way to fix it was to be constantly tracked by my location with some giant, you know, to send all of my data and where I am at any one point in time to California to be kept in a big giant database. And I think the, but the kind of the network distance, like the geographic distance on a peer to peer network could actually be utilized in a way like that. Which is just so cool to think that you could bring back the locality of the Internet. You could search, you could search your network near you first before going out. [00:55:58] Speaker A: It's more real. [00:55:59] Speaker C: Yeah. [00:55:59] Speaker A: But what's also so cool is that the city of the town is very incentivized to make that network run really, really, really well. You know, we talk a lot about how peer to peer can be used, you know, to take control of your own data. But a community taking control over its data collectively is also really, really interesting. [00:56:12] Speaker E: Yeah. [00:56:13] Speaker A: You know, so it's kind of, you know, when you give that community its keep room, you know, it is in control of its communication. [00:56:18] Speaker C: Yeah. [00:56:19] Speaker A: And in the future we can add, you know, we'd like to add lots of customization there to make it so that those communities can actually control the experience that they have. [00:56:26] Speaker C: It's exiting at every scale, you know. [00:56:29] Speaker D: And I also think people started using citywide networks because cities are really small. [00:56:33] Speaker C: Yeah. [00:56:34] Speaker D: And if you know anything about physics, the reason why I think is real is because speed of light, just the latency in these rooms, like how fast things are going to be, people are going to think it's all fake. [00:56:43] Speaker C: It's going to be so fast because. [00:56:46] Speaker D: It'S like, you know, stop. Millisecond latency anyway. And I find it really interesting also what you said because it's kind of like there used to be, when you went to a town, if you wanted to find something, maybe you could get a local guide or like get a book, like a guide book. This bar is good, restaurant is good. And then the Internet happens. For some reason, we put all that on these like centralized sites when we rate things very centralized in this centralized opinion. But at the end of the day, when I go to a new town, I still ask the locals. I'll hit up my local friend. [00:57:15] Speaker C: Yeah. [00:57:15] Speaker D: Where should I actually go? Right. Yeah. It's this very old school, super old school way of doing it. Technology just ignores that. [00:57:21] Speaker C: And it's just. [00:57:22] Speaker D: I know, you should just look in the big phone book. [00:57:23] Speaker C: Yeah. What's, what's Yelp's 10 best things to do in this City? [00:57:27] Speaker D: That's such a big disconnect from me and I actually do it to. [00:57:30] Speaker E: So there's a lot of places that still aren't online as well. [00:57:33] Speaker C: They don't have online. [00:57:34] Speaker E: They don't have the investment into, you know, SEO and Google SEO and all. [00:57:39] Speaker C: That kind of stuff. [00:57:39] Speaker E: So. [00:57:40] Speaker C: And it's an uphill battle anyway for most people. It's become so saturated. And because you don't have that sense of locality, it's, it's, it's work without even a payoff anymore. [00:57:50] Speaker D: You know, just when you say it out loud, I think it's so funny how we say it because it's a wrestling, right? [00:57:56] Speaker C: Exactly. [00:57:57] Speaker D: They're making food. [00:57:59] Speaker C: It's like, well, what's the SEO guys? It's like, I don't Know, we make food, I cook. You can have your steak medium rare. Oh yeah. So the question was, what's the vision around implementing bitcoin payments in the app? [00:58:18] Speaker D: Well, so actually we used to do lightning in the app and then when we went Alphabet and beta, we actually disabled it temporarily. [00:58:29] Speaker C: And it's still temporarily. [00:58:31] Speaker D: Yeah, just because, actually because we wanted to iterate it when we're not really happy with the experience and stuff like that. And now we internally we revived it and it's coming back a wallet in early next year. Is the plan where you can do different kind of payments, like tether payments. I think it's super, super important. [00:58:50] Speaker C: Yeah. [00:58:51] Speaker D: Bitcoin payments also super, super important because obviously. And also tying into the thing we talked about before with like local networks, city networks, being able to go to your restaurant and just pay by opening a chat with your local person. Again, the same thing. They should not be technology experts. [00:59:08] Speaker E: You should just have a chat app. [00:59:10] Speaker D: Pay for your meal, get a picture of the receipt if that's what you need in chat. That's so powerful. So it's really, really needed and something we really want to come back. We just wanted to have the experience we wanted to do. So we actually have a team now working on this internally. It was actually part of a big talk also at the conference about just wallets in general. So we're trying to solve this for everybody. So we're making like a development kit for wallets that run everywhere. So that's super exciting. So we're going to have tons of announcements about that soon. We're also still working a bit on our internal credit systems we talked about a couple years ago. We're going to have some stuff on that soon also. So that's exciting. I shouldn't say anything more about that because then people will, yeah, me more here. But that's super exciting stuff and it's really, really important to us. That's actually, I think that's the main point. People might think that's not important to us. It's really important to us. It's something we talk about a lot. So coming back soon. [01:00:08] Speaker C: Sweet, Sweet. If you want to save in better. [01:00:13] Speaker B: Money and you want 20,000 sats for. [01:00:15] Speaker C: Free, right now it's about 20 bucks. [01:00:17] Speaker B: At $100,000 of bitcoin check out, fold. Every time I swipe this card, I get 0.5%. Sometimes one, sometimes even one and a half percent. I can get two, three, five, even 10% on gift cards with major merchants between the roundups. [01:00:32] Speaker C: The gift cards the auto stacking, the. [01:00:35] Speaker B: Sats back on every single swipe fold literally does all of the work for me and it is denominated in bitcoin. And I have more savings just by using this card than like 90% of the United States normal consumer. I've got a referral link for you right here. Shout out to fold for sponsoring my work and honestly being the most important service for my being on a bitcoin standard. [01:01:04] Speaker C: So what are some of the considerations regarding privacy and anonymity when using Keat? [01:01:10] Speaker D: Great question. [01:01:11] Speaker A: Yeah, I mean I think anonymity is kind of a separate. Very separate from privacy. So the technology of using KI is extremely, extremely private. All your chats, all your communication encrypted, encrypted at rest, encrypted in transit. [01:01:24] Speaker E: There's no option for it not to be. [01:01:26] Speaker A: Yes. Yeah that's. [01:01:27] Speaker D: And also like. And also I think people because again it's technology. It's hard to understand messages you send Anki don't have to come from the person who's sending it. Anybody can relay it which creates this layer of like you can actually go anywhere, write some messages, turn off your network and I can still read them later. So you can actually be insanely private if you want to than just like default privacy. So it's really, really important for us to have that bait in. You should always, you know, if you're. There's always a level of like paranoia. I guess it's the wrong. I don't know what's the positive word for paranoia? I don't know, like just like security sensitive. Yeah. [01:02:03] Speaker C: Careful, careful. [01:02:04] Speaker D: No, just more like, you know, like I think it's always, it's always. It's a hard discussion. Right. Because it's kind of like I'm in Switzerland, people know I'm in Switzerland. I don't need to hide that. [01:02:13] Speaker C: Yeah. [01:02:13] Speaker D: Myself. But a lot of people definitely need to hide it because of other reasons. That's what I'm trying to say, you know. So if you care a lot about that kind of stuff, you should still always use a vpn I think. You know, it's always a good, good thing. [01:02:24] Speaker A: That's the anonymity side of it. [01:02:25] Speaker D: Yeah. [01:02:26] Speaker A: About anonymity. [01:02:26] Speaker D: Yeah. Defense and depth is always a good idea. Like just use. Use all the tools to stay. They actually very much stack together. We do all the stuff we can do all the time to make everything as private as possible. Like and assist the technology is insanely proud of it. Adding more stuff on top doesn't we're. [01:02:42] Speaker E: Working on a php, Right. [01:02:44] Speaker D: I don't know if we should talk about that. [01:02:48] Speaker C: Allegedly. Allegedly. [01:02:52] Speaker D: But yeah, sure. But I think it's like. Yeah, like Andrew said, it's actually, it's, it's, it's interesting. It's one of. Again, as a non English speaker, it's hard for me because two words sound the same, but they mean very different things. [01:03:07] Speaker A: But I mean, I mean to get on the technical side of it, you know, when you, when you're in a key room with people, you're. You're sharing data with them and. Very similar to how people is done. To how it's done in BitTorrent. [01:03:17] Speaker C: Yeah. [01:03:18] Speaker A: And in BitTorrent, you know, once you actually establish a peer to your connection, you have that IP of other people who are in the rooms. [01:03:22] Speaker C: Yeah. So if you really, really care about. [01:03:23] Speaker A: IP secrecy completely, that's not super important to you. Other tools would be needed. You'd have to layer that on yourself. But that's stuff that we can expect. [01:03:32] Speaker C: Explore more also in the future. Yeah. And I think it's important to, to distinguish. I mean I think the cypherpunks really explained it better than anybody. And I think it's Timothy May who talked about like anonymity is about secrecy. It's about like not wanting anyone to know something and who you are. Whereas privacy is simply the choice. You, you