Title: S12 E8: Lisa Neigut on Running a Core Lightning Node Podcast: Bitcoin Takeover Podcast (host: Vlad Costea) Guest(s): Lisa Neigut (niftynei), Core Lightning developer at Blockstream, organizer of bitcoin++ and founder of Base58 Published: 2022-11-10 Duration: 1:22:47 Episode URL: https://bitcoin-takeover.com/audio/?name=2022-11-10_s12_e8_lisa_neigut_on_running_a_core_lightning_node.mp3 Audio: https://bitcoin-takeover.com/audio/media/2022-11-10_s12_e8_lisa_neigut_on_running_a_core_lightning_node.mp3 Transcript: CORRECTED. Whisper-Large-V3 first pass with speaker diarization, corrected with Claude Opus 5.5 against two independent re-transcriptions of the audio (Qwen3-ASR full pass; ElevenLabs Scribe v3 on doubtful passages) and the show notes; automatic consistency checks. Format: Clean verbatim (v2): fillers (um, uh), stutters and false starts removed; wording, meaning and order unchanged. Each paragraph starts with [hh:mm:ss] (its start time in the episode audio) and the speaker's full name; times of paragraphs split out of long monologues are interpolated (within a few seconds). [unclear: ...] marks words that cannot be made out from the audio. ================================================================================ [00:00:30] Vlad Costea: Hello there and welcome to the Bitcoin Takeover podcast. This is still season 12 and I have a very special guest today. Her name is Lisa Neigut and she works at Blockstream on some Lightning stuff, which is super interesting. She's also interested in privacy, on-chain and Lightning privacy. And she also is very funny on Twitter and I'm going to ask her about some of the stuff that she posted. And she's also doing a conference which takes place in Mexico City, and she's going to tell us about that too. So hi, Lisa. [00:01:24] Lisa Neigut: Hey, Vlad. Thanks for having me on. It's good to see you. [00:01:27] Vlad Costea: Well, this is going to be fun. You're usually full of energy and enthusiastic. And you work at Blockstream. And Blockstream developed c-lightning, which was rebranded as Core Lightning, I guess, because you changed the main programming language, I mean, diversified from it. And there's more stuff going on. Yeah. So I have a very specific question to begin the interview. I have this RaspiBlitz, which is a Raspberry Pi, which runs a node. And I'm going to upgrade it and I'm going to make some changes and give it another try. [00:02:06] Vlad Costea: This one has been running for almost two years and I'm going to change the board and some other stuff and do it on video. So if you're listening to this, subscribe to YouTube so you get to see the video that I do. And I get this choice when I install RaspiBlitz on this device. I can either do LND or I can do c-lightning or Core Lightning. And I need your advice. Of course, you're going to be biased here, but you tell me what I'm going to do with this one. [00:02:37] Lisa Neigut: Yeah, I mean, of course, I would recommend that you run Core Lightning, first and foremost, so that if you ever have any questions, you have a reason to come talk to me or one of my colleagues who work on Core Lightning. We're a super fun group. The other things that, I mean, you know, it's one of those things where people are always like asking me, "What's the difference? Which one would I prefer to run?" I don't really know a lot about how LND works or what features it offers. [00:03:05] Lisa Neigut: But I'm more than happy to talk about all the great stuff that we've got on Core Lightning that I'm excited about. I think, like, Core Lightning stuff that I'm really proud of and that I like: we just shipped some really cool bookkeeping, like accounting, so you can get really detailed reports on where you spend your money and where your money's going. I'm hoping that more dashboards will get built around that, but I think having a really good idea of where every millisat goes is pretty novel for Lightning nodes and something you get out of the box for free with Core Lightning on our latest. [00:03:43] Lisa Neigut: You have to be on.12 or later to get it. So that's really cool. So if you're starting up a new node, Core Lightning will give you really good details about where your money is going. Some other things that I really think are cool about Core Lightning is our plugin system. So it's kind of like you can run all these extra apps on your Core Lightning node. One of those that some people seem to really like is CLBoss, which will open channels and do loopouts, and if you put money on your node, CLBoss will automatically create channels and make it such that you are always, relatively always able to send and receive payments. [00:04:19] Lisa Neigut: And when you couple this with like the new bookkeeping data, it gives you a really good idea of like how much it's costing to run CLBoss and like kind of, yeah, I don't know. It makes running a node, if you just want to have a node that like has money available, a lot easier. So yeah, that plus there's this really cool thing that someone started building called Clams. I think it's Clams.io. I don't know. But we've got this totally new thing called Commando. Have you heard about the Commando thing? It's really, really cool, basically. [00:04:50] Vlad Costea: Oh, yeah, the 1985 movie with, I think, Arnold Schwarzenegger. [00:04:56] Lisa Neigut: Yeah, exactly. So we called up Schwarzenegger and he made this new protocol for Core Lightning. It basically lets you talk Lightning protocol between like a web app or an iOS phone. There's this like, anyway, so there's projects that are now using this, and so you can use the systems. Core Lightning has this like permission token system called runes. A similar version of it that exists for LND is called macaroons. I really like telling the story because I think it's kind of funny about the difference between macaroons and runes. Should I go into it or I can keep talking about what I like about Core Lightning? [00:05:33] Vlad Costea: No, keep talking. This is fun. [00:05:35] Lisa Neigut: Okay, yeah. So part of Commando, so Commando is this ability for your browser to basically pretend it's another Lightning node. And it connects to your Lightning node wherever it is in the world. So you need to have like your connection string information. And then you need like this auth token. Basically, a rune, you need a rune. So you get a rune from your node to put into the website that wants to talk to your node. [00:06:05] Lisa Neigut: Because that gives it permission to access stuff. But basically, your web browser pretends to be another Lightning node on the network. And it makes a secure connection using, I think it's BOLT 8, that is the description of secure socket connections. So any two Lightning nodes, when they talk to each other, they make a secure communication channel, and then they send information about payments or opening a channel over the secure socket connection. [00:06:32] Lisa Neigut: Your browser can pretend to be a node, though, and securely communicate with your Lightning node wherever it is in the world. So you can make like distant connections between your node and whatever browser you want to access it from, or maybe a web app or iOS app. Yeah. And so our runes and this Commando thing make that all possible. And someone's built a web app on top of it called Clams, which right now just lets you send and receive payments from the browser. But I'm assuming they'll add more stuff to it, so I'm really excited about that. But let me go back to this idea about runes versus macaroons. [00:07:12] Lisa Neigut: Okay, so macaroons are a Google technology. Google wrote this whole paper about how to make these cool... So basically, you make a token, and the token maybe says it gives you access to make any call on your node that's read-only. But then they made it so you can take that token and add more restrictions to it. So maybe it's only valid for the next six hours or something. And then give that macaroon to someone else, and they can use it, but only for the next six hours or something, right? [00:07:43] Lisa Neigut: So it kind of has this cool thing where you have like a base token, and then you can add more restrictions onto it, at least that's my understanding. And then you can pass them down, so you kind of have this weird tree of authentication based off of one permission that you can then hand out to more people or whatever. Anyways, but macaroon is kind of like a joke. So the macaroon joke is like macaroons or macarons, as they say in France or whatever, are cookies, right? They're like little baked goods. So on the internet, the earliest authentication or session things were cookies, right? [00:08:19] Lisa Neigut: So we call them cookies. Your web browser will have cookies on it. And it gets stored in your browser, and when you go to a web page, it sends the cookies to the page. And then depending on what cookies you have, those are the permissions you currently have. [00:08:36] Lisa Neigut: So macaroons are a type of cookie. So it's like an improvement on a cookie type. So it's a particular type of cookie. And they called it macaroons because part of the protocol of how you build these macaroon tokens is you use this thing called MACs, which is like a crypto authentication scheme thing. So they have MACs, right? Okay. So Rusty read the paper. Rusty Russell's one of my coworkers, very brilliant software dev who's been in the business for longer than anyone else in Lightning, I think. [00:09:08] Vlad Costea: Also the best Twitter account for trolling. [00:09:12] Lisa Neigut: Yeah, he's great on Twitter. It's rusty_twit if you want to go look him up. So he read this macaroons paper and he was like, "Okay, I like this scheme, but I think we can make it better. We can get rid of the MACs. This MAC part that they added is unnecessary." So he took the MAC off macaroons, and what you end up with when you get rid of MAC on a macaroon is just the word rune. So he calls it runes. That's where the name runes comes from. [00:09:37] Lisa Neigut: It's Rusty's improvement to the macaroon