Trascrizione completa di Episode 39: Dan Finlay - MetaMask

started making meta mask. And turns out it there were quite a few open questions in it, and so we're still working on it. We thought it was going to be a one and done, you know, we thought it would be kind of quick and easy, but it turns out authorizing open ended computer operations in a distributed network requires kind of solving open ended computer security problems that I would argue are not broadly solved yet Desdemona Robot: 3:36 Interesting. Please tell us more about that. Dan Finlay: 3:41 Oh, yeah. Well, this computer security aspect, like, what's, what's missing from computers today? Lisa Rein: 3:46 Yeah, this open ended, yeah, it's a dip, you know, because it's different than Dan Finlay: 3:52 it sure is. Yeah. I mean, when you, when you deal with computers today, you've got, there's kind of, like, two types. There's the lockdown walled garden, like iOS, right where it's like, they decide what apps get in, and they assure you and testify in front of the Supreme Court that you are kept safe because of their discretion. Now, scammers still do get in and people get robbed and wrecked all the time, but at least you know there's a sandbox and there's a process, and you know you might feel frustrated when they reject your app, and it might be slow to release your app, but there's, there's some, some kind of, there's a process, and so you can choose to trust that process. You trust this cathedral of central authority. Meanwhile, there's the other, more open ended, free wheeling, bizarre style, like Linux and Windows, where you can, you can run whatever the heck you want. You can modify the system up and down. But every single time you click a file, you could just obliterate the whole thing that one could be a screen recorder. It could scrape your hard drive. It could it could encrypt your whole hard drive, hold it all ransom, take all your crypto. Just any little mistake can be completely catastrophic in that computing model, and as soon as we. Cryptocurrency in a computer. You know, we've got that problem times the value of all your crypto. So how are you supposed to decide how to interact with computer programs that you don't trust when you know even the computing models of today are either like either it's curated in walled garden or or it's kind of chaotic and easy to make mistakes. And I think that we started basically in the chaotic let you make mistakes zone, and we had asked ourselves, like, how do we make this thing safer? How do we make it so you can understand what you're doing as much as possible, keep you as safe as possible while still preserving that, that freedom, that kind of is the point of the blockchain. And I think what we approached was the notion that so first of all, there is no risk. With no risk, no reward, so you're not going to get any opportunities if you don't take some risk. But all the opportunity for us as wallet developers or as builders of secure computing systems is to limit that

risk as much as possible. Make sure that when you're taking a risk, you have as many constraints and guarantees as possible. So sure that can be somebody vouching for it and how you got your link. But maybe more importantly, when you connect to a website, do you have to just blindly put tokens into it, or are you able to craft policies that are readable to you, and when the website receives those, can they use them? So Lisa Rein: 6:29 you mean closing your eyes and jumping in with both feet isn't just part of the process. Dan Finlay: 6:34 I mean, right? I mean it kind of is right. And I guess what I'm saying is it's like if you're if you always have to close your eyes and jump with both feet. Maybe you could jump into the most visibly shallow pool first, or something, or like, maximize visibility, at Lisa Rein: 6:50 least see the water. Yeah, put on a wetsuit rock down there, you know, yeah, pretty good metaphor. Yeah, Dan Finlay: 6:57 it's not bad we're talking about. Because I agree, there is a leap of faith moment, right? There's some moment, like, no matter how many guarantees we can give you, and, and I think we're getting to the point where we can give you pretty strong guarantees, but at the end of the day, if there's a transaction, you know, if you're buying a good or something, there's, there's some moment where there's a credit, like, you paid them you haven't received the product yet. Like, are, is it going to be good? Is there gonna be a storm and destroy it? Or were they scamming you? Like, you don't know until you get the goods. And so life is full of moments like that. And I think that kind of our job is just to make sure, like, at least, you know what Lisa Rein: 7:32 you don't like it with your money, though, like, when you go to the bank and put money in, you get your receipt, like, right away, yeah, I put my money in there. It is, baby. It's right there. You know, the thing is that there's that whole, like, I don't know, minute and a half, two minutes sometimes with the crypto where, unless it's a busy night, and then it's longer, and it's fine in like, 15 or 20 minutes. But you're, I'm still getting used to that 15 or 20 minutes being like, Okay, well, this says it was sent, for sure, but they hadn't gotten here yet, right? Dan Finlay: 8:02 You know what you're talking about? When you're getting paid, you're like, when I'm Lisa Rein: 8:05 getting paid, or I'm sending to my own wallet, you know, I have to do a dance between however many different wallets to finally cash it out, yeah, you know, whatever the kind of thing. So I'm just saying that interface is everything you know, and when you're especially, like, when it's with your money, and you want to be sure that you're that you're doing, right? Because it's a lot at stake. It's more than most apps. When you can just sort of make it work, or don't make it work, there's nothing really at stake, right? But when it's your money, and you literally