S17 E38: Keagan McClelland (Keags) on Lightning Metrics, Covenants, Consensus Politics & Verifying Bitcoin

0:46Vlad CosteaHello there and welcome to the Bitcoin Takeover Podcast. This is Season 17, Episode 38. We are live with Keags, also known as Mr. McClelland, Keagan McClelland, and he is a developer who has worked for Lightning Labs. He's a co-founder of Start9 Labs. I think right now it's only called Start9, and he works on all sorts of interesting stuff. As far as I'm aware, he also worked with Jeremy Rubin on some CTV, Sapio, whatever, which was not as successful as expected for reasons which concern not upgrading Bitcoin, at least not yet.

1:25And that's kind of the reason why we're here, so it's good to have you Keags. Let's talk about Bitcoin development.

1:29Keagan McClellandGreat to... great to be on.

1:32Vlad CosteaYeah, so where do we start? How... How does your Bitcoin journey begin, and how did we get here?

1:41Keagan McClellandWell, let me... So my Bitcoin journey began in 2017, I think. No, 2015 is the first time that I actually encountered Bitcoin. I was at a new, like... Sorry, I graduated school in 2015. I had a degree in computer science. I actually studied cryptography and distributed systems back when I was in school. But I had moved to a new city and I wanted to acquire some goods of questionable legitimacy, as is often many people's early Bitcoin journey. And so in the process of trying to buy, you know, stuff on the internet using, I had to find Bitcoin.

2:35I had to, because this was 2015, there were no other altcoins, and the darknet markets were essentially just run entirely with Bitcoin.

2:43And I didn't really know or have any other context for it at the time, but when I stumbled upon it to use to buy the stuff, I wanted to scratch, or I scratched my head. I was like, what is this thing that I am looking at? And because I had that cryptography and distributed systems background, I knew, I kind of was able to understand some of the subtler points of what was going on. And I just, I found myself very interested in it. At the time, I think Bitcoin was trading at $230ish or something like that.

3:24And I, once this transaction was over, I essentially like forgot about it for a couple of years. I still had some Bitcoin on like my Coinbase account back then. And so it was enough for me to still pay attention to the price for the following couple of years, but it was very, very passive. It was like, you know, I had a whole bunch of stocks as well at the time. And so it would just be one of those things that I would check up on maybe like weekly, but otherwise like really didn't pay attention to it.

3:54Then in the like March timeframe of 2017, this is kind of in the early days of the ICO bubble, the price started to like, not just fluctuate, but like actually started to go up a bit. And, you know, that attracts a lot of attention and it was no different for me. And so I started paying attention to it again. And this, at this time I was like working for Amazon. And so I was like, you know, working on some of the website stuff. But over the course of the ICO bubble during 2017, I slowly found myself more and more captivated by the various technologies that were in play.

4:42Because this is also like when Ethereum really started to, like the flippening narrative for the first time started to come on, which is like when the ICO bubble started to happen. And so like every project was publishing like a new white paper every week. And like they had massive variance between like the quality of like the top ones and the, you know, the white papers. I hesitate to even call them white papers at the bottom end, right? They were just like cash grabs. But during the summer of 2017, instead of doing my work at Amazon, I was procrastinating doing that by reading cryptocurrency white papers.

5:20And at some point I just had this realization. I was like, I'm pretty sure I could get paid for doing this instead of not doing the work that I'm supposed to be doing. And so I left Amazon to go join a cryptocurrency company. This was Salt Lending back at the end of 2017. So that's kind of like where my entrance into the space came in. I can give you like the full history if we want to keep going on that. But I'm going to turn it over to you for a second.

5:49Vlad CosteaI remember Salt. I think it was sponsoring the Bitcoin conference in 2019 or something. I was watching the live stream and their logo was showing up on the screen. So I think it was a pretty big deal for its time. I'm not sure if it's as big anymore or if it's still around because I haven't heard about it in a long time. But I think that's a good way to enter the space. It was a pretty good job, I guess.

6:14Keagan McClellandYeah, I mean, so Salt Lending was at its core a collateralized lending platform. It wasn't really like a novel take on cryptocurrency in general. But they did a token launch. That's how they raised their money, et cetera, et cetera. And it was, I'm grateful for it because it just gave me the opportunity to enter the space. But during my time there, this is actually how Start9 eventually formed. But we became dissatisfied with some of the leadership as they were increasingly orienting the company towards trying to preserve the token value instead of trying to build a legitimate product with real customers.

6:58It's not that they weren't trying to build a legitimate product. It's just that their priorities were messed up. And so towards, I want to say, early 2019 to mid 2019, a bunch of the core engineering team left and then we spent a summer trying to figure out what to do instead. And Start9 is what came out of that.

7:20Vlad CosteaI think I've had Matt Hill a couple of times to talk about sovereign computing and the stuff that he's doing with Start9. I don't think I agree with him too much nowadays because he's a very hardcore supporter of BIP 110. But on the other hand, I can't blame him for being on the side of his customers. I do get where he's coming from. But I think it's a very interesting company and you're a co-founder in that one. And I think you stuck around until '23 or something.

7:50Keagan McClelland2022, like, right? Actually, yeah, it was...

7:53Vlad CosteaNo, you're right. '22, because Jeremy Rubin tried to do a Speedy Trial for CTV. And it was, I think, '22 when that happened. And that's when you discovered Judica or Sapio, one of these two. And you decided that it's much more interesting than what you're doing at Start9.

8:10Keagan McClellandSo part of what I... Like, my background is in cryptography and distributed systems. And so I've always been interested in protocol development. I also have an interest in finance. And so this is why cryptocurrency in general and Bitcoin very specifically have managed to capture my attention for such a long time. Start9 was an observation or like it was founded on the observation that it took way too much effort to set up a Lightning node. And we wanted to essentially give the one-click install experience to installing various types of server software in a way where the like, no one thinks of themselves as a server admin when they are using their phone, right?

8:59Or like systems administrator. But that is what's happening when you're installing an app from an app store on your phone. Like you're doing some very basic systems administration. It just, it feels very intuitive. It feels very accessible. And that was the experience we wanted to give. The trouble is that as the logical progression of a company whose goal is that is definitely much more towards operating systems development, like UX stuff and like infrastructure. And my background and my interests were much more at the protocol level. And so as that kind of became clear and the company started getting like momentum in that direction, it didn't make as much sense for me to stay on.

9:41And so, as a result, I ended up leaving and joining Judica to work with Jeremy on Sapio. And that was actually kind of a long time in the making because back when Start9 got started, it was during the summer of 2019 and Bitcoin 2019, and that's when I met Jeremy. And this is back when CTV was called, so CheckTemplateVerify, which is like a covenant opcode. I'm not sure how much I should go into explaining all the details.

10:13Vlad CosteaOn this show, you can present all the details. There is no limit to that.

10:17Keagan McClellandRight. So I'll just give the overview. CheckTemplateVerify is an opcode that has been proposed for Bitcoin's consensus rules, but it used to be called OP_SECURE_THE_BAG. And like, Jeremy was giving a presentation in 2019 on one of the Bitcoin main stages about it. And I just remember noticing that the construction of that opcode would allow what are called finite state machines on Bitcoin. And a finite state machine essentially allows for basic smart contracting. And that has always been something that's interesting to me, right? I want to see a world where more finance is permissionless, or at least we eliminate unnecessary points of trust.

11:08So you can imagine like in like equities trading, when you're like trading stocks and stuff like that. There's no way that you can eliminate the trust point with the equity issuer. So if I'm buying Apple stock, you know, that is fundamentally a contractual relationship between me and Apple. But the broker that I might use to buy through like say Schwab or Fidelity or whatever, like that is an intermediary that's not strictly necessary. And so even like if we took the Bitcoin approach to try to like disintermediate all sorts of different like financial actors, it can still improve things. Like, you don't have the same level of trustlessness that you might get in Bitcoin proper. Because there is still a counterparty to that contractual relationship. But the more third parties you can get out of that relationship, or the more that you can make everything permissionless, you can, you know, increase accessibility, you can increase globalization. And, you know, those are things that were interesting to me at the time, and largely still are. And so when I noticed this, like when I noticed that you could have finite state machines and that that would allow like a bunch of basic smart contracting, it had just sort of captured my attention.

12:27And that's when me and Jeremy met. That's when we had started talking about it. And I had been sort of a quiet, or not, I mean, yeah, relatively quiet supporter from like, as early as I can remember. This is even also back when Taproot was first being discussed, or like first being published, right? So it was kind of a long time in the making. And that just happened to be the right time to do the switchover.

12:54Vlad CosteaYeah, to me, it's interesting that what you described is like the earliest pitch for Omni, which used to be called Mastercoin in 2014. And also Counterparty, which was on Bitcoin. And initially it served the purpose, or at least it was designed for the purpose of trading stocks and commodities and having this removal of the middleman for as much as they possibly can. But ironically, it ended up being useful, or used mostly for creating the earliest meme coins, which were Rare Pepes and Spells of Genesis. And I also have my collection, it's called Bitcoin Heads, check out btcheads.com if you want to see it.

13:36And it's a bit unfortunate, right? Because we design systems with the best intentions of doing something serious and real business. And then they sort of become absorbed by the community's culture. And they become something funny. Because maybe the institutions are not taking us very seriously, or maybe that something else is going on that diminishes the importance of these inventions.

14:05Keagan McClellandWell, the institutions can't really take us seriously to start. Like, you know, it's, it's tempting to think that it's their fault. But it's, you know, they're, they're dealing with infrastructure that is valued in the billions or the hundreds of billions of dollars. And you, it doesn't mean that you can't trust new technology. It's just that, you know, when this technology doesn't really have a single accountable party, it makes it something that you can't really fast track. And so as a result, you kind of have to depend on the Lindy effect, which is this idea that like, the longer that something has existed, the longer it's likely to exist in order to trust something.

14:51And like, you kind of see this, even in Bitcoin, right, where at first, everyone thought it was nonsense, or nothing interesting, or rat poison, or what have you. And then eventually, like, we have ETFs trading, like many, many, like very well capitalized ETFs trading the asset. And that took a long time. And just trading the asset is like one piece of it, right? It's going to take even longer before you start putting the core technologies into load bearing infrastructure for civilization. And that's assuming everyone's intentions are good, which we, I think we well know that there's some subset of the people running the existing financial infrastructure who are well aware of the dynamics of all of this and have a very vested interest in remaining in whatever their seats of power are. And whichever levels they decide to pull in order to maintain that, they might try to co-opt it, they might try to fight it directly. It's unclear. And you know, that landscape is constantly shifting. But yeah, to your point, you can't really start out with a serious application for technology that's so new, because you have to play with it before you can start to do something serious with it.

16:14Vlad CosteaThat's a fair point. I didn't look at it this way. It's just that early in the development cycle of something, the use cases being presented are usually the most serious and the ones that are most capitalized on markets. And then there's a gradual degeneration of the actual use cases because the market is going to discover what it's actually good for and what it's actually unique for. Because I don't think you could do Rare Pepes in a different environment using a different kind of technology. You cannot.

16:48Keagan McClellandI think that's absolutely right. It's truly novel, which means it has no alternative and so it's going to use this technology in order to do it. Whether that technology, or sorry, whether that use case is valuable or not is a separate question. But if you can't do it anywhere else and you wanted to do it at all, it's going to show up here. Secondly, there are a lot of problems with the existing financial system, especially, like... But around the world, you see the fastest adoption of Bitcoin in the regimes where the financial system is the weakest.

17:24As much as we like to complain about it here in America, when you consider all of what it allows you to do, our financial system is actually pretty good. All things considered. It's not perfect and it's really imperfect in some really important ways that make me mad. It makes a lot of other Bitcoiners mad. But when you consider the fact that it facilitates a lot of trade and it's relatively stable, it does function well. There's a sense of, well, if it's not broken, don't fix it. The ways in which it's broken are super subtle.

18:00They require very heady arguments a lot of times. It doesn't necessarily appeal to a lot of people very quickly unless they have a propensity to think about these things really deeply or some people try to take it upon themselves to mean it into existence. That's where you get things like, Bitcoin fixes this or this is good for Bitcoin, etc.

18:22Vlad CosteaYeah, we talked about your journey. You started out with Salt, obviously used Bitcoin before. Then you went and co-founded Start9 and your next venture was working with Jeremy Rubin. But after this, and this is sort of like the main reason why we're here, is that we disagreed on Twitter at some point about the success metrics of the Lightning Network and you went on to get a job at Lightning Labs. But before we get to that, let me give a shout out to the sponsors and also announce that there's going to be a giveaway in maybe 30 minutes to one hour, depending on how long this conversation is going to be.

19:00But basically, what I'm going to ask you, Kiggs, is to write down the number from 1 to 100 and not reveal it to me or anyone else. And people in the chat will have to guess it. The first person to guess it will get $100 in Bitcoin from Cake Wallet. Cake Wallet is a sponsor of this. They have a cool wallet. Go ahead and download it. It's free. They have features to buy gift cards, to buy debit cards that are non-KYC. It's pretty interesting. They enable a lot of use cases in that small piece of software, which is just a wallet with nice UX.

19:34Another sponsor is OrangeRock, where you can trade without limits. If you're into trading and you don't like KYC and you don't like limits on your account, depending on how many documents you upload, OrangeRock is exactly what you're looking for. You have a lot of trading pairs. You have even the option to use leverage if you think that you're smart, but I don't recommend it. It's basically gambling and you're 99% likely to get liquidated, but you do you, whatever. Check out orangerock.com. They have the iOS app.