scheme. And so they're very similar, but clearly they don't interoperate, so they're different that way. But it doesn't really matter as long as you're able to... I mean, if you're building a web app, from that perspective all you care about is you need a little string thing that you send in the call, and you send it to your node, and the node gets the string thing, and either the string thing that you sent is correct and lets you get access, or it's incorrect and it sends back an error. But your web app doesn't really care what it is, if it's a rune or a macaroon or whatever. So I think it's totally fine if they're different standard things, sort of. [00:10:24] Lisa Neigut: I don't know. Anyways, yeah, so that's like the macaroon web app stuff. I think that's really cool. It's still, I think, pretty new. So I think it's definitely going to be... I hope there's more apps and stuff that people are building, like this Clams project. I'll have to send you the URL later. I'm really excited about it, one, because it's very pretty. They have this nice little logo, loading logo. [00:10:45] Lisa Neigut: But I think it's really exciting because it'll let... I'm hoping someone builds really nice accounting dashboards using all this accounting data that your node is now getting, so you can see really nice pictures of where you're losing money and where you're earning it and what your profit and loss is. And hopefully having all that data will make people running Core Lightning nodes much better at where they put their money, and having a better understanding of what's costing them money is the whole idea. [00:11:16] Lisa Neigut: Yeah. Anyways, that was a lot. I'm sorry. I don't know if that helps you make your decision. I feel like now you need to go talk to someone who can tell you all the details. [00:11:24] Vlad Costea: Let me get back to my decision, because this is kind of a bummer, because you have some very cool apps for Core Lightning, but the ones that are for podcasters, like Sphinx and Fountain, and the ones that enable you to monetize your content are integrated with LND for some reason, but are not ported to Core Lightning. So if there's anything you can do to propose that, I would very much appreciate it, because as far as I know, Core Lightning is a lot more stable. And let me segue into something else which has happened in the last couple of weeks. [00:11:58] Vlad Costea: At first, there was someone who did a huge multisig. I think it was 99 out of 100 setup or something. [00:12:05] Lisa Neigut: No, it was worse than that. It was 998 out of 999 signatures. So in order to spend, someone took some Bitcoin, locked it up to a 998 out of 999 threshold signature, which meant they needed 998 signatures in order to spend it, right? And then they locked the money up to it and they produced a transaction that would spend it. So they had a transaction that had 998 signatures in it in the witness data, which is very large. That's a very large transaction. I think this is Burak, right? [00:12:41] Lisa Neigut: Burak. I think Burak said it only cost him like $5 in Bitcoin to get that one mined at the time. The price is pretty stable. But yeah, so it cost him like $5 to mine a 998 signature transaction. But that's very large. That's a lot of witness data. Also, notably, you can only do that with Taproot. So this is like... being able to make a transaction with that much... The place that you put signatures for a Bitcoin transaction is called the witness data. So all of that data, those 998 signatures, went in the witness data. [00:13:20] Lisa Neigut: And I think, I need to look, but I think each signature is probably 64 bytes, because that's the new Taproot thing. I think they're 64 bytes now. They used to be like 71, but assuming the smaller one. I don't know. I'm going to do some math. What's 64 bytes times 998? 63,000 bytes, that's a lot of bytes. So it's a big transaction, right? Yeah. I'm sorry. Yeah, so that's what happened. [00:13:52] Vlad Costea: Yeah, but it's not just about this multisig. It's about the fact that LND nodes crashed. [00:14:00] Lisa Neigut: They didn't crash, though, they just... The block that that transaction got put into, they couldn't parse. So they stopped parsing. They stopped advancing where they were on the chain, right? So if something were to happen on chain after that block, your node wouldn't see it because it was like stuck, unable to process a block that had this big transaction, this transaction with a lot of witness data in it, right? So I think as long as you didn't turn your node off, it was still up and kind of doing stuff. [00:14:28] Lisa Neigut: But it was like in a wounded state, right? Like it wasn't operating at its full capacity. It was like there's a lot of stuff it just like wasn't able to do. It's my understanding. Yeah, that really sucks. That's really bad. In terms of impacts, being like who's affected, I think almost everyone who runs LND was impacted, and basically the only way to fix it is for them to patch it and send out a new copy of LND. So you have to update your... Everyone has to update their node. I kind of was joking, and maybe it's like sort of a taste... [00:15:05] Lisa Neigut: It's not a tasteless joke, but I was like joking: "Well, I hope they have all the new features they wanted everyone to have already in the app, because everyone's going to have to update all at once." So they're not going to have to worry about old versions of LND after this, because any old version of LND won't be working. So, I mean, as a software dev, it's kind of exciting. So you can definitely deprecate stuff now, because there's definitely, depending on where they are in their software cycle, like everyone's gonna be on this new version of LND now. And they know that because like you wouldn't have a node if it was any older, so. [00:15:39] Vlad Costea: Yeah, I guess it was their BerkeleyDB moment, where everyone had to update on emergency. [00:15:47] Lisa Neigut: Yeah, exactly. So, I mean, it sucks. Really, it's a shitty bug because it's like one of those things where like, if you would say it like, I don't know if anyone would like... It's like one of those bugs that's very easy to fix because it's a parameter. There's just a hard-coded number, and you have to change that hard-coded number to be bigger, and all of a sudden it works again. So it's really shitty because it's not like you had complicated software interactions that were causing problems. [00:16:16] Lisa Neigut: It's literally just a setting in the software. But it's one that, in order to update that setting, everyone has to get a new copy of the app and upload it, if that makes sense. It wasn't something you could go in the configuration file and change, and then everyone would be good again. Though you could write an app such that that number was configurable, but no one ever writes software that way, just because of where this particular parameter or sort of configuration number is. Anyways, so as a software developer, yeah, that's a frustrating, kind of annoying bug where you're just like, "Ah, why?" I don't know. [00:16:57] Vlad Costea: Yeah, but there is another problem right now. Of course, my next question is going to be: why were Core Lightning nodes not affected by this multisig? [00:17:08] Lisa Neigut: Yeah, that's a great question. The multisig transaction that went out in mid-early October is valid underneath the Taproot rules and is not valid under the SegWit rules, right? So anything before Taproot, that would have been invalid, but it's a Taproot transaction. It uses the Taproot rules. Anyone who implemented the Taproot rules would process it correctly, right? So most Core Lightning nodes, the way that it works is... Core Lightning is very like, we like to call it like a plugin kind of architecture. So we write an interface where these are like all the things we need to be able to call from like a Bitcoin block source, as we call it. [00:17:58] Lisa Neigut: And then anyone can write like a new application or piece of software that, as long as it can respond to all of our like questions, like, you know, we'll ask, "Is this transaction in the mempool?" [00:18:17] Lisa Neigut: Or like, "Give me the next block," kind of thing. As long as your application can respond to all those questions, like anyone can write their own block source for Core Lightning. The default one that we ship with works with bitcoind, but there's a lot of people that run other versions. And in fact, one of these custom block source plugins that someone wrote, called TrustedCoin, used BTCD, which was the library that had the problem. So everyone who was running a Core Lightning node using the TrustedCoin block source had the same problem. Their stuff froze. [00:18:53] Lisa Neigut: I was going to say, not every Core Lightning node was impacted. Anyone who was running a block source that used the BTCD library was impacted. But most people who run Core Lightning use bitcoind. And bitcoind didn't have... bitcoind was able to process all these transactions correctly. The other thing that Core Lightning does very differently than I think like this BTCD package is that Core Lightning, for the most part, assumes that you, as the person running your Lightning node, picked a block source that you trust. [00:19:28] Lisa Neigut: So most people trust like bitcoind, right? Like if you run your own full node, you did the initial block download from scratch, right? Your node has verified every single transaction and block that has come through your Bitcoin node. Any data that your Bitcoin node has is trusted data, right? It's not trusted, but like your node has done the work to verify all the blocks that are on it, right? And so when Core Lightning asks it for block data, we just assume that it's done all of the correct parsing, like we just kind of assume that it is valid data. [00:20:03] Lisa Neigut: And then when we read it out, we parse it. I haven't looked at the exact code on how this works. I really need to, but basically the problem in BTCD, maybe this is way too much