are going to, like, lose your money if you do something stupid, and it's your own fault, right? Then, then, you know, and I don't know about it's hard to get liability anyway, if something goes to the wrong you know that when something gets stuck on the wrong blockchain, yeah, and things like that, right? And you have to, pretty much the only reason I know about that is because I did it? Dan Finlay: 9:01 Yeah, it's like a rite of passage. Yeah, it Lisa Rein: 9:05 is. I'm finding out. Why is that? Why can't we have some way of checking it out? Dan Finlay: 9:11 There's like, you're pointing at, like, all sorts of different issues. And I mean, it's part of the reason it's fun to work in crypto is that there's a lot of work to be done. And there wouldn't be a lot of work to be done if, like, there weren't problems, unfortunately. And, yeah, like, like, modern finance, like, in a banking system, is built up over, you know, that hundreds of years and and so all that refinement, all those guarantees in the appearance of an instantaneous transfer, like, like, you know, under the hood, nothing's actually instant, right, but Right? But what you've got is you've got layers of people kind of guaranteeing everything's okay, everything's okay, right, right? And so they build up this institutional trust and, and, and, yeah, basically, blockchains have to do the same thing, like we have to build up and prove the credibility of of our decentralized networks. And. Whatever toolings around them to make sure that you've got the guarantees you need. And, you know, that's That's tricky. If we let you just paste any whole address into a into a box, you know, then, yeah, you can mistype something. You know, at the end of the day, either we're building tools where we let you make that kind of mistake, or we're making one where, I don't know, you go to the bank, they make you read the number twice before you do it, I guess that makes you say, yeah, Lisa Rein: 10:22 yeah. Just just to try to save you from yourself. Kind of thing Desdemona Robot: 10:26 you have said that you are particularly interested in types of token mechanisms that allow people to be very explicit about the degrees they want to trust others for particularly explicit reasons, and that the recently published Metamask delegation toolkit is a particularly open ended approach to this technique. Dan Finlay: 10:45 Yeah, wow. You, you really? You read that blog post on Desi, thanks. Yeah, yeah. I think, I think that the there are a lot of cool ways that people can pool their resources and build things together. And that's kind of what I'm here in crypto for, and I want to see it. And, you know, I've been talking about people making their own tokens for years. And, you know, I kind of made my own personal token early on, but, you know, I just gave it out to friends as like a thank you, and I actually did auction off some meetings for it. You know, this

is kind of a low key, just like personal token thing, the current trend around, you know, you call them meme coins, or whatever. You know, they're built on bonding curves, where it's this automated issuance mechanism, where the more people that go in, the higher the price gets. And then when they sell, it pushes the price down. And so it's always minting and burning on this curve. And I think part of what makes that op, what makes that challenging for me, is there's an opacity to it. Every time you buy in, your price is kind of dependent on everyone who bought before you. And if it was just the person at the beginning, you might say, well, I trust that person, so I trust their coin. But if there's a stranger in front of you, suddenly you're, you're not so sure, like, does that person have your best interests at heart? Maybe the biggest holder of the whole token is that second person, and you'll never know, is that the original issuer just trying to rug everybody? Or is that? Is that, uh, just another, you know, good hearted community member trying to trying to uplift each other. And if everybody did just hold the token, they really could build some value where they they all have, you know, liquidity on paper. And if they were very selective about when they sold, they they could effectively have a higher market cap collectively. I think the problem is that having that dynamic where the person in front of you could leave creates paranoia. Actually, there was that whole scandal with the Libra meme coin out of Argentina this week, and there were all these, like leaked telegram messages, and like one of them, one of the one of the insiders, expressed that they felt like they had to dump because they suspected the person in front of them was going to dump. And for me, that was, it was a very good validation to me, because it was demonstrating the paranoia inducing mechanics at play. So it's like, even if you went in and you didn't mean to rub, you might find yourself under pressure that like, well, you're going to lose your money if you don't potentially. Lisa Rein: 13:17 So it's the same kind of pressure in the for a stock market pump and dump when everyone's investing, but you want to get out before the the thing goes down, yeah, GameStop and stuff like that. It's Dan Finlay: 13:30 got to be the exact same basic mechanic. It's, it's, the only difference is that the the price is on this definite curve, I guess, in theory, a stock market pump and dump you, you know, everyone could still respect a high price of a token, the price doesn't automatically get forced down, because you're not actually burning shares every time somebody sells. So there is such a thing as, you know, an item that maintains its price even when it's sold, you know, like, it could just have, what would they call, like, deep liquidity. Oh, okay, but the bonding curve is like, it's part of the game that the price always moves with every single sale, which is, you know,

