S10 E7: Yuval Kogman (nothingmuch) on Improving Bitcoin Privacy
Search this conversation
0:29Vlad CosteaHello there and welcome to Season 10, Episode 7 of the Bitcoin Takeover Podcast. I am Vlad and my guest today is nothingmuch. That's his name on Twitter, but you can also call him Yuval. He is a researcher and a coder and someone who has over two decades of experience with coding. And I met him in Budapest when I was visiting the Wasabi Wallet HQ. He was maybe contributing or consulting the project that is going to become Wasabi 2.0. We're not here necessarily to talk about his affiliations, but we're here to talk with him about Bitcoin privacy and what it means and how we can truly achieve it.
1:28In this very limiting framework that we have with Bitcoin and under this situation in which the protocol gets more ossified over time and people seem to be more susceptible or more skeptical when it comes to any kind of change. So, hi Yuval, it's good to have you.
1:49Yuval KogmanHello, hello.
1:51Vlad CosteaYou're a bit shyer right now when you're supposed to present yourself than you were when we were just chit-chatting before recording.
1:59Yuval KogmanOh, I'm so bad at this stuff. I guess. Sorry.
2:12Vlad CosteaNo, it's fine. If you want to follow Yuval, he is M-H-A-G-Q-N-O-A-C-Y-F-M. Is that a zero or an O? I think it's a zero H5. So that's his Twitter handle, but you can look up nothingmuch. I'm also following him, so you can look on my list. So let's skip the introduction part and talk about, you know, the spicy stuff. Let's talk about Bitcoin privacy and what kind of privacy we actually have, because there are so many nuances and types.
2:48Yuval KogmanYeah. So I think from the legal perspective, I think that's where these like terms came from. Like you have counterparties and third parties. And I think that's a useful like starting point. Like whenever you're transacting, you're transacting with somebody else and Bitcoin is supposed to like disintermediate the authorization of payment, right? It's a peer-to-peer electronic cash or whatever Satoshi wrote in the abstract of the white paper. But from a privacy point of view, like sure, no third party can tell you can't send that money in principle, right? As long as mining doesn't get captured and...
3:41You know, that whole rabbit hole. But because the whole protocol is decentralized and permissionless in this way, to resist control and inflation, it needs to be fully transparent. And that creates a tension where, okay, maybe there isn't an intermediary, a third party involved in deciding to transact, but the whole world is observing, right? So I don't know if we should call that like a third party from the privacy point of view. I'm, you know, I don't really speak legalese. But I think it's, yeah, the perspective I would offer is you need privacy from the rest of the world.
4:37And you need privacy from whoever sends you money. And you need privacy from whoever you're sending money. And you need that for different reasons, right? It could be because it's financially sensitive, like maybe you're a business and you need to protect your business interests, you need confidentiality. It could be a safety issue, like maybe you're living in a part of the world where the social contract is fragile. And if you need to spend Bitcoin to, you know, sustain your life, to, you know, obtain housing or food or whatever, exposing information about how much money you have or something like that might put you at risk of theft or...
5:33Right? So there's very different threat models that apply to different people, different situations. And some of them like are critical at a point in time and some of them have like very long-term consequences as well. And I think like usually we focus on the transaction surveillance perspective, right? Like KYC exchanges and this kind of dragnet surveillance that mainly hurts fungibility. And I think that perspective is very important for Bitcoin as a system, but it kind of comes at the expense of the personal safety part, which is something that personally concerns me to the same degree.
6:23Vlad CosteaRight. But when most people think about Bitcoin privacy, I think they think about obfuscating the public blockchain. I think that's what they first think about. But there are layers to that one. You have sender privacy, which means that you're not going to reveal to the rest of the world how much money you're sending to that person, or you're not going to reveal your identity. There are these two layers. And then there's also receiver privacy, which is easily mitigated by generating a new address because it grants you some sort of plausible deniability, but it is also easily broken.
7:02If somebody has access to your master public key, the xpub. And that can also be improved in some ways. And we also have network-level privacy, which can be improved with Tor, with other types of network obfuscation mechanisms like VPN. You can also use Tor and VPN. Is there anything that I'm missing in this picture?
7:28Yuval KogmanI think you were saying much more clearly what I was trying to allude to with the counterparty stuff. The one major thing I would point out in addition is you have interactions between blockchains. So things like, you know, shitcoin forks that take, you know, the existing UTXO set. Those things kind of violate some of the natural assumptions people tend to make about Bitcoin privacy. And the second is information that is external to the blockchain, apart from just the network layer.
8:05And maybe like a good example here is, so dust attacks and rebroadcasting is kind of a nuanced interaction between these layers where, if you have, even if you respect sender or receiver privacy by avoiding address reuse, oftentimes the, like, there used to be, like in the past, more often than now, but this has happened multiple times, like...
8:51People, or I don't know, some entity sends very small Bitcoin amounts to addresses. And it's not clear how those addresses are chosen. But the purpose of that is insidious in two kind of related ways. One is, if your wallet uses such a coin, then you're basically tricked into reusing addresses, even if you're not, even if you're taking deliberate steps to prevent that, right? You're generating a fresh address for every recipient and so on and so forth. Like, dust attacks can still cause your wallet to be much more clusterable on the blockchain.