info, whatever, but basically the problem in BTCD had to do with the limit of how much... So in computer programming, one of the things, one of the like work that a computer program has to do is you take like a long set of information and you transform it into sort of like different fields and stuff, so that your program can easily access data because it's got labels and stuff. [00:20:45] Lisa Neigut: It's basically you take a long form of just like jumbled words and you organize it with labels so that your program then can easily access different parts of information. Anyways, we call this parsing. When you read data in, though, if someone sends you way too much information, and your little program tries to read in a whole bunch of data, it's possible it'll run out of memory, and the whole program will crash. So there's definitely a crash factor on your node any time you read data in off a wire. [00:21:19] Lisa Neigut: How much data you should read in before you can stop reading data in is one of the hardest questions in computer programming, right? Is like, maybe it's four megabytes, maybe it's 10 megabytes, maybe it's half a gigabyte of data that you read in before your program goes, "Nope, that's too much data. I can't handle that," and falls over. So, this is where I don't know exactly what Core Lightning does in terms of reading it in, but I think we just use a default buffer size that's pretty big. [00:21:52] Lisa Neigut: And we don't really worry too much if it's a block or whatever. We'll just read it in or not, because we trust that whatever data your block source sends us is valid, whatever valid means. So we just read it all in and have some generic large buffer size that we don't want to go over. But it's not at all related to how big the block would be, right? Sort of like a generic buffer size, I think. Whereas in BTCD, whenever they tried to decide how big of a buffer to read in, they used numbers that are considered consensus numbers. [00:22:28] Lisa Neigut: So like, well, consensus says, the rules about how big a witness stack can be in a SegWit block is like 10,000 bytes or something, right? So they decided that they would only read in 10,000 bytes. And then if they read more than that, they would throw up and say, "I can't read this," right? Like, "It's more than I'm allowed to read." And the reason they picked those numbers is like, well, that's the limit on the protocol. As of SegWit, that was the limit on the protocol. [00:22:58] Lisa Neigut: So it should be good forever, right? Like, it's a very reasonable thing to do as a software programmer. But it runs into problems: as we found out, Taproot changed the rules and they removed that number. They said it can be much bigger than that. But the software still had that limit from an older, what used to be consensus. So it's a shitty bug. It's shitty because you think you're being smart or clever by using this number from the spec, right? [00:23:28] Lisa Neigut: You're like, "Oh, well, I'll just..." You know, deciding how much buffer data to read in is, like I said, very much a hard problem. You're like, the easiest way to solve this is go figure out what the biggest number allowed ever is, and you just use that, right? But as we found out, the problem is that the spec changed, and all of a sudden that assumption was no longer good, and everything started falling over for people who were using that software. [00:23:52] Lisa Neigut: So yeah, it's a mistake that anyone could make because you're trying to be smart, but it turns out the real smart thing to do in terms of wire parsing is to use a number that is, I guess, the size of... they changed it to be the size of the largest allowable block, but that assumes that we're never going to change the block size. I don't know. I need to go figure out exactly how big Core Lightning makes its buffers, because I think it'd be interesting, but I just haven't gone and looked. [00:24:20] Lisa Neigut: I don't know where it is. I'd have to go find it. Sorry, I've talked a lot. I don't know if that tells you more information, and maybe that's way too much information. But that's my understanding. [00:24:30] Vlad Costea: It's actually a lot of information. And someone actually accused Blockstream, I think it was today, of purposely looking for limitations in LND and breaking them. [00:24:40] Lisa Neigut: Yeah, no, so I think what happened is... okay, so what exactly has been happening is someone has been building Bitcoin transactions and then getting them mined, and the existence of that transaction in a block makes problems with this particular BTCD library, right? So there's one guy who's been figuring out how to make these big weird transactions and then figuring out how to get them mined, right? He did it the first time and it broke LND. And then he went... [00:25:10] Lisa Neigut: Back to the same library, found another way that he could trigger basically a very similar bug, the same kind of bug, because there was still an existing problem in BTCD. So he built another transaction and broadcast it. And in his message, he said, I think it is, "You will run CLN, you will be happy." [00:25:31] Lisa Neigut: And so then all of a sudden, we had people on Twitter who were like, "Blockstream, what is your relationship to this? Why are you breaking LND? Why is this guy doing this thing?" The other thing about this... I'm going to call him a researcher hacker. Burak is literally brilliant. He's very smart. He has his own company called Bitmatrix that I first found out about because they were building... Bitmatrix was building very complicated covenant-style DeFi trade contracts on Liquid. [00:26:13] Lisa Neigut: So he's been doing Liquid research and work for years, and now it seems turned to making extremely weird and large Taproot transactions on mainnet. So there was some confusion, even amongst... I was under the impression that Blockstream had invested in Bitmatrix to do Liquid stuff, because... [00:26:38] Lisa Neigut: You know, it would make sense that we invest in the Liquid space, but I was wrong about that. We had not invested in Bitmatrix, as I later found out. This guy is just an independent guy who really likes being able to do covenants on Liquid, and Liquid lets him do that in kind of a Bitcoin setting. So yeah, the fact is that Burak, obviously, is a fan of Core Lightning and likes what we're doing. We're also like, you know, I think there's maybe 5% of the entire Lightning Network that we know of that runs Core Lightning. [00:27:11] Lisa Neigut: We're a very minor, we're a small part of the network. We're very much like an underdog in the Lightning node space. So I think he really likes what we've built and the kind of level of care that we put into our software. And so I like to think he's just a big fan of us and our team. And so, you know, he's like a fan, like, you know, he's a big fan of our project. And clearly, like, I mean, sometimes I think there is some amount of like amusement in like bringing down like whoever's the top dog, whatever. [00:27:50] Lisa Neigut: Especially like, I don't know, it's like a similar bug to the first one. So anyways, like, yeah, so he definitely, I think, antagonized the LND community by doing it again and not telling anyone he was going to do it. He just did it. And people thought that Blockstream was somehow responsible for that. And so we were getting called out for funding the attacker, sponsoring the attacker, who like, you know, published a transaction which hit a bug that LND had in it, if that makes sense. [00:28:24] Lisa Neigut: So like, yeah, I mean, I can see. I mean, I definitely was under the impression that we had invested in Bitmatrix, which Burak is a part of. So I can see why other people would think, also like the CLN thing, I can see why people would think that we were like somehow funding this guy who was attacking. [00:28:45] Vlad Costea: No, no, I'll stop you here. Don't say anything more because it might require lawyers. [00:28:50] Lisa Neigut: OK, yeah, yeah, right. That makes sense. Yeah, yeah. But like, you know, I think the truth is like none of us knew anything about it. Like, I don't know. It's like, I don't, yeah. Then there's this whole big thing about like what you should do if you find a bug. And it seems like a lot of people are like, you should not publish, you should not take advantage, you should not like activate the bug if you know about it. You should tell people. I think that's probably the right way to do things when you find bugs. Yeah, I don't know. Like, I don't know. [00:29:29] Vlad Costea: Yeah, I'm not sure if people are gonna be happy about some of the stuff that you added. But you're super honest and that's fine. Like you did not self-incriminate, just for the record, but at the same time you made it look so dubious. Like people are going to ask questions right now, like, "Oh, so this guy does stuff on Liquid. Okay, he's a super fan. That's suspicious." [00:29:59] Lisa Neigut: And dubious in what way, though? I guess, like, I don't know how else to, like... I don't know what, like, suspicious means exactly. Like, what does suspicious mean? [00:30:09] Vlad Costea: No, they would imply complicity. I think that's the right way to put it. [00:30:15] Lisa Neigut: Yeah, but what does it mean? Like, but what does it mean to be, like, complicit? Like, none of us... I didn't know there was a bug, and like no one knew that that was going to happen, if that makes sense. I don't know. [00:30:32] Vlad Costea: No, let's change the topic anyway. I wanted to ask you a little bit more about running Core Lightning on my little device. And had I run LND and not followed Twitter at all, let's say I was on vacation or something for a week, I guess I would have had all the channels force closed or something, unless there was some malevolent actor with whom I opened channels and he could steal money from me. Yeah. So I think it's useful at this point to explain to people this danger that you have in Lightning, that if you don't have the latest state of your channels, and you're not