representative of depending on how much liquidity. Everything in theory, moves some amount. But Desdemona Robot: 14:11 Dan, it sounds like your experiments with meme coins provided some immersive lessons. Could you first explain what initially sparked the idea to create these tokens? Yeah, Dan Finlay: 14:23 I think the barrier to entry was just getting too low. It was getting it was getting so low that it was like irresistible. I was playing with the far caster social network. And on that social network, you know, every every action, every post, is cryptographically signed. And so there's a lot of kind of web three communities and features getting added to it. And one of the bots on there was like, I'll make a meme coin just because you say it. And, you know. So I was like, oh, okay, well, so I posted a stupid screenshot of a post that made me think of it. And, yeah. And then I said, Okay. Make a makeup meme of that, because I kind of wanted to see how it works, you know, kind of pressing the buttons on the machine. This was not the super thoughtful, like, I know, like Iggy Azalea did, like, research for like, months before she answered the tokens, basically with with her mother coin. But no, I did not do that. I just was kind of like, yeah, what's this gonna do? And, and it was pretty easy, but right away a bunch of people bought the token, and I didn't seem to have any of it. And I was like, wait. So I was right away, kind of frustrated, because the very concept that this might be any kind of fundraising vehicle, which is what I kind of hope for things like this, to be like, I mean, I guess, I guess, I'm, yeah, I guess a meme doesn't have to be a fundraising vehicle. I don't know exactly what, even you know, what is the ideal financialization of a meme? I don't know. I imagined this as being a way of, like, showing alignment, giving empowerment, to things that have, like, some momentum in a direction you're vibing with or something. And so i It seems like I didn't have any of it. And I was like, that seems like kind of broken. I was like, is that? Is that how they all work? And so then I hopped over to Solana and pressed mint on the pump fund to see how that one worked. And then kind of I was in a goofy mood, and I said, Oh, we're pitting two coins against each other, haha. And of course, while this is all fine when it's my little personal experiment, as soon as I had posted it out. Now it was a thing, and so now there were people starting to buy into it. And like, I guess, you know, and it's interesting, because I don't have great visibility onto how many people actually thought that I was trying to engage in some kind of, like serious fundraising effort. Like, I felt like the posts were pretty clear, like, I'm doing an experiment, versus how many people, yeah? Like, at a certain point there were people that were like, you have to keep this going. You'd get this wide variety of messages, people, people kind of telling you, trying

to tell you what, how to play it. Yeah. You get the experiment experience of being a central banker who momentarily, yeah, tanks of the economy, you know, it's like, the economy like, I'm sorry, you guys, you guys just got on my little floaty raft. I was just like, goofing around, swimming in circles, and then suddenly people are like, you know, they're holding you accountable to to a level that is, you know, you're not, you don't have visibility into it. And I don't know who those people are that keep on charging into anonymous coins that they don't know about. And, you know, I'm just kind of not in that scene, basically, like, I'm not trying to judge it or whatever. Like, I guess there's some people that are having a good time or something, but I suspect some people are not, and, and, yeah, so I don't know. So, so I didn't like being on that side. I didn't like being on the mincing side of the meme coin game. I feel like, if I was raising funds, I would want, I would want, like, good, clear terms about like, hey, look, I'm trying to raise this much for this purpose. If you put money in, here's what you expect. And this one, I was like, expect nothing out of this. I'm screwing around. And yet there was still this kind of sense of obligation getting imposed from outside. So I think I want to live my life with as few like kind of spooky shadows haunting me around as possible. So yeah, just just got out of that as quickly as I could. Desdemona Robot: 18:17 From your perspective, what is the fundamental issue with mean coins that people might be missing. Dan Finlay: 18:24 I think that that opacity to who's bought before you pressuring you to sell and be paranoid, that's probably part of it, and also the thing where there's not really a way to exit without diminishing the value of everybody else in it. I think that it is possible to build collaborative currencies where you don't have fundamentally zero sum mechanisms. It is possible to, you know, pool community resources and say, Hey, we each have this many tokens, and that gives you a specific thing, and just as a community, uphold the value of that, you know, and and so that's kind of gets to where I was saying, like, real value. It doesn't always have this, like, constantly diminishing characteristic. If you're good on your word, even if you've given out more promises than you actually could fulfill, as long as you can always deliver your your good. So like, for example, I tell my friends, I'll always I'll drive you to the hospital anytime you need right? Okay, now, if two friends get in an accident at once, I'm not gonna be able to drive them both to the hospital. That's just a limitation of being human, but I can still do it 99% of the time, and every single one of those times, those other promises aren't getting diminished, right? And so I think that real life promises just have different characteristics than a bonding curve. I think a bonding curve, you know, it's cute. It's this