having the power to selectively reveal yourself to the world. [01:03:58] Speaker D: Yeah. [01:03:59] Speaker E: And that should just be a default. [01:04:00] Speaker C: And that should be the default. Exactly. [01:04:02] Speaker E: That's just basic human rights. [01:04:04] Speaker C: Yeah, yeah. So anything. Anonymity is a question of, okay, well, do you want to take. Go to the network level and do whatever you need to do to do something of a tour or whatnot. But otherwise private is the default. You are like when we sit here and we're having a conversation, nobody's listening. We're not sending the conversation to some server somewhere else and then routing it back to Matthias. And that's true in Keat. Like when we're chatting, I have a connection directly to him. Obviously he can hear my half of the conversation and vice versa. And we are encrypted. But like it's between us. It's between us. It's between us. [01:04:41] Speaker E: We lived in a world where microphones were in every room recording everything that we said. That would feel very uncomfortable. But digitally that kind of. [01:04:51] Speaker C: Wink, wink. There's the tin phones in here. [01:04:55] Speaker D: I also find, you know, again, just going back to how we use keep today because at least to all these fun conversations internally where we like, because people had to, you know, because people used to automated names and they had the default avatar. So we have all these conversations, always internally. Like the dog said this. The dog said this. Yeah, like the bear bog or the penguin bar. [01:05:17] Speaker C: I was about to say the emperor penguin. I was like, you gotta keep that one. [01:05:21] Speaker D: And just the fact that we do that just to me tells you everything. It's kind of. We have no idea. [01:05:26] Speaker A: We have no idea. [01:05:27] Speaker D: We make the app. We have no idea other than like. No, we talk. Maybe we talk in the same time zone. [01:05:32] Speaker C: Yeah. [01:05:33] Speaker D: But it's like, you know, that's what you know. But it's like we don't even know. And the app doesn't require you to put in a phone number. I find it very interesting how quickly apps, even privacy apps, ask you to do that argument. [01:05:43] Speaker C: Oh my God. It's always a phone number. [01:05:45] Speaker D: We do nothing. Right. We know nothing. It's just a key. We just a key. The key is just to last degree, a random bunch of strings and that's really, really powerful. And I totally understand why other apps don't do it because it's really hard. And also people want to have that list of users. People often ask us how many users we have and you're like, it's anywhere between one to whatever the next key is. Stars in the universe. We have an idea obviously based on activity, but that's the same idea you have. We don't know more than you do. And that's the thing that's actually the point of this thing. Like we don't even know if we knew. It was not very private. [01:06:25] Speaker C: Yeah. Okay, so question is, what are your plans around identity management and web of trust and are you looking at what pubkey is doing and. Or nostr? [01:06:35] Speaker D: I can just add a little thing to the first part because it's actually something we're talking about a lot with. Also with Wallet, which we talked about earlier. I actually think entities work amazingly. They're very seamless. Andrew made all that code. It's amazing. Andrew still has hair. It's going to be my current thing. [01:06:52] Speaker A: It's getting great. [01:06:54] Speaker C: Give it a couple of years, but. [01:06:56] Speaker D: It'S like it's one of those huge unlocks where I use Keat across multiple devices and I'm the same guy and I know how hard it is. [01:07:04] Speaker C: Yeah. [01:07:04] Speaker D: Just seeing the few works. [01:07:05] Speaker C: Yeah. [01:07:06] Speaker D: All the time. I think the weakest part of that in Keith, just to be real, is actually the setup. There's a setup phase here where you. Because we Want people to store your seats and stuff like that. That's too complicated just to be real. And we talked about that also a lot and that's something we're going to be streamlining, having more options where the app can hold you your seat for a bit so you don't have to. You're not forced to do that step immediately. I think that's way too hard right now. And that's the stuff we want to do and that's luckily a very easy thing. But it's easy technology but hugely important for ux. So if we ever want to onboard the next level of people like our families and stuff, that just has to be non intense and it is a bit right now. So that's really, really important for us. And that's something we're going to be tweaking also with Wallet. So it can be much, much better in terms of looking what other protocols are doing. And with pubg. So we're always looking at everything because it's shared efforts. We're actually using five IDs right now. We're using. I can't remember one of the bips from Bitcoin 85, 39. I don't want to, I don't know, I don't want to say the wrong number. [01:08:17] Speaker C: Somebody like what bip85 is. I think it's 85 is the one with the derivation key. [01:08:22] Speaker D: Yeah, we're using. So we're just like. Again like that's great. Yeah, reviewed Sometimes we're all for using all the stuff we've been choosing. Why not? So we're always looking at everything. I'm not never familiar with all the details but we're looking at everything. So we're always looking at that. So that's basically like dialogue on that I think. [01:08:44] Speaker A: But in terms of adding higher level things on top of this for Web of Trusting, that's very interesting. I think it's the kind of thing that probably not super short term for us to be hitting that now that we have like a really nice identity experience, we're going to mainly at least in the near term future be focusing just on the core room experience. Making sure that works, making sure you can recover your IDs and make IDs work really nicely, but IDs are definitely a really, really strong primitive for doing Web of Trust stuff. [01:09:06] Speaker E: Later between apps as well, you build another app and then you derive identity from your key to identity. And now we've got that all. [01:09:13] Speaker D: It's actually funny because these guys saying, because Android is like Saying web of trust. Right. Weber Trust to me is also just a contact book during key and literally just I'm going to add Guy. [01:09:25] Speaker A: I mean it's adding the contact and saying, hey, these other contacts know this person. That's really all it is. [01:09:30] Speaker D: And then it's opening up. [01:09:32] Speaker C: I was about to say it's open source. When are you going to build it? [01:09:35] Speaker D: And we're actually going to release a ton of our ID stuff next week actually. [01:09:40] Speaker C: Just open source. [01:09:41] Speaker D: Open source, yeah. Our parent stuff also because we reared it and we're happy with it. It's been working really well. [01:09:48] Speaker C: Nice. [01:09:49] Speaker D: We want more people to build on this stuff. Like I would love to see more work also in the community on web across stuff because I think actually one of those things where we don't know if we're still fully settled on how that should work at a very technological level and we'd love to see some iterations on that. Yeah. And like David is saying, because David works on platform mostly super interesting for us. They also like, you know, you should be able. I should, you should send me an ad and I should say Guy from your contact book made this app. [01:10:19] Speaker C: Yeah, yeah. [01:10:20] Speaker D: And then I would immediately. [01:10:20] Speaker C: It's just my little signature. Like, like my little, little. Do you trust guys in pub or you know, whatever your thing is. Yeah, yeah. [01:10:27] Speaker E: Then you can maybe use any of guys. [01:10:29] Speaker C: Yeah. [01:10:29] Speaker D: And we just, we actually added this to platform already. We just haven't shipped it because we're waiting to. Oh, that's cool. And if you think about it, right, that's a bajillion times better than what they make you do now. Right now to do that. Apple makes you pay $100 to get a, to get a key. [01:10:45] Speaker C: Yeah. [01:10:45] Speaker D: Windows makes you pay a thousand dollars. [01:10:46] Speaker C: To get a key as a certificate authority. [01:10:48] Speaker D: And then the certificate says random entity that Guy runs sign it. And like that means less to me. Yeah, it means less to me than like Guy actually land is on his laptop and it's Guy from key room. It's not like compromised on something. So that's, that's. I think that's actually also the next level of trust for us. Rolling that on. We, we have all that. We're just sitting on it. [01:11:09] Speaker E: Very exciting. [01:11:10] Speaker C: Yeah, sweet. That's awesome. And one element of all of that is I found in like what we're trying to do with a pair drive is that like all of it is agnostic. Like if you have a way to communicate or just attach data to a profile, like my key identity can Also just have my end pub right there. And then if you want to load the nostr ID and that data, you can just ask the like ask the end pub data and it can literally be the exact same thing. Like, so it could just be a pub key blob that we're passing around that just has the whole, the whole stack of. It's basically the. Oh, I posted on Twitter to prove that I'm. This account over here, it's like. Well, it's agnostic. Like, you know, you've just signed something so you can just carry it with you everywhere you go. [01:11:57] Speaker D: We should probably add that, huh? Like a little blob you can attach to the id. [01:12:00] Speaker C: Yeah, yeah, that's done. Yeah. If you want to build it for me, that would be great. Yeah. [01:12:06] Speaker E: Focus on the web. [01:12:07] Speaker C: Yeah. [01:12:08] Speaker D: Also, Andrew said, how does easy as hard as easy. [01:12:13] Speaker C: So the question is what features are. [01:12:15] Speaker E: On the roadmap of Keats or pair? [01:12:18] Speaker C: All of it. Just give me the whole list after this. [01:12:21] Speaker D: I'm gonna go take a nap. Seriously. Like obviously dms, we wanted to ship at DMS a little bit earlier, but we wanted to iterate a little bit more. So we're shipping that. [01:12:33] Speaker A: Yeah. And in the DM's release, we're also going to have a whole