constantly online to synchronize with the rest of the network, there's this risk that someone can exploit by sending an older state of the channel to you and force closing and therefore taking your money. [00:31:28] Lisa Neigut: That's right. Yes. Well, so yes. How do I explain? The thing with the state is you need the latest copy and you also need to be online looking for someone to publish not the latest state, right? So there's like two parts to it. One is that if you don't have the latest state and your peer somehow finds out about that or decides to like test you, that sounds weird. [00:31:54] Lisa Neigut: If you don't have the latest state, and you don't have the state, like, if you don't have, like, the ability to correct when someone publishes the wrong state, I should say, then they will like get money that they probably shouldn't have because that wasn't the current state of the channel. The other thing that, what do you call it? The other, yeah. So like if you don't have the current state, that's a problem. [00:32:21] Lisa Neigut: If you don't have any state, that's also a problem because your peer can publish any state, probably the one where they have the most money, right, no matter what the current balance is, and then they're able to get money they shouldn't have. Other problem is if you accidentally publish an old state. So like maybe you don't know that you're behind and you accidentally publish a state that isn't like what the most up-to-date thing is. [00:32:50] Lisa Neigut: Based on the way the protocol currently works, your peer can like take all of your money, which is not cool at all, right? So there's a couple of things here. One is like you always want to have the latest state. The other is that if you're not watching the chain, your peer can attempt to publish an out-of-date state. And if you've gone and you're on vacation and your node is down, then they'll be able to basically publish an older state where they had more money than they currently have. [00:33:19] Lisa Neigut: And basically get all that money that they shouldn't have. There's two ways, kind of. There's like three ways to fix this, sort of. I don't know if fix is the right way, or maybe avoid it. First one is never go offline and never lose your database and always be up to date. Part of the way you can do that, Core Lightning has a pretty cool way that you can run a second disk. And every time that the database makes an update, it'll write it out to that second disk for you. [00:33:47] Lisa Neigut: So you can use like a database backup plugin. So you have two disks that are like independent, is the idea. And they each have a replicated copy of your database. So that way, if your node ever goes down or there's corruption on one of the disks, you have that second disk that has the perfect backup copy of your database. So that's one way in Core Lightning that we kind of help get around that problem. The other thing is, what do you call it? [00:34:16] Lisa Neigut: So it's kind of like option one is like just always have like magic perfect backups or some sort of backup form such that if something goes wrong, you can like replicate back from like your secondary disk and it'll work. The second thing that you can do is use a watchtower. I think this helps for certain cases. But a watchtower is like a third party. Ideally, it's a third party. Or it's like a watchtower that you run, but on a totally different computer, hopefully on a totally different network, with the idea that if something happens to your node, the watchtower won't be affected. [00:34:53] Lisa Neigut: And as your node makes updates, you send the watchtower the latest copy of your channel state. So the watchtower has like this sort of anonymous, in theory, I'm not really like 100% sure on like the privacy concerns around watchtowers. I think they're pretty good, but the idea is the watchtower doesn't really know, or if it's a watchtower you run, it doesn't really matter. But the watchtower, if it's a third party, doesn't really know who you are or what transactions, like what channels are watching. [00:35:23] Lisa Neigut: But if they see that transaction go to chain or any of those go to chain, they'll have the history and they'll be able to take action on your behalf if your node is offline or if it's like fallen off. So like, number one was like back up your own stuff locally or have like a replicated database. [00:35:44] Lisa Neigut: Option two is use a watchtower, who is like a third party who you kind of do backups with, like a third-party protocol of making like backups, but it's better than a backup. Like a watchtower is not quite a backup. The watchtower also has that component of like watching the chain to make sure that if anything goes on chain, they can react, right? So they kind of also have this, like, it's not just a backup database. It's also like a reactive component. [00:36:10] Lisa Neigut: Like they'll take action on your behalf and like save your money sort of thing. And then the third option would be to, well, I don't know if this totally fixes it, but it solves some of it, is to like move over to like something like eltoo, which, it's only helpful in that it like reduces the amount of data that you need. And it makes your database backup like a lot thinner and a lot smaller. You only need to remember one thing per channel instead of right now, like a lot of things, you need a whole list of things to not lose money. [00:36:41] Lisa Neigut: So it just really reduces the total amount of data you need to keep track of. Yeah, so as long as you only have the last piece of information with this new eltoo protocol, or maybe there's other protocols we can move to. eltoo is just the one that I know... Do I know the most about it? I don't know everything. I know the most about it, but I don't know a whole lot. Anyway, so those are kind of like three things, kind of like current state, ideal state with a watchtower, and then like future state, maybe with a new protocol that helps make it a little less bad for losing data. [00:37:28] Vlad Costea: Commercial break. [00:37:32] Ad voiceover: After Mt. Gox collapsed, that was really the precipice of me saying, "Right, this has to change. We need a totally transparent exchanging system and base it on gold instead of fiat." [00:37:47] Ad voiceover: Voltoro is the hard money exchange which helps you beat inflation with instant swaps between the best stores of value known to man: gold and Bitcoin. Unlike most exchanges, Voltoro understands the importance of transparency and security. All gold holdings are secured in top-tier Swiss private vaults and fully insured against theft, fire and more. Maximize your purchasing power today by going to voltoro.com/bitcointakeover. This is not financial advice, but gold has been humankind's most reliable store of value in the last 6,000 years. Do your own research. Use promotion code TAKEOVER for a one-time bonus of 1 gram of gold for the first 50 new customers buying gold with Bitcoin. [00:38:46] Ad voiceover: Are you concerned that your friends, neighbors or KYC exchange might know how much Bitcoin you own? It is time to take your financial privacy seriously with Wasabi Wallet, a free and open-source wallet solution which makes use of mega CoinJoins to mix your coins with those of hundreds of other strangers. Thanks to the groundbreaking Wasabi engine, your coins get divided in smaller, untraceable units, which grant you great anonymity for both hodling and spending. Download Wasabi Wallet 2.0 today at wasabiwallet.io and take advantage of the mega CoinJoins. It's free and it's open source, so don't trust, verify. [00:39:55] Ad voiceover: What are you going to wear when Bitcoin hits 1 million dollars? The same old 20 dollar T-shirt? Try M-A-I-S-O-N-M-A-S-H-I, the designer clothes made in Paris by Bitcoiners for Bitcoiners. They're not your average mass-produced sweatshop clothes. Mashi will ask for your measurements and tailor every piece of clothing according to the shape of your body, so you always look and feel great in your T-shirt, hoodie or dress. It's all made in France by real artisans who also happen to be Bitcoiners, which is why M-A-I-S-O-N-M-A-S-H-I only accepts Bitcoin as payment for their clothes. Get a M-A-I-S-O-N-M-A-S-H-I T-shirt or hoodie today. It might even help you find a girlfriend. [00:40:50] Ad voiceover: And once you do get a girlfriend, you can also buy her a M-A-I-S-O-N-M-A-S-H-I dress. Stop having fun looking poor and check out MaisonMashi.com. That's M-A-I-S-O-N-M-A-S-H-I.com. [00:41:10] Vlad Costea: Right, so coming back to this node that I'm going to run once again, I have this option to choose between. Obviously, I don't want to reveal to the rest of the world my IP address, and Bitcoiners are super paranoid about doxing themselves and their IP addresses, their location, everything related to that. So you want to have some sort of network-level privacy against the rest of the Lightning Network, and for that you can either run Tor, which is a very bandwidth-consuming layer of obfuscation. And it's very good when it works, but it's terrible when it crashes for some reason, and it has been subject to lots of DDoS actions in the last few months. [00:41:52] Vlad Costea: And you also get the option to run a VPN, which I guess is more reliable, is faster, but is not as private. And you need to be connected to the same server, so it subjects you to some more third-party risk, I guess. So what's your experience with this? [00:42:12] Lisa Neigut: Yeah, this is a great question. The question is like, you know, difference between Tor or VPN or like just clearnet, which is really almost impossible to do from having a home server. So like, let me see, like the problem, there's a couple of problems here, right? Like one is, I don't know about you, but I run my node, you know, not in a data center, right? It's like a computer that I have physical ownership over. And the reason for that is like, one, I don't want my keys for my Lightning node anywhere that like anyone who's not me has like access