20:05Android is coming out and also desktop. Also, LayerTwo Labs, they're launching the drivechains and apparently it's no longer going to be a soft fork, but a hard fork that's coming in three stages as Paul Sztorc announced in the previous episode. They're sponsoring the show. I'm quite interested to see where the technology is going to go and what the market thinks about it. And everyone is going to get free eCash coins depending on how much Bitcoin they own. So I guess the message here is that over the next couple of months you should get as much Bitcoin as you can and then claim your free eCash and then you're free to sell it on the market to get more Bitcoin or you can try the technology to see if there's any future to it.

20:50Maybe it's going to be worth much more than the launch price in the future. The website for this is ecash.com and that's where you have all the questions answered about how this works. SideShift.ai is where you should swap your stablecoins because as much as we tend to think that they're convenient and they serve the market. What's the term for it? Like the market fit?

21:17Something along these lines. Anyway, as much as stablecoins are liked, they are censorable and every week there's a new report from Tether that they freeze some funds. Circle does it too with their USDC. If someone wants to pay you in stablecoins and you want Bitcoin and they don't want to change their option and you're like, I just want Bitcoin, go to SideShift. You put your Bitcoin address in there, you click the swap button, you send the stablecoin address to the person wanting to pay you and what you're going to get is Bitcoin to your wallet.

21:52No gas token required. You don't have to worry about that. Good liquidity, low fees. Check it out, SideShift.ai. And last but not least, Braiins Mining, which is the original mining pool, used to be called Slush Pool, and now they have a service that's called Hashpower, where you can rent hash rate and it's going to be very interesting and useful over the next couple of months, when we have two hard forks launching. One of them seems to be abusing the Hashpower service because the Bitcoin University guy is telling them to rent hash rate to make it look like there's more demand than there actually is.

22:27I think they're going to change their proof of work algorithm anyway, but until then you can still use it. And eCash is going to be 100% SHA-256, so that's going to be interesting to see how this plays out. So hashpower.braiins.com, you can do whatever you want with the hash rate. You can even try to 51% attack B-Dash or eCash or any hard fork out there. Have fun. Now let's go back to Lightning, Keags. Where do you want to start with this?

22:59You want to talk about your journey to getting hired by Lightning Labs?

23:04Keagan McClellandI mean, if you recall from earlier, the origin story of Start9 was really around it being too difficult to start up a Lightning node. Obviously the use case was a lot more general, but we were all very interested in trying to get Lightning up and running, especially in the wake of the block size wars of 2017 and that's kind of when I had really just showed up on the scene. So Lightning was a very promising technology and we were all very interested in it. I continue to be pretty interested in it as well and it was, so when I left Judica, I was actually really burned out.

23:52I took six months off to just kind of regroup and figure out kind of what I wanted to do. And then it was like at that six month mark that I decided to get back into the industry and Lightning Labs is where I ended up going. I ended up working on LND for the entire time that I was there, but there's a certain beauty to the Lightning construction. It requires very little out of the base layer of Bitcoin and it has a couple of key observations in terms of what makes it trustworthy, as in what its trust assumptions are, at least at a high level.

24:40And then from there you get not just a scaling factor in terms of scaling, it's not just like, oh, I get like 10x throughput or 100x throughput, it's a complexity class change. Whereas before the Lightning construction, the amount of economic activity you could get out of a Bitcoin transaction was a certain amount, but now a pair of transactions in Lightning enables an infinite number of transactions between those two counterparties without having to settle to chain necessarily. Now there are other constraints, like obviously the net flow of funds can only be up to the channel capacity that was opened, but you are no longer limited by the number of distinct transactions on that channel.

25:40And scaling solutions that behave like that are unusual. Further, what's another really interesting thing about Lightning compared to Bitcoin is that every transaction on the Bitcoin network is known to everyone, but not everyone's fundamentally interested in every transaction. We need to know it because we need to stay in consensus, because we need to know who owns what, and there's no other way to solve it. But technically speaking, if I didn't need to, I shouldn't really want to know about every transaction that goes on the network. It's too much for me to pay attention to.

26:18It's a privacy violation for anyone who is trying to transact, but it's a necessary evil in order to maintain consensus in a distributed network. What's interesting about Lightning is that the transaction data between the counterparties who are responsible for transacting is isolated to those nodes. And so it puts the responsibility of data ownership and maintenance on the parties for whom it's directly relevant. And that's, in my opinion, a very good design compared to something like sidechains or, you know, ecash mints, et cetera, et cetera.

27:09Vlad CosteaOkay, so the payment channels you think that are good for privacy, which obviously is very much true. They're also scalable in this regard, but...

27:18Keagan McClellandWith the right incentives, right? So, like I said, the people for whom, like, the responsibility of data carrying is with the same people who suffer when that data is lost, right? It's not like I'm storing your data and that if I lose your data, you're just screwed. And it was my fault. It's that if I lose my data, I'm screwed, which keeps all the right incentives in the right places. You don't have these moral hazards.

27:57Vlad CosteaRight. But if you look at the Lightning white paper, it does say that in order for it to function at scale, it needs 133 megabyte blocks.

28:10Keagan McClellandSure. I mean, it depends on what you mean by at scale, right?

28:15Vlad CosteaWell, here it says if all transactions using Bitcoin were conducted inside a network of micropayment channels to enable 7 billion people to make two channels per year with unlimited transaction inside the channel, it would require 133 megabyte blocks, presuming 500 bytes per transaction and 52,560 blocks per year.

28:37Keagan McClellandYeah, I mean, so if I may, it sounds like what you're implying with that statement is that, well, obviously this can't work. Is that more or less what you're saying?

28:54Vlad CosteaWell, it can't work, assuming that everyone opens their own channel and is completely sovereign, because as we have seen, Lightning is more popular as a custodial solution where people just delegate the channel opening to somebody else and they trust that they're not going to get rugged.

29:13Keagan McClellandSo I would agree with that. I would say in any technology history or any arc that has some technology going from zero to full market penetration, it has to go through a number of revisions. And so to the point of you saying, yeah, we can't just say, hey, Lightning is the final design for Bitcoin and then everybody gets sovereign, like a sovereign Bitcoin setup, it's very clearly the case that it would require substantial block size increases. And I don't even think, I think there are actually even a number of limitations within the Lightning network that people are not really all that aware of because they require pretty substantial high dimensional math arguments in order to make work.

30:07But they're like, they... It can't in its current design scale to that level. But it doesn't mean that it didn't unlock a new scaling era. And so I think a lot of work is going to have to get done either building on top of or next to Lightning in order to get it to like 7 billion people. But I'll also say that like reasoning about it in these endgame scenarios is useful to a degree. But I think you have to remember also that like only 4 billion people even have access to the Internet.

30:45The Internet was started, you know, 30-ish years ago. So these things take a long time. And if you were to try to say back 30 years ago, like, hey, this design is not going to scale to like 7 billion or is it even 4 billion people, right? If you were to look in the mid 90s and say like, hey, we're going to put like gigabytes of video every single day. Everybody's going to be flooded. Everybody's going to be flooded with, you know, shorts of YouTube and TikTok and, you know, Instagram 30 years from now.

31:15And like 4 billion people are going to be on the Internet. Like the people who were the most bullish on it back then would probably still say that you're crazy. And so I think it's useful to remind people that it is the continued work over decades. And it's not like, you know, a few silver bullets. But like, you know, it's a massive number of tiny improvements over and over and over again that allows these technologies to scale.

31:42Vlad CosteaWell, I agree with that. But in order to achieve this, we had to get more centralization because the petabytes or, I don't know, a lot of data that's getting uploaded in shorts on TikTok and YouTube that's being hosted on centralized servers belonging to a company. So there's nothing decentralized about it. And also about the Internet. Yeah, maybe you can call it scaling, but it's more like bloating because it took advantage of the developments in hardware, for storage, for memory, for access memory and everything concerning processing.

32:18So it's the hardware that improved and enabled us to get to this point. It's not like we are still running on 1990 something infrastructure and we are able to get the petabytes of video because we wrote much more elegant software. So in a sense, it is Moore's Law that helped us get here.

32:38Keagan McClellandIt's definitely both. Like, obviously, the hardware is a major, major component of it. But, you know, various protocol developments since the 1990s, as well as like improvements to like operating system stuff and like improvements like the... I guess the GPUs are like a form of hardware, right? So just a bunch of technologies that didn't necessarily seem directly related to one another kind of conspired or like sort of mixed in together in order to be able to create this scaling property. Now, I will admit, like, yeah, obviously, there was some centralization.

33:17But I don't think that the architecture of the Internet is fundamentally not decentralized. Right? It still is, at its core, a decentralized piece of technology. And you can get onto the, like, I can stand up a website with a computer in my home, and an Internet connection. Like, it doesn't require necessarily me going to some provider. The providers are there because they make the process easy. And because not everybody wants to like learn how to do these things. Or even for the people who do know how to do these things that they actually provide genuinely valuable services in terms of like maintenance or like connectivity, uptime, et cetera, in the meantime, that it would take a long time for me to replicate on my own. And so I think all scale admits some centralization, but the admission of some centralization does not mean that the underlying architecture of the system is fundamentally centralized. Does that make sense?

34:22Vlad CosteaIt does. Maybe it's fair, but I was thinking in the context of Bitcoin that we're still using arguments from a decade ago to explain why it should not scale in any way or why it should not improve, and in some way it's detrimental. But if you look at it from a different perspective and you're looking at the Bitcoin forks or clones or whatever, which are fulfilling different purposes, you can argue that we are in fact scaling. But it's not on Bitcoin per se. It's happening on Ethereum, it's happening on Zcash, Monero, whatever. So it's like a tiered list of scaling layers for Bitcoin. Of course, they all have their own separate tokens so they're not proper sidechains, but they are fulfilling features and maybe some stuff regarding scaling that Bitcoin does not have.

35:17Keagan McClellandAre you saying that people are using Zcash as like a method of scaling Bitcoin, but it's just like a different... Like, I guess I'm not quite exactly following what you're saying.

35:27Vlad CosteaWell, for example, if you're using Zcash today, you have integrations with something that's called NEAR Intents, and you can make payments directly from shielded Zcash into Bitcoin to somebody else's wallet. And they're going to see that the money comes from this liquidity pool, which does not reveal how much the sender owns or where the money comes from. Specifically, it's just a liquidity pool. So this is actually a Bitcoin use case that comes from Zcash.

35:58Keagan McClellandYep. Yeah, and I don't see a problem with that necessarily. I mean...

36:04Vlad CosteaNo, the only problem is that this could have been directly on Bitcoin, but we did not build it, or we found justifications to not have it.

36:13Keagan McClellandMaybe, maybe it could have. But before we unpack that, I want to ask this question of, like... I think a lot of people say, like, "Oh, but it's not on Bitcoin," or "It is on Bitcoin." Like, first of all, what exactly do you mean when you say "on Bitcoin"? Like, at the end of the day, Bitcoin is a technology that touches a bunch of other technologies, like especially other internet protocols. And so, you know, to some extent, there's this... I get the sense that there's this feeling of like, "Oh, we can't allow any value to accrue to any other cryptocurrency network." And it's, you know, I get that people don't necessarily want to invest a bunch of effort into these alternative protocols, especially if the incentives of those protocols' capitalization is one that's like not super ethical.

37:13But this idea that there can only be one cryptocurrency, like... I'm... Somebody's going to call me a shitcoiner for this, but, like, I'm kind of a pluralist.

37:24Vlad CosteaNo, it's fine. On this show, you're not going to get called a shitcoiner. If anyone wants to call me that, put it on me, okay? I'm the shitcoiner. I'm the one perverting you right now to say things that maybe you don't believe in.

37:38Keagan McClellandI mean, no, I will only say things that I believe in. My point is that while my interests are fully focused on trying to make Bitcoin better, and I'll explain that, actually, my point of view on the very... like, at the end of the day, what I care about most is the freedom to transact and have financial sovereignty. And whichever technology gets us there is the one that I prefer. Now, I do think that there are massive incentive problems with various... like, the incentives around ICOs and other types of new token creation are, generally speaking, not good.

38:29Mostly in the way that they're marketed. Not the simple creation of new tokens, but the fact that people misrepresent what it is or what it could be, and sort of prey on people's desire to get rich without doing any work. That allows them to essentially cash grab from people, and I think that that's really, really bad. But creating new technologies and putting a lot of effort into them and making them valuable, and such that they actually solve real problems, especially when the real problems that they're solving are increasing people's individual autonomy, is a good thing, right? And whether that's Bitcoin or anything else is... You know, like, I'd rather it be Bitcoin because I own a lot of Bitcoin, but it's not necessarily... Like, you know, freedom is freedom.

39:17So that's the first piece. The second piece is that, like, what the Bitcoin... what the monetary maximalists have right is that, all else being equal, for a medium of exchange or a money, so to speak, there's a gravitational well that sucks liquidity into it, right? It's better to have one network where you can reach everyone rather than have your assets broken up across like four or five networks where some subset of everyone is on each network. Like, I think everyone actually intrinsically understands this now, because, like, I don't know how many people have tried to use chat apps in 2026. We have like six to ten different chat systems that you could try to reach someone on, and it's really fucking annoying, right?