bunch of new moderation tools also. So you can kick people and ban people and if it's somebody spamming, boom. [01:12:40] Speaker C: That's cool. [01:12:41] Speaker D: The DMS adds like. [01:12:42] Speaker C: Yeah, that's an important thing. Just because of the nature of peer to peer. [01:12:45] Speaker D: Yeah. Like you can lock your room, but it can come in. DMS also adds a capability where you can actually add people without sending invites of how DM works. So like, you know, if I know you're in my contact book, I can add you to my. [01:12:59] Speaker C: Just go ahead and sign for me. [01:13:00] Speaker D: Yeah, like just be like I'm any guy. And that's. I think actually that's one of those things. [01:13:04] Speaker C: Oh, that's cool. A push rather than a pull kind of thing. [01:13:07] Speaker D: Yeah, we're missing tons of UI on that. It's going to roll out in iterations, but it's actually one of those things that's so much bigger than I realize. I think because right now every kid experience starts with a text message. Like I send you a link. [01:13:22] Speaker C: So I'm telegram. Yeah. [01:13:24] Speaker D: Some other service. [01:13:25] Speaker C: It's out of band. It has to be out of band. [01:13:27] Speaker D: Yeah. With this, the entire experience can exist in the pew pew world. So actually like my text message, we're gonna, gonna go cold right? [01:13:38] Speaker C: Yeah, you can't tell you how many times I've sent my wife for April text message check. [01:13:44] Speaker D: Here's the string of words. And I think actually that's really, really important. And also just the privacy. Also just for leaking it, controlling the experience. So stuff like that is rolling out. Super important. We're rolling out, talk about that. Also media mirroring. So we took, we completely take mirroring. It's actually car, same problem. Just one is a little bit bigger than the other one. Yeah, works flawlessly most of the time. Like very few bugs. Media mirroring is just the same but with images and stuff. Once we roll that out, people are going to be like, oh, this is how I expect it to work. It's actually really, really insane if you think about it. It's just fully peer to peer. [01:14:26] Speaker C: No, I still have like as someone who used peer to peer for like literally like since 2000 and 2000 or 97 probably. I don't even know I was young. But who's used it the whole time, like every once in a while just like open up ketone and shoot, like. [01:14:41] Speaker A: How the fuck does this work? [01:14:43] Speaker C: This just does not make sense, you know? Yeah, it just seems undoable. Especially when you get to the point where like it's always been the user friendly last mile. It's always been the user friendly last mile that was never there. The work had never been put in. [01:14:59] Speaker D: But yeah, totally, totally agree. And just to add the last thing on that, obviously then I don't know if this is a feature, but we're really pushing a huge performance upgrade the next month. [01:15:08] Speaker A: Just there's a huge amount of focus going on. [01:15:10] Speaker C: Just like core experience, user experience, feature number four better. [01:15:14] Speaker B: Yeah. [01:15:16] Speaker D: And it sounds very man. Always when I talk to marketing guys, they're like, how do I explain them? [01:15:21] Speaker C: I don't know, just say it's better. [01:15:23] Speaker D: It's just so important. Like the experience just has to be so flawless. And I get asked this a lot where people like, are you worried about like they were being, you know, catching up with the performance on Slack? I'm like, you don't realize all our stuff is local. [01:15:41] Speaker C: It can definitely catch up. [01:15:43] Speaker D: Well, it's kind of, it's gonna surpass it by a magnitude. Yeah, because like local is just really. [01:15:47] Speaker C: Really fast, really smart right now. Just so you're focused. You spent so long focused on the network stuff. The local stuff is just been ignored. Yeah, but local is really, I mean how fast can you read off your. [01:15:59] Speaker D: Hard drive, you know, you know, not to put too Much pressure on it. But I also think it's going to be a little bit of a concern because I don't think people are used to just writing things and they appearing on the screen and stuff. Like, we're very big into latency, like for it to feel like it went through. [01:16:13] Speaker E: We have actually stopped optimizing at certain points because we're like, if we go further with this, people are going to think it's not real and they're not going to trust it. Like, we got the distributable size down to like a very small size and we're like, people are just going to think it's a bug. [01:16:28] Speaker D: Yeah, exactly. The downloader is so true. Like the downloader you get on the website. We had a discussion about this because that could be like two megabytes. And then we're like, if you download a two megabyte file, some people are going to be like, that's bad. Because it doesn't feel like an app. [01:16:44] Speaker C: It's too small. [01:16:45] Speaker D: It's like it's malware. So we made it 20 megabytes and add a bunch of cash to it. Like actually, like actually useful data. [01:16:52] Speaker C: Yeah, yeah. [01:16:53] Speaker D: And it's like. And it feels more natural. [01:16:54] Speaker C: Yeah. [01:16:55] Speaker D: But it's like maybe, you know, over time we can do more of that stuff. But it's so interesting, so natural and sometimes we're just so wired. [01:17:02] Speaker C: That's crazy. That's crazy. You know, there's. There's one thing just on the nature of how fast and responsive, like when you, when you have that connection. So, like in one of the rough drafts, the architecture wasn't at all right, but we were just kind of like proving that we could do this. Had a version of a Pear drive that I did keep to Linux and I have my list of files and then on the Linux machine I booted up so I can see the list of files. And then I was just adding and deleting and I joined from both things. So I had a chat thing. I could see both desktops right next to each other. And I shit you not. Like, I could not, I couldn't. Like my brain could not calculate the delay between deleting the file from the list on one computer and the video updating, which was not just the list updating, it was the list update being sent to the Linux machine, which then clears the list and then sends it back over Keat with the video frame back to my main machine to clear it. And I was just like, delete, delete, delete, delete delete. And it's just like the both lists are like, like right in line with each other. I was just like, oh my God, this is going to be so cool when we're done with this. [01:18:15] Speaker D: Latency is crazy. We had one also where we were chatting to the office. I remember in the office latency, obviously we sit close to each other for sure. I had one time where I was like, it felt like the message was sending before I was sending it. That's because when you click the enter button, I did registers before it goes all the way down. [01:18:31] Speaker C: Yeah. [01:18:31] Speaker E: Yeah. [01:18:32] Speaker D: That latency for me to push it down. Arm was longer than. [01:18:35] Speaker C: Yeah, it was before the sound. It was exactly. It's like when the sound. You see the thing and then the sound hits you and your brain makes that uncanny. [01:18:43] Speaker D: That kind of experience you normally only get in a test environment where you're like, because we're testing a local survey. This is just because the connection was as good as a test connection. That's crazy. [01:18:52] Speaker C: That's nuts. The question is. So he was. He was very like, they should have been doing TypeScript. We had this long conversation about it. They should have been doing TypeScript. So many young developers are Typescript only. And he was like, I wish they just wouldn't do JavaScript. And then we had the conversation not too long ago that he was like, okay, I think. I think they made the right choice. Like he has shifted away. So the question really is, are you going to be moving to a TypeScript compatibility so that people can build TypeScript within the environment, or is it going to be a situation where you expect developers to come to the same realization that Hope did that this should just be built in java. [01:19:31] Speaker E: So the JavaScript. So the thing is compilation. TypeScript requires compilation. That's a step that takes time and that time affects your iteration speed. And the longer that takes, the more it affects it. Because then you start going, oh, just check my emails. And then you get distracted and you do something else. It's a productivity issue for. That's one of the reasons. But in terms of what's next, we have something that's already out in Pair, but it's not very well like communicated about currently. But it's called Pair interface. Currently that's just JS doc. But we're going to move that to TypeScript so you can suite so that just declares the pair global and we'll declare the pair global in there as well. But another cool thing of a feature of that is that when you. When Pair updates on Your system, it'll update your interfaces in your. That you're using, like in TypeScript. [01:20:24] Speaker C: Okay, cool. [01:20:25] Speaker E: We're going to, we don't want to alienate a TypeScript audience. [01:20:29] Speaker C: Yeah. [01:20:29] Speaker E: And like, there's nothing wrong with using TypeScript for certain situations. We're very small module focused and small service focused on that kind of stuff. That's the way we like to do things. If you're, if what you're building is a massive monolith, which some applications might be, then types can be a decision that makes sense there. If the language had types, it wouldn't be an issue. Right. But like. Yeah, so we're gonna, we're gonna support, we're gonna support TypeScript in terms of users and making sure that like, you know, like we can add DTS files to our modules and all that stuff, but we wanna decouple it from our own development process and that's why we didn't address it until now. [01:21:08] Speaker C: Okay, sweet. [01:21:09] Speaker D: It's also as soon as we add that, you know, when I change anything else. [01:21:12] Speaker C: Yeah, yeah, yeah. It has to be after, you know, you're not going to break it. Yeah. [01:21:17] Speaker E: API to remove. Before we can even take the API as small as possible