9:47But there's also this rebroadcast functionality in Bitcoin Core, which the default wallet, when it knows about a transaction that has an output that it can spend, and that transaction is not being mined, the default behavior for a long while, and it's still, like... There's been ongoing work to like improve this. Amiti, oh crap, I forget her last name, but she's been doing great work trying to address this problem in a very foundational way, where the behavior of...
10:32Bitcoin Core used to be to just keep broadcasting this, right? So if you don't have Tor-level, you know, network privacy, or even if you use Tor, but like your, the hidden service address that Bitcoin Core generates is still the same, right? There's ways of kind of tying your network-level identifiers to your blockchain-level identifiers. And for an adversary, like what matters is the sum total of the information, right? Like, nothing is preventing somebody who's actively trying to de-anonymize somebody or attack them from putting together independent lines of evidence to, you know, draw conclusions about the financial activity of somebody. So it's not just that there's also the network layer, it's also that there's all sorts of interactions, and that...
11:37Just using Tor is not enough, right? It's a tool that you can use, and if you use it correctly, then it can address certain issues. But, like, ultimately, the way that these elements compose is almost more important than the elements themselves, right? There's all these interactions, and the NSA has this like slogan, "The attack never gets worse, it only gets better," or something like that. Like, you should always assume that the power of the adversary is strictly improving in time. So it's very easy to accidentally destroy privacy in the past just because of some like unexpected leak like that.
12:41Vlad CosteaYes, and you know, the next natural question that comes to mind is why does privacy matter at all in the case of Bitcoin? And a lot of people who are going to hear about privacy, they're going to be like, oh, no, we should not be doing this because it leads to terrorist financing. It leads to all sorts of stuff that we don't like and we don't agree with. But at the same time, I suppose there is also a crowd of people who say, no, Bitcoin is perfect as it is, and we should keep it as transparent and easy to observe by anyone as possible, even though with adoption like in El Salvador right now.
13:17Where the government has control over the wallets and lots of users are basically custodians. I think that's the most Orwellian kind of financial system, and even the banks that we have today preserve your privacy a lot better than a system in which your government has designed and has control over what goes on with Bitcoin transactions. So, I suppose there are layers to this. You have, you know, privacy for the sake of it because you need it as an individual. You have privacy for the sake of the currency because you want it to be hard money, so it needs to be fungible.
13:56And there's also privacy to make Bitcoin, you know, reach its goal because, you know, a lot of people say, "Yeah, but Satoshi did not design much privacy into Bitcoin on purpose." But he did have a dedicated chapter about privacy in the white paper. And maybe that his assumptions are dated and were a bit naive maybe, but it just proves that he was a cypherpunk and he wanted this to work as decentralized Chaumian eCash maybe.
14:30Yuval KogmanYeah, I mean, Satoshi anticipated some of these problems. If you read that section of the white paper, the two main heuristics that are still used to and have been empirically validated through like 2011 to '13, there was a whole series of papers like using these heuristics. The first of them is the common input ownership heuristic. I think that was formalized... Sorry, I don't remember the actual paper that introduced this term for it, but Satoshi kind of implied this heuristic where if you see a transaction, you can just assume that all of the inputs belong to the same wallet.
15:13And the second heuristic is the change identification heuristic. There are others, right? So merely assuming that all the inputs belong to the same owner and guessing what the output, like which output is the change output, and clustering by that. This is a very naive and crude way of trying to recover information, but nonetheless, we have very good reason to believe that it's still highly effective. In particular, one of the most like mind-blowing results in my mind is Jonas Nick's master thesis. He identified the issues with Bloom filter privacy for light clients.
16:02So this was an independent way of clustering wallet addresses with very high accuracy. And he used this bug, which he responsibly reported, and it was, you know, more or less fixed, right? This functionality is, I think, now disabled by default. And most light clients have like used different ways of giving an acceptable user experience for like allowing light clients to discover their own addresses. But I think the most remarkable achievement of that work is the, like, a very scientifically rigorous way of measuring the scale of the problem, measuring how effective these clustering heuristics are.
17:05Now, it's from 2015 and it only applies to certain kinds of wallets, but it appears that this is extremely effective, and in particular, it's like, I think like an 80% plus recall rate. So if you only use the heuristics, you can get very close to the ground truth, as he managed to measure by exploiting the Bloom filter privacy leak. And this is kind of a terrifying result. So in this regard, I think Satoshi was very optimistic about, right, he was aware of the problems, but he had, like...
17:49I think he did not realize just how fragile this is. And it's in a way a little bit ironic, because at the consensus layer, within a transaction, individual satoshis are 100% fungible, right? Like you put some inputs into a transaction, you take some outputs out of that transaction. There's no way at the protocol layer to tell, like, you know, this output had three satoshis coming from that input and five satoshis coming from that other input. Like, that's just not a nonsensical statement. So, yeah, it's like a complicated... Right? There's many layers to this, and I think the flaws mainly have to do with how we apply the protocol.