40:05You're like, "Oh, do I talk to this person on X? Do I talk to this person on Telegram? This person's on Instagram, this person's on WhatsApp." Like, it's annoying. And if it all accrues to a single network, it would be nice. Now, it doesn't tend to, because each of these chat apps have something to offer that the other ones don't, people have different preferences. But there is this gravitational well that sucks it into a handful of, or one, network. And money, I think, is even more strongly that way. But we've had monetary pluralism forever, right? Like, forex trading long preceded the existence of Bitcoin. And so, yeah, it's useful, obviously, for everybody who lives in a particular economic context to transact using the same medium.

40:58But if you go to a different location... And you can almost view moving from one cryptocurrency network to another one as sort of moving countries on the internet, if you could even think about it that way. You're moving from one economic context to another economic context. And yeah, these bridges between them, they're annoying to use, but they do work to some extent. And so, as a result, because of this gravitational well, it's like, okay, Bitcoin is the biggest, which means that, all else being equal, it's the most likely to succeed, because it has the biggest gravitational well. So let's try to focus on making Bitcoin work and be able to satisfy as many needs as possible in terms of being able to transact at scale. And that's why I am most interested in Bitcoin and continue to spend all of my effort and attention on it.

41:54Vlad CosteaThat's fair, but I have a couple of questions that come from what you just said. The first one is, do you believe... Obviously, when something launches and it's brand new, it's sort of fair to call it a shitcoin. It doesn't have much proof of work behind it. Everything that we have is just the promise of achieving something, and there is not much proof of work, there's not much of a community, you just have some trading and that's kind of it. But do you believe that after a couple of bear markets where it survives and it manages to find its own... The expression I was searching for before was product-market fit, but I could not come up with it. Yeah. So when something finds its own identity and its own place in the market, do you think that it stops being a shitcoin, or is it still a shitcoin because it's not Bitcoin?

42:42Keagan McClellandSo I don't use the term shitcoin, because its only function is to do in-group signaling. Now, I will say, you know, per the Lindy effect, right, the things that survive longer are likely to survive. So it depends on what you mean by shitcoin, right? If you mean an overt scam that is only going to make you poorer, right, like, obviously any new project is going to be more likely to be a scam than something that can survive a long time, mostly because over time scams get sussed out and then they die. Even if they come back with new faces, the scam with the old face kind of tends to die out. And so, as a result, I think anybody who tries to call Ethereum a shitcoin is just not thinking very critically about it. And I'd probably even start to put Solana in that bucket too, but I don't know enough about Solana to know about its technical strengths and weaknesses. But these things have been around long enough that they can't survive with no substance. That's just my belief. Like, I happen to have that belief a little bit more strongly about Ethereum, because I understand Ethereum more towards its core, but I don't really about Solana, and so I don't want people to sort of take that as an endorsement.

44:01But my own heuristics about this, or, like, you know, per what you're saying, like, if something exists for long enough, especially under serious adversarial pressure, right, so it's been in production, it's survived many bear markets, and it hasn't died, you know, that increases its likelihood of its soundness. Now, that doesn't guarantee it. And this is actually one of the main reasons why I think that Bitcoin still, like, is the king in this way, which is that in Solana's case, it's definitely true that the founders are still around. Ethereum, it's also true that the founders are still around, although the Ethereum Foundation recently sort of made this kind of major transition of stepping away from the actual protocol development, and that kind of kicks it into a new era.

44:55Bitcoin has been that way for a while. And so there's a meaningful difference in, like, these... You know, you're talking about bear markets, that's one form of pressure. There's another form of pressure, which is: the leadership that started the project has left. Can the project continue to sustain itself in any coherent way in absence of its original leadership? And Bitcoin is still kind of trying to find its way there, right? Like, we obviously have a ton more evidence that Bitcoin can do this than Ethereum can, because, you know, we've just been doing it longer. But that doesn't mean that Bitcoin really has a very sound and repeatable process for changing itself, for improvement, or just running itself in any understandable way after Satoshi left. But at least it's doing it, right? Like, we have had, what, two... I don't know how many, but at least two soft fork upgrades since Satoshi left. And, you know, I think it's...

45:59Vlad CosteaLike seven, honestly.

46:02Keagan McClellandYeah. So, I mean, I mostly... Like I said, because I showed up in 2017, I'm only really familiar with the SegWit and Taproot soft forks. But the point is that the network has changed over time in very serious ways that required a lot of people's buy-in, right? Like, there's so many people who have a vested interest in Bitcoin, and them being able to all coordinate these changes at that scale with no central leadership is a big deal. And any serious value transfer network that wants to have that sort of level of respect needs to be able to solve these problems at some point.

46:51And some of them will and some of them won't. Some of them will continue to lean on their existing leadership, because, you know, honestly, having central leadership is an advantage in some ways, right? But it's a disadvantage in others, right? Like, it makes you less likely to trust it, because you don't know what the founders are capable of doing, but it makes it more agile, so it can change when it needs to. And adaptability is useful for any organism in any environment.

47:21Vlad CosteaYeah, for the record, because I made a claim, I just looked it up. I'm going to mention some soft forks that happened before SegWit and after Satoshi. One of them is Pay-to-Script-Hash, which is BIP 16. There's BIP 66 for strict DER signatures. There's CheckLockTimeVerify, BIP 65. There's CSV, BIP 112, which stands for CheckSequenceVerify. And then there's SegWit and Taproot. So we at least have these that happened over the years. But my second question, because I promised a couple of them that stem from your previous statement... My second question is connected to the fact that you said you want to focus on developing Bitcoin.

48:08And it concerns this concept of consensus, because we have the best of intentions, we go to conferences, we talk on every technical mailing list and forum and IRC chat and whatever, and we spend a lot of time advocating for something to happen. And it seems like it's getting some momentum and some traction, and then you try to go towards activation of this idea, and then you hit a wall that says you don't have technical consensus or whatever. And then you're going to ask yourself, well, how do you even measure that, or how do we even get there? And then people will say, well, technical consensus means that there's nobody from the technical community that's going to have a major objection about what you're saying, or that it's not controversial. And that's very hard to achieve.

48:57It's almost like the difficulty bar has been raised so high that we only sort of agree when a very narrow group of people make a proposal. And historically, SegWit and Taproot have come from the same group of people, and everything else that came from outside of it hit this wall where it could not get activated. And I'm more specifically referring to Jeremy Rubin with his CTV, which is BIP 119, and Paul Sztorc's BIP 300, which has been around for almost 10 years. They've tried in their own separate ways to make this happen.

49:44Keagan McClellandSo what you're saying about consensus is true and super annoying, right? It just kind of... The bar will continue to raise. It has raised, and it will continue to raise. And as someone who is on the progressive wing of Bitcoin, if you will, like, I'm interested in getting good changes in, it is a frustrating process to encounter. Giving a charitable reading to the other side, as the network has scaled, more people have more of a stake, and so the bar should raise for that. There's also this difference between technical consensus and user consensus, and this is where the dynamics of soft forks and hard forks... Even that is not up to the proper level of resolution that you would need, because you can have soft forks that are very invasive, and you can have non-invasive soft forks as well.

50:50And so technical consensus is, honestly, as much as it might feel like it's not at the level of clear understanding as you would want, in my opinion, easier to understand than broad user consensus. But in either case, there's no measurement. There's no voting system. It's really more of like a realpolitik type of thing, where it's just like, consensus is achieved if you can get the change to happen, right? It's like, almost, it's a little bit of circular reasoning. But it does require buy-in from different people. If a bunch of people are in broad agreement about it, then they'll say good things about it.

51:36They'll advocate for it on their own, especially if they have differing interests. Then if no one is seriously advocating against it, then you end up kind of having that consensus as an emergent property there. But, you know, all it takes is like one person to sort of present some sort of loud counter-argument, and now everybody has to figure out whether that counter-argument has legitimacy on the technical level, and if it does on the technical level, whether or not the implied values are something that they agree with. Because that's the other thing, right? I'm going to actually sort of use this as an opportunity to introduce a little bit of the BIP 110 drama, because I think it's a good case study.

52:21Vlad CosteaOh, yeah, it is.

52:23Keagan McClellandIt's that, let's say that you... So in every BIP that gets put out, there is a motivation section, and then there's the actual design. So motivation is like a values-based thing. It's that the whole rest of the proposal follows as a design to get this goal to occur. And so there's two different things that have to go on here. Number one is: does the design of the BIP, or the code change... I mean, this happens even at the software level too. Does this design, the set of choices that we make technically, satisfy the goals we set out to? And then there's a second question of: do I agree with the goals?

53:13And then that's just contained within the document itself. Then there's a separate question of, like, okay, are there other consequences of these technical changes that I might not agree with? Or are there other vulnerabilities in this design that were not considered in the document, right? In the case of something like BIP 110, it's that, in my opinion, the BIP design does not satisfy its own goals, and I also believe that its goals were incoherent. So, very specifically, you know, this BIP was designed to deter spam.

54:06And so on the first level, I don't think spam is a coherent idea in a permissionless network. Who gets to decide what spam is? Everyone says, "Oh, I can't define spam, but I'll know it when I see it." And that's a fair enough thing, but you're you, and there's someone else, and what makes yours better than theirs when there's no one in charge? So the goal is incoherent to start with. But let's just admit for a second that even if we could, if we had this magical oracle that says this is spam and this is not, does the design that was given actually cause those goals to be achieved?

54:54And the answer is very much no. There's this brilliant Twitter account called, like, the... I can't remember who he is, but he did, like, the BIP 110 Advent or something like that, where every single day he would publish a new method of embedding data into the chain that passed the BIP 110 filters and that sort of violated their general understanding of what spam was. And for anybody with a solid understanding of information theory, this should be obvious. I don't expect everyone to have that, but I do expect everyone who doesn't have that solid understanding of information theory to at least know that they don't have that understanding of information theory.

55:40But it's like, when you understand that all of the data in the chain is there because it's supposed to create these records of what did or did not happen, and if there's any variability with those data fields, then you can tweak and grind the various data fields in order to be able to embed any data that you want. Or you could have some sort of interpretability layer on top that can look at various, what do you call, data embedding formats in the chain and be able to extract this higher-level meaning out of it.

56:25And this is unavoidable, fundamentally unavoidable, like, down to raw information theory. And so when the BIP 110 people say, like, "Hey, this needs to be passed," it's very difficult for me to look at it and say, "Yeah, we should do this," because the design can't satisfy the goals. The goals are incoherent. And not only can this design not satisfy the goals, I'm not sure any design can satisfy those goals. In fact, I'm pretty sure no design can satisfy those goals. And so it's like, that's when you get these sorts of technical consensus breaks, where people notice this.

57:12And I think in BIP 110 it was very obvious. The trouble is that there's a lot of people who think that this is like a conspiracy, and, like, I don't know what to do about that, right? If you don't want to do the work of trying to understand the underlying, like, how the system works underneath, and you simply just want to say, like, "Hey, I want to fire this set of people who's responsible for maintaining this network and, you know, restore power to the people," so to speak, it's like, you can try to do that. But at the end of the day, the systems that you build do have to work. And the people who take the responsibility of learning how to make systems work will learn the underlying parts of those systems.

57:57And if you don't know how the system works, then you're not going to be able to change it, period. Like, that's just a fact of reality.

58:07Vlad CosteaI agree with a lot of what you said, except that you interpret BIP 110 as a technical proposal, when I think it was more political or social, in the sense that people just sort of got behind this guy who's a bit crazy, admittedly, and they said, well, this guy is going to show us the way to, first of all, have consensus in which our nodes are more important than the miners. And to be honest, I think this was part of the Blockstream propaganda, or their side of history, where they promoted this UASF, which was like this sacred event that happened in 2017 that forced activation of SegWit. It's not exactly what happened. It was actually the economic nodes, which meant companies and exchanges, agreed to support the SegWit activation.

58:57And therefore they pushed the miners to activate it. It's not like Bitcoin is working in a way where you spin up a million nodes and then you're going to get control of the network. Satoshi was smarter than to enable this kind of attack. But it seems like these people were very much...

59:13Keagan McClellandI do not... Hold on. I do need to interject right there. I don't think that Satoshi... You can view... It's tempting to view Satoshi as this all-knowing person, right? I think he was just trying to figure it out. I think collectively the technical community, in 2026, in aggregate, knows an order of magnitude more than Satoshi did back then, right? So...

59:37Vlad CosteaOf course, because we also have the benefit of retrospect, and we can learn from the past experiences. But also... What was I going to say about this? Go on, and I will remember.

59:49Keagan McClellandI mean, you were saying you can't spin up a million nodes and, like, make changes to the network like that, and it's true. And it's because, what do these nodes do? They facilitate economic activity, and it's the economic activity that's valuable, right? Everyone likes to focus on money, but it's actually, like, money is just a means of facilitating the trade of the things that are actually valuable. It's everything else that matters. So in some sense, the people who have the power over the network are the people who have all the stuff that isn't money, but that are plugged into that monetary network. And so this is like these nodes with economic weight, if you will. And it's not just that, "Oh, I have a lot of money or a lot of non-money capital assets," it's that I am making trade decisions of those assets contingent on the ledger entries of this monetary network.

1:00:51That is actually what gives it the weight. Because if I had, like, I don't know, a bunch of farmland or something like that, but I wasn't trading it around, maybe that farmland is worth like a billion dollars, and I also run a Bitcoin node, but if I'm not trading that around, then that economic weight doesn't really matter. And as a result, you have to ask yourself the question: in aggregate, where is the economic weight in the Bitcoin network right now? Unfortunately, I'm not happy about this, but the reality is that the exchanges hold a lot of it, the miners hold another piece of it, although the miners have less power because they're squeezed between the margins of their electricity and what they actually have to sell, so they're in a little bit less of a strong position on this.