and crunch it down and when we know, okay, this is where it's at, we're not going to make any changes to these APIs then we want to keep, get those files out. [01:21:28] Speaker D: We had one module that didn't have any documentation and one of the guys and the pair channel. I don't know if it was Hope. Maybe somebody else wore a very nice PR sending documentation. [01:21:39] Speaker C: Hope is a documentation like freak. He loves Doc. There's probably hope. [01:21:44] Speaker D: One of our guys is saying, you don't realize it's undocumented for a reason. Don't use it in a friendly way. But that's kind of like, that's kind of like we don't want to send the wrong signal. Like we're still in beta. We're iterating really hard. Like things don't change much now. [01:21:59] Speaker C: Yeah. [01:22:00] Speaker D: But as soon as things are not going to change at all and we're very close to that point, that's when. [01:22:06] Speaker C: You'Ll see things like that. [01:22:07] Speaker D: And like we're all for. And like David says, you know, we're, we're all farts. You know, we have our own way of doing things. Iteration speed is really important for us. People should do what they feel is best. Yeah, always. [01:22:17] Speaker E: And there's no point in, you know, we want to support more and more people to build on top of this. Of course, in the future this goes beyond JavaScript and TypeScript as well. [01:22:25] Speaker C: And to be likely implementation. [01:22:27] Speaker D: Yeah, like completely different code basis. Already did a ton of stuff there and all disabled parts. The rest will come. [01:22:34] Speaker C: Sweet. Sweet. Okay, so the question is, is there a pair phone on the roadmap? [01:22:42] Speaker A: Well, can I say first, I think dedicated hardware for peer to peer that's designed for peer to peer is very interesting. There are a lot of things that you could do. There are a lot of limitations on mobile phones today and even laptops that are just not optimal. [01:22:56] Speaker D: We've been so frustrated about the fact that we all run around with our computers in the pocket and they're turned off. Yeah, but they're not even turned off because every time you get a notification they turn on and we're like, so they are turned on. So why can't we just have them be peer to peer friendly all the time? Just get a little signal, get our messages. It's the ultimate peer to peer world. [01:23:15] Speaker C: Yeah. They're just built for background. Like built for that type of environment rather than. Except expecting a centralized entity to be there. [01:23:24] Speaker D: It's really easy. But the hardware just has. There's a bit somewhere that's like you can't do that. So we, so we almost, from the beginning we've been talking about we need to do that, we want to do that. We are doing that. So we're gonna. We are looking into that right now. We don't have a timeline for it. [01:23:38] Speaker C: Sure. [01:23:39] Speaker E: In 2025, two weeks. [01:23:43] Speaker C: No, just shut up everybody. [01:23:47] Speaker D: So but it's like. But obviously that's the future we want to get to. [01:23:50] Speaker C: Yeah. [01:23:52] Speaker D: And we're at Terrain right now. We're going to talk more about it, but for sure it's like Angry says it's like the easiest idea ever. [01:23:59] Speaker C: Yeah. Besides Keat as a chat app, what are other applications you would like to see? [01:24:07] Speaker A: Really smooth, seamless file sharing apps. Really, really cool. [01:24:10] Speaker C: Hey, hey. I know a guy. [01:24:15] Speaker D: That'S so true. That integrates with Kit also. It's also like where we can have file apps, browse them, stickers, collections, things like that. [01:24:27] Speaker A: I'm really interested in seeing the room model that Keat has generalized to different kinds of social applications, different kinds of rooms that are specialized to different things within KYT and also as separate apps. [01:24:36] Speaker C: Would be really cool. [01:24:37] Speaker A: So like media sharing rooms that kind of look like a Netflix ui, either standalone or in Keat. [01:24:42] Speaker C: That'd be awesome. [01:24:44] Speaker E: I'd like to see that. [01:24:44] Speaker C: May or may not be part of the reason that we're doing something. [01:24:50] Speaker E: But yeah, we don't even know. I mean there's a lot of apps that could be like replaced with peer to peer technology. And something that's frustrated me about apps that win marketplaces with good features and then unfeaturing those. [01:25:04] Speaker C: Oh my God, it dude. [01:25:06] Speaker E: Worse for users. [01:25:08] Speaker C: The inshification of the Internet is exactly right. [01:25:11] Speaker E: I'd really like to see a whole load of apps that kind of disrupt what's going on. [01:25:17] Speaker C: Unshittify. [01:25:20] Speaker D: Is that biased? [01:25:23] Speaker E: But also, I don't know. I think for me one of the apps I'd like to see is something I've not thought of as in I want to be surprised by something that can't be done in a centralized way. And it's like, oh wow. And then I mean just really just generally just things that, that interact with the real world, with what's happening in real life. You know, whatever, whatever that is. [01:25:47] Speaker D: I think there's a lot of scary apps like that. Scary in the way of like disruption. You know, I think the first, you know, when there's an ad out there that's just. And this could be a feature in keep, like the live streaming app where you can just go around filming on. [01:25:57] Speaker C: Yeah. [01:25:58] Speaker D: Go to football games. Somebody's gonna shit their pants. High quality. [01:26:04] Speaker C: Yeah. [01:26:04] Speaker A: It's a double edged sword. [01:26:06] Speaker D: It really is. But I think it's also very exciting. Super exciting also. I also think just with the initiative, like there's a lot of social apps out there, you know, dating apps for example. I think it's interesting because they all, that's, they're just all ruined by the wrong incentives. Yeah, yeah, let's match people. But we only make people, we only make money if people aren't matched. So that's not made up that good. Peer to peer takes that out. So you know, people can make good apps like that. [01:26:31] Speaker C: Yeah. [01:26:32] Speaker D: And that's a whole range of categories like that. [01:26:33] Speaker C: Yeah. [01:26:34] Speaker E: So. [01:26:34] Speaker C: And from a personal perspective, I'm, I'm a huge wholesale junkie which is, you know, built on top of the pair stuff and the ability to just run services. Like I, you know, one of the things before I leave, I always run a bunch of services that I use on my Linux machine. Then set up a key and then I can just throw it in my phone and I can access it. I can do my AI stuff, get my transcript, like those sorts of things. Like just having that direct connection so easy is just crazy. And then something that I saw I think this was in like the pair babies room, the one where like everybody just talks about ideas or pair development. I'm not sure. But this one was like so. Oh snap. I never even thought about. That was a find your friends app. One that's just literally share your location. Because like I use you know, find my. With my wife and sister and you know, like when we're trying to be like okay, how far away are they or whatnot. But I hate it because you know, I'm. This is all going to Apple and then going to them and I hate that I have to use that service and I'm constantly going back into settings, turning off location, going back, turning it back on or whatever because I, I want it to be useful when I want it and I don't want it to be tracking me otherwise. But it definitely is if I forget and just the ability to just have a peer to peer where I'm literally just into our little network just sending a ping of my location. I know the only people who can see that are those people. And. Yeah, and you could, you could do that like in a museum. It's like how the hell did say just. They just left. Bring up your app. He's like, oh, he's looking at the dinosaur again, of course. So question is, are there any plans for grants or investments in other peer to peer apps? [01:28:21] Speaker D: Totally. We already do a little bit of both actually. We started out a small grant program that it's very manual right now just because we see a lot of people building really cool stuff and we want to help support that. So if you have good ideas for ways to improve the ecosystem and things like that, please reach out to us in our channels. Like I said, it's very personal. Right. You could hear people talking. We're like, yeah, let's figure it out. We got a bunch of good documentation guides out of that already and a few example apps and stuff. So that's really exciting for us. We're going to do more of that stuff going forward for investments. Yeah, we want to see this thing go freaking mainstream. I take over everything and that's going to require money, like to be real, like people, I guess people have to eat. And so we're, we're working really hard to also set up, you know, framework for ad investment vehicles that, for people to make apps and get support like that. And that's something we're very, very interested in. Just because we want to see this succeed. Yeah, we can't build all of it. We don't want to Build all of. [01:29:29] Speaker C: Already just a candidate. Like, I just don't want to until I can't get that time. [01:29:35] Speaker D: So always reach out to us. We're in all the channels. We're gonna be more stuff about this as we grow and it's gonna be more structured. [01:29:44] Speaker C: And that's also the big unlock too is you talk about like you want to see the app that you don't expect. I want to see somebody bring something to you that you're like, oh, that in hindsight is obviously. And it's just going to be some unique perspective or unique situation or some 14 year old who's just like, oh yeah, it could all just be peer to peer who's like come up with some idea. [01:30:07] Speaker E: We have to be moving from one frame to the other. So we're like, we have to like. [01:30:11] Speaker C: We still have to remind there's unlearning that we have to do. And there are some people who won't in those paradigms. [01:30:16] Speaker E: They're in different kind of scary. [01:30:19] Speaker C: Good scary. Really? Yeah. Yeah. [01:30:24] Speaker D: So many English words. [01:30:26] Speaker C: Awesome. Okay. Is there a plan to integrate a machine learning. I'm still bugged by the AI. Even as someone who has a podcast, AI Unchained. Is there a plan for AI machine learning in Keat? [01:30:41] Speaker D: Yeah, no, but it's actually one of those cool things where actually almost immediately when we launched Keaton were talking about. Because Kit is really fun because it's very international already. There's people from all over the place, China, Taiwan, you name it. Right. And as an old Star Trek fan, you know, the universal communicator. It's always interesting to me, we're like, immediately like, we need to be able to talk to people in their own languages. [01:31:12] Speaker C: Translation. [01:31:13] Speaker A: We did translations from the very beginning. We did tons of translations, dude. [01:31:17] Speaker C: Translations are wild. Like, LLMs are so much better than the traditional. It's crazy. Like, and it's funny, I actually think about LLMs as like, they are a universal computer translator. Like, that's really what it is. It's not AI, it's literal translation. It's translation from English to Python. It's translation from English to Portuguese. You know, like, it's a model of contextual patterns that can translate between any type of computer data. Yeah. [01:31:45] Speaker A: And I mean the real challenge for us with peer to peer, it's like, you know, we have. We need it to run on the end, on the device. All on device. [01:31:51] Speaker C: Exactly. [01:31:51] Speaker A: Real constraint. [01:31:52] Speaker C: Yeah, yeah. [01:31:53] Speaker D: Most of these models, I think. Yeah. Because it's kind of like it's A cool feature. Not on device. Very dystopian. [01:31:58] Speaker C: Not on device. It's awful. Yeah. [01:32:00] Speaker D: And so, yeah, like we immediately was like, we need to do this on device. Huge constraint set up. A team at some point realized this can be very generic and Tether took a huge initiative on that and they're making Tether AI now just that's like, you know, local first AI. And we're going to be deploying a ton of that stuff in key soon. Like translations is the first thing you want to do is almost there. So. And it's AI that can run anywhere because it's running on the engine. So it can run, you know, on your watch, on your phone, on the desktop. Same model. It's just all the same JavaScript script and these models. So that's super exciting. And I think translations is very easy to understand, but there's so many more things once this starts. [01:32:43] Speaker C: You know, that's just kind of the naive first thing I always say. [01:32:48] Speaker D: Centralized AI, I want to feed zero amounts of my data. It already has a lot of my data, like resource data. I want to give it nothing else because it's really scary to me. Local AI, I want to give everything. Yeah, I want to give it everything. [01:32:59] Speaker C: It's going to be better the more I give it. Like the local. It's my AI. It's totally. [01:33:05] Speaker D: Yeah, so. So it's really, really important. Just for me personally. [01:33:08] Speaker C: Yeah. [01:33:08] Speaker D: That's also why I'm like always talk pushing about this internally. Because I really want it. [01:33:11] Speaker C: Yeah, yeah. [01:33:12] Speaker D: I want to watch movies, can find the best movies, etc. Etc. So that's, that's definitely coming here. [01:33:17] Speaker C: Sweet. Okay, so are there any proof of concept apps for things that are coming? [01:33:23] Speaker D: Yeah, so actually this one of those things where like Andrew talked about earlier, we make everything generalized. Collaborative app. Generalized. We actually had an internal process earlier where we're like, the best way to explain these things is actually just make an example app. Because I hate reading. I hate reading about code. [01:33:40] Speaker C: Yeah. [01:33:40] Speaker D: I just want to run it and be like, how does this work? And how that works is just like 100 lines of code. I understand immediately rather than reading about it. I'm not a very big reader. [01:33:49] Speaker C: Sometimes I can help. I can help if you need. [01:33:54] Speaker D: So we're like, let's make some example apps. And then the problem is as soon as we start making example apps, we realize, wow, this is actually just a really good app. So right now we're making like a password manager app. As a test, we actually had a one the guys who applied who made a password manager app also by himself. I think it's one of those cool concepts where it's not that hard when you have technology super useful like no sharing, very private, sensitive. That's great. [01:34:20] Speaker E: Will never close down. [01:34:21] Speaker D: Will never close down and also access a good example which is kind of scary because like how can something that complicated be an example app? [01:34:28] Speaker C: No honeypot. [01:34:29] Speaker D: Yeah. [01:34:30] Speaker C: As someone who's now has to. Has to change everything again but recently just had to do a run through for like a whole week because LastPass got hacked like God. A local. Local freakin between my devices. Password manager is. [01:34:47] Speaker E: It's amazing how much hassle there is with a central point of failure. [01:34:50] Speaker C: Yeah like enormous. And there's no reason too like it's just because like the payoff of being able to hack all of LastPass and then be able to just like take your time and seeing who's got a weak enough password in this massive database. And it's like this is. It's so antithetical to the idea of security. [01:35:13] Speaker A: Right. [01:35:14] Speaker C: You know so yeah that's a huge. [01:35:16] Speaker D: One and that's our example. I think that's actually part of crazy. [01:35:19] Speaker C: Yes. [01:35:20] Speaker D: And that's directly applicable also into to do apps. So you know we're going to make some example apps on that and also those are useful and things like that and actually just having a little framework for that. So anybody can just. If you want to make an app like that we can just start with this thing. No cheating. It's just the data structures but just like underlines the proof up and running. Super cool. [01:35:40] Speaker C: Hell yeah. Okay. So how I guess resilient or whatever is peer to peer that if I can't remember if a government or wi fi decides that, you know. Yeah. ISPs decide that they don't want to allow peer to peer traffic. Well what's. [01:35:59] Speaker A: Well I mean ISPs right now, you know it's very asymmetric the upload download. They're not really used to. To people having huge amounts of upload which is really useful for peer to peer. [01:36:08] Speaker C: Yeah. [01:36:09] Speaker A: So I think it's possible in the future it'll definitely get their attention. Whether that gets cracked down on I think it's unclear. Maybe in certain areas it could be problematic but I don't anticipate any massive issues. [01:36:21] Speaker D: So it's a technological discussion and then also not like Andrew was saying because it's like the technology itself just. It's just encrypted traffic. It's not like you can just be like block PGP traffic. You can block a ton of traffic, including PGP traffic, but you have to apply a very, very big net that's going to be very disruptive for people. [01:36:40] Speaker A: Yeah, that'd be a really future. I think there's a massive, massive scale crackdown on that. Then other things have gone wrong. [01:36:46] Speaker C: It's hard to be very targeted about exactly what kind because I'm just connecting to an IP address and sharing encrypted information. So if that becomes difficult for me, it's a question of what have they targeted because they can only paint very, very broad strokes. And that's a, that's just a, that's a place that I probably will just need to leave just to get out of it. [01:37:09] Speaker D: Just to refer back to someone earlier because I think this is also about getting serious because part of when we were just working on this as open source, like small things like, just like hackers getting small funding, I think then it's a real concern. Yeah, that's why we're like, we don't want to do that. We want to make something where this becomes like a behemoth that's uninstallable also this sense, you know, so we have to deploy new infrastructure. We'll talk about that and make a plan for that. This technology has to be installable in every way and it's not going to be tomorrow or somebody's going to try to block it. [01:37:41] Speaker A: But I think, you know, maybe our goal, we can say it's to get to that point where it actually does become a talking point. [01:37:46] Speaker C: Like it's something to discuss. [01:37:47] Speaker A: You know, if there's a world with millions and millions of people uploading tons of data to each other, self sustaining these networks, it'll get people's attention and like that'll be something to deal with that. [01:37:55] Speaker D: And also just the way the world is moving right now, I'll be more scared on as a person on the immediate timeline. Just they're going to block much more centralized things very fast. Yeah, like more parts of the Internet. I think we'll probably run into a future soon where you know, we already had that but like certain countries can connect to certain countries and things like that which is by just literally something across the cables. [01:38:17] Speaker C: Yeah. [01:38:18] Speaker D: So it's a general problem, it's a general concern. But peer to peer is definitely the way to also fix this because like. [01:38:26] Speaker E: And to work around it if we need to. [01:38:28] Speaker D: Yeah. And it doesn't, it's peer to peer because inherently it doesn't care where the traffic comes from. [01:38:32] Speaker C: Right. [01:38:33] Speaker D: You can send it to a satellite and beam it down, it's fine, doesn't matter. You can, you can beam an outgrade radio, it's fine. [01:38:40] Speaker C: It's about a connection, not any particular way. [01:38:42] Speaker D: Exactly. And I think actually that's actually the huge unlock. That's where it just becomes so hard. And Having worked with BitTorrent in BitTorrent's heyday, I just remember, I always tell this, so many resources went into blocking that because that was actually where I was at that scale. And it was so hard. [01:38:59] Speaker E: It was so hard. [01:39:01] Speaker D: And they do all kinds of stuff to block it. [01:39:02] Speaker E: Right. [01:39:03] Speaker D: And that's all the