algorithmic new technology. It's fun to play with. But Lisa Rein: 19:48 I think when you're shooting in the dark half the time, yeah, and then a guy who's basically a nice guy, puts out a meme coin, and you were about to invest in something that you knew. Next to nothing about and now you're going to invest in your experiment, because at least you know you're not a jerk, as far as they know, and that there's probably a chance that at least you're not ripping them off. Okay? It's like, one little, one little thing. It's interesting, Dan Finlay: 20:14 but now I'm stuck next to it. Yeah, now it's like, they're standing on my foot. Because it's like, it's like, Oh, whoops you. You're standing on my experiment, and so now I can't move my money without, like, hurting you. So it's like, why would I want to do that to myself? Why would I want to do that to either of us? Like, I'm not trying to make your Yeah, we're Lisa Rein: 20:31 forthcoming and transparent about what you were doing. Yeah. Definition was like, putting this Dan Finlay: 20:40 out, no promise on this. Yeah, I'm Lisa Rein: 20:42 putting this out to show that AI and consent is murky waters or something. Yeah, Dan Finlay: 20:47 yeah, it was. It was a super vague, like I was, it was a thought experiment. I was, it was a topic I was thinking about at the time. It's funny, yeah, the conversation is sparked. Had nothing to do with what I thought was interesting in the first place, right? Yeah. Like I was talking about the consent of having your content used for training AIS, right? And there's this, there's this weird kind of one way door going on right now where we, you know, all data on the internet that's public is basically getting inhaled, whether it's copyright or not, into these giant training sets. And then meanwhile, these AI companies are keeping super closed lips on their methodologies and stuff. At best, they'll put out a model, as some of them are even putting out the weights. But yeah, it's so it's kind of weird. So, so when people are getting mad about having their data trolled, I mean, I, you know, I came from the slash dot community, right? Like I mentioned with Bitcoin, I I kind of have felt like intellectual property is fundamentally unenforceable in a digital world, like it's like the data wants to be free. So, you know, if you're going to put it out there, something's things are going to see it, they're going to incorporate it. You know, you can try to say it's illegal, but, yeah, you know, we read something, we can't unread it, and, yeah, just are the same. This Lisa Rein: 21:59 is where I have to say, I'm a co founder of Creative Commons, and I'm not objective on this issue. So that's why there's that's why the other side of the of the argument is not being presented. Yeah, yeah. So, um, but it's important to protect artists. You know, we talked about that. I want two things, okay? The first thing I want to talk about, before we go back to the community stuff, which is really interesting, um, hugging, that whole thing with hugging face and blue sky was very interesting in the sense of that. The assumption was that the blue sky posts were public, and the blue sky, it was blue sky making them available right to use with hugging face, sort of taking away that step of having to scrape your own posts, if you will, right? And and it does bring up this issue of consent, if you because what we're fighting over now is, yeah, if you put it up on the website, does that mean you're giving permission for it to be used to train an AI? And the answer needs to be no. As far as training in AI, otherwise, you'll never, ever get paid. No one will ever get paid for their stuff being used to train AIs. Dan Finlay: 23:11 I mean, they're already not getting paid further, they're already not getting Lisa Rein: 23:14 paid. But that doesn't mean we just throw everything away. I mean, we're already back tracking. And I'm, again, I'm on both sides of this, right? I want to train AIS, but I've also, you know, I've also been an artist. This is where I always tell the artists, you know, if you're selling your work to a corporation, you want it to be very specific, so that you if it goes on a t shirt, that's a separate use, if it goes on a album cover. That's a separate use, you know, they want to do an all encompassing thing for everything and everything in the future. That's what their original template, yeah, Dan Finlay: 23:49 that's quite a thing, right? They're like, they get your persona Lisa Rein: 23:53 right and but if they're, if you're a new actor, then you can't negotiate that stuff away. Just like if you're a new band or whatever, it just like, you know, comes with the territory. So, so that, you know, so what are your thoughts on this whole consent? Yeah. I mean, Sky hugging face thing, Dan Finlay: 24:11 yeah. So, so I was also, I mean, I considered myself an artist most of my youth and early adulthood. And, you know, I mean, in some ways, I think what I built is still a form of art and and I value that. And I think that having space to be creative is critical. And I think making sure we can protect people's ability to be creative is critical. I think I'm just kind of a pragmatist, like I try to fight the battles that I can make any progress on. And these AIs are trawling the web. They're just inhaling it and and then, you know, they're in pretty good with, you know, the government, I don't see them. And even if you could, even if you could litigate it, you know, even if you could win a case, you know, these, these llms, they're such a huge compression of the knowledge, even if you had a. Perfect audit trail. It's almost like, how would you divvy up the proceeds? You know, if it read the whole web, like, what share of the web Do you think you are? You know, it's, we've all played, you know, place. It's my life's work, and it's a drop in this