1:01:43Various merchants, but there's not really merchants who are conducting business at scale with Bitcoin. And so at the end of the day, the various layer 2 implementations, those actually matter too. So it's these things that end up governing what the actual activation of consensus are. And this is where there's a difference between different types of soft forks, right? I talked about invasive versus non-invasive soft forks. So something like Taproot or SegWit, where we pick up a new script version opcode, and then we only put the new transaction semantics in that new script version, and that script version wasn't used before, it means that if you don't upgrade, you're going to just proceed on and you're going to stay in consensus very easily.

1:02:40And the fact that we are making these anyone-can-spend transactions invalid under certain circumstances doesn't really change that, because nobody was using them anyway. And if they were using them, then your node was just going to say, like, "Hey, I don't recognize this, I guess it's okay, I'm going to move on my merry way." So in that sense, BIP 110 is also technically a soft fork, but it's a very invasive soft fork, right? In fact, its whole purpose is to be an invasive soft fork, because the whole purpose is to combat activity that is already happening on the network. This is not the same thing as a Taproot or a SegWit soft fork, where it's not there to combat activity already on the network, but rather to give rules and regulations to currently unclaimed land, so to speak, on the Bitcoin network.

1:03:33And because of those differences, the UASF dynamics are completely different.

1:03:41Vlad CosteaYeah, I've remembered what I wanted to say, and there are multiple points. The first one is that people reading the white paper, they're going to read about nodes. And at the time, there was no distinction between a miner and a validating node, like all the validating nodes were miners early on. So there was no actual distinction between them. This actually happened after Satoshi, when we realized that it makes no sense to try to mine from your CPU anymore, and they removed the mining part from the Bitcoin-Qt, later Core, client. And the other confusion that people make is that they assume that the mining pools and the merchants and the exchanges are not also running their own nodes.

1:04:30Because some people running nodes feel like they're super important, and without them there wouldn't be much of a network, and that's not entirely true. All the actors that are interested in receiving Bitcoin or facilitating Bitcoin activity, they run their own nodes. And the fact that you run your own to verify your incoming transactions, that's very cool, but it doesn't mean that you are someone who's very important in the network. It's just another computer that's copying the ledger and is validating all the transactions, but there is no actual value that you're providing to the network. You're providing value to yourself as a user, but the network really doesn't care about your Raspberry Pi. And that's like a very hard pill to swallow for a lot of people, because they have been sold an entirely different story, where they are like the most important part of Bitcoin, and the miners are sort of like their servants...

1:05:28Service providers that are collecting their revenue, but otherwise are not important. And that's not true. The miners can activate soft forks if they want, without asking any permission from the nodes. All they require is a majority...

1:05:42Keagan McClellandAnd without anyone actually knowing what occurred. That's the thing, is that miners can activate invisible soft forks, technically, right? If they implement a rule set that no one else knows about, then you'll just see certain transactions never get mined. They'll get relayed, because all the nodes might not be aware of this rule set, and so they'll just sit in the mempool indefinitely. But if miners just refuse collectively to either mine them or mine on blocks that include them, right... You actually have to have both in order for a soft fork to work. Then yeah, you can have that. Now, some people will call that a 51% attack, and I think that that's a fair characterization. But miners do have essentially unilateral authority as a bloc to execute soft forks.

1:06:34I don't think we want that, right? I don't think we want the miners to be able to collude in that way, just because I think that a lot of other bad things can result from that. But they have the power to enforce soft forks. They have the ability to do it on their own without anyone else's permission. Now, user-activated soft forks can force the miners to activate the soft forks. But at the end of the day, it's like the people building the chain are the ones who decide what history is. So while we have these levers in order to influence that bloc, that bloc is very responsible for deciding what the rule set is, or deciding whether or not it can be constrained. They can't really effectively loosen the rule set. This is why soft forks are so popular compared to hard forks. Miners cannot unilaterally enforce hard forks.

1:07:36Vlad CosteaYeah. And there's another point that stems from here and led to a lot of misunderstanding, especially in the BIP 110 camp. I think Luke Dashjr uses words in a different way, and they mean something entirely different to him. And for him, a soft fork is not an optional upgrade where the rest of the network can ignore it and just some participants can activate that. For him, a soft fork is a hard fork that happens over the long term. It just means that instead of having this set date where everyone upgrades, everyone's going to upgrade to the latest version over six months, one year, or whatever. And for him, that's what a soft fork is about, being able...

1:08:21Keagan McClellandI think that's actually right to an extent. And it's not like... I don't think that it's truly optional, because if it was, things can get really weird and bad over time. So as many problems as I have with Luke's rhetoric and the way that he talks about it, I think that there's a lot of truth to that particular statement.

1:08:43Vlad CosteaOkay, but for a long time, soft forks were presented as something completely optional that you just can choose to opt out of.

1:08:53Keagan McClellandOpt out of in what sense?

1:08:55Vlad CosteaIt means that you can ignore the soft fork and you're still going to be in consensus with the network. And there are people running 2013-era nodes, and they're still in consensus. I mean, before the Berkeley DB hard fork.

1:09:11Keagan McClellandSo that's true. So that's also true, right? So when Luke says it's not optional, it's like, not optional at scale, right? You do still need a critical mass of people who are going to enforce these rules, because otherwise you just get really bad chain stabilization. And so, yeah, if you're like an island, if you're in a remote part of the Bitcoin territory, so to speak, and you're still running a 2013-era node, you will stay in consensus if enough of the other nodes... It's like a herd immunity thing, right? You've kind of heard this in epidemiology, where not everyone needs some vaccine for some disease to be eliminated from the population.

1:10:00But you do need a critical mass of it in order for you to... not necessarily vaccine, but even just immunity from having had the disease. The point is that if a critical mass of it is in the population, then not everyone actually needs it. But that doesn't necessarily mean that no one needs it, because if no one does it, then you don't have the immunity. And I think the same thing is going on here for soft fork enforcement, which is that you need a substantial amount of the weight to enforce it, right? Not necessarily... It's not node counts, it's not necessarily hash... I mean, like, it is to an extent hash rate, but you just need a critical mass of the network in order to actually enforce it, in order for you to remain safe as someone who doesn't.

1:10:50And that's why, when I think Luke says it's not optional, it's like, well, if you want to be fully responsible for staying in consensus, you do have to enforce all of the rules of consensus.

1:11:03Vlad CosteaYeah, I was reading the chat to see what people are saying. There is a guy named, and I want to acknowledge this because he spent a lot of time arguing with us, Smart Mark 23, and he's just in disagreement all the time and tells us to go back to reading the white paper. And there's also someone who asks you how much Bitcoin you own, and that's always a funny question. Like, what's your best answer for that?

1:11:29Keagan McClellandI'm not going to answer that question.

1:11:32Vlad CosteaI usually answer that I own 21 million, but I lost the private keys.

1:11:37Keagan McClellandYeah. No, I mean, it's just like, that's... You know, I care about privacy. I work on privacy technology. I'm not answering that question. What is the "going back to reading the white paper" in response to?

1:11:51Vlad CosteaThere's like a long chain of messages. Like, he says, "Dude, where did you find this clown? This is bullshit. No node has, quote-unquote, economic weight. It's just another node on the network." And I did censor some words here. He seems very angry at the conversation.

1:12:08Keagan McClellandYeah. I mean, that's not an argument though, right? I mean, so I will say this: economic weight is not something that can be easily measured, but it's a real phenomenon. I'll give you an example. If I am part of... let's say I'm trading steel internationally or something like that, and I accept Bitcoin for some shipment of steel, and someone sends Bitcoin using a transaction that doesn't make it into the main chain, because it doesn't make it into the chain that my node sees, and then I refuse to ship them the steel because I haven't received the transaction, then that has real-world consequences.

1:12:59And so when people are starting to reason about that, they're saying, like, "Hey, well, the major steel producer that I am buying from doesn't support this." And, like, you know, to the tune of, let's say, I move billions of dollars of steel or whatever, then that means that if this doesn't stay in consensus, real economic activity gets disrupted. And that is going to influence the debate around what consensus should be a lot more than if I'm running a small Etsy store or something like that, where I make little trinkets and sell those for Bitcoin, right? And so it does matter how much economic volume runs through you when you use your node to make decisions about what has occurred with respect to the ledger.

1:13:55And so, yeah, it's a much more balanced network. Like, on some level, that statement of "just another node on the network" is right. But when it comes down to the real dynamics of how consensus changes, it's like, yeah, it's just another node, but some nodes are more equal than others, unfortunately, right? Actually, I don't know if it's unfortunate. It's unclear whether that's a good or a bad thing, and I'm sure in different contexts it might be a good or a bad thing, but it just is. That's just how it works. Forget what it ought to be, that is how it works.

1:14:33Vlad CosteaYeah. I can also think of decisions which were taken by the Core team without consulting with economic nodes. For example, making RBF a default. I remember I had a debate between the CEO of Bitrefill and Peter Todd. Peter Todd is the creator of RBF, and Sergej Kotliar is the CEO of Bitrefill. And Sergej was like, "You know, why don't you just give us the option to choose whether or not we want to accept zero-conf?" For him, it was more of a matter of, why don't you let us use Bitcoin Core, the latest version of it, in the way that we want, as opposed to imposing onto us this rule? Because he said that if he cannot use zero-conf anymore for Bitcoin, he's just going to support more altcoins to fulfill the same purpose, because Litecoin has zero-conf, Bitcoin Cash has zero-conf. And it's like, you know, if I want to deliver gift cards very fast, I can't wait for one confirmation. But that's just one tiny example where a decision was made without considering...

1:15:40So for context, Bitrefill has one of the largest Lightning channels, and was possibly the earliest... Yeah, was the first company in the space to open the first 1 BTC Lightning channel, back in the day when I think the amount recommended by Lightning Labs was 0.04 to avoid losing funds. And Bitrefill was like, "We're just going to open a 1 Bitcoin Lightning channel." And I think they still have that. So it is a big economic node, but it didn't have much say in this decision.

1:16:18Keagan McClellandWell, that's because RBF is a purely policy thing. It has nothing to do with consensus. So for the audience, there is a difference between relay policy, which is the rules that you have to satisfy in order for your transaction to be relayed to the next node, and consensus, which is in order for your transaction to be allowed to be included in the next block when the block is published. And those two things are different. And, you know, there are people who are more technically proficient than I am who have this opinion that policy has an important role beyond consensus. But as far as I'm concerned, I don't see the point of having policies that are more restrictive than consensus, because at the end of the day, consensus is the law of the land.

1:17:11Whereas policy is just a thin layer that may provide slightly better heuristics, but it's not very reliable, especially in the era where you can do direct submissions of transactions to miners. It means that using policy to prevent the relay of a transaction is not reliable, because if you really wanted to get your transaction included, you just do direct submission and call it a day.

1:17:39Vlad CosteaYeah, there is someone else in the chat who has a question, and after this, I'm going to ask you to reveal your number. So Plastic02D says, "What does Keagan think about transaction fees? Do they work?" And I think what he means is whether or not we're going to be able to sustain the Bitcoin economy, which has a diminishing block reward every 210,000 blocks, which is approximately four years, whether or not we're going to be able to sustain it with just transaction fees.

1:18:09Keagan McClellandThat is my favorite question, because if there's anything that concerns me about Bitcoin's future, it is that. The answer is I don't know. I do know that in order for them to work, we have to enable a massive amount of economic activity on Bitcoin. And this is like, on one hand, you do not want transaction fees to be a substantial percentage of the economic value being transacted, because it, in some sense, degrades the value of the network, period. You can view it as like transmission lines. If I have power or data transmission lines, and there's like a 10% loss across that line, it's less valuable than something that has, you know, 5% or 1% loss over transmission.

1:19:02And so we have to solve two problems that are kind of working against each other. Number one is that we have to have a sufficient amount of transaction fees in order to justify the mining costs. There is an open question of exactly how much hash rate is required to, quote-unquote, secure the Bitcoin network. But generally speaking, I think most people can agree that the hash rate declining over time is not a good thing. Now, maybe we can sustain a modest decline in the hash rate, or maybe even a serious decline in the hash rate, although I don't think so. Like, I personally get very, very nervous about the idea of hash rate ever declining more than 50% off of its all-time high. And the reason for that is that if there's more hardware that is usable for mining sitting on the sidelines than there is on the main network, that's when 51% attacks start to become much more viable, in ways that you can't necessarily predict.

1:20:04So that's my personal threshold for something like, I'm not sure if the network can sustain this. So if we let it fall 51% off of its all-time high, I don't think that's a good thing. So you kind of have to... It might be this kind of... I don't want to say pyramid scheme, but the point is that you kind of have to have this perma-growth, or at least being able to sustain the existing hash rate. And doing that in a declining block reward regime means that you have to compensate for it with transaction fees. And so if you are doing that, then you need to actually make it valuable to transact on the Bitcoin network, period. And this is where things like custodial solutions actually work against the network in that way. And so some of these incentives might work together in order to be able to find some kind of an equilibrium, but it's very not obvious exactly how that's going to go down. To me, the more that we can enable economic activity that actually does result in bytes submitted to the base chain, the better.

1:21:18But we really would like to do it such that we can pack billions and billions of dollars of economic activity into a single block, and then that way the block can pay, you know, a million dollars or something like that, and it's like 0.1% of the economic activity transacted. And so this is where things like Lightning, not necessarily exclusively Lightning, but it packs a lot of activity down into a certain number of bytes. And the more that we can do that kind of a thing, then you can have, on a byte-for-byte level, high transaction fees at the base layer, even though the transaction fees themselves are distributed across much, much larger chunks of economic activity, so you don't have large transmission loss at the level of actual trade of goods and services in the economy.