stuff we work around today. And so now we have solutions for all that stuff. And I'm sure that's going to be. It's like it's warfare, it's going a bit back and forth. But they could, you know, they couldn't. So it's kind of the same story, I think. [01:39:16] Speaker C: Yeah. And I think it's a, it's a good point to just remember that centralized is going to, they're going to block the centralized first. Like, like inevitably it's an up. It's going to be harder to do the peer to peer than it is every other avenue that we have. [01:39:32] Speaker D: But also people don't even realize how east is the block centralized. People are like, the thing is like, oh, we got to call Facebook and be like, no, they're just, they go to that DNS record. Happens all the time. [01:39:43] Speaker C: Yeah. [01:39:44] Speaker D: So it's scary. [01:39:46] Speaker C: Yeah. [01:39:46] Speaker D: There's probably just one guy who can do it who works at that office. So. [01:39:50] Speaker C: Yeah, yeah, yeah, sure. I can give you I guess probably a general overview. Obviously Hope would be much better informed about this specifics but we've had to kind of go back to the drawing board a couple of different times. But the basic idea is built around that problem between getting from my MacBook to my Linux. And then also the. One of the really cool things about Pair is that to get files or to be able to build this for all platforms at once. Because one of the things that gets me and every time I've tried to use nextcloud or the long list of just frustrating and not user friendly options is, you know, we're out here in the conference, you know, I pull out my phone and I take a picture of. I was like, well we'll do that as soon as this ever and I want to be able to put this in my Pair drive. And then the second it gets a connection to any of my other devices, it uploads like it just goes there like in my Linux machine is at home waiting for a connection and it just uploads to that computer. And if I lose my phone, but it's been 20 minutes and I've been connected to the Internet since I took the picture, it's there, I didn't lose anything. And the number of things that I can do. And so it's interesting the way we've built it this latest time, which I think we've finally kind of got the architecture right and it's simpler. The whole thing is just kind of an abstraction is basically everybody has their own hyperdrive and then each device just has a policy for what to do with those hyperdrives. So like every device is a peer and obviously we'll abstract a layer of people on top of the device peers. But we want to start with just like local networks, as in like my devices like my MacBook, my phone, my Linux, like that sort of thing. But just make it so like the user experience is literally just you go, you scan, you scan the key or whatever, boom. Ask to join the network, approve it if it's something where you're not sure if it's not your id, you know, and then you just, you see the list of files and you select a couple of different devices. Like let's say my Linux machine, I say I'm an admin and the policy is any device that puts any file in any of the hyperdrives, download it, download it, just be there. And then everybody who joins automatically has a read only so everybody can see the whole list. And it's just kind of like a normal bittorrent or whatever is that you have your help. Like okay, well this device, there's two people online, two little check marks or whatever it is in the thing. And this one's grayed out because it's only on the phone and the phone's not online right now. And then you can just selectively go through and we have a little metadata packet on top of it for thumbnail info and all that stuff. And that's just an index that syncs through the swarm automatically. So everybody has it locally and you can look through and you can look at the images, you know, you can. It could be a Netflix or YouTube sort of thing. Like the interface is completely agnostic. You make it look however you want it to look. But the idea is just file sharing is easy. It's just easy, you know, and there's been. [01:43:17] Speaker E: And unlimited. [01:43:18] Speaker C: Yeah, and Unlimited, Unlimited and obviously the long term. And I expect tons of different problems as soon as we start getting into the whole public network side of it. And that's clearly where we want to go. Like, I think what you guys have built is such a strong foundation for this that I think the end goal of making BitTorrent look old and stupid is not out of the question. And that's kind of where I want it to go, is that BitTorrent always felt like piracy, whereas Netflix felt like legitimate. And it's like, why do these things feel differently? You know, they shouldn't. Like it should be like there should be the same protest protocol. There should be one good protocol for getting files or one good foundation for getting files from computer to computer. And if you want it to look like Netflix, make it look like Netflix. If you want it to look like you're pirating, I don't know, I don't even know if you want it to look like a developer environment. Make it look like that. But it shouldn't, it shouldn't change, you know, it should just work. So that's, that's kind of the idea. And I guess I haven't really gone into it publicly yet, but hopefully. Two weeks. Two weeks. [01:44:34] Speaker D: I always love it. I love it when we have these peer to peer apps and just explaining it from a personal scale because even a personal scale is hard because we have so many devices, we have all these problems. [01:44:44] Speaker A: Also as you were talking, I was thinking about the file sharing challenges. We fit in key also with UX experience around FIL while downloading and just getting info out about how much you have on your device and how much other people you might have uploaded. And so as we solve those problems also in the UX and Keat and like all this stuff we're adding, you. [01:45:00] Speaker C: Know, that'll just help makes my job a lot easier. Yeah, saves me a lot of bitcoins. [01:45:07] Speaker D: Hope, who was also working on this, I think. [01:45:08] Speaker C: Right. [01:45:09] Speaker D: He texted me the other day and he's like, how do I do progress management? And that was like five minutes after one of our guys just pushed it. And I was like, go look at. [01:45:17] Speaker C: The latest company. [01:45:22] Speaker D: Going back to this game. We talk about like solving things generally so other people can build. [01:45:27] Speaker C: Yeah. [01:45:27] Speaker D: So powerful. [01:45:28] Speaker C: Really nice. [01:45:30] Speaker E: The other, the other thing is there is like, I'd like to see with that the flow sharing or if you're like media streaming or whatever it is like I want to see con, I want to see a new resurgence of content creators. [01:45:41] Speaker C: Yeah. [01:45:41] Speaker E: You know, who are who can get paid paid for their content? [01:45:44] Speaker C: Oh yeah, right. Oh yeah. It's. That's, that's 100 a part of where we want to take it. 100. [01:45:51] Speaker E: We're all getting to that point. [01:45:53] Speaker C: The question is essentially long. And this is, this is kind of close to my heart as well because like one of the things that I've been most excited about since LLMs and since this peer to peer stuff is this idea of these single serve apps is things that just do something very specific, do not require me any of the overhead and the centralized and the account and all that bullshit is that I can just make a function and it can be delivered to all of my devices if I need it. So the idea is how are you thinking about and how will people be able to build maybe plugins for Keat or compatible apps that can take a Keat link and use it in a different way? Or that I can have a PEAR Drive link with a file and I can send it to somebody over key and it loads right in because it's a pair link, you know, it's. It's the same as how everything else talks. [01:46:44] Speaker A: So I think there are multiple sides to that. So for the, for the parallel stuff, the pair drive links or the paired drive links, I mean, yeah, the. Because it's all just hyperdrives internally under the hood, including inside of Key. That's how we share files there. [01:46:55] Speaker C: Yeah, there might. [01:46:56] Speaker A: Those might just be automatically compatible in the sense that any drives that you generate inside of PEAR Drive, if you post a link to that in Keat, the preview system would be the exact same as the preview system that's used inside of Keat itself. That's one of the cool things about peer to peer is because it's all the same data structures, so you might just get automatic compatibility. That would only go so far though. I think if you add a lot more custom stuff to Paradrive, then you might start to diverge and then you would need actual plugins and we actually implemented an early version, like a prototype version of plugins and Keaton as part of the identities push when we got IDs out. So it's one of these crazy things where the way IDs are implemented, it's all just a room. Like in Key, you have a little hidden room between all your devices and that's how your IDs to share it basically with yourself. [01:47:38] Speaker C: So it's just another room on top of a rooms all the way down. [01:47:44] Speaker D: It's actually really cool because when you add a device you're actually just inviting it to that room. Seriously. [01:47:51] Speaker A: So that's close to the commercial. [01:47:53] Speaker C: That's awesome. [01:47:54] Speaker A: But as part of that, we implemented that as a plugin system inside of the room code itself. So we have plugins, but we're going to revisit that. I think we're going to kind of scrap this version of plugins and do them in a much nicer way on top of the new room database that we're making. And so as part of that, yeah, you could definitely expect us to start making a kind of plugin interface for Keat that'll let you create custom UI elements and all the things you might expect chat plugins to do. But that's also a little bit TBD. We have to define the APIs. [01:48:20] Speaker D: But I think it's kind of like the essence of this is like peer to peer data is always self described. So if you post something that's self described, it's just fancy words for like there's not like a server or something it's just describing. So it's actually a really easy problem because there's