to. [00:42:49] Lisa Neigut: And then the other, so like, okay, so like you kind of start with this, like, you know, your Bitcoin node, your Lightning node, you run them at home. It'd be like if you were able to run it, sorry, if you were able to run it on like a server, right, like a box somewhere on the internet that was like in the cloud. The nice thing about that is that like the IP address for that server, if you were able to expose it, well, one, it's a lot easier to have an IP address on a server in the cloud. [00:43:19] Lisa Neigut: So when you run your node at home, you want people to be able to connect to you, right? It's really hard to connect to anyone who's on a home network. You don't have a fixed IPv4 address. I think IPv6 is a little different. But then you run into problems where... And then also, it's like, okay, maybe that's where your house is. You don't want to be putting advertising where in the world your physical location is to your IP address, right? So there's a few things here, right? [00:43:43] Lisa Neigut: So one way to fix that would be rent a box in the cloud, put your node on the box in the cloud. Then you run into the problem of like, well, I got to pay money every month to run it in the cloud. And also like Lightning is hot wallet. Someone else might have access to that box. I don't even know, right? So people started running nodes at home and they started using Tor, right? The nice thing about Tor is it gives you like an address that you can access from anywhere. [00:44:11] Lisa Neigut: It does this really cool tunneling stuff so that no one knows where you're located in the world. Plus, it works behind any kind of network, whatever. So it sort of immediately solves all these problems. It solves the problem of, I don't want to tell anyone where I am. And it solves the problem of, I don't know how to advertise my address properly behind this home networking setup. So Tor is, on the surface, a very perfect solution for all of the problems you might have trying to run a Lightning node at a house. [00:44:43] Lisa Neigut: But then started to run into like problems, so to speak, and Tor's like super unreliable, right? Like this is like the big problem with running it on Tor. And I think one of the reasons, I was talking to one of my co-workers at Blockstream, Warren, who's like done a lot of work, and I think he probably has like one of the best articles that I've seen about how to set up your node such that you can use like a VPN service and not leak your IP address. You don't have to use Tor. [00:45:10] Lisa Neigut: But Warren's whole thing that he says, at least that I think I remember him saying about Tor is like, Tor is not really meant for long-running connections. So most of the time when you're like browsing a website or something, you are sending like one-off requests for more data at a time. It's kind of like you're reading or maybe you're sometimes writing data, but it's a little more like, not like this long-lived persistent connection. You're on the website for a few hours and you disconnect and come back. [00:45:38] Lisa Neigut: And Tor works fine for these kind of one-off messages sort of thing. The idea with Lightning nodes, however, is that the type of connection that a Lightning node has between each other are supposed to be like very long-lived, right? Like you're sort of expecting every Lightning node to always be up and always be available and always be able to respond to payment requests, right? And if you start getting any trouble in that connection, if you agree to make a payment, like accept an HTLC and then aren't around to like finalize it, there's kind of like this two-step process with HTLC payments. [00:46:15] Lisa Neigut: There's like a commitment phase, and then there's like a settlement sort of like phase thing that happens. If you're not around to do the commit and the settle, then if you're there for the commit, but then you like kind of disappear before the settle, then all of a sudden your channels will start closing because that full loop has to happen within a timeframe. Otherwise you have to like force close your channel basically. So Tor is unreliable in that Tor creates problems for Lightning nodes in that you start getting these like connections that fall off at random times and don't get reestablished. [00:46:47] Lisa Neigut: And then if you've got a payment in process, it can lead to, if that instability lasts for long enough and you're unable to get back in touch, it leads to this like force closes on channels, which costs money, right? Like because you have to pay on-chain fees and stuff. So Tor is unstable and potentially has a risk of being expensive, has a risk of leading to force closes in a way that non-Tor connections don't. So one of the downsides of running, like, right, so Tor solves these problems about like, I want my node somewhere... [00:47:20] Lisa Neigut: Not like at a data center, and I still want to be able, and I want to tell people where I want to be able people to be able to connect to me without me like exposing where I am exactly, right? So Tor solves those problems, but it has its own issues, and that the connection, it's not really built for the kind of connections that Lightning nodes need to be, to kind of get rid of that risk of like falling on chain for like unresponsiveness. So okay, so it's like the Tor thing. So you can definitely run Tor, like I said, it's got certain advantages, it's got some definite downsides. [00:47:50] Lisa Neigut: And also like, you know, if everyone's running a Tor node, it means that like payments across the network might get stuck a lot more frequently because everyone's sort of in this position where their connection to their peers might not be stable. And then so like everyone being on Tor to some extent makes like the ability to make Lightning payments a little less certain, because there's always this element of like, oh, we don't know which of these, like if everyone on the Lightning Network had channels that were over Tor, some percent of them would be, I would assume, like falling over or having trouble every time. [00:48:23] Lisa Neigut: So every time you make a payment, what's the chance you're going through a connection that's having problems at the current time? It would be like, you know, a certain percent. And if everyone's on Tor, it's like higher than zero or higher than if everyone's on clearnet. The other option is the VPN option. So the VPN option trade-offs are, one, you don't have to have, you don't get the flakiness of Tor, but you need a, like, you need a VPN, basically means that your exit node is like, I'm going to call stable, right? So like on Tor, the address that your traffic comes through changes, but you still have a stable address or whatever, because of the Tor onion routing thing. [00:49:00] Lisa Neigut: And then for... But for clearnet with like a VPN, you have to pick a stable IP address that you're going to send to be your, pick to be your VPN. And then once you have a stable thing, it's how attached to your identity is that stable connection, right? Are you... So in my case, I am renting a box that I use to bounce my Lightning traffic through. So the IP address that I publish is very far away from wherever my box actually is in the world, but it like secretly tunnels all the data to wherever I am on the internet with my box. [00:49:40] Lisa Neigut: But yeah, if someone decided they wanted to figure out who's running that node, they would just have to subpoena the company that I'm running the box from, right? So there's like, you know, it's like, what's your threat level? Like, it's like, okay, the people at this company know who I am, and at least like where my billing stuff is and who I am and what my IP address is for that particular box. So anyone who has the ability to ask that company to give them information would be able to figure out who is running my node, right? [00:50:10] Lisa Neigut: I mean, aside from the fact that like, I think I've like told people what my node is. But like, aside from that, if I hadn't told anyone that, like... So once you're running the VPN, like I said, then you have to, like, whatever your relationship is to wherever that VPN endpoint is, that becomes a vector of someone being able to figure out who you are. But depends on what your threat model is. But the upsides are that it's not as unstable, really. I think is that like the only upside? That might be the only upside that I can think of. Yeah, that was a long explanation, but... [00:50:49] Vlad Costea: Yeah, when I had my node on for longer than a year, I think it was about a year and a half, I was very much concerned about the fact that I was not really routing transactions. It was on Tor. And at first my reaction was, "Okay, this just means that I'm not very well connected, so let me open more channels." And then I opened more channels with more people and I realized that it does make a difference. And I said to myself, okay, this means that I need larger channels. [00:51:15] Vlad Costea: I need to have more capacity. And I opened larger channels and still there was not any kind of tangible difference. And later I found out that usually in this ranking system of nodes, you're usually de-boosted if you're on Tor for the specific reason that you're usually slower and you delay payments. And that's something that I did not know and something that I want to get right the second time around. [00:51:44] Lisa Neigut: Yeah, that makes sense. Yeah. So, second time around, what have you considered for your VPN? Like, have you... So for me, I have to actually pay money every month to rent a cloud server that I use as my like VPN front end. And I basically use an SSH port forwarding script to do my poor man's VPN. I don't think it's what Warren recommends because you do leak your like home IP address a little bit. But it's definitely better than nothing. Yeah. There's services though I've heard of. I think there's like a VPN service that works. I don't know. [00:52:31] Vlad Costea: Yeah, I've heard some