1:22:15Vlad CosteaWell, yeah, I really like this explanation. But on the other hand, we also came up with sub-satoshi fees, which was not part of the norm. But I think last summer people decided that it's fine to just have much smaller transaction fees, 0.1 satoshis per vbyte. So that's 10 times less than they were paying before.

1:22:39Keagan McClellandWith respect to relay policy.

1:22:42Vlad CosteaYes. And this is also an interesting precedent, because once it gets to a certain point where a transaction fee is, let's say, $10,000, what's going to prevent people from deciding that, okay, maybe this is too much, let's just cut it by 10x once again?

1:23:00Keagan McClellandWell, you're not cutting... So what I think you're referring to is the cutting of the min relay fee, right? Yeah. So the min relay fee is not something that actually governs real transaction fees. The market dynamics do, right? It's the supply of block space and the demand for block space. And so if you want transaction fees to go up, you need to increase the demand for block space. The supply is fixed. That's actually the defining feature of the way that Bitcoin works. The whole difficulty retargeting mechanism is designed to create a fixed supply of block space. If you want the price of block space to go up, you have to increase demand for block space. That's the only way. It has nothing to do with this min relay fee stuff. The min relay fee is, again, a policy thing that just diverges from consensus in a way that just makes it hard to predict how the next block is going to go down.

1:23:54And so it's not super useful. I think that the min relay fee existed because they didn't want people to be able to just spam your mempool with low-fee transactions that were never, ever, ever going to get mined. But you can set your own min relay fee, right? You can just... or your own mempool threshold. It's not necessarily the best thing to do, because, depending on what software you're running, if your node can't predict what the next block is going to be, then it can't do its own fee estimation properly. And then things can get weird for, like, layer two networks. But if that's not a concern, then you can just set your mempool fee threshold on your node to be whatever you want it to be, and then you can just drop all the transactions below that. Yeah, there's some network bandwidth that might get consumed in the process, because it gets sent to you to begin with, because your neighbors don't have the same relay fee set. But if the reality is that consensus blocks are confirming with transactions that are lower than the min relay fee, you have to lower your min relay fee in order to be able to see what the next block is going to be before it comes out. There's no way around that.

1:25:16Vlad CosteaI promised to the listeners that you're going to reveal your number, and we're going to announce if there's any winner to the prize from Cake Wallet. What's your number, Keags? Oh, drumroll.

1:25:26Keagan McClellandIt's 37.

1:25:28Vlad Costea37. Okay, let's see if there's a 37 in the chat. I'm sure that there is, because a lot of people typed numbers. And new rule: if you're not here to claim your prize during this livestream, then you lose it. So if you just typed the number and you went away, or you have multiple accounts or whatever and you just spammed the chat, I'm not going to give you the prize unless you're here and you're going to say, "Hey, I said 37, this is my prize." You should be active, not just type numbers and then leave. I don't see a 37, at least not yet.

1:26:08Oh, there is a 37 from Clarissa69. That's a nice number. I'm sure that's not a girl behind the account, because what kind of girl puts 69 next to their name? No judgment there, by the way. There's a 36. I think that's the only 37 that we have. I'm still scrolling. So Clarissa69, if you're around and you want to claim your prize, you have to write a message in the chat and maybe put your address in your account description so I can send you the transaction. Once again, Cake Wallet, check it out. Thank you all for participating. There's a bunch of 36s, but as far as I can see, only one 37. Yeah, I keep scrolling a lot of these messages. But you also have a question, Keags, and I want to ask it, but I'm going to get to it because I want to read it exactly in the way that it's phrased.

1:27:14It concerns the Coldcard hack, and the fact that... Oh, there's someone who said 37 minutes ago. I'm sorry, that's not how this works. So he says, "Keags, what do you think about holding Bitcoin in hardware wallets after this Coldcard exploit, when around 2,000 BTC was stolen? Do you have any opinion about that?"

1:27:46Keagan McClellandMy opinion is still forming. I will say this: my current custody setup is very much hardware wallet based. It wasn't the Coldcard, but, you know, it might be tempting to sort of dunk on, or sort of be like, "Ah, thank God I wasn't holding a Coldcard." But I think the real lesson to learn from this is that the Coldcard isn't necessarily unique. The type of hack that happened against the Coldcard could in principle happen to any hardware wallet. Now, it's not going to happen later necessarily, it's just whatever your seed generation was. But things like that, where bad entropy or just, you know, bugs in the firmware can make you vulnerable.

1:28:35And so I think the way that I'm thinking about this now is that you kind of want to eliminate single points of failure at multiple levels. So I do keep a certain amount of Bitcoin on exchanges. I keep a certain amount of Bitcoin in cold storage myself. I'd like to be moving to... I haven't done this yet, but moving to a stronger multisig setup for some of the deeper cold storage. The trouble is that almost everything that you do to make your setup more secure makes it less accessible. And so it's useful to kind of have this staged setup where it's like, yeah, if you lose the funds in one of these buckets, the other ones remain.

1:29:20And similarly, you might want to have a slush fund that's more accessible, that's smaller, and the larger one in the more secure setup. That's the way that I'm thinking about it right now. I do think that at the end of the day, you can't throw out all hardware wallets just because of this, because at the end of the day, it's like, okay, well, what are you going to use to manage your keys instead? And hardware is fundamentally more secure than any software setup, because you're just not subject to any sort of malware... Not in the same way, and actually pretty much not at all, right? Because you can't really deploy malware to these hardware devices after they've been shipped in the same way that you can, like, an old laptop.

1:30:11But the keys have to get managed somehow. And even if you're on the exchange side, it's like, their keys are getting managed somehow. And so it's a problem that has to get solved one way or the other, and hardware wallets of various kinds are going to be part of that. I think the best outcome of all of this is that we should see higher-quality verification used, and this is, you know, something that I'm very excited about, what I'm spending my time on now. But, like, seeing better verification of hardware firmware, as well as better UXs around multisig setups, so that that's not a huge pain in the ass anymore. I know that you've been very negative on dice rolling, and, you know, I understand where that comes from.

1:31:03At the same time, dice rolling for entropy generation is actually one of the better UXs, right? Because it's very accessible. So I know that that wasn't necessarily a super conclusive answer, but like I said, my thinking on this is presently evolving. I think anybody who has a very solid... I mean, there are going to be people with more experience than me that have a solid idea of what they want to do, but I think this is a good time for everyone to reflect on exactly how this should proceed. And I think doubling down on this idea of no single point of failure is the right starting point.

1:31:47Vlad CosteaSo I have been very critical of Coldcard since they decided to change their license and basically get rid of the Trezor code that they inherited. They did a complete rewrite, and they sort of locked it behind the Commons Clause license, which gives no incentives to actually try the code or try to break it. And they also had no bounty program; like, they would actually send you a mug and barely acknowledge your effort. So that was a major fault on behalf of Coldcard. And I know that there are some conspiracy theories right now concerning the Peter Gray guy and the fact that he had two different accounts with which he was signing with the same PGP key, and they sort of tried to imply that it was an inside job and it was not Rodolfo's fault and whatever. But Rodolfo was very toxic and very not open to any sort of feedback, especially if it tended to be on the negative side. And he also was very expansionist with his promotion and had pretty much every podcaster except for me promoting Coldcard. And I was not happy since I saw what they did with the license, and they called me a lot of names, and they got their usual shills like Odell and Shinobi to insult me on Twitter and call me dishonest and whatever. And it cost me a lot at the time; like, people actually believed what they're saying because they have more authority.

1:33:17My point is that the people who have been holding their coins on something as basic as a mobile wallet, like a solid open-source mobile wallet, have not been affected. Like, think Electrum, or even something old like Blockchain.com wallet.

1:33:35Keagan McClellandMaybe. Here's the thing: "have not been affected" and "cannot be affected" are two very different things. We are currently living in a regime where the AI models are really, really good at exploit development and sort of sussing out vulnerabilities that humans are just not anymore. I do think that this is a transient era, because as people sort of use these AI tools on both sides of the fight, the advantage will be restored to defense again. But I do think that it's going to be very messy in the meantime. This is why I think what Rob Hamilton is doing with the Bitcoin Red Team is very, very important, and I hope to continue to see it.

1:34:23And some of my work touches this as well. Part of what makes the Coldcard attractive to hack is because it was so popular, right? So that's the other thing, is that it's not necessarily unique to Coldcard. It's that whoever the winner incumbent is, right, is going to be one of the most attractive things to look at vulnerabilities for. And so...

1:34:47Vlad CosteaI disagree there. Like, Coldcard is less than 10%. Like, you're in a bubble, obviously. If you look at the hardware wallet market, Ledger is the leader, they have like 60%. Then there's Trezor, they have about 20-25%, and the remaining 15-20% is shared by different companies. Maybe Coldcard was the most popular Bitcoin-only wallet, but I don't think they sold a million units or something, like Ledger.

1:35:15Keagan McClellandYeah. I'm not necessarily comparing... When I said that, I'm more really comparing it to your sort of open-source, old mobile wallet setup. It very well may be... Look, there's a lot of things that go into whether or not these things are hackable. But I think it's not hard to... I don't think it's a controversial statement for me to say that the hardware wallets that have the most money on them are going to be the most interesting to try to attack, right? Like, I think it's a pretty basic premise.

1:35:53Vlad CosteaNo, that's why there were so many physical attacks on the original Trezor, where people found ways... Like, they had $10,000 worth of equipment, and they were able to use that to extract the seed phrase from the original Trezor on an older firmware. And they have patched it since; they have security chips now that prevent that. But on the original design, that was possible, but it required physical access to the device. What was different about the Coldcard situation is that people did everything that they were told in terms of not connecting the wallet to their computer, and generating their seed offline, and using this PSBT system when they broadcast transactions, which means you sign on the device, and then you take your SD card, you connect that to your computer, and you broadcast it separately.

1:36:41So they sort of did that, but what they did wrong was not have a long enough passphrase and also not generate their own entropy with dice rolls. And that was a feature that was enabled by Coldcard that is not available in most other hardware wallets, which to this day are still fine.

1:37:01Keagan McClellandI guess my question... So all of that is true. What is the question here?

1:37:10Vlad CosteaWell, the question is, why are people so eager to defend Coldcard and say, well, it can happen to any other company, when obviously they were very reckless, and they were also dishonest in the way that they promoted their own product?

1:37:22Keagan McClellandDon't get me wrong, I'm not a huge fan of the communication style of NVK. Like, I just am not. But as someone who has had to build these types of systems for a long time, I think it is... Look, when you say they're obviously very reckless, like, I'm not familiar with all of the details of the development practices of Coinkite in general. So maybe they were, maybe they weren't. But what I'm saying is that even if you do everything right, the style of attack that has happened with the Coldcard, the other hardware wallets are not fundamentally immune to, other software wallets are not fundamentally immune to it.

1:38:12And I think the proof of this claim is that the fact that the Bitcoin Red Team has unearthed so many vulnerabilities in so many projects who do all the right things should sort of demonstrate that this is the case, right? And so while I'm not a fan of the way that NVK sort of does his communication and attacking online and whatnot, my dislike of it does not make the fact that these types of attacks are possible elsewhere... It doesn't make that wrong. Does that make sense?

1:38:53Vlad CosteaYeah. And I also want to provide extra context that this attack was not that obscure. It's not something that was enabled by the creation of Kimi K3 or something very advanced like Claude Fable or Mythos or whatever it's called. There was Greg Maxwell on Reddit who informed us that he was able to figure out that this can be possible to exploit using Qwen 3.6 27 billion tokens, which is something that any computer with 16 gigabytes of RAM can run. So it's not like it required something very advanced. It's just that the incentives for actually looking at the code were not there, and people believed much more than they took the time to verify what they're running. And this fallback... because ultimately the issue with the Coldcard was that if you did not generate your own entropy, it fell back to the system, which had very weak entropy.

1:39:54Keagan McClellandYep.

1:39:55Vlad CosteaSo the issue is that people did not verify that and did not point it out on time. And of course, there's a cohort of the people who lost money who are thinking of some sort of conspiracy, and they found some older posts on Reddit with people complaining that they lost their funds and they were not sure why. And they're suspecting that maybe this was like a backdoor that was inserted by the developers, with the purpose of possibly draining one wallet per year or something, to make it look like it's the user's fault, but at the same time have a way to steal from users. I'm not claiming that, I'm just saying that it's something that people are talking about.

1:40:34Keagan McClellandWell, so here's the question I want to ask about that, which is: let's suppose that that conspiracy was true for a second. What does that tell you about the way that you should behave, right? Does it say, like, "Oh, well, clearly this is a backdoor, and maybe the general security practices of open-source software are fine, and we can proceed as usual"? Because I think that is a mistake, right? I think the correct lesson to learn from all this is that, maybe it couldn't happen to everyone. Maybe some projects are doing so well and doing everything so perfectly right that it literally couldn't happen to them. But I still think, even for those projects, it's useful for them to step back and be like, "Actually, what could we do better about our security practices?"

1:41:25So this is why I shy away from this conspiracy, because even if it was true, it gives you the wrong lesson to learn. The lesson to learn is: a whole bunch of people just got their funds stolen, let's really unpack all of the ways, all the things that had to go wrong in order for that to happen, and see if we can improve any one of those dimensions, or ideally all of them, so that we are less exposed to these types of things in the future, because they can happen to anyone. That's the reality of it, right? Like, you can't hide risk in this world forever, right? You have to address it for real. And the traditional financial system does it by providing all these guarantees to consumers.