no privacy lost in this. It's not like all of a sudden you can read all of our stuff by posting this link because it's just data. And then the problem just becomes how do we render this data? Which has, you know, complexities, but it's, it has implications. Yeah, but it's. But it's. But it's not that hard. Yeah, it's not that hard. Like compared to all this other stuff we do is not that hard. [01:48:54] Speaker C: Yeah. So it's like somebody having like a special function in Markdown and not being able to see it displayed in a different Markdown compatible app. It's like, well, if you just tell them they can just be like, oh, display it this way. [01:49:06] Speaker D: Exactly. It's pretty close to that. [01:49:07] Speaker C: Yeah. [01:49:08] Speaker D: And then it's just the mat ops, you know, having a way to express that so you can plug it into Keat. That's a system Android describing. So. And that's also should be peer to peer, right? Yeah, should have. Not. Not asking the loop stuff that. So, yeah, it's super exciting actually. Yeah, awesome. [01:49:25] Speaker A: Lots of possibilities there. Yeah. [01:49:27] Speaker C: All right. Okay. So question is, I guess technically you think about it from the context of broadcast rooms as well. But how can you like cross post so that I could like tag or pull something from one room to post and share just that in another room and. Or if I've got my RD memes and the Guy Swan network and you Know this thing, a bunch of different broadcasts. How could I put these into an aggregate feed so that I can just see all of that content from in one location? [01:49:55] Speaker A: I think one, one way to tackle it. So as part of identities, as we talked about this, the room, the shared room that's you know, goes between all your devices, this is very private. This is like for sharing all your private data between your different devices. Yeah, but it'd be very natural to also as part of having an identity, to have a public profile. You know, something you can write Twitter like posts to that you can add to from all your devices. It would also just be a room, it'd be a public one that you could share and we might want to render that as more of a feed view. So if you just want to broadcast content for yourself, you would just add that to your public profile and then people could, could see that that'd be a supernatural thing to add. So then for each of your contacts in your contact book, you could see their public feeds and maybe you could follow them and do all those other things you expect. So that would be one form of aggregation. [01:50:40] Speaker D: Really good. [01:50:41] Speaker A: Yeah. [01:50:41] Speaker D: I also think just for doing any kind of thing that is not a feature, peer to peer is just code, just code running your device. There's no endpoints. So we have APIs for all this stuff. Very easily instrument all this stuff. We haven't released many of these APIs because we're not really satisfied with how they work from people outside, just like new users. I is actually way too easy to just nuke your own data, which you're always very concerned about. Because we invest a lot. You invest a lot in chat. Right. We want you to make sure that you don't accidentally corrupt that. So but obviously once we have that, you know, because we're doing a lot of stuff like we talked about, that's. Then we're going to roll that out and then you can actually just script whatever you want because anything you can see in Keat is just peer to peer operations. There's nothing hidden. Like there's not. There's nothing that the app does that you can't do with a piece of JavaScript. And I think actually that's really, really powerful because there is no API. It is an API. [01:51:39] Speaker E: Like there's an exciting implication here as well in that once we start talking about aggregation, then you have to start talking about, well, how is that, how and when and where and is that information presented to me and whoever else. And so then you start getting into like well, algorithms, like what algorithm determines what's shown? And at the minute, like in the centralized space, you get the algorithm that you're given, right. And that's it. But a marketplace of algorithms. So you can select like different people have done it in different ways so that you can kind of get something that works for you. There's a whole other ecosystem that evolves from this totally. [01:52:16] Speaker D: It's actually funny because when you say that also, I'm just thinking it totally revolutionized how we build software internally because when we release a keed and the runtime, we make new releases, obviously that's pewter peer. We do it all through keed. So we have a bot running that's just a key user. It's just a robot that's just using this key API we have and it's just watching the swarm. So every time we make a test build, it watches that and then it posts an hour key saying, hey, there's a new build. And then you can adapt with that bot and say, we want to release it. Releasing is just multisig. It's just a matter of. It's just a matter of we go to our cold keys and we tell that bot, hey, make a signing request, just messages, David and me and your signers and then we just do our thing. Async, reasync, right. Sometimes David is somewhere, takes two days, doesn't matter. We just post the signature back to the bot. Then the bot is like, oh, I got the quorum. Now I'm just going to compute the signature for you, fully trusted, put it back in the room. I literally just take that, paste it into a script that just publishes to this form. It seems very scary, but it's like, it's just a signature. Then your phone goes, ding, dude. [01:53:29] Speaker C: You need like a formal set of that. Like that's. Yeah, that's its own ecosystem in itself. That's. That's like private git. [01:53:37] Speaker D: Yeah, I know, right? [01:53:38] Speaker C: Like that's crazy. With a multi sig, like publishing. That's my God, that's crazy. [01:53:43] Speaker D: And what I love about that is like it's. It's just a chat app. Yeah, it's just a chat app. Like, it's one of those things, if you went to Oracle, whatever, they'd be like, we have a solution for that. Yeah, it's going to be $100,000. Get the multi six solution. No, we're just using the freaking chat. [01:53:55] Speaker E: Yeah, but it's so cool because you've got this, that you've got such sort of like you've got Keith and you've got pair. And then you're like, oh, I need to do this thing with pair. And then you're like, ah, but I need to do this. Oh, we can use key to do that. And then like, so key facilitates pair and pair facilitates key just keeps going round and round. [01:54:10] Speaker D: It's very powerful. [01:54:11] Speaker E: Yeah. [01:54:12] Speaker D: It's one of those things where once we had to restart the build, I remember, and we didn't have pair running. We're like, it's so circular. It's actually really hard. [01:54:21] Speaker E: Once it takes off though. [01:54:23] Speaker D: Yeah. You need pair to build pair. [01:54:24] Speaker E: Yeah. [01:54:25] Speaker D: So I think just like the TLDR is actually all possible because it's all just, it's just code. [01:54:31] Speaker C: Yeah. What can you write and what can you look at locally on your machine and how. What can you not. [01:54:38] Speaker D: The only reason actually I think is why we have. The question is because we haven't published it all yet. Once we publish it, that's when it's going to get really scary because then there's going to be a lot of stuff going that's going to be redeemed. [01:54:49] Speaker C: My question is, and this is something we've talked about in previous episodes because it's such a huge element of this is the complete, essentially lack of infrastructure costs and especially with really popular apps or anything. Like I've had this conversation actually with Free Market Kids and Scott and Tali because they're trying to build a game and they don't want to have to pay thousands and thousands of dollars to have people log into a game on their servers. So the question, the heart of the question really is that let's say somebody has an enormous amount of infrastructure costs to give this app to these users. How difficult and or possible is it for them to take what they already have and to make it work in a peer to peer environment or do they have to kind of rebuild it within the peer to peer framework in order to get that benefit of zero infrastructure? [01:55:42] Speaker A: I think there's this really interesting property of peer to peer where some of the centralized apps that would be most expensive to operate infrastructure wise and bandwidth wise, like media sharing apps, for example, they also have really, really minimal collaboration models. Like they might not be collaborative, they might just be publish broadcast. That's very simple to do peer to peer. [01:56:01] Speaker C: Yeah. [01:56:01] Speaker A: So in some ways, like the easiest ones to do peer to peer are also the most expensive. [01:56:04] Speaker D: They're the most expensive here means like millions of dollars. [01:56:08] Speaker C: Yeah. [01:56:08] Speaker A: Like if, you know, if you're somebody sitting on your couch and you want to create a TikTok clone or something like this, if that's something you want to do, I don't know if you would. That'd be incredibly expensive for you to start prototyping and start building just because you'll immediately be blocked by bandwidth costs just everywhere. And peer to peer, it's a pretty simple thing to do, you know, just at least the broad strokes of the design. [01:56:28] Speaker D: And it's also, I think it's like, like Andrew saying, it's one of those things where everything, every time you move media into peer to peer, it's the same process because it's like there's very established patterns and centralized systems for making media sharing. So you just do it once and you can more or less put over all your things. [01:56:43] Speaker E: From a more technical perspective, if the thing is implemented in Node, for example, Pair on top of Bear supports the same things as Node. So you can just kind of put stuff across. [01:56:57] Speaker C: So it's a question really of how they have built it so far to a certain extent. [01:57:01] Speaker E: But the front end is obviously going to be integrated maybe with HTTP or GraphQL or something like that. You just need to take that piece out and then just integrate it with the