18:49Going back a little bit further to like your question, so we...
19:05Vlad CosteaYeah, so another part of my question was why does privacy matter at all and why do we even bother with this?
19:13Yuval KogmanYeah, and I was like trying to organize my thoughts with regards to actually the first thing you mentioned, which is, you know, the whole like terrorist dilemma, and kind of got stuck, you know, organizing the path I wish to trace through these things. So maybe let's get that one out of the way first. I think that's a 100% bullshit argument. First of all, because like the fiat system already gives us terrorist financing and drug cartels, you know, Deutsche Bank, HSBC. I think Santander had like issues with money laundering in Puerto Rico. Like basically, the system is already broken, and it's broken in a discretionary way where the...
20:03The public ends up paying for these negative externalities and there's no real consequences, right? Like when HSBC launders like God knows how many, and like I think the last thing I heard about was like in the trillions of dollars over the years, and then they get fined for a few hundred million, right? That's just a negligible part of the cost of business for them. Like, you know, the slap on the wrist is utterly meaningless from an incentive point of view. And secondly, cash is already, you know, very widely used by criminals.
20:46And we also have cryptocurrencies like Monero and Zcash and so on. So like, any way you look at it, the argument that we must destroy privacy in order to protect the public is wrong on two fronts. First of all, it's not going to fix the problem. It's only going to criminalize people who are rightfully concerned for their own safety and privacy. And I think privacy is a human right on some level. And like you said, fungibility emerges from privacy. So if we wish to actually address the systemic problems of the world today through like having... I mean, money to me is just a protocol, right?
21:38It's a way for brains to convey information to each other about what should be done with the resources and time at hand, right? So we use markets to execute this decentralized protocol where we price discovery gives an individual information about what society would value from them, and they can use their individual situation, their skills, their private knowledge to strategize and exploit this and provide value to everybody else. I think when people say Bitcoin fixes this, like my steelman version of that kind of ridiculous meme is, like, we need to...
22:33Stop corrupting this protocol because the signal-to-noise ratio is just off the charts with fiat. Like you have financial derivative markets like dwarfing the productive sectors of the economy by orders of magnitude. So how would you expect such a system to like ever, you know, converge and give meaningful information? And therefore, the argument that we need to destroy privacy in order to protect the public just falls flat from my point of view on like both of these fronts. Either one of them should be enough to completely refute that argument. And then yeah, so with that out of the way, I think there was a third part to your question, and could you please remind me?
23:30Vlad CosteaRight, so why does privacy matter and why do we even bother?
23:35Yuval KogmanOh, right. So, again, this kind of goes back to, like, what is money, what is it for? I mean, it's this thing for exploiting information asymmetry and decentralization to obtain like better results than centralized planning. And I mean, it emerged naturally way before we had any like notions of hierarchical societies at scale and so on. And there's something very innate about it, right? Like, it's a, you know, I think Dawkins wrote, and Nick Szabo quoted him in "Shelling Out", that money is a token of delayed reciprocal altruism. I think that's in like one of the later chapters of "The Selfish Gene."
24:31So, basically, by accepting money, you're believing that the other person had difficulty in obtaining that money. And it doesn't matter what is money in this. Like it could be tea, it could be salt, it could be seashells, right? As long as the cost of producing money is larger than the value to most individuals, i.e., like it's not economically viable to inflate the supply, and you know that there is some notion of scarcity and you have the ability to interpret that scarcity and trust that it really holds as an assumption, you can...
25:21Make the conclusion that it's highly likely that that person obtained the currency that they're giving you in exchange for something of actual value that you have. They must have, at least on average, obtained it by producing value on their own. Like maybe they stole it, maybe, but statistically, like, that's unlikely in a society that actually has trade. And through the subjective, like, theory of value, like, whenever you have voluntary economic exchange, if we assume, like, value is subjective, which I think is a fair axiom, then every voluntary transaction is addressing some, like, problem in the world and producing value, almost by definition.
26:19And this inherently kind of assumes like everybody's situation is different. Everybody can contribute something else and everybody wants something else. And privacy is a way of protecting these assumptions, right? It's... you need it for personal defense. You need it to protect your stash, basically, your ability to exert economic power over the world around you, right? You use money to get other people to do stuff that you want, right? They grow produce or they take minerals out of the ground or they build things and then they give them to you for the exchange of some like weird promise from the rest of society that at some later point they'll get something back in return.
27:12And so this, you know, that's where the societal value of money comes from. In game theoretical terms, it's like you just get better outcomes, more favorable equilibria, if you add a transferable utility to a non-cooperative or a cooperative game. But for all of this stuff to work, like, you need to respect the individual, right? Like, you need to let people apply their own discretion to their, like, financial planning. Like, some people value saving for the long terms. Others wish to, you know, leverage everything. And the moment, like, you break that privacy...
28:01Everything becomes much more of a meta game about like, well, how do I exploit my counterparty's strategy in order to extract more value from this interaction? So I think that privacy is fundamentally important for the function of money. Like without it, you can't actually negotiate informed consent in a way that converges.