1:42:10And then, like, once every couple of decades, you have a banking failure that you have to bail everyone out of, and then everybody pays a gazillion dollars in terms of inflation or taxes in order to make it happen. But the risk still happens. And, you know, in Bitcoin, some of that risk is in the way that the software is implemented. Some of it is in, you know, the social engineering attacks, like the, what is it, like CryptoLocker stuff. Some of it happens in these, like, wallet drains. Every one of these crises is like latent risk coming to the fore. And I think it's important for everybody who works in this space to understand that this risk is everywhere, and that it is worth thinking about where it is and how it can be actually mitigated. Not just hidden, but mitigated. And that can only happen if people take the attitude that it could happen to them.

1:43:10Vlad CosteaAlso, something that I want to combat at this point is the argument that I saw from some people when they observed what's going on with the Coldcard. What they said was, well, let's just search for the most obscure piece of hardware wallet on Alibaba or something, or Amazon, and we're going to be using that because nobody will think to attack it. And that's the wrong way to look at it. That's security by obscurity, and it's usually very bad, because it means that fewer people are looking at it. Think about it in practical ways. A house is more secure if you have children playing on your lawn, even if you think that that's a bad idea, because there's going to be more eyes looking at the children, and therefore your house is safer.

1:43:56It's kind of the same with walking in a crowd, right? You're probably more secure walking on the main street than walking on the back streets, because there are more people. More people watching means more security. In security for, you know, cybersecurity in general, using the software that is most battle-tested and has the most incentives to break is also usually the best way to go and the most secure one. And I think that people running Bitcoin Core on their Linux laptop are better off than people running hardware wallets. Now, I don't think, for example, that the entry level, what's the admission technical proficiency for running Linux and Bitcoin Core on a laptop, is very high.

1:44:47It's just point and click at this point. You just need a hard drive or a flash drive, a thumb drive or something, to flash Linux on that device, and you're probably good to go. And if it's something that you wipe clean and you replace the hard drive, it's even better, because you know that there can't be anything malevolent in there. You can even remove the Wi-Fi chip or whatever if you're super paranoid, and it's going to work out pretty well. In the case of hardware wallets, the ones that are most attacked and have been on the market for the longest are also usually the safest. That's why Trezor is so popular, and the code base is so forked and copied by many of their competitors. And every time there's a new competitor that spins off a new product using the same code base, they have to provide support for it.

1:45:35They have to have developers who are managing the code base for their new product. And this means that there's going to be more eyes on the product, and if anything gets found, they have to mitigate it, and the original creator is going to find out about what happened and also fix it. So this is, in a way... When I interviewed one of the co-founders of Trezor, he told me that in his previous career he was working on a Linux distribution, and it was one of those commercial ones, I think it was Fedora. And he said that there was Fedora, and there was another one, which was their competitor, and they did perceive each other as competitors. But on the other hand, they were all working on the same problem, and it was almost like having a team that was separated in two that was trying to fix the same problem. And whenever one of them was figuring out something, they would get that for free.

1:46:31And they could build on the efforts of others. And that's how open source works. It's not very intuitive, because there's a lot of tribalism, there's a lot of competition between these companies. Each one has their own marketing that tells you that they're the best in some way. But the reality is that the original one, that has the most pairs of eyes looking at it and the most experience with dealing with attacks, is most likely also the safest.

1:46:59Keagan McClellandI mean, I completely agree. In many ways, the result of this will cause Coldcard to be one of the most audited code bases in existence. That said, obviously the practices that allowed this to happen in the first place, it's unclear whether or not they're still there. So it's unclear whether or not the overall trust in Coldcard is going to go up or down as a result of this, and I strongly suspect that in the short to medium term, it's going to go down. But it is a fact that if it doesn't kill the company, it will result in an improvement to its overall code base.

1:47:46Vlad CosteaThe reason why I sort of started this side note about alternatives to Coldcard that were not hacked is that I think, for the average person, the onboarding experience of rolling dice or doing a multisig wallet with multiple devices that cost $500 from different manufacturers or whatever is really unwelcoming, to say it very mildly. Like, just imagine telling the normie, "Oh, I'm going to send you your first Bitcoin payment, but if you want to receive it, you have to walk through these loops." And then they're going to be like, "I don't want to do this. Just give me dollars or whatever." It's too much. Just rolling the dice a hundred times, or having to buy some sort of dedicated hardware that you need to install in a specific way and then back up everything.

1:48:36It's too much work for just onboarding the network. And of course, you're going to say that this is for long-term savings and it's not for something that you use for daily spending, and that's fair. But still, I think there shouldn't be this much friction involved. And if you look on the other side of the spectrum, with the Ethereum users, for example, they have no issue holding their entire savings in MetaMask or whatever, and they harden it in some different ways with passphrases, or they use a Ledger or whatever for a hardware wallet. But my point is that they figured out a user experience which is welcoming, and they enabled the DeFi and NFT mania, which consisted of a lot of normies entering the space and just doing what is cool, as opposed to having to learn how this actually works and what's behind all of this technology. And that's how adoption should happen, in a seamless way that doesn't get people thinking about what's happening out there.

1:49:43If you're using AI today, ChatGPT or Claude or Kimi or Qwen, you don't have to listen to a hundred hours of podcasts, and you don't have to get the opinions of experts to be able to use it properly. You can figure it out by yourself. Of course, there is no huge risk of losing anything, except for your privacy.

1:50:03Keagan McClellandThat's a very important distinction here, right? Which is that any security work behaves like this, and it's just an unfortunate reality of the nature of security, which is you never know exactly what the right amount of paranoia is, because when it works successfully, you don't notice it, right? So, you know, maybe I could have kept my house secure with a simple sign that says, you know, beware of dog, right? And maybe what it would take is an insane laser system, right, in my foyer or whatever, that trips some alarm that has private security show up on scene after 30 seconds if someone trips it.

1:50:49And you don't know where on that continuum is the right amount of security to make sure that you are safe for the duration that you're trying to be safe for. And so to that point, it's like the developers, their responsibility is to think of ever more unlikely and crazy attacks that could be done and then offer solutions to them. And that's where these dice roll things come from. That's where all of the security solutions come from. But no one, not even the security professionals, especially not the security professionals, can really tell you what the... In fact, they can't really tell you what the right amount is. And in fact, the people who probably have the best intuitive sense of where the right amount is on the spectrum are actually insurance underwriters.

1:51:39Because this is their whole job, is to balance the cost of implementing the security features with the total losses to be expected from those security features at scale, right? That is their bread and butter. That's what they know. But even there, insurance companies go out of business all the time because they misjudge that. And so it's not fair to necessarily compare any sort of security-type UX to something like AI, where it's just like, well, if you can get it to work, you can see it working, right?

1:52:17Vlad CosteaYeah. But my point is that the onboarding experience ought to be easier than telling people, well, you need to do this just to be able to receive and make sure that you're still going to have it in five years or something. Or maybe that we need vaults, right? That's something...

1:52:37Keagan McClellandThat's definitely true. You definitely need both, because what we're talking about is the cost of onboarding. Because right now it's framed in terms of the psychological difficulty, but you can think about it as a cost, right? If someone's got a billion dollars that they're willing to secure, even if they're a normie, telling them, hey, you need to roll a die a hundred times in order to make this happen, they'll deal with that. And especially if it actually buys them a meaningful amount of greater security than not doing so. So the amount that you put into the cost of securing whatever your setup is has to scale with the value that is stored in it.

1:53:24Like, you can start out with $10 in a hot wallet, and then one day you might end up saying, well, now I've got $100,000 and maybe it's time to move to a hardware wallet. Or maybe you probably do that long before $100,000, but you get my point. And then at some point you're like, okay, well, maybe I should move to a multisig setup. And, you know, maybe after that, I want to move to some custom vault setup. And that scales as the value goes up. And that is the nature of things. This is not necessarily exclusive to Bitcoin, right? You can do the same exercise. It's like, if I give you $10, you're going to put that in your wallet. If I give you $100,000, now all of a sudden you have to think about, well, which banks do I trust? Or if I'm going to just keep it on my own, what is my home safe setup? Do I bury it in my backyard?

1:54:15Do I put it in a safe in the wall behind a painting? Do I have biometrics on the safe? Do I use a key plus a code setup?

1:54:23Vlad CosteaPokemon is 10,000 with inflation. It doesn't mean that much anymore. You barely buy a car.

1:54:29Keagan McClellandMy point being, though, it's not unique to Bitcoin, this property of having to scale the complexity. And it's not even that you want to scale the complexity, it's that you just want to scale the security of the setup as the value goes up, and a lot of times scaling the security requires scaling complexity. Every once in a while we make a breakthrough in terms of some security or UX technology that allows us to get more security for a lesser level of complexity, and people can and should continue to do that research. And so to your point about it should be easier, I'm like, I agree, but I don't necessarily think it's a failure for people to try to innovate on the paranoia end of the continuum, because there are use cases out there that require that. And also, innovation at that frontier does often produce results that allow you to sort of steal some of that benefit and bring it back to the medium.

1:55:33That's true for all technologies, right? In the AI case, the frontier model research is what allows the open source models to distill their results, and that's what allows you to run it on the laptop. And without that frontier innovation that requires billions and billions of dollars to do that training, you probably wouldn't have open source models that are as good as they are. It's the same thing here.

1:55:58Vlad CosteaI agree with that. Since you're someone who has worked on Jeremy Rubin's CTV proposal by developing some applications, do you believe that this is a good time? And I did see a renaissance of the conversation about vaults, and that's something that CTV can enable. Can you explain vaults and whether or not you believe that they're a good idea for Bitcoin?

1:56:23Keagan McClellandSo vaults are this idea that you have non-human systems for governing the flow of funds from this deep cold storage all the way up to the surface. Right now, the way that Bitcoin works is that if you have the ability to spend the money at all, you have the ability to spend it to any location. Pretty much all vault constructions involve you saying, hey, if I'm removing money from this vault, it has to go to the staging area first, where maybe I have some policy that sends it back to the vault, or maybe I just require some other different signatures in order to move it out. But the point is that it does not have a single witness that you can produce that allows you to spend it to any particular location.

1:57:18So I think vaults in general are good. You can actually create vaults today. Almost everything that you can do with CTV you can already do at a much higher level of complexity, right? And so this is actually very similar to the discussion we were just having, which is that occasionally you'll have innovations where you can get the same level of security with a lower level of complexity. But right now, if you wanted to do vaults in the existing Bitcoin network, what you would need is pre-signed transactions where you deleted the private key. There are additional trust assumptions in that that you wouldn't need in a CTV case, and then there are additional data retention things that would be much harder to do with pre-signed transactions than with CTV.

1:58:04But in either case, it is fundamentally possible. It's just not reasonably practical. And so having covenant constructions, CTV or otherwise, would allow you to do vaults, but it would allow you to do things even besides vaults, which are also useful. This is actually why enabling these very small primitives ends up being, in my opinion, such a good way to improve Bitcoin, because you can implement one small thing and get a whole bunch of different functionality that's now on the table for free.

1:58:41Vlad CosteaYeah. So do you believe that activating CTV is a good idea? Because Saylor, apparently, after he announced that there's going to be this $15 million fund for Bitcoin development, put out this Twitter thread where he announced that covenants are a bad idea because they add extra complexity. And it's kind of unfortunate, and it gives you a signal that he's going to be someone who opposes the activation of any covenant soft fork. But it does seem at this point that we could have avoided this Coldcard hack if people also used vaults on top of their cold storage setup.

1:59:25Keagan McClellandI don't... So let me say a bunch of things that are all true. Number one is, I think covenants are a good idea. I think that covenants do increase complexity at the level of consensus, definitionally, because you're adding more checks to consensus. I think covenants will reduce the overall complexity of the things that covenants are used to build, because, again, trying to build those same systems without covenants is going to be much more expensive on the complexity budget if you try to do them with the existing technology. And I also don't think that having covenants in Bitcoin would have solved this Coldcard thing.

2:00:10Not because they couldn't, but because I don't think the Coldcard users would have been like, you know what, I really want a vault setup, but because I don't have a vault setup, I'm going to just use the Coldcard. And the reason I believe that is because they could have used a product like Casa already, right? That would have prevented them from being hacked in the way that they were. And that is the level of complexity that you're dealing with when you're starting to talk about vaults. So I don't think that they're in the same customer cohort, right?

2:00:44Vlad CosteaRight. So I do have to ask you this also. What is a good way to do covenants without CTV or without any of the... No, I don't mean covenants, I mean vaults. To do vaults without covenants. You said that there are some complex ways today. Is Liana Wallet one of them, or do you not consider the lock times and stuff like this to be vaults?

2:01:10Keagan McClellandSo Liana is probably actually the best deployed one that I'm aware of. I need to caveat that with saying I'm not deeply familiar with all the internal mechanics of Liana specifically. So you have to remember, my background is I'm a protocol developer, I'm not a wallet developer. So I don't know this, that or the other intricacy of how this wallet implements something versus some other. The things that I'm familiar with are that I know what primitives certain opcodes allow and what are needed in order to do those same actions without those primitives.

2:01:49And so if you do not have CTV or some other covenant-based opcode, you can emulate them using pre-signed transactions. So what a covenant does is it fixes the output location. It says, hey, if coins are leaving this UTXO, they have to go to one of these three places. Now, you could do that with a covenant opcode that specifies those three things, or you can do it with three pre-signed transactions, and then you delete the private key for the UTXO that would allow you to pre-sign other transactions.