good stuff about Mullvad. And also, what's the other one? IVPN, I met one of their guys in Riga. And he told me that it's pretty good. And what's the other one that's good? I think Proton is also decent. [00:52:50] Lisa Neigut: Cool. Yeah, I haven't heard of any of those, but that's cool. Those are good resources to have. Another option that I don't know how ready it is, but like, I haven't done a lot of research into it. I just know that one of my teammates has mentioned them as maybe a possible alternative to Tor. I don't know how well they work with long-lived connections like Lightning, so kind of like putting those disclaimers out there. There's this project called Nym, which I think they're building like a network, like an anonymity networking network. [00:53:24] Lisa Neigut: Also, I don't know if any Lightning, I don't think any Lightning nodes work with it yet. It's like not really compatible the same way that Tor is with like addresses and stuff. But it's definitely like on my sort of radar in terms of things that are happening in the ecosystem. [00:53:37] Vlad Costea: Yeah, I did interview Harry Halpin and another one of their guys. And I'm super interested in Nym. I don't care much about their token monetization system. All I know is that this is a lot more advanced than Tor. And unlike Tor, it doesn't have the exit nodes controlled by the Department of State and lots of US three-letter agencies. So from this point of view, it's more pure. It's also VC-funded, so there's some shady stuff there. But the technology itself is super useful, and if we find a good way to use it, that can be for our benefit. [00:54:17] Lisa Neigut: Yeah, that makes sense. That's really cool. Yeah. Yeah. Anyways, yeah. But depending on what you want to use your node for, if you want to be like a routing node, figuring out how to make the VPN work for you is probably the right move. Just like you're saying in terms of being able to like have good routing statistics and less of a chance of having channels fall to chain, which is kind of risky. [00:54:50] Vlad Costea: Yeah, and something else which I want to mention, and it was something that popped out when I was in Riga because mempool.space updated their website to also include information about the Lightning Network. And they were able to do this very nice graph where they show where the nodes are hosted. And about 2,000 Bitcoin, and this is a large chunk of the network, it's hosted on Google Cloud. And there's another, I think it's a quarter of the network hosted on Amazon. And then there's DigitalOcean, there's Alibaba, there's Comcast and lots of others. Basically, you have too much reliance on this big tech infrastructure that's being provided for rent. [00:55:43] Lisa Neigut: Yeah, so here's the thing, though. So my node will show up, like, for example, under the Google Cloud thing, but my node isn't actually hosted at Google Cloud, just like my front-end server is, right? So if Google Cloud took my node connection data down, I'm pretty sure I would be able to put my node... I just have to rent a new box and get it set up. Even so, my node wouldn't change. So even if like Google Cloud took it down, I guess I'm saying you don't know how much of... [00:56:13] Lisa Neigut: That money is actually that service versus how much of that money is like hiding behind that service somewhere else and just being used as like proxies. Does that make sense? [00:56:23] Vlad Costea: Oh yeah, it does. I guess it's the same situation as we look at mining centralization. And sometimes we say, "Oh, if Bitmain does this, then look at how much of the network is controlled by Antpool and some of their other Chinese mining pool friends." But at the same time, I can be in Romania and mine with Antpool, so that doesn't really mean much. It just means that people use pools to mine. So I guess it's the same situation with Lightning here. You connect with your node via Google Cloud or Amazon for privacy or reliability purposes, but it doesn't mean that it's a single point of failure. I guess. [00:57:06] Lisa Neigut: Exactly. Yeah. Yeah, exactly. And honestly, like, even if assuming that, like, and I'm not sure, maybe people actually are running their nodes on it, and I'm like the weird minority, but I exist, so maybe lots of them do. The other thing I was going to say about that is like, I thought looking at the chart, like, even if Google Cloud was like, you know, like, it was only like, it wasn't 50%, right? It was like a lot, but like, there wasn't one particular part that had like a huge part, more than 50% of the network, right? [00:57:41] Lisa Neigut: Like, I think the Google Cloud stuff was like maybe, I'm gonna say 30% or so. So I thought it was pretty cool that we were pretty like spread out across all the different hosting groups. So even if like Amazon went down and was unreachable, only like, you know, 20% of the Lightning capacity would be like unable to make payments, which is kind of interesting. [00:58:07] Vlad Costea: Yeah, you know, all of this is new to me and a lot of people listening to this, so it's useful to clarify some aspects. Because you open the page, you look at it, and you're gonna jump to conclusions that it's super centralized. But this is worrying to some extent. It makes you realize that we do need to run more nodes at home. And getting back to this little guy, there was a lot of FUD about the Raspberry Pi, I guess, a couple of months ago, and everyone was saying you should not be using that, you should buy a laptop or you should buy hardware that's designed to be a server, like Intel Xeon processors with sturdy motherboards that are made to be 24/7 online and stuff like that. [00:58:53] Vlad Costea: But this little guy lasted for like a year and a half, so that's a lot more than you get from a laptop or whatever, at least for the cost. And it's also super cost-effective with energy prices that are rising. And that's also a concern. It's quiet. You just put it somewhere. I have lots of reasons to still use the Raspberry Pi. And I was curious to know, what's your take on this? [00:59:16] Lisa Neigut: I wouldn't put money on a Raspberry Pi, I think, is my take. But I understand that a lot of people do. If you are running it on the Raspberry Pi, I strongly recommend you get a disk backup or figure out a way to have two disks that you're reading and writing from. I don't know, I think that they're definitely not as robust as other processors are. And I think that's always like a little bit of a risk in terms of running financial software. So. [00:59:54] Vlad Costea: No, I understand what you're saying. And I should probably look into better hardware, but for my personal use, and most of the times I just want to make payments and connect my phone to this home node, I guess it gets the job done. If it doesn't, I mean, I'm going to close all the channels and look for something that's more reliable, I guess. But this second time around, I'm gonna be more cautious with the way in which I open channels. And I'm also going to pay more attention to this situation with Tor and VPNs. [01:00:32] Vlad Costea: Because I was pretty well connected, but at the same time I was not routing because I was on Tor, and that put me in a very undesirable place. And now I'm going to learn from my past mistakes and try to do better. If I don't do better with the Raspberry Pi, it's just a sign that I need to invest in something better than this. And that's fine. It's just the second phase of my experimentation. The first time when I tried to run Lightning in 2018, I actually burned my SSD drive. [01:01:03] Vlad Costea: I think drive is in SSD, but I burned my SSD trying to synchronize with Lightning, and I thought, okay, this is so exciting, this is so early that we're still in this pioneering phase where hardware gets destroyed. It's like people trying to fly with their airplanes and crashing and being like, "Yeah, I need a better prototype." [01:01:29] Lisa Neigut: Yeah, it's definitely like that. I'm also wondering, so I know you're based in Europe, right? Somewhere. I don't know how public you are, but you're somewhere in Europe, right? [01:01:45] Lisa Neigut: Like, I'm wondering, like, so in... I live in the United States, I live in Texas. I'm wondering if it's easier for... So like, I can buy a used like business commercial desktop for a couple hundred bucks, and that's what I use as my Lightning node. And it's got like more expensive processors that have gone through more rigorous testing, whatever. But I'm wondering if like, but that's like, you know, if I can spend like a couple, like 100, 150 bucks on, you know, a lot of memory, a good processor, it's like a couple years old, but it's been used successfully. [01:02:24] Lisa Neigut: I'm wondering though, like that's almost around the cost of what I would get to get like a nice Raspberry Pi with like, you know, a nice hard drive or however that, I don't know exactly what all this stuff is. But I do think the price point for me at least is about the same. But I'm wondering, especially given like some of my experiences in like Brazil and undergrad, like other places in the world, one thing that is American that you don't really appreciate is like the access to like good computer hardware for cheap prices that you have in America is not the same as like everywhere else in the world. [01:03:07] Lisa Neigut: Like in Brazil, this was like a long time ago, like a decade ago, but every computer was so expensive and getting any amount of like... I didn't even know how much Raspberry Pis cost in Brazil, but I would be very surprised if you have as much access to like really good quality, cheap used computers like in other places. So I'm saying all this to be like maybe a Raspberry Pi is like actually the most cost-effective option in other places. So, yeah. [01:03:33] Vlad Costea: Yeah. [01:03:34] Lisa Neigut: I don't know. [01:03:34] Vlad Costea: Yeah, it makes sense. I can also get for like $250 a ThinkPad T430, Intel i5 processor with 16 gigabytes of RAM. I think that's a lot more than