hyperstack and with things like HyperDB on the way, which is really, really, really compelling. It's like a MongoDB type interface. These are familiar interfaces. We have Hyper B, that's just a key value stor or it's all familiar interfaces and those sorts of things will be being used already. So it's just a case of swapping certain pieces out. But certainly if it's on Node already, like we're not node, Pair is not node at all. It's a totally different thing. But on top of Bear, you can set it up so that it just runs all the no code straight up. [01:57:43] Speaker A: A big part of this, it really just kind of depends on the collaboration model your app uses. I think that's actually what dictates it. So if you have a very, very advanced patterns for how, you know, data is collaborated on with the different users in the system, that kind of determines how difficult the porting process. [01:57:59] Speaker D: I think if you just, if you just boil back to the essence of the question, it's like cost and I think it's actually cost. Like you said in the, in the talk, the bigger the cost, the easier it is, which is, can you kind of find. [01:58:11] Speaker C: Right, Yeah. [01:58:11] Speaker D: I guarantee you if the company comes in, like we're spending millions and millions of Nvidia we're going to be. [01:58:17] Speaker C: Yeah. [01:58:18] Speaker D: And if they're like we're spending thousands and thousands of dollars on something, we're gonna be like, it's hard. And so it's actually inversing the Internet because Andrew said that the truth in the first two minutes of this answer. And then we just started talking about the harder things, which is the less cost. So it's probably like, you know, it. [01:58:35] Speaker E: Justifies it as well. [01:58:36] Speaker D: It's really funny. [01:58:36] Speaker E: You've got 800 million a cost year. You just, just put the money in and do the migration keeps happening. [01:58:42] Speaker C: Yeah, there's the tiny fraction. It's also an inverse like cost benefit analysis. You know, like if you're spending that much, like it makes Sense to spend 20 million. Even if it did cost 20 million, spend it to get the, that infrastructure cost gone. [01:58:57] Speaker D: It's almost like you could have built it this way in the beginning. [01:58:59] Speaker C: Yeah, yeah, yeah. Awesome. So bots inside of Heat and in just kind of generally the pair ecosystem being accessible that way. And then also I just because of what's happened with my computer and the break in and everything, it's now suddenly relevant to keys being held in RAM and hardware being used to sign or open things up. [01:59:25] Speaker D: Totally. [01:59:26] Speaker C: So bots and security, what are naive thoughts on that? [01:59:31] Speaker A: Yeah, I mean so we have bots in some of the key rooms currently, but they're, they're very manually done. Like it's, it's, it's kind of a. [01:59:38] Speaker D: Tricky process to get kind of cool way I think because it's just like a messy script. Because it's just using the API. Yeah, exactly. [01:59:43] Speaker A: We, you know, but going forward we do want to have a bot API that makes it incredibly easy to add these things into rooms in general. And we're thinking about building, you know, based on the pair terminal system. We want to make a, like a little framework where you can just super easily and very, very safely write a bot that you can then deploy. Deploy and it'll handle all the nitty gritty. [02:00:01] Speaker D: Like obviously a keyboard should be running on your laptop, like not on the server. It should just be like a bot and you click in Keaton. It's like also run this bot for me. Yeah, it's just running. Yeah, there's no reason why not. [02:00:12] Speaker A: What's also cool is there are kind of, there are two kinds of bots. You can have, you can have a bot that you're running somewhere as a service in the network that is external. That's like looking at the chat and then writing Chat messages as a user. Then you can have something like a. It's a very different kind of bot bot that runs inside of the room code itself and kind of like as a plugin. And there are really interesting things you could do with that that kind of look like bot behaviors. Like you can have a plugin that maybe is going to play a game of like tic tac toe, that's kind of like a tic tac toe bot, but it actually runs in the code itself. And you could just deploy that just as a little pair mini app kind of thing. Different ways to approach it. [02:00:48] Speaker C: Yeah, super cool. [02:00:49] Speaker A: Yeah, lots of possibilities. [02:00:50] Speaker D: Kind of like installing an app inside your app. [02:00:54] Speaker C: All right, so second question. [02:00:55] Speaker D: Yeah. [02:00:56] Speaker C: Was security keys held in RAM potential hardware devices for like really cracking down or making sure that you're not vulnerable if you're using an untrusted machine? [02:01:06] Speaker D: You know, it's funny you say that because it's actually we're doing a really big push on that right now because so we know we're beta, we iterate hard. So some things we get too fast. Something we don't get too. Actually we've been talking about this for a long time. It's really, really important to us. And as part of. So we've been doing this rocksdb rewrite of some of the internals where we switch a new storage and we actually was like as part of that, let's think this in from the beginning because then we don't have to do migration. So we're thinking that in from the beginning so we can have much more easily pluggable keys and stuff like that. And I think that's going to be really, really important for just anything in the future that you have. You know, all you have should just be to put a thing in your computer and run it and take everything out. Nothing can run. So that's definitely coming. Something being worked on very aggressively right now. The only reason why we haven't done that aggressively before is we have a lot of stuff to figure out. We want to get stuff out first. And while this is insanely important, making sure that the chat runs and stuff is also really important. [02:02:15] Speaker A: Yeah. [02:02:15] Speaker D: But it's just to say like we're actually working on this right now. And I think that's. That's going to be really cool. And I also think it's like having keys on hardware devices, it's probably really important for some people. But also just having keys on my phone and not on my computer already makes it a lot better for normal People that don't have all these extra devices can play. So we want to have like a modular system where you can kind of like pick your. Pick what you have. But it's always. [02:02:45] Speaker A: It's always a tough balancing act because again, you know, Keyed is targeting consumers that might not have some technical knowledge of this. And so it's kind of like you know, you want to both. [02:02:52] Speaker C: Optionality. [02:02:54] Speaker D: Optionality without it being like exactly like that. All the non key users are far. Yeah, yeah, exactly like security for. But you can. You can tune that extra. What was the positive word for paranoia? I don't know. [02:03:08] Speaker C: Do we ever come up careful. [02:03:09] Speaker D: Carefulness. Yeah. Like so we. We have an extra step there. But yeah, that's. That that should be coming in like very short time frame. [02:03:17] Speaker C: Sweet. All right. Well, shit, guys. I just want to say, you know, it's been a pretty serious trip watching this, all this stuff evolve and especially getting to the point where I felt like I could do something with it just out of pure selfishness, I guess that I can build something and I've never been in that position before. And it's only like I said, talk about standing on the shoulders of giants is I'm using a platform built on top of a set of protocols and things built on top of a set built on top of a stack and then using an LLM that's probably trained on all of Yalls open source code in order to produce something built on top of it that I can just deliver to friends immediately which is just bonkers. So. But this is the first time I we actually got to hung out. Yeah thanks with actually meet in person so thanks so much guys. Seriously. It's been an absolute pleasure. Yeah man. So cool. Appreciate you for joining me man. Always tons of fun. [02:04:27] Speaker B: All right guys, I hope you enjoyed that episode. And don't forget to check out fold. Don't forget to check out Keat. I will have links to everything in the show notes so that you can start exploring this stuff if you haven't yet. I'm sure you have heard me talk about it way too much. I feel like I bring it up all the freaking time. In fact, I don't have. I had my Keat hoodie on just a little while ago. I don't have it on anymore. Missed opportunity here. But definitely don't forget to subscribe to the YouTube channel to rumble. I've been trying to make sure I publish everything on Rumble still like there's no activity over there. But I. I like having that as a backup even though YouTube still seems to be for as for better or worse, it still seems to be the place that you can reach more people. Follow me on X follow Math, David and Andrew. All of their details and links to their stuff can be found there as well. And start playing around with pairs, especially if you're a developer or you're interested in peer to peer tools and you want to start exploring what you can build with these things, they are, they're getting really, really close. And keep an eye out or an ear out for Pear Drive because the main foundation of the module that we built, of course we built it for. [02:05:45] Speaker C: The app that we wanted to build. [02:05:47] Speaker B: But the main foundation and the reason we have put together the module like we have is so that it is as utterly simple as we could think to make it to build with a peer to peer foundation and peer to peer file hosting and sharing between devices. So if any of this stuff confuses you and you're not sure how to build with it, I hope, I hope fingers crossed that we could make this easy enough that anybody who can code can build with it. So stay tuned. All of that very, very soon to come. In fact, I might get an update on it today, I'm not sure. So stay subscribed, check it out, follow me, you'll hear all about it. And until next time, everybody, this is Bitcoin. Coin audible I am Guy Swan. [02:06:38] Speaker C: Take it easy guys.