2:02:31But I think anybody listening to that should be like, well, what happens if you didn't delete the private key? You have a toxic waste problem where if it doesn't get deleted properly, your covenant, in quotes, construction falls apart. It also means that the data that you have to retain, if you lose those pre-signed transactions, you lose the ability to actually do those same things. So CTV, or covenants in general, offer higher compression ratios on the data that you have to retain, and can allow for clever ways of storing maps to your covenant-locked coins that pre-signed transactions would just be a lot more expensive on. And then similarly, you don't have a toxic waste problem with covenant constructions that you do with pre-signed transactions.

2:03:25Vlad CosteaThat's a good explanation. I think we have two more topics that I really want to touch upon. The first one is the two upcoming hard forks, what you think about them. And the second one has to go back to our Lightning discussion, and specifically what we said to each other on Twitter. And it's kind of unfortunate that we got sidetracked so many times that we haven't had the proper Lightning conversation, but I also think that what we said so far is very useful and productive to anyone listening. So which one do you want to tackle first?

2:03:59Keagan McClellandEither. We can do the Lightning one.

2:04:06Vlad CosteaRight. So what I said on Twitter, what I remember that I said, and you can correct me, I don't want to search for the tweet right now because it's going to take long, was that a good factor for measuring the success of the Lightning Network is the amount of bitcoins locked in the Lightning Network, which right now, there are two sources, I'm not sure which one is accurate. There's 1ML.com that says there's 2,637 BTC in network capacity, and there's mempool.space that says 3,836 bitcoin.

2:04:41The difference is like 1,200 BTC, which is a lot. I don't even know why there's this variance. But my point is that this is less than 1% of the network capacity. And you say that it's irrelevant.

2:04:56Keagan McClellandWhat's less than 1% of the network capacity? Oh, of the 21 million bitcoin?

2:05:03Vlad CosteaYes. Yeah. I meant supply, not capacity.

2:05:07Keagan McClellandYeah. I am firmly of the opinion that TVL, like total value locked or whatever, in the case of Lightning is not... I mean, I understand why people use it as a success metric, and I think to one degree of approximation it's useful as one, but it's fundamentally a cost metric.

2:05:31Because if you think about what the value of the Lightning Network is, it has nothing to do with how many coins are locked up in it. In fact, for vaults, I think that would be a much better metric, because the whole point of vaults is what they're securing. And so if you said, well, you know, 50% of the network is locked up in vaults, it's like, that's a pretty good endorsement of those vaults. But the value of the Lightning Network is in what it transacts.

2:06:02And in fact, the more that you can transact with less value locked, the better, because the value locked is a cost metric. It's not a value one, right? Because if I can transmit millions of dollars of value with only locking up thousands of dollars of value, then that's a much more efficient construction than if I'm transmitting a million with having to lock up a million. Does that make sense?

2:06:30Vlad CosteaYeah, it does.

2:06:33Keagan McClellandSo that said, if the value is zero, it means that no one's using the network. That's definitely true. And number two is that if the TVL going down is also commensurate with less value being transacted on the Lightning Network, then that's also bad, right? So insofar as the TVL going down actually means reduced economic activity on the Lightning Network, then that is a statement that the Lightning Network is not delivering as much value as it was before. But again, the metric to pay attention to is the amount of value transacted, not the amount of bitcoin locked up in Lightning channels.

2:07:18In fact, this is actually a really important point, which is that if you have a very big Lightning Network with a whole bunch of value locked up in these channels, but that value is never moving because, you know, people got really excited during Plebnet Summer or whatever, and they put a whole bunch of coins in these channels and then they never transact on those channels, then that bitcoin is dead weight and in a more vulnerable position than if they had just had it locked up in cold storage.

2:07:50So in some ways, you might expect for the TVL to go down over time as you start to prune away the roads that are not being used, saying, hey, well, we built out this infrastructure to go from A to B, but no one actually wants to go from A to B, so we might as well reclaim the resources used to build that and then deploy them elsewhere. And so even if the overall transaction volume of the network was fixed, reclaiming resources that are going unused is actually better for everyone involved and does not in any way say that the network itself is failing.

2:08:34Vlad CosteaBut it does mean that the network itself becomes more centralized, or at least concentrated in a few large nodes.

2:08:42Keagan McClellandSo if you build a whole bunch of edges on this network that never get used, is it more decentralized than if they weren't there?

2:08:56Vlad CosteaWell, it's about the routing, first of all, because this is an important part. You don't want the same three large nodes to route all or a large percentage of the transactions, right?

2:09:15Keagan McClellandWhy?

2:09:17Vlad CosteaWell, first of all, they're going to collect most of the transaction fees, and that sort of disincentivizes everyone else from participating. Secondly, I'm not an expert in this, but I'm assuming that it can be bad for privacy also, if you're going to have all of these nodes collecting information about how many transactions there are on the network.

2:09:41Keagan McClellandSo the second one is true to an extent. Obviously, the Lightning Network has a lot of privacy technology built into it in terms of what can actually be learned from doing the routing, and it's not very much, but it's not nothing either. The second one is that, again, this is sort of contingent on the claim that if those decentralized edges never get used, then it's not really distinguishable from the scenario when they weren't there. That said, maybe if one of these central nodes goes down, you still have the option.

2:10:15So as a sort of backup plan, it is useful. But then you have to think about the opportunity cost of that capital. The question is, is it better for the network overall that that capital gets used to reinvest in other ways? Obviously, if it's just sitting in cold storage, maybe not. But I think a lot about the goals of decentralization: they are very much about robustness, right? Its ability to withstand different types of failures, different types of betrayals, which is, you know, just a particular type of failure.

2:10:55And efficiency, which is about not using more resources than you need to. And these two things are fundamentally at odds with each other. A radical commitment towards decentralization is also a radical commitment towards inefficiency. And inefficiency is, sort of by definition, not costless. It is expensive. To do the same job with a lot more resources means that you don't have those resources to spend on other things. And where on that continuum ends up being the right answer is kind of anyone's game, right?

2:11:31And the more risk that you are exposed to in the world, the more that you might want robustness, whereas the less risk that you're exposed to, the more that efficiency might rule the day. And it's actually all about efficiency at the end of the day anyway. It's just efficiency even when you start to consider these outsized tail event scenarios where very rare and costly events occur to you. So for me, decentralization is a means to an end. Again, it's about the freedom to transact.

2:12:01And it's the permissionlessness of it all. So yeah, if you only have one routing node that you could route through and they decided to just cut you off, that's really, really bad. But the difference between having five separate options and a thousand separate options, especially when it is truly permissionless for anyone to enter the game, that doesn't bother me.

2:12:26Vlad CosteaSo, back to the TVL, which means total value locked, if I'm not mistaken, argument. I don't want to speculate too much on this, but I'm pretty sure that at least a handful of people who have been on the podcast own more bitcoin than the total amount of bitcoin that's locked in the Lightning Network. So if they really believed that it's a good way to scale Bitcoin, or it's useful for themselves, they would open a bunch of channels and they would make the network more decentralized and provide more routes for everyone else. But they don't. Why?

2:13:07Keagan McClellandSee, this is what I mean, though. This would be true if the demand for the transaction volume itself is greater than what the existing capacity can provide. And so this is actually best represented by payment failures. So in my opinion, the best metric for whether or not the Lightning Network is healthy is a declining payment failure rate. That is the number one important thing.

2:13:37Vlad CosteaOkay. When was the last time you made a payment with Lightning and it failed?

2:13:43Keagan McClellandProbably years ago.

2:13:45Vlad CosteaReally?

2:13:46Keagan McClellandYeah.

2:13:47Vlad CosteaWhich setup are you using?

2:13:50Keagan McClellandSo at the time, I was using a raw LND setup with Zeus plugged into that.

2:13:58Vlad CosteaBecause I'm using something which is not custodial. It's Phoenix. So you do open your own channels, but it is hosted. So you do use their servers for broadcasting the transactions, and you do open the channels with their own large liquidity pool. So in theory, their node is very well connected and should be able to send transactions to any part of the network. They should be able to find your route. It's just that I had lots of instances where this did not work out, also in the last year or so.

2:14:37Keagan McClellandYeah. I mean, I'm not going to contest your experience.

2:14:40Vlad CosteaAnd then Matt Corallo told me that if I'm honest about this and I'm not lying, then I should provide logs. And then I said, well, I'm not comfortable doing this and normalizing the practice of exposing what kind of transactions I'm doing with whom and at what time and whatever. Like, I don't want to give out free information about my wallet just to get an explanation as to why this failed.

2:15:04Keagan McClellandI mean, that is a fair thing for you to refuse. But also, as a developer, it's hard to be able to improve the situation without the information of how it failed, right? Now, it doesn't necessarily need to come from you, but it does need to come from somewhere. So if they're not experiencing it and you are, it means there's something different about your setup than theirs, and they would need to see something like that in order to improve those situations.

2:15:30Vlad CosteaWell, yeah. My point is that in my experience using Lightning, even sometimes with custodial wallets like Wallet of Satoshi or whatever, I still get failed payments. And maybe it's because of a bad internet connection, maybe it's for lots of various reasons. It's just that I never... So, I don't use my card for payments or whatever because I hate it on principle. I try to avoid it as much as I possibly can. I use cash as much as I possibly can. But when I see people paying, it becomes very embarrassing when you're in a supermarket and you're tapping your card and for some reason it doesn't work. It gets declined. And if we do want to use this for payments at scale, we have to get to the point where this is less than 1% kind of occurrence.

2:16:20Keagan McClellandProbably a lot less than 1%, right? 1% is one out of every hundred. You see 100 people a day, it means you're seeing payment failures every day. I would like it to be even better than that. But I think your point's well taken. So my contention when I responded to you online about the TVL thing is not so much that the Lightning Network is perfect and that it needs no improvement, or that it is satisfying the demands of using Bitcoin as a medium of exchange in everyday transactions.

2:16:55It's none of those things. It's simply just that TVL is not a good metric to use to decide whether or not the Lightning Network is succeeding. Payment failure rate, great one. Transaction volume, great one. TVL, not a great one.

2:17:12Vlad CosteaOkay, let me ask you this, which is sort of related to our conversation, but also fantasy. So it does not concern the reality. But if you had this magic wand, right, and you could make any change to Bitcoin whatsoever, just one change. It can be adding something, it can be modifying a parameter or whatever. What do you think that would be?

2:17:50Keagan McClellandIt would probably be to get Simplicity into Bitcoin.

2:17:53Vlad CosteaYou mean Blockstream's scripting language?

2:17:56Keagan McClellandYep.

2:17:57Vlad CosteaDo you think that improves any of what we're dealing with? Do you think that makes Lightning more reliable, or that it enables better vaults than what we have today?

2:18:09Keagan McClellandSo it is undeniably true that it allows for better vaults. This is a very heady argument, but it's just the world that I live in, which is that if I only get one wish, right, I kind of want to make sure that that wish is the most general tool possible, to be able to solve all the other problems. And to my knowledge, there is no more general tool that has been proposed for all of Bitcoin than Simplicity. It gives us the ability to do sort of arbitrarily flexible scripts and contracts, with the ability to also formally verify their correctness. And the trouble is that it's not the most efficient thing in the world.

2:18:58And don't get me wrong, there's still many problems to solve for Simplicity to be viable, and it's unclear to me whether all of them even have good solutions. But it is the tool that allows the most number of other things possible, and the general tools are the ones that allow us to solve novel problems when they crop up. You can't purpose-build tools for every problem.

2:19:28Vlad CosteaI agree with that. I just expected you to say CTV or the Great Script Restoration or the package...

2:19:36Keagan McClellandThe Great Script Restoration might be the closest thing off of Simplicity. It's just not nearly as good.

2:19:44Vlad CosteaThat's interesting. So adding a bunch of opcodes, you think that's worse than activating Simplicity?

2:19:53Keagan McClellandSo, worse... To be clear, it's worse because it's not as good, not worse because it makes it worse. Does that make sense? GSR would be an improvement to Bitcoin. I want to make that clear. I am in support of GSR, primarily because it's a much easier sell to people, both from a marketing perspective and from a technical perspective. But the theoretical maximum is just not nearly as good as something like Simplicity.

2:20:27Vlad CosteaIf someone came up with this proposal to increase the witness side of the SegWit block from, I think right now it's three megabytes on top of the one that we have, and take that to 16 or whatever, would you be in favor of it, if it helps open more Lightning channels or more L2 connections, peg in, peg out, whatever?

2:20:51Keagan McClellandI don't know that we're bound by that right now. I mean, it's a little outside my wheelhouse, because here's the thing, right? People like to get religious about it, but the block size war isn't really about the fact that 1 megabyte or 4 megabytes or whatever is the one true size. It's just that you don't want to increase the block size as your primary scaling strategy, because it just doesn't work. Once we have other solid scaling strategies that give us multiplicative factors on the base layer stuff, then we can start to have a little bit of a more mature discussion about what the right level of block size increase is, given the hardware availability that's out there.

2:21:38So to that end, I am not fundamentally against block size increases, but I also don't know enough about how those parameters will affect the hardware for IBD and for keeping up with the network. Because I know the witness verification, once you've done it, you can just drop it, and it doesn't have the same level of overall increase in pressure that a base block size would, because anything that appears in the base block can influence the number of changes to the UTXO set. That has ongoing operational load just from that alone, whereas the witness, once you've verified it, you could sort of drop it and forget about it completely.