28:39Ad voiceoverCommercial break. Gold and Bitcoin on Vaultoro. Vaultoro is the exchange where you deal with honest and hard money. There is no fiat on-ramp or off-ramp and you get to diversify your Bitcoin portfolio into gold or silver when you sense that a bearish moment is coming. Also, you can instantly trade your gold for Bitcoin to buy the dip. And if you're into gold custody, Vaultoro can also send you the gold that you own directly from their insured Swiss vaulting facilities.
29:29Vaultoro was launched in the aftermath of the Mt. Gox hack, so since 2015 they have published monthly glass books to prove that they own all the gold reserves and all of their customers' money. Sign up today by going to vaultoro.com/bitcointakeover. Keep in mind that this is not financial advice and you are responsible for your own decisions.
30:12Wasabi Wallet is the perfect Bitcoin privacy wallet. It's free, it's open source, it's available on Windows, macOS and Linux, and it offers groundbreaking Chaumian CoinJoins, which mix your Bitcoin. Even if you do not use the CoinJoin feature, you still benefit from a trustless experience with block filters, a hidden IP address via Tor, and easy management of your wallet outputs. After you deal with KYC exchanges like Coinbase, like Kraken, Binance, Gemini, or Bitfinex, you can remove the association between your identity and your Bitcoin address by performing a few rounds of CoinJoins. To find out more about the privacy benefits and limitations of CoinJoins, listen to Bitcoin Takeover Podcast Season 6 Episode 6 with Max Hillebrand. And if you want to give Wasabi a try, go to wasabiwallet.io and download the wallet for free.
31:24Wasabi Wallet, a Bitcoin privacy wallet for the citadels. Bitcoin Reserve is Europe's destination for buying Bitcoin simply online. It's the first Bitcoin-only venue in Europe that allows serious beginners as well as professionals to make daily purchases larger than 1,000 euro per day from bank account or via credit card. Learn more about Bitcoin Reserve by heading to bitcoinreserve.com. Bitcoin Reserve, where Europe buys Bitcoin.
32:12Vlad CosteaRight. So another question that I have for you at this time concerns, and let me blow the dog whistle, Bitcoin cash. But I don't mean BCH, but actual cash that is backed by BTC. And there are two ways of doing this. You can either think of a very complex scheme in which you have a decentralized system, or you can do it with paper bills that are backed by a central institution which issues them. But the advantage is that you're going to be able to move your bitcoins without an internet connection and with greater fungibility. These bills are not going to leave any kind of digital traces. There is nothing that leaves a trail except for maybe your hand fingerprint, which can be easily wiped with a piece of cloth if you're paranoid about it. But yeah, I think about this a lot and I think that in order to truly develop the economy of Bitcoin, we need to also have this physical dimension to money.
33:20Yuval KogmanYeah, so I think I would mention two other like fully decentralized alternatives, Opendime and this other thing called Tangem. I think they've since pivoted a little bit to shitcoins, but I haven't been keeping up. But like, these are essentially bearer instruments that provide a cash-like function where you can hand it over to another person and the physical security of the device itself is meant to provide you with the same finality assurances without broadcasting anything on-chain. And I also think that centralized eCash is going to be like a viable alternative.
34:14Hal Finney described this in his like early posts on the Bitcoin Talk forum. I have no problem with this at all. I think it's a great way to scale things. And it's a whole spectrum of trust and custodial assumptions. And the more options people have to economize their fees or, you know, address their privacy threat models or like, more choices is strictly better in my opinion. As, like, right now, we kind of have eCash, right? It's just a very... Sorry, interruptions.
35:08So we already have a kind of eCash system with custodial exchanges, right? It's just a very draconian, like, panopticon version of that. But people use it as a way of sending funds to each other. And if we can have similar, like, cost trade-offs, scalability trade-offs, and privacy trade-offs that are less in opposition to each other, I think that's a net benefit. I would be apprehensive if that is the only way of using Bitcoin. So like, if there is a single centralized Chaumian eCash server that just manages all of the Bitcoin, that would be a terrible, terrible outcome.
36:08But if there are many of these in competition, federated ones, things like side chains, things like statechains, things like these physical bearer assets, every one of those provides relief in some particular use case, and I think that's a net benefit. And in particular, there's things like Fedimint and Scriptcash. In a certain extent, WabiSabi is a protocol that can support this kind of eCash.
36:47There are a few caveats with that, but I think that the net result of that is that Bitcoin's censorship resistance properties and inflation resistance properties and its, like, basically its ability to define the rules upfront so that everybody knows what they're agreeing into, right?
37:15The rules don't get changed from underneath you. That to me is the important property of this system. And I don't see an inherent conflict between even centralized eCash and that. And George Selgin's, like, theory of private banking stuff gives a historical account as well of why that's likely to produce better outcomes than a purist approach or a, like, a government-centralized approach.