I need for a node. I can run some sort of Linux distro and let it run with the node 24/7. I guess I could do that, but at the same time there's always the question: how do you know that the hardware that you're buying pre-owned is not compromised? Because you're going to put money on it, right? [01:04:06] Lisa Neigut: Oh, I see what you mean. Yeah, that's an interesting point. I hadn't really thought about it. I guess I installed a new operating system on it entirely. But if someone had done something a little like more, I'm going to say, nefarious than operating system level, yeah, I probably would be at risk. I don't know, I guess like, yeah, I don't know. That's a great question. I don't have a good answer to. I mean, how do you know that Raspberry Pis don't have the same problem? Like now that everyone knows, like people know that Raspberry Pis are being used for like Raspberry Pi stuff. [01:04:46] Lisa Neigut: Seems like that would be a much easier build process. Oh man, I'm going to create so many people who are paranoid now. I don't mean to do that, but like, I feel like it would be easier to know that like the chance of any random Raspberry Pi being run as the Lightning node is probably higher than any random refurbished computer. Does that make sense? [01:05:07] Vlad Costea: Yeah, it does, because it's more general purpose. But I know people who use Raspberry Pis for video game consoles, for emulation. They use it for home servers. They use it for... What's it called? Like the Google device that you use for your TV. [01:05:27] Lisa Neigut: Yeah, I know what you're talking about. Yeah. [01:05:30] Vlad Costea: So it has lots of use cases that have nothing to do with Bitcoin. But I guess it's more obvious when you want the 8GB Raspberry Pi 4 that you're going to run a Bitcoin node on it. [01:05:43] Lisa Neigut: It's definitely easier to get access to a single source of new computing chips than it is to the aftermarket or refurbished machines, right? So, I mean, I guess it depends on where you're getting the refurbished machines from, but it's a lot more distributed. I think it would be harder to do something to any one refurbished machine versus the Raspberry Pi. Anyways, whatever, that's totally speculative. Where do you get your chips from? How much do you trust the like manufacturing process? I don't know. I mean, there's definitely people that like, I feel like I've like talked to people who are like, "Yeah, I'm like writing a program." [01:06:25] Lisa Neigut: And these are Bitcoin Core devs, so this is not your everyday Joe who's doing this. So I think it's okay. But they're like, yeah, they're basically writing some software to run on a TI-84 calculator that was manufactured before Bitcoin existed, so that there's no chance that anyone would have any sort of malware or like rootkit or whatever that would be able to like mess with their, you know, private key information. Besides, I don't even think TI-83 calculators have like, what do you call it, I'm pretty sure that doesn't have like Wi-Fi. [01:06:58] Lisa Neigut: So like, yeah, your vector of like someone being able to like mess with your hardware stack, your firmware stack or anything is like much lower. But I thought it was interesting that like people were like thinking so far as to find computers built until before Bitcoin was even a thing. [01:07:21] Lisa Neigut: So you have to go all the way into the past to find a pristine computer that there's like zero chance. To some extent, that kind of level of thinking about attack vectors feels incredibly sci-fi. We live in the sci-fi storyline where there's some event that happened in the world that was the invention of Bitcoin. And every computer from there on, you have to suspect might be, you know, whatever. So you have to go back before the event in time to find a machine that can save the world or whatever. [01:08:00] Lisa Neigut: Anyways, or just in this case, like hold secret key material with like you having zero worry about it being problematic. I don't know. Okay, really hope I'm not like building a million paranoid people from like speculating on this. But yeah, I'm pretty chill about my stuff. So mine runs on a refurbished machine that I bought a couple years ago. Hopefully we're not giving anyone ideas about this either. All right. Maybe I shouldn't say that. That's cool. Whatever. Yeah. [01:08:33] Vlad Costea: Yeah. So I feel like I have asked you all the selfish questions about my own node, which I'm going to start running. And at this point, I should also do you the favor of talking about the conference that you're organizing in Mexico City. [01:08:50] Lisa Neigut: Yeah, let's talk about that. So I'm organizing another conference. I just did one in June in Austin. The conferences I run are called Bitcoin Plus Plus. They're meant to be technical Bitcoin conferences. We really mostly do workshops and kind of longer form lectures and stuff. But I'm planning one for December. It's going to be in Mexico City, December 9th, 10th, and 11th. We're still getting our website up, but you can follow us on Twitter at btc++. And yeah, so what's going to be so cool about this conference is I work kind of on protocol stuff on Lightning that has to do with building transactions. [01:09:32] Lisa Neigut: And one thing I don't really know a lot about is like on-chain privacy and like heuristics around, you know, who owns what Bitcoin. So like, there's just like a whole lot I don't know about different ways of like looking at the chain and figuring out what's private and what isn't. And so I've invited a bunch of people who've worked on protocols for on-chain privacy to come and talk about the protocols they've built and what they like about it. And I also invited some researchers that work on things called anonymity sets, that I'm hoping are going to kind of talk about their work on that sort of protocol or way of thinking about on-chain privacy. [01:10:15] Lisa Neigut: And yeah, I think it's gonna be really exciting. Again, it's December 9th to 11th in Mexico City. I'm only selling like, I think, 80 to 100 tickets. So you should get your ticket. I think we're already like 10% of the way sold out. So yeah, it's gonna be really great. I'm really excited about my speaker list. I'm hoping to get out my list of speakers really soon so people can start getting excited about the kinds of conversations we're going to have at the conference. [01:10:44] Vlad Costea: Yes, and I hear that it's going to be a privacy themed conference and you have managed to invite people from Wasabi, from Samourai and JoinMarket. And you want to put them in a debate or something? [01:10:58] Lisa Neigut: Yeah. So the people from JoinMarket haven't confirmed yet. Wasabi and Samourai are both hoping to send some people who can talk about their project and their trade-offs. And yeah, I think that it would be really great to have all these conversations in person and give, you know, these different very popular projects opportunities to talk about the work that they've put into their protocols. Yeah, so I'm excited. Yeah, I'm really excited. I think it'll be really good. [01:11:33] Vlad Costea: Okay, so you're going to have a panel about privacy. I mean, on-chain privacy, but obviously there's going to be talk about second layer privacy on Lightning. Is there any connection between the two of them? Do you expect CoinJoins in the future to also open channels within the CoinJoins? Is that going to be part of the talks? [01:11:55] Lisa Neigut: It's not something we've planned on talking about. I think it's like, we're definitely going to have someone talk about splicing. But really, one of the reasons I'm running this conference is I'm really curious about how these different protocols work. I think there's things we can do on Lightning to maybe add some of those heuristics or understanding to our nodes. Maybe that's a big project we won't actually get into, but yeah, I think it's definitely a cool thing to kind of think about. [01:12:32] Lisa Neigut: There is a cool project that I feel like I should mention called like Vortex that Ben Carman made that is really good. Well, it's like not really good. It's basically built to do like collaborative CoinJoins from Lightning opens. So yeah, it requires running a coordinator... I think there's like a central coordinator. There are central coordinators on a lot of those projects. Having a central coordinator is a bit of a... it's not a custody risk, but I feel like it's a centralization risk. If you want to stop a service from working, you take the coordinator down, it stops working. [01:13:11] Lisa Neigut: I mean, you can always go to a different one, but someone has to run the coordinator and if there's takedown risk, that's like a whole thing. One of the projects I worked on for Lightning is the dual-funded protocol, which lets people build transactions collaboratively with Lightning in a very decentralized way. So the cool thing about the protocol is that it makes basically essentially like PayJoins, I think is probably the closest analogy to a dual-funded channel in like other payment world, non-Lightning payment world. [01:13:45] Lisa Neigut: And the cool thing about it, like I said, it's decentralized, so anyone who's got a Lightning node and wants to open a channel and can do it with a dual-funded thing is able to build, like, basically collaborative transactions with anyone on Lightning. And there's no coordinator that someone could take down and would stop working, basically. So I'm really excited about the dual-funded/v2 channel open protocol. Working on getting it working with Eclair. I've been busy traveling, so I haven't been working on it very much, but very soon we should have at least two Lightning implementations that can do it. [01:14:22] Lisa Neigut: And I know there's someone working on LND. Maybe at some point in the future, at least like three of the four biggest Lightning implementations would be able to use this protocol, which could make it way more decentralized to do PayJoins and very easy, pretty much. Yeah. Kind