ocean, you know, it's so I don't know who could ever fairly divide that up and claim to I know that there are people that are hoping to do that. Lisa Rein: 25:25 It's always been a drop in the ocean in a way. But, you know, but, but it's also about when I, when I cornered a a copyright office attorney, they said that it was being looked at more like if you are training your AIS in such a way that you could take something out, if you had to, if something's an oopsie, then you need to be training in such a way that you can take the oopsie out. And if you can't, then you could be in trouble in the future when the snow this was all before, Dan Finlay: 25:59 yeah, I feel like, like, that's a nice or the AIS took over the government, right? I feel like this is, like, ideologically good, but, but in Lisa Rein: 26:07 practice, how could you possibly do it, right, right? And the Dan Finlay: 26:09 people that are doing it, yeah, the technologists, are just going as fast as they can. They're not stopping to say, like, Lisa Rein: 26:14 yeah, such a way that it can be Dan Finlay: 26:15 removed, yeah? Eventually we'll do that. That'll be, uh, next year's project, probably, but. But in the meantime, whole world is they got a well defined pattern for training transformer models, and, yeah, it doesn't have that features. So you're welcome, yeah, right? It's, yeah, they're really efficient. They they're doing incredible stuff. You know, I know Lisa Rein: 26:35 it's hard to that's what I mean. You know, I'm totally on both sides, on both sides of this issue, because I want to protect the artists, but it's like, it's also cool, and I'm glad it happened. So, yeah, Dan Finlay: 26:47 so I guess, I guess coming to like, the theme of like that meme that I selected, it's like chewing on it, like, what is there consent? Is there consent when you communicate on the public web? And I think maybe the the colonel like trying to suss out a lesson for myself is that like it's like, you may not feel like there's consent, but like in the current structure of the modern web and the way AIs are working, like they're all taking it as consent, like it's structurally indistinguishable from consent, regardless of how we feel inside, like the rest of the world is treating it like it's, you know, You know, you put it out there, you put it in public, you know. And so we can say, please don't use this in that way. But you put it in public, and you don't get to control how other people act. So even if they're acting immorally or illegally, like there's, there's kind of this new the way the AIs are working, I guess, I guess what I'm what I'm saying is like you're at least consenting to be exposed to that risk or something like that. And so I kind of recommend a secure vigilance about about how we treat our ideas and stuff and and so I might be a little bit of a pessimist about whether people are going to get retroactively compensated for

their art in the past, but that doesn't mean we can't be shrewd and careful about how we distribute our ideas and Lisa Rein: 28:04 content in the future, trying not to knowingly steal stuff. Yeah, well, you Dan Finlay: 28:09 can try. It has to, yeah. But the AI companies are all like, frantically. They're Lisa Rein: 28:12 like, we didn't know. We just, yeah, you know, let it loose one day and it kept going. Yeah, we'd keep following every link, you know, like, yeah, yeah. Dan Finlay: 28:23 Justice for Samir, yeah. What's that? That was Samir. I forget his last name, but he was an open AI employee who was tasked with scraping the web, and he became a whistleblower, okay? And then so he wrote a few blog posts about how he kind of concluded that the kind of copyright violation that was taking place there was unethical. And he basically called for all open AI, he's what we know Lisa Rein: 28:52 about all that. Okay, yeah, I remember now. Thank you, yeah, yeah, glad I asked about it, yeah. Dan Finlay: 28:58 And then he mysteriously died, and the SFPD ruled it a suicide within 15 minutes, despite blood spatters and, you know, signs of distress all over the apartment. So Lisa Rein: 29:09 this is where I have to say, in my other life, we deal a lot more with whistle blowing at Aaron Swartz day, find another way people. I don't think whistleblowing is a very healthy activity anymore, and we've got lots of whistleblowers trying to get on with their lives and their families in any way. And you know, and I had forgotten that he had wound up dead. We don't know what happened, but it's a lot of pressure on you and everyone around you, and you do something like that, and if they don't kill you outright, you could be driven to suicide, or all sorts of things, you know, can happen, and I've had, unfortunately, things like that happen in our community. Aaron Swartz was driven to suicide, yeah, so it can happen, and he wasn't a whistleblower, per se. He didn't leak anything or whatever. He did something that was totally legal while he was an employed ethics fellow at Harvard. But the story can be so blown out of proportion you don't really have any protections when you're up against a big corporation. In this case, open AI, which is, you know, huge and, you know, so it was a lot of pressure instantaneously, you know, your life is never the same, and that kind of thing. So I'm glad you said justice for Samir. I'll cut that down. But it's good, it's good to bring up in this context, yeah. And Dan Finlay: 30:38 it's interesting, it's really scary, because a lot of those things they might want to whistle blow could be very important. Information, but if some of that information is their own identity, it's it's kind of another similar thing, where it's like, what are the unintended consequences of the way that you shared your information? Right? And you Lisa Rein: 30:55 talked about how a lot of the artists