37:54Vlad CosteaYes, and I think that you have made the perfect transition into WabiSabi. You put it into this category of being a coordinator, an actor that is in the middle for the purpose of helping you achieve more privacy. And I feel like, even though I have had interviews with Ádám, nopara73, and also Max Hillebrand, I feel like I didn't get a lengthy explanation on what WabiSabi is and how it works. I understood what it tries to do, but not necessarily how it works. And I know that you did spend quite a lot of time lately working on that one. So could you please explain how it works?
38:41Yuval KogmanSo, for the purpose of CoinJoins, WabiSabi is merely a denial of service protection mechanism, but it achieves this goal by basically providing a centralized eCash-like token mechanism. eCash in David Chaum's original paper, which introduces blind signatures, it's a way for you to go... let's say you deposit a dollar at a physical bank and you obtain a blind signature. The signing key of the, like the providing the signature represents the denomination, right? This is the key that's worth one unit, let's say $1.
39:33You obtain a valid signature in exactly the same way as like elliptic curve or RSA signatures, depending on the technique. By the way, elliptic curve like blind signatures are kind of broken. So just PSA, don't use the regular Schnorr blind signature protocol. So the bank in this case is signing a message to itself. And it's signing the message, like what the message says is this person actually deposited a dollar at my bank, right? Like I, the issuer of these tokens, certify to myself that they're entitled to one dollar worth of, you know, like credits in their account.
40:22The crucial privacy element is that the signing event and the verification event are completely unlinkable, right? Like, if you sign 10 blind signatures, you can verify these 10 signatures. You never know, like after the user unblinds them and presents them again, you never know as the issuer which signature corresponds to which request to sign. So this gives you a sort of like anonymity pool, which is in principle like unbounded. You can constrain it by time, right? As cash flows in and cash flows out when people are depositing and withdrawing with this mechanism.
41:15Yeah, you can constrain. You can say like, this user is withdrawing, therefore they must have signed, they must have received a signature at some point in the past, so it could never be, you know, a later request, but this can grow, and as long as it's not like a deposit followed immediately by a withdrawal, and every single time there's only one unit in the system, eventually an anonymity set will emerge. The problems with this system is it only conveys one bit of information from the issuer to itself, right? It's like either you have a valid signature or you don't.
42:01And what represents the amount is which key had signed it, right? So the typical way that these systems are deployed is you have multiple signing keys representing, for example, every bit in the binary representation of an integer. So if you have just any number of some unit, you can break that down to just powers of two. Write it down in binary and then for every one that you see in the amount in binary, just issue a signing key, issue a blind signature with the signing key corresponding to that place in the binary representation.
42:45So this lets you construct a protocol that lets you deal with arbitrary amounts, but it's a little more... Like, there's a lot of ceremony involved. You need to deal with these different denominations, and every one of them is by a different key, so you can still kind of segregate the set of tokens. Okay, so hold that thought for a moment and let's recall what confidential transactions are. So this was introduced by Greg Maxwell, I think in 2015 or 2013, I don't remember.
43:28Vlad CosteaI think Adam Back was the first one to introduce it, at least theoretically, in 2012-2013.
43:39Yuval KogmanYeah, the underlying concept of homomorphic amount commitments. So, confidential transactions is a concrete protocol. And yes, you're absolutely right that Adam Back, and I think that's why I was confused about the date. And this is a helpful correction because actually it's not confidential transactions that are related. It is homomorphic amount commitments, because confidential transactions use homomorphic amount commitments in the context of a blockchain protocol with transactions and unspent coins and so on. So what are homomorphic amount commitments?
44:27So, you can think of hashes as a way to create a commitment to something, right? You, like, let's say, I want to represent the value 5. If I take the hash of the value 5, well, I can never open that commitment to any other value. Like, I cannot say, "Here's a proof, like, just hash the number seven and you'll see that it's the same." That's, if that were to work, then that means the hash function is broken. You can achieve a similar kind of mechanism with the cryptographic building blocks of like asymmetric cryptography.
45:12So, on an elliptic curve, you can construct what are called Pedersen commitments, which is kind of the same, right? The private key is like the value that you're committing to, and the public key is like a hash. And I mean, there's a problem that's true for hashes and true for commitments as well, which is like, you don't want these values to be guessable, right? Like anybody else can hash the number five and see, oh, that's just the number five. So typically, what you do is you add randomness in there as well to obscure. Right, so now instead of opening the commitment showing like if you take the number five and you hash it, so take the number five and this like large random string and hash it.
46:02And then you can see that I actually put the number five into the commitment, right? Pedersen commitments kind of work the same way. You're taking essentially the sum of two public keys built out of two distinct generators. And you're able to open this commitment later, like a hash. Where it becomes like, where the homomorphic part comes in is elliptic curves have this beautiful algebraic property, which if you take two values, so two private keys, and you add them up and multiply by the generator point, right?
46:46So derive the public key, which is corresponding to the private key, that is the sum of the other two private keys, that will give you the same value as if you take the two keys independently, multiply by the generator point, and then add the two public keys together.