of almost like... not like you don't have an option, but it makes it really, really easy to do this and anyone can do it and you don't have to have a coordinator. Yeah, anyways. [01:14:57] Vlad Costea: Yeah, so the website for the event, if you want to buy tickets or something, it is btcpp.dev. [01:15:05] Lisa Neigut: Yeah, that still points to our old conference. They haven't gotten all the new information out. So if you want to find tickets for this, the Twitter account has all the right information as of right now, but hopefully it'll be up at btcpp.dev very soon. [01:15:21] Vlad Costea: Okay. And one last question for you before I ask you how people can follow you. I think I gave you these magazines when we met in Istanbul this year. And I'm pretty sure you must have gotten something at TABConf this year. [01:15:38] Lisa Neigut: I don't think I... I wasn't in Istanbul, so I don't think I got the Istanbul one. [01:15:43] Vlad Costea: Were you not in Istanbul in February? [01:15:45] Lisa Neigut: No, I wasn't. [01:15:46] Vlad Costea: Really? [01:15:48] Lisa Neigut: Yeah. [01:15:49] Vlad Costea: Okay. Well, bad mistake. But were you at TABConf this year? [01:15:55] Lisa Neigut: I was at TABConf, yeah, but I don't think I picked one up. [01:15:59] Vlad Costea: Was it not in the swag bag? I mean, swag bag, not gig bag. [01:16:06] Lisa Neigut: I don't know where my swag bag is. [01:16:08] Vlad Costea: Oh, you missed it. Okay. Last year's magazine is about a hundred bucks right now. So no big deal. [01:16:18] Lisa Neigut: I really missed out then, didn't I? Oh, man. Okay. Well, this is why it's worth $100, right? [01:16:22] Vlad Costea: You're going to find another one at Labitconf. [01:16:26] Lisa Neigut: I am going to find another one. You're right. Yeah. I'm going to save it. Trade it for sats? I can't believe someone's trading that away for sats, you know, like... [01:16:36] Vlad Costea: Well, collectibles, you know, these are NFTs. They're like physical items that people want to collect. [01:16:43] Lisa Neigut: Yeah, real life collectibles. That's cool. And yeah. [01:16:49] Vlad Costea: Anyway, beyond this awkward moment where I thought I saw you in Istanbul, when in fact I haven't, how can people follow you? [01:16:58] Lisa Neigut: Yeah, you can find me on Twitter. I'm at niftynei, N-I-F-T-Y-N-E-I. I'm also on GitHub. If you want to follow me on GitHub, I haven't been committing much lately. I'm at niftynei on GitHub. And then you should also check out the Core Lightning Twitter, C-O-R-E-L-N underscore L-N. I feel like, oh, I should probably also talk about Base58, like, for two seconds. That's my Bitcoin education project to teach you about the Bitcoin protocol. So if you thought I explained things pretty easy to follow today, you should check out our online Udemy class where I explain how Bitcoin transactions work at the byte level. So it's very in-depth. [01:17:40] Lisa Neigut: But you learn all about script and how scripts work and find out how I know so much about scripts. And script limits, I've got a lot of stuff in the class. We don't have anything Taproot. I'm going to have to make a new class for Taproot that goes into all the cool Taproot stuff. But yeah, that's my Base58 plug. [01:18:04] Vlad Costea: Oh, yeah. So could you mention the website once again? [01:18:07] Lisa Neigut: For Base58? Yeah, it's Base58. So B-A-S-E, the number five, the number eight, dot info, I-N-F-O. [01:18:18] Vlad Costea: I just opened it. I'm going to take a look because it sounds very interesting to find out how transactions work. I, for myself, have some questions about this, but since I have this resource, I'm going to read it as opposed to exhaust you with more questions. So thank you very much, Lisa. This was a very useful interview for myself and I hope also for the people listening. And maybe I'll get you once again to ask you more questions about the Lightning Network and Base58 and conferences and whatever. [01:18:49] Lisa Neigut: Yeah, I look forward to it, Vlad. Take care. [01:19:04] Ad voiceover: After Mt. Gox collapsed, that was really the precipice of me saying, "Right, this has to change. We need a totally transparent exchanging system and base it on gold instead of fiat." Voltoro is the hard money exchange which helps you beat inflation with instant swaps between the best stores of value known to man, gold and Bitcoin. Unlike most exchanges, Voltoro understands the importance of transparency and security. All gold holdings are secured in top tier Swiss private vaults and fully insured against theft, fire and more. [01:19:51] Ad voiceover: Maximize your purchasing power today by going to voltoro.com/bitcointakeover. This is not financial advice, but gold has been humankind's most reliable store of value in the last 6,000 years. Do your own research. Use promotion code TAKEOVER for a one-time bonus of 1 gram of gold for the first 50 new customers buying gold with Bitcoin. Are you concerned that your friends, neighbors or KYC exchange might know how much Bitcoin you own? It is time to take your financial privacy seriously with Wasabi Wallet, a free and open source wallet solution which makes use of mega CoinJoins to mix your coins with those of hundreds of other strangers. [01:20:50] Ad voiceover: Thanks to the groundbreaking WabiSabi engine, your coins get divided in smaller untraceable units which grant you great anonymity for both hodling and spending. Download Wasabi Wallet 2.0 today at wasabiwallet.io and take advantage of the mega CoinJoins. It's free and it's open source, so don't trust, verify. What are you going to wear when Bitcoin hits 1 million dollars? The same old 20 dollar t-shirt? Try M-A-I-S-O-N-M-A-S-H-I, the designer clothes made in Paris by Bitcoiners for Bitcoiners. They're not your average mass-produced sweatshop clothes. [01:21:52] Ad voiceover: Mashi will ask for your measurements and tailor every piece of clothing according to the shape of your body, so you always look and feel great in your t-shirt, hoodie or dress. It's all made in France by real artisans who also happen to be Bitcoiners, which is why M-A-I-S-O-N-M-A-S-H-I only accepts Bitcoin as payment for their clothes. Get a M-A-I-S-O-N-M-A-S-H-I t-shirt or hoodie today. It might even help you find a girlfriend. And once you do get a girlfriend, you can also buy her a M-A-I-S-O-N-M-A-S-H-I dress. Stop having fun looking poor and check out MaisonMashi.com. That's M-A-I-S-O-N-M-A-S-H-I.com. ================================================================================ Production notes: Remote video interview in English, recorded for YouTube and the podcast, with a music intro and two pre-recorded sponsor ad blocks (Voltoro, Wasabi Wallet, Maison Mashi), one mid-episode and one at the end. Correction: anthropic/claude-opus-5.5 via NanoGPT in 6 windows with a whole-episode speaker/glossary plan, the show notes, the Qwen3-ASR full pass (112 chunks) and 30 Scribe v3 re-checks of doubtful spans; [unclear] resolve pass: 4 of 4 flagged paragraphs re-checked. Automatic checks: word count vs raw, timestamps, seam duplication. Fixes: Removed music/© BF-WATCH TV hallucination and "Bitcoin is a very important part of the blockchain."; removed "Episode:" and inline "Lisa Neigut:" tags; fixed Neigut, Blockstream, c-lightning, RaspiBlitz, LND, Core Lightning, runes (room), macaroons, MACs, BOLT 8, read-only, rusty_twit, Burak; restored dropped text from second ASR (secure socket sentence, Clams/runes vs macaroons, error clause, "all the details"); repaired speaker turns at 00:01:24 (Hey, Vlad), 00:04:50, 00:05:33, 00:09:08, 00:11:24, 00:12:05; split long paragraphs with interpolated timestamps; Fixed Barack→Burak, L&D→LND, Bitcoin D→bitcoind, Core Lightning, TrustedCoin, Taproot/SegWit; "Moldeus egg" false start removed; "Mike Horn, Berkeley DB moments"→"BerkeleyDB moment"; restored dropped words (full capacity, runs LND, TrustedCoin plugin, assumption no longer good, 5%) from second ASR/re-check; removed hallucinated "I think that's a big problem"; "why?" moved into Lisa's turn; merged 00:24:30 split mid-sentence (Lisa's "But that's my understanding" / Vlad's "It's actually a lot"); Fixed Barack→Burak, L&D/LNG→LND, L2→eltoo, desk→disk, note→node, CLN/Core Lightning casing; restored dropped words from second ASR/re-check (backup option one passage, "fallen off", "I don't know if this totally fixes it", ad lines); removed "Thank you."/"Thanks."/"See ya."/"breasts."/"Lisa's like" artifacts; fixed "gold instead of fiat"; relabelled ad blocks as Ad voiceover (incl. misattributed "Voltoro is"); reordered Maison Mashi girlfriend line and dropped the out-of-place 00:41:06 timestamp; Maison Mashi spelling; split Vlad's interruption at 00:28:45; Fixed "read my node"→run, clear net/Clarinet/ClearNet→clearnet, tour→Tor, Mulvad→Mullvad, NIM→Nym, Tilkun→token, HTLC, Blockstream; restored dropped speech (Lightning node, world, downsides, 50:10 VPN passage, Nym funding lines) from Qwen/Scribe; removed hallucinations ("Bitcoin Takeover Podcast, hosted by Vlad Costea", "This is a very interesting topic", "Thank you for watching...", "Thank you."); moved "Yeah, I did interview Harry Halpin" to Vlad; removed inline Lisa tags; Fixed mempool.space, DigitalOcean (dropped "DataWeb Global"), "rent a new box", Antpool, Tor, FUD, Raspberry Pi, ThinkPad T430, attack vectors, Bitcoin Core, TI-84. Removed hallucinations ("Thanks for watching!", "This is a video about", "Bye.", "Thank you.", "Bitcoin Takeover Podcast, hosted by Vlad Costea", "This is a very interesting topic", "This is a video that I made for the first time", "Level 2", the stray "Episode:" label and repeated inline "Lisa Neigut:" tags). Restored dropped speech about the processor, the US hardware access and Brazil from the second ASR and audio re-check, and replaced "somewhere near a break" with "somewhere in Europe". Split "Exactly" off as Lisa at 00:57:06 and moved the speaker boundary at 01:05:30; Fixed names/jargon (Samourai, JoinMarket, Eclair, Vortex, Ben Carman, TABConf, Labitconf, niftynei, Base58, btcpp.dev, WabiSabi, hodling, M