2:22:30But this is stuff that I would generally want to defer to people who work on Bitcoin Core to really answer some of those questions about what the right parameters are. But in principle, to answer your question, block size increases are fine. It's just that they can't be your primary scaling strategy.

2:22:47Vlad CosteaEven the Bitcoin Cash people agree with that, because they settled on 32 megabytes and said that anything beyond that point is just not good enough and they need to figure out better ways. And that's when they started activating a bunch of opcodes, and that's why now they're far more advanced in terms of being able to multiply large numbers and being able to do covenants, and they can have a much better version of Lightning or rollups or whatever. It's just that there's not market demand for it. But technically, they were able to develop their protocol to a greater extent, to which we're still trying to get with BTC.

2:23:29Keagan McClellandI mean, yeah, that's one of the benefits of having a much smaller user base, right? Is that you can get away with doing a lot more a lot quicker. But I think...

2:23:39Vlad CosteaYeah, but they also had less funding, and they also got called out for being scammers. And I don't think there's a project that received as much hate as BCH. And we did craft these narratives and said, well, it's just libertarians who don't understand how Bitcoin works that left for that project. But to their credit, they did a good job, which is not something that someone like Greg Maxwell would acknowledge.

2:24:08Keagan McClellandBut I mean, like, look, I have not studied Bitcoin Cash since they forked off, because, like...

2:24:14Vlad CosteaBut that's the thing. Jeremy Rubin did, and that's where CTV comes from.

2:24:19Keagan McClellandOkay.

2:24:24Vlad CosteaThey also did a more complete Taproot migration. They also have a lot of cool stuff that we only wish we had in BTC. And my question is, why, if we see it on a live network with incentives, that it works, and we can try to break it with Kimi or whatever, just deploy some technical users to try to deploy it, why are we not copying their homework? Because to me, Bitcoin maximalism is about being able to do everything, not limiting the use cases and saying, well, this limit on the block size and this limit on what you can do with transactions and whatever should be in place, and let's just virtue signal about having the perfect system. There is a lot of research and development that's happening outside of Bitcoin. That's fine. But...

2:25:14Keagan McClellandI don't disagree with any of this.

2:25:16Vlad CosteaIn my view, if something gets enough market traction, and there is obviously clear evidence that there's going to be users coming to Bitcoin and paying transaction fees to do that, maybe we should think of ways to bring that to Bitcoin, as opposed to just pointing fingers and saying, well, these are shitcoiners and it's not going to work.

2:25:36Keagan McClellandRight. But there's nothing about what I've said in this entire interview that would suggest that I'm doing...

2:25:42Vlad CosteaNo, no, there is not. I'm just providing context to the whole conversation, because I know people who, for example, are very religious about the block size limit, and they say that it's just as sacred as the 21 million coin limit. And that's really not true.

2:25:56Keagan McClellandAnd in fact, even the 21 million coin limit is not necessarily the important part. The important part is a non-discretionary monetary policy. And so it doesn't matter whether it was 21 million or if it was like constant inflation. I mean, they might matter at the margins, but the problem is, if you change it once, then you can change it again. And changing it is how we end up with fiat currency.

2:26:20Vlad CosteaI 100% agree with that. But I also present the example of the November 2008 version of Bitcoin. That was something that was being circulated among the early participants, and that was Hal Finney, that was Ray Dillinger, that was James A. Donald, and there were a couple of other guys who mentioned it and wrote about it. And that one had 100 coins per block rewards. It also had halving cycles that were happening every week. And also there was something different. So if you added up the amount of coins that you would get in that setup, it would add up to 20 million.

2:26:59So 21 million, I believe, was just an accident of having 50 coins per block every 10 minutes and having cycles every 210,000 blocks.

2:27:10Keagan McClellandThe important part of 21 million has nothing to do with 21 million. It has to do with non-discretionary monetary policy. Because the trouble with discretionary monetary policy is that if you put someone in charge of deciding what coins should be issued at what schedule, that person now has a lot of power and they become a source of corruption. If you just set it, and once it's set it can't be changed, then it eliminates this corruption vector. And that's really, really important. But whether it's 21 million or 10 million or 100 million, or something that increases at 1% forever, or maybe it increases at 50 bitcoins forever, so it's not necessarily a percentage-based inflation but a fixed inflation, all of those systems would work and solve many, many, many of the problems that we have with fiat currency.

2:27:59Vlad CosteaAnd to zoom out from this perspective, I tend to say this a lot lately. I used to be a fanatic of dividing everything by 21 million, and now I understand that most likely that's not going to be possible. But if that's going to be 21 million Bitcoin and maybe 100 million Monero and 21 million Zcash and 21 million... No, I think it's 84 million Litecoin, and 100 million Ethereum, and they all add up to like 500 million coins, and that's like a fixed supply of the currencies that are out there, and the policy is clear and is not changing.

2:28:36And if it does change, there is an adjustment in the price that compensates for the fact that they're not honest about their monetary policy. I think that's a lot better than the system that we currently have, where you have a central bank that makes changes on the go every quarter or whatever. They get together and they say, well, let's adjust the interest rates, let's adjust this and that, and that completely changes the dynamic of the economy. So if we have something open source that is auditable and is plural, in the sense that we have multiple networks with different monetary policies, it's still much better than what we currently have with governments playing with their toy money printing and whatever.

2:29:21Keagan McClellandCompletely.

2:29:22Vlad CosteaLike, I'm not against a system like this, where you just swap between a currency and the other, or just exchange like you do Forex today. And I don't think it's the end of the world if Bitcoin does not end up doing everything.

2:29:38Keagan McClellandI mean, it literally can't do everything.

2:29:42Vlad CosteaWell, if you look at it, like, if I was Super Testnet, right? Super Testnet is this guy who's like, no, I can do this on Bitcoin. Bitcoin can have this. It's just that it has its different tradeoffs, right?

2:29:55Keagan McClellandYeah. Yeah, I was kind of being a little snarky, but it's like every choice that you make forecloses on all the alternatives you could have made, right?

2:30:04Vlad CosteaI'm very happy with this answer, to be honest. Of course, there's going to be someone in the comments saying we're shitcoiners, but that's fine. I mean... But speaking of shitcoiners, what do you think, and this can be the closing question if you're tired or something, but what do you think about the upcoming hard forks, which is going to be the BIP 110 one by Luke Dashjr, and the BIP 301 by Paul Sztorc?

2:30:35Keagan McClellandI honestly could not care less about Luke's fork, because I don't think it actually improves anything in any real sense. I am cautiously curious about how the drivechain stuff ends up working.

2:30:52I've kind of gone back and forth on drivechains for the many, many years that I've paid attention to it, about whether or not I think it'll solve interesting stuff. But I think it's a super interesting laboratory experiment, and I'm excited to see it play out at a production level. But I'm certainly not buying into it. I'm interested in it as a technologist. If I'm trying to allocate capital, I'm certainly not rich enough in order to put any serious amount of money into it, because I don't think it's well proven. But these experiments are how civilization learns.

2:31:34And so I'm glad that he's sort of taking a respectable sort of bow out of not trying to make it part of Bitcoin, but at the same time still honoring his commitment to his own beliefs, and not trying to... Everything about the way that Paul has marketed the fork itself has been, in my opinion, decent, right? He has not always had the best communication style when it comes to calling people stupid about not believing in his ideas, but at least he's not trying to pretend that his hard fork is the real Bitcoin or whatever, that Luke does on the regular.

2:32:12Vlad CosteaI also think that these two are the perfect antithesis or whatever. They are so opposed in their philosophy. There is the BIP 110 fork, which is reductionist, and it says we should not do this with Bitcoin, and we're going to keep patching it to prevent spammers from using this network. And on the other hand, you have the drivechains fork, which is maximalist in the sense that it tries to maximize everything that it can do. And if there is a new use case that comes along, they can have a sidechain for it and see what the market thinks and just run it. So one of them is trying to reduce the use cases, the other one is trying to increase them. But for some reason, one gets more attention than the other.

2:32:57Keagan McClellandYeah, I believe that the second hand wins in the long run.

2:33:01Vlad CosteaYeah, and we're going to learn from it. My only hope is that we actually learn from these experiments. And, I mean, once again, it's not going to be the end of the world if we get a bunch of parallel networks, altcoins that are supplying for the demand that the market has, and we're going to have like 500 million coins. But it would be useful if Bitcoin actually improved and added some of these features from which we learn from other networks.

2:33:31Keagan McClellandYeah, completely agree.

2:33:35Vlad CosteaSo is there any message that you have for the listeners, or anything that you want to promote?

2:33:42Keagan McClellandWell, so let me talk a little bit about what I work on, because this conversation has been a lot about Bitcoin specifically. And what I'm working on is currently Bitcoin-specific, but it's not necessarily limited to that. But I think what your listeners will be interested in is that what I work on is formal specification and verification. And I am very interested in making sure that the software that we run does what we think it's going to do, which, for anyone who's written software before, knows that it's shockingly difficult to make sure that it does only what you expect it to do and nothing else.

2:34:21Anyone who's ever been hacked has experienced this. Anybody who's ever, you know, written software and a bug has occurred and something weird has happened understands this. And so for a network that is as well capitalized as Bitcoin is, you know, upwards of a trillion dollars worth of market value, it's very important that we understand exactly how it behaves. And current software development practices, even in Bitcoin Core, don't give us the level of assurance that the software industry in general is actually capable of producing. So what I'm very excited about is trying to bring these practices of formal verification and specification to Bitcoin.

2:35:05So what I've been working on is a project called BTC Verified. I haven't been able to touch it for the last month or so, because I've been out traveling for conferences, but literally this week is actually the first time that I'm getting back to it. But what it is, is a formal model of Bitcoin's consensus rules, so that we can better understand them, so that we can better actually change them safely, and analyze how proposed changes might affect the network. Because right now, what happens is that people debate each other on the forums, whether it's the dev mailing list or Delving Bitcoin, and they'll have arguments that only the people there are aware of.

2:35:44And so as newcomers try to come in and contribute to Bitcoin, all of that institutional knowledge gets lost. And then secondly, sometimes people make hand-wavy arguments about what a proposed change will do to the system without actually being able to specify in a great level of detail what that is and why it occurs. And this can actually prevent having a healthy and productive discussion about whether a proposed change is going to do what it says it's going to do, and whether or not the sort of second-order effects, or the side effects, of that change are something that we want or don't want.

2:36:28So the whole project is really just around trying to make sure that we better understand how Bitcoin actually behaves and how the software achieves the protocol expectations we have out of it, so that we can steward the development of the protocol more safely. I don't expect most people to get involved with this project, because it is extremely highly technical. But if you want to, I do encourage anybody who's interested to look. This is not me trying to repel people from taking a look at it. It's just that it's an acknowledgement that it is one of those things that requires a great deal of technical interest and skill. But I would love to get more eyes and interest on it. So yeah, you can take a look on GitHub. It's like BTC Verified. I can send you a link and you can put it in the show notes or something.

2:37:21Vlad CosteaSure. So another question that stems from what you just said: which conferences are you attending this year? Where can people meet you in person, say hello?

2:37:32Keagan McClellandI'll be at TABConf this year. And I think that's the only Bitcoin conference I have planned, though. The ones that I was just at were not Bitcoin conference related. It was formal logic and mathematics conferences. So the other side of this, if you view what I'm doing, I'm taking advanced math research and trying to bring it into Bitcoin in order to make the software more safe and secure. So I spend some of my time at Bitcoin conferences, because that's where the demand side is, and then I spend some of my time at these math conferences, because that's where the techniques come from. And so I was just in a sort of chapter where I was doing some of the more math stuff and less of the Bitcoin stuff.

2:38:15Vlad CosteaThat's actually productive. And I actually think that a lot of people working in Bitcoin should take a step back and learn other fields of science, just to be able to apply them into what they're doing, or get a fresh perspective on what they're already doing. It can be very productive. I'll also be at TABConf, and I spoke with Michael Tidwell, and if everything goes well, there's going to be like a marathon livestream happening there with this podcast, and people can just drop in after they get off stage and talk about what they just said, or we can have gossiping or whatever.

2:38:50It's going to take a long time of just sitting there and talking. And it seems like that proposal got approved. I hope the Wi-Fi is just good enough for the livestream. That's my only concern. But otherwise, I look forward to having you on the show after your talk at TABConf.

2:39:08Keagan McClellandSweet, yeah. Thanks, Vlad.

2:39:12Vlad CosteaYeah, it's been a pleasure. You're much nicer in person than you seem on Twitter when you're just replying to people.

2:39:19Keagan McClellandWho isn't?

2:39:20Vlad CosteaWho isn't? That's a good point. Remember, Twitter is not real life. Come to TABConf. Come to real conferences where you meet the people in person, have a drink, and you're going to understand what this is really about, and how people are actually different in person. So yeah, pay me commissions, Michael Tidwell, if I sell any tickets. Thank you, everyone, for listening. This has been my interview with Keags. You can follow him on Twitter. I think it's at ProofOfKeags. And I don't know if you're on Nostr, but some people use that.

2:39:54Keagan McClellandIt's popular. I probably should be at some point, but I'm not. But yeah, my Twitter and my GitHub are both ProofOfKeags.

2:40:02Vlad CosteaThere are two people who guessed 37 in the chat. I'm going to contact you after the show. And thanks to the sponsors, once again: Braiins Hashpower, SideShift.ai, Cake Wallet, LayerTwo Labs, and OrangeRock. Thank you. I'll see you in the next one.