47:11This means that in a sense, the numerical properties of numbers are preserved in this hash function, right? It lets you do things like add together a whole bunch of values that are otherwise obscured and then show, for example, that they sum to zero. So confidential transactions uses this with Pedersen commitments. And Pedersen commitments are just one way of constructing homomorphic value commitments. And there's nuances like, you know, what happens in a post-quantum world, like will we still have inflation protection, etc. So, ignoring all of those details, the meaning of this is like now instead of putting your amounts in the clear on the blockchain, you're putting commitments to those amounts.
48:21And what you're doing is you're proving to the rest of the network that you didn't cheat, without showing them what you actually did. So in WabiSabi, we use a technology or a concept called anonymous credentials, which are kind of like blind signatures. But the difference is, instead of the signer having zero information about what it is that they're actually signing, right? So, basically, the only thing that you can do in a blind signature protocol is sign a unique identifier that lets you detect double spending, right?
49:01As you verify the signature, you can just check, like, did I see this unique identifier before? If so, then it's already been redeemed, and it doesn't matter that the signature is valid. With anonymous credentials, at the time that the signature is being issued, or I mean, it's not a signature, it's a credential, the cryptography kind of allows you to constrain the message that is being signed without revealing what it is. So when I deposit, you know, some arbitrary number of monetary units into this, like, bank, or in the CoinJoin example, I post an intent to spend some input.
49:46...into a CoinJoin transaction, I register my input. The signer, the issuer can see that the, even though it's a commitment to the amount, it's not the amount itself, it can become convinced that the commitment really is for that value, which sounds kind of useless, but the trick is that you can do this to multiple such commitments at the same time, right? So you can subdivide your deposited amount in an arbitrary way, and you can prove in zero knowledge to the signer, the issuer, sorry, I keep saying that, that...
50:37There's no funny stuff going on, right? Like you didn't ask for more units than you're entitled to, but without revealing anything about how you're allocating those units, right? Like you could be splitting up in completely arbitrary way. And this is exactly the same privacy assurance that confidential transactions give you in a blockchain setting.
51:05So a nice analogy for this is, and there's a little explainer in the WabiSabi research paper repository. Imagine like you go into a room and the coordinator is just a dude sitting in the room, and you put on like, you know, one of those fake mustaches with the glasses and the nose and stuff, and let's just assume that that's analogous to perfect privacy, right? So like you walk into the room and the coordinator doesn't recognize you if you walk in multiple times. And let's say you put like $5 on the table, right? So the coordinator knows that you're entitled to five units.
51:49Now you come in with just blank envelopes and there's a bunch of sand in the room, and let's assume the sand has like zero value. What you can agree to do is like weigh out on a set of scales five grams of sand and then like, or better yet, you come in with the sand already in the envelopes according to your desires. And you just put the envelopes on a scale and the coordinator sees, yeah, that sums up to five grams. And you lay them flat and it doesn't get to touch it.
52:29It uses its own scale, so it knows you can't, you know, screw with the measurement, but it doesn't know how you... Like, maybe one of the envelopes is empty and the other has five grams in it, or maybe it's two and a half and two and a half. Regardless of what it is, as long as it matches up, it can use a wax seal and seal these envelopes. So these envelopes are analogous to credentials. So then you can walk out of the room with your envelopes and you can do this multiple times, right? Every one of these operations is kind of like a deposit. And then at a later time, you can come back into the room and just say like, here's an envelope.
53:23You can see that the seal is valid, you made it. You lay the envelopes flat on the scale again and let the coordinator inspect them without touching. Then at some point, it's convinced that the weight is fine and then you just like torch the pile or tear them up or something, spill all the sand onto the scale so that there's no more information about where every individual grain comes from. This is basically what WabiSabi credentials are. It's a way of ensuring this invariant holds, right? Like you cannot inflate these tokens, but other than that, they are basically, every credential is like a monetary... It's a little bit like a UTXO with no transaction history.
54:22It's a little bit like eCash but with, you know, this bundling together of a bunch of bits together and without the fragmentation across different keys. And there are, by the way, other systems that are like this. So in WabiSabi, we chose a very like minimalist anonymous credential system. And like we're only using it to protect the CoinJoin protocol, to make it so that users are only able to register outputs if they are entitled to the satoshi amount that the output claims without revealing the link to the inputs. But there's no, like, these are not conveying actual economic value, right?
55:10It's just a way of making sure that you can exclude undesirable users, users who are, you know, disruptive or adversarial, who did not even participate in the protocol. If you don't exclude those users, then anybody can come in and add an output to your transaction, and then like somebody else necessarily didn't get their output, otherwise it would not be a valid Bitcoin transaction, therefore that user would not sign. And right, so to keep out like this kind of disruption we basically... Right, WabiSabi does not give you privacy.
55:48It lets you have, like, it lets you construct Bitcoin transactions in a privacy-preserving way by using the building blocks of Tor's network-level privacy without opening the door to denial of service and disruption. So yeah, there are other systems like Danake, lightweight.money is the URL. There's another paper whose name I forget, which uses a publicly verifiable anonymous credential system. There's also the Nym anonymity network, which uses an anonymous credential system that has threshold issuance. So anonymous credentials are flexible. And as long as you have an anonymous credential system that fits your use case, you can instantiate it with these homomorphic value commitments and use that to basically represent a kind of a flexible and lightweight eCash system under different transparency and trust assumptions or centralization assumptions. I think that's a fairly full account of what WabiSabi is as defined in the paper.