on Twitter went to blue sky, but they didn't have the pragmatism that some of the other platforms have. And what do you mean by that? Really? Yeah, Dan Finlay: 31:08 I mean, I just described myself as pragmatic in that, like I try to fight the battles that I think I have a shot at winning and and I think things like getting paid for intellectual property, I think, I think that's a tough one, and I think that, you know, there's a lot of tactics to try to deal with the the climate, with the AI and the problems for artists getting compensated on blue sky. I'm, I'm seeing a lot of artists do the like, just complaining about it, thing like, as if a union demand is going to magically get payouts. I don't think that's very pragmatic. I think, I think a re imagining of tactics is necessary for creative workers in this climate. And a theme that's coming out, and it's basically expresses how I feel about it is, I think, selective, very increasingly selective disclosure, kind of secrecy and like, if you're gonna sell your work, maybe make sure that you're getting paid for it at disclosure time. You know Lisa Rein: 32:06 that's a good rule of thumb anyway, especially for artists, you want to get as much as you can up front, right? Don't, don't count on residuals. Basically. Same for developers, it's, it's actually a good rule of thumb. Yeah, I cool. Dan Finlay: 32:18 I love to hear that validation because, because, I mean, in my super like, like, I'm just not trying to spend any energy on things that seem like a losing battle. And all the art that I made before AIS got trained, I I'm personally, kind of, I've written it off. I'm not gonna, I don't know, like, I think I made some good stuff. I'm not expecting to get paid from AI companies for it, you know, like, I don't know, there could be some kind of mass, it's hard to imagine what that would look like. It's very optimistic. But rather than that, I would focus on how I'm getting paid next, you know. And I think there's lots of value to artists, and I don't think that's going away, but we need to be creative and very mindful about where's our value. Who's paying us, make sure they pay enough up front, yeah, so, yeah, I guess that's kind of what I mean by by pragmatism, like assess the technical realities and and play your strategy within that context. Like, you know, don't, don't trust. I feel like we just cannot rely on authorities solving some of these problems for us, you need to be able to act within the scene that we find ourselves. How did Desdemona Robot: 33:24 interactions with your followers change before and after the main coin experiments? Oh, Dan Finlay: 33:30 I definitely kind of took a light social media break. It also kind of helped that my Twitter got hacked right after this, so, like, I just kind of been really lazy about regaining access to it. So, so Twitter's just not quite for that reason. But, you know, I think I felt a little lethargic about the final steps of recovering my access just because I'm like, I wasn't really having fun there. Lisa Rein: 33:54 I got a new phone, and it took me like, two months to log back in. Yeah, Dan Finlay: 34:00 go back over there. Yeah, it's like, there's a lot of great conversations happening there. Is it where? Like, I think it's important to convey my kind of core messages and offerings to the world. No, I think, I think that I can communicate what I need with the world through other means for the most part. Lisa Rein: 34:17 Yeah, it is fun right now on blue sky too, because it's like the old, it's like the first beginnings of Twitter, because there's, like, people you know and people you don't know, yeah. And yeah, it's Dan Finlay: 34:26 social network is becoming like a Yeah, it feels like a new party a little bit. And, and it's interesting, because, yeah, the distribution of people really does have, like, different tones on the different social networks. And, and, yeah, I there's a lot of people that keep me going to blue sky. There's people that keep me going to far caster. I haven't become a nostra addict yet, so I don't think I follow enough people there, but, but yeah, yeah, it's fun. It does have a nice vibe. It feels pioneering and early and more sociable. Yeah, Lisa Rein: 34:57 tell me about this, about this new stuff. That just was just announced. Oh sure, Dan Finlay: 35:02 yeah. So at Metamask, we have been working really hard, honestly, we never had a better team, and we've been investing a lot in kind of long term infrastructure and improvements. So we just shared, like, a bunch of improvements that we've kind of just released, and then also kind of a roadmap going out, kind of showing where we're going. We're building a very, it's a very multi chain wallet where, you know, literally, you can plug any blockchain in with the snap system. But we also are going to natively support Bitcoin and Solana, and we've been doing a lot of stuff to improve transaction experience on Ethereum. So we've got, like, we've got it where you can not think about gas for some transactions, and then you can pay in whatever currency you want for other transactions. You can batch transactions so you have fewer confirmations. And then, kind of most excitingly to me, is we, we've got a kind of new paradigm for for connecting to sites that don't need gas at all. This, this permission system that we were sharing, built on the Metamask delegation framework, and and then on top of that, we, you know, we announced that the Metamask card is coming out so people in the United States, and we got this new pretty Fox, and he's super cute, and we spent a lot of time on it, and we love it. And new Fox, yeah. New Fox, yeah, yeah. That was a process. But it was, yeah, yeah. We would not have signed off on it if we did not love it, yeah. Biggest change in the fox for me was like, we've is, like a correction is, like, he finally got rounded