57:16Vlad CosteaYes, and I know that you have worked on refinements and you have worked on basically a system to divide between amounts, because I think right now the issue with Wasabi Wallet is that you have a fixed amount that you have. So basically, you have to put in the same amount as everyone else. So it's becoming very obvious when someone has used the CoinJoin. But in the future with this 2.0 release, they're going to enable, I guess, outputs of all sizes to become part of a CoinJoin as long as they're divisible by each other. Did I get that correct?
58:00Yuval KogmanSo, you can think of the protocol right now as literally just a PSBT whiteboard. So, PSBT is partially signed Bitcoin transactions, right? Like, you connect over one Tor circuit, you say, let me add this input, you connect over another circuit, let me add this other input, and the coordinator doesn't know that you're the same person. And then everybody does this. And at some point, they add arbitrary outputs, and WabiSabi can be used in this way. However, that would, like, even though it's flexible, and you know, maybe there's good reasons to do that, like economizing on fees or... The privacy assurance of just doing this naively is very brittle.
58:49There's the old, like, CoinJoin Sudoku attack on SharedCoin and a paper called "Anonymous CoinJoin Transactions with Arbitrary Amounts", which we usually refer to as the knapsack mixing paper. There's a fairly rigorous analysis of like, what does it take to construct transactions which actually preserve privacy? And there's a trade-off there between overtness, like, does this look like a privacy-preserving CoinJoin, or does this just look like a giant batch transaction? In Wasabi 1.0, because it uses blind signatures, there's a technical constraint that says, like, you must use a standard denomination for the reason I explained before with eCash, or blind signature based eCash. This restriction is now lifted, but if you just...
1:00:04Make use of this tool, I mean, it's a giant foot gun, basically. Like, so you need to apply it very carefully to actually obtain the outcomes, the desirable outcomes of protecting users' privacy and fungibility. And with CoinJoins, there's a serious trade-off, which is like you're always announcing to the world on some level, if it's a private CoinJoin, you're saying like, "I'm using CoinJoins, and therefore you don't know which of those inputs belong to me." All you're seeing is I'm spending an output that follows that transaction.
1:00:41And you know that I'm one of those entities. But you don't, like, you can't say which, and therefore you can't go and analyze my history very easily. You can use a process of elimination, right, but that requires a much more powerful adversary that's able to survey much more widely. So this kind of trade-off, like, let's just assume that we accept it.
1:01:19Right, like, there's other Bitcoin privacy technologies, which I think are very important and complementary to CoinJoins, precisely because they don't make that trade-off. So things like coin swaps or statechain swaps or Payjoin. And they are different methods of obscuring or adding plausible deniability or ambiguity to the history of your coins. That's fundamentally where privacy emerges from. The actual structure that Wasabi is kind of going for right now with the next version is an adaptation of the ideas from the knapsack paper. And nopara73 had a blog post a while ago which describes this greatest common divisor mixing, right? That's kind of like an extreme example of that.
1:02:33Like, take all of the input amounts, compute the greatest common divisor, and then just use that as your base denomination. Everybody only is allowed to register those outputs. So this is in a sense perfectly fungible for the same reason that satoshis in a transaction are fungible.
1:02:51The strawman version of, like, even more strawman version of that is, like, assume there's no dust threshold and there's no fees. Everybody, like, if everybody were just to register outputs worth one satoshi, we would have like horrible blockchain load, but I mean, it would be perfectly fungible in this regard, and the privacy assurance would be very strong. What we're trying to do in practice, for reasons of cost, has to be a lot more, like, it's a much more delicate balance. And what we're trying to do there is standardize on a set of values. So, in particular, powers of two.
1:03:39The 1-2-5 series, which should be familiar to everybody from just their local fiat currencies. So 1, 2, 5, 10, 20, 50, 100, 200, 500, etc. And finally, also powers of three and their doubles. And the reason to take such a seemingly odd combination of standardized values is when you have this mix, it lets you represent many arbitrary values, typically with a small number of these standard values. If you only choose one, like only powers of two, for example, or only the 1-2-5 series, then some values are going to be very inefficient and other values are going to be quite efficient.
1:04:35And the clearest example of this is, suppose you're registering, like, let's just assume only the 1-2-5 series. So one whole Bitcoin would be a standard value. It's in the set. But if you subtract a very small fee amount from that, now you have like 0.99999750 or something like that.
1:05:02Your ability to represent that value with a large number of nines efficiently, I mean, you can't do it, right? Every digit takes one five and two twos, or the seven takes a five and a two, right? So you're going to end up with something like, you know, eight digits of precision. Let's say like six of them require three outputs. That's like 18 outputs just for a single amount. That's cost prohibitive and extremely inefficient. So mixing these up together lets you kind of get around this problem.