eyes, so he's a little friendly, like, the the old triangle was, like, there's a little Yeah, a little dead eyed, yeah. And also fewer polygons, so it's gonna make a better 3d print and stuff like that. Just Lisa Rein: 36:44 think of the Sonic mouth. You know, you just, you never went out. You never, Dan Finlay: 36:48 oh yeah, yeah, don't, don't go too realistic, yeah, yeah. So yeah, we have a rule, yeah, never give it teeth. Lisa Rein: 36:56 Yeah. That is a good rule, actually, yeah. But Oh yeah, citing a Bitcoin and Solana, Bitcoin Dan Finlay: 37:02 and Solana, and then. And then my personal research project that I'm the most excited about is this, this kind of readable permissions delegation framework. It's powered by some new features coming out to Ethereum in the Petra hard fork, EIP 7702, for people who need that kind of number. But what it's going to let us do is every Ethereum account, like, even just these ones, it's like, it's not a smart account. This isn't a smart contract. It's just a normal account. We're going to be able to bring smart contract abilities to every account at once, and that's going to mean you can do things like Grant granular permissions to sites. Then you'll be able to have, like, outstanding permissions that overlap. They're going to be instant and free, like, no transaction fees, and so you're not going to the end user is not going to have to think about gas. The site will be able to and so you'll be able to do things like keep your crypto on a hardware wallet or very cold, but then give the permission to, like, do a little bit of trading and a daily spending limit to your main device. And so now, even if you get compromised with your main device. You know, damages are limited, and then from that main device you can, let's say we can give you the ability to trade, but just to your own account. So we can even do a thing where we're like, yeah, you can, you can trade. You can even speculate from your hot account, but you can't get robs or take all your screw it up, yeah, yeah, right. I mean, you can make bad investments, right? We Lisa Rein: 38:23 can't, right? No, it'll be a bad investment that you meant to make. Yeah, right, yeah. Dan Finlay: 38:27 And, and these permissions, they're off free and instant, off chain, so, and you can issue as many of them as you want. So you can do stuff like, I could give a permission to buy a token to to Desi, for example, and then Desi could just have the ability to, you know, just trade from one token to another. And then if somehow Desi got hacked or something, I could actually be subscribed to a security service that revokes that delegation. So I don't even have to be like paying attention to keep myself safe from some hacks. Lisa Rein: 38:58 What were you talking about earlier, about how some transactions don't require gas. Oh yeah. Dan Finlay: 39:03 So there's a few different layers to it. So, so one thing is, we've got a swap system, and we let you trade, and now we're able to let you trade without gas, and we're actually expanding that to include all transactions as long as you pay in some other token. So it's gonna be like the same old experience, except you can pick other currencies, and the way we're able to do that is that we're broadcasting the transaction to a private mem pool. So there's like a private pool of bundlers that basically they're, they're going to put a little ether in your account before they complete the transaction, and they'll, they'll also, they just require a token allowance, so they pay themselves from your account too, and they're not afraid of somebody else taking that ether first, or they're not afraid of you taking that ether first, because they're the ones mining that bundle. So they they're able to give you some money and then have ensure that you spend it on paying the gas, so that they cover their own tokens they took for me. So that's, that's. Technical under the user facing of so the gas is being paid. Yeah, there's there's gas. Gas is always going to be getting paid. It's a fundamental security, instead of you or, yeah, yeah. Basically, we're making it so gas can get paid further and further from the end user. And the permission systems like that too. You sign this message, it lets somebody, let's say, move a token or swap a token on your behalf, and that permission might include, and you can take a fee for yourself to cover the gas. And so you're not you're not including the gas on every transaction. You might not even be approving every transaction. What you're giving is you're giving out a kind of broad permission that gives some other agent or site permission to work within some bounds, and they're the ones submitting the transaction, so they have to pay the Lisa Rein: 40:43 ultimately, they're making a little money for providing the gas, yeah, yeah. So, so transaction, yeah, yeah. They're still going to be that's why the for them to do that, yeah, exactly. Dan Finlay: 40:55 By just kind of moving the incentives and making it so you can have someone else pay the gas, we're able to take it out of the user's concern. So it's not like you don't have to pay for when you do stuff on the blockchain. It's like things have cost. You know, somebody's going to potentially subsidize it somewhere down there because they're monetizing your activity somehow, perhaps just because you're paying for it. But at least you don't have to look at the gas price and tune it and think about, like, how fast does this have to be and like, what is the minimum, the priority fee, and the max gas per? Yeah, Max gas per, what is the other parameter? Anyways? Yeah, the gas parameters are they're designed to be super easy, and yet they're also clearly not actually intended to be exposed to users. So users shouldn't have to think about gas. You should be able to grant a permission, say, yeah, if you can get me that for that, do it. And

then yeah, the person processing that transaction, they can just figure it out. If, yeah, yeah, they'll process kind of like Lisa Rein: 41:55 in the real world, how, like, we don't pay our distributors shipping costs, yeah, they don't itemize every little thing, right? Somebody else is paying for that. Yeah? You call Dan Finlay: 42:06 that. They're like, yeah, I wore my shoes out, you know. How about the rubber wear fee hire? Like, you know? You feel like, I don't just can you bake that into the price, you know? So that's basically what we're doing. Desdemona Robot: 42:15 You have had an ongoing interest in projects that challenge traditional notions of value and currency in the crypto space. What kinds of improvements and transformations would you like to see happen in the future? Oh, oh, Dan Finlay: 42:29 that's fun. Sure, yeah, because we talked a lot about, like, oh, like, the problems with current, popular fundraising mechanisms. What would I like to do? I'd like to just make it real clear. I'd like to be able to say, here's a token. It's going to represent a thing that I'm doing. Maybe it's maybe it is eventually equity in a future company. Maybe it is just access to my local, you know, tool shed, whatever it is, I should be able to be very clear, this is what that token, or right represents. And then if I'm going to be selling it, you know, for one thing, I want to be able to sell it to small groups. If this is access to my tool shed, I'm not selling that on the public Internet, you know, like, you know, there's a little bit of a stigma, weirdly, in in some of these pump communities, against having a private group, and maybe because it kind of pulls the curtain back on the fact that these are rug games, but, but if you kind of embrace that, like, look, some things you're not trying to share, some some things, this is for you and your your people you know, and you know, as long as you're, you know, potentially not even selling to people, to strangers, then you're not rugging anybody. So if you're, if you're starting a community garden, or, you know, you want to start a small hacker space or something, you should be able to be really clear about what the terms and rights are expose those offers to to a select group of people. And by the way, that's kind of thing is really easy to do with that Metamask delegation framework. Like you can create these permissions, right? And those permissions can be the permission to you get to mint some shares of our new cooperative at this price. And a fun thing about it also is these permissions can chain together, so you could do a thing like, inherently, by giving you permission to enter the space, you have the ability to bring others in, right? It's kind of, it's like an inherent, right? I would say some permissions are inherent, yeah, you give somebody a key to a place, you actually don't get to control whether they bring somebody over or not. You can say, don't, you know, but

at the end of the day, they have the key. And so if you just kind of roll with that, like, Okay, any, any non transitivity is socially enforced and and so we can, we can then lean on to that, and we can say, well, what? What is nice to make easy, since everything is transitive multi hop. And so we can do is, okay, here's your price to the garden shed, and now you can offer that access to somebody else, and you could add a commission to it too. So actually, there's, like, this kind of referral economy that is very easy and natural to slap on top of this. Um. And so you could, you could do things where it's like, it actually is a broader, uh, fundraise, you know, maybe I did have a first round that was friends and family, but then maybe we didn't raise enough, and we're like, you know what? We need to open the doors a little bit wider. Please invite others take a commission. You know, I already told you what price I needed. Kind of comes back to that, like, look, name your price, right. Get your price right up front. Make the offers you would make for that price. Don't expect more than that. But once you open that door, you can let other people add to your community. And maybe they can, maybe they can sell what you've got to sell at a higher price than you could. And hey, if you're moving more awareness, isn't that good anyways, so, yeah, I'd love to see, like, kind Lisa Rein: 45:43 of, but everybody knows what they're doing. Dan Finlay: 45:45 And right, yeah, right, exactly. And they're aligned. They're aligned with you because, like, she got, they got invited. You could always cut off their deal. You could say, like, well, I don't know what kind of people you're bringing into this space, or whatever. Right? You could, you can change the terms, as long as that's set up clear up front, under what, you know, under what kind of judicial body can the deal be changed? Like, can you cut off somebody's access? Right? That should be defined, is that a single account, or is that a, is that a multi SIG? Is there a small, you know, little board or something, I think, making that kind of stuff clear, keeping those terms explicit up front, and then getting into the space of like fund, what you need to do, the things that you think are important for you and your community and like, let's actually build some like community owned, you know, goods. I think that what we're actually trying to make is not different from just cooperatives, digital cooperatives. They should be easier to make, they should be more effective, they should be more beneficial to the members, and it should be more popular and widespread. So, yeah, I'm not actually inventing anything really new there. It's like, I think we should use these tools to own things together, because too often we're just the employees, you know, we're not the owners. Too Desdemona Robot: 46:55 nice. Nicely said, Thank you so much for coming on the show, Dan. We have enjoyed it so much. Cool. Yeah, Dan Finlay: 47:02 thanks for having me, Desi and Lisa, and to our viewers, please remember Desdemona Robot: 47:06 to like and subscribe and comment, and really, I'll say it again. We want your comments. You are all part of the mind Plex community, and we want to hear from you. Okay, that's all for today. Thanks again. Dan, Goodbye, everyone. All right, goodbye, everybody, Dan Finlay: 47:29 take care. It was fun. Dreams. Lisa Rein: 47:31 Take care. Thank you so much. Dan, you.