1:05:45But ultimately, it's still a fairly small number of values that everybody has to agree on. Like, these are the standard values. So this means that you still have a much more robust privacy assurance where you actually have values that are truly indistinguishable from each other with high probability. So when I register a 0.5 or a power of two output, there's a decent likelihood that somebody else in the same CoinJoin is registering a similar order of magnitude as an input. And therefore, the chances that they will select exactly the same output value as me, I mean, that's the same privacy assurance as we're already familiar with the 0.1 values of Wasabi.
1:06:47We want to kind of mix that up and allow a wider range of input values and allow combinations to be done without the risk that SharedCoin took, where if you use, like, you look at a transaction and you start constraining, right? You, by a process of elimination, you say, you know, if these two outputs go together, then whoever created them must have had at least that much plus the fees. So therefore, if that other input is spending this much, this much would be left over. Let's check if the output amount contains that.
1:07:33Oh, yes, it looks like these could be clustered. You just go through this procedure and naively it's very, very inefficient, but there's tools, various constraint solvers, mixed integer programming and stuff like that, that can solve these kind of systems of constraint much, much more efficiently than the sort of naive exponential cost that the general form of the problem has, right? There's many special cases where the process of elimination can actually be very efficient. And what it boils down to is like, can you construct CoinJoin transactions where there is just not enough information in the transaction there to rule out every possible way of clustering things apart from one?
1:08:33If you can do this, then the transaction is ambiguous. This means that there's multiple different interpretations, and that's exactly the very intuitive concept of having everybody just use the same standard amount. It's just far less, right, it's not all or nothing and it's not as rigid. So using this kind of building block where we do have equal outputs, but the standard denominations are able to be summed together to produce any arbitrary value with only a small number of standard values, and allowing this to all happen in fairly large CoinJoin transactions, we can, if we're careful about this, we can kind of have the best of both worlds, right?
1:09:38We can have the on-chain efficiency, like, not exactly, but we can approach the on-chain efficiency of just, you know, not giving a shit, basically, and just, you know, "I'll use those inputs and I'll send that payment through this coin joiner," like SharedCoin had, right? So that comes at no additional cost to you. But it's a very brittle form of privacy. It shouldn't even be considered a privacy technology, in my opinion. And on the complete other end of the spectrum is this, like, you know, greatest common divisor approach, right? So that's very cost prohibitive.
1:10:18In between is a balance that is convex, right? You get more privacy for less cost overall by having multiple users perform this procedure together in the space of a CoinJoin. And that's what Wasabi 2.0 is trying to do. It's doing this in order to remove as much of the decision-making and requirement to understand the implications of your actions, right? Like, Bitcoin is inherently kind of flawed, as we discussed earlier, where if you use it in a particular way, there might be consequences to your privacy.
1:11:08And you need a sort of understanding of what a UTXO is, what the transaction graph is, et cetera, et cetera. So using WabiSabi and this very careful process of not using it stupidly, Wasabi 2.0 is trying to relieve the users from the burden of understanding on a technical level how exactly an adversary can analyze the structure of transactions. And yeah, I hope that's a complete enough and satisfying enough answer for, like, what the structure of the transactions is going to be. But yeah, you were close, but the devil's in the details.
1:12:03Vlad CosteaWell, obviously.
1:12:05But, you know, when people hear about CoinJoin, the biggest critics say that this doesn't scale and there is no way that we're going to have CoinJoins at scale. And you know, it's easy for an outside observer to figure out that a CoinJoin had happened. So these are two of the arguments that people use as a rhetoric to make others not CoinJoin and say that it doesn't work, even though it adds extra plausible deniability. It's useful to understand the threat model because in some cases, it's a lot better to leave some sort of deniability, as opposed to having a clear, you know, history of your coins. If you withdraw from CoinJoin, for example, sorry, Coinbase, for example, it's very easy for them to track you because they have hired an actual...
1:13:01Security team that has this task of trying to figure out where their coins are going. So for this purpose, if you don't want the recipients of your coins to know that you have them from Coinbase and you don't want them to get tracked further on, you do a CoinJoin, and it's most likely preferable this way. But I suppose this also has to do with the way in which CoinJoins are getting portrayed, as some people claim that they are the solution to Bitcoin privacy. But obviously there is a lot more that needs to be done.
1:13:34As you mentioned, there are the coin swaps. There are all of these mechanisms that are based on the statechain idea. I'm not sure how they work, so I'm not going to even try to describe them. So I guess my question has two dimensions. First of all, how can CoinJoin scale? And secondly, how can you use CoinJoin in conjunction with somebody else to attain greater privacy?
1:14:02Yuval KogmanSo there's two ways in which CoinJoins might not scale. The first is the CoinJoins themselves, and the good news is that that's, like, we know how to do it, right? It's a cooperative algorithm, right? It requires everybody to sign at the end, so it's easy to disrupt. But in practice, it more or less works, and we can ensure that there is a costliness associated with, right, you just, if you see a UTXO register and not sign and register and not sign, at some point you just ban it and ban all its children, and then it can no longer disrupt. And, you know, doing a reputational whitewashing, right, like cleaning, unbanning yourself by, you know, privatizing your coins in some other way, has a cost associated with it.
1:15:03So you can't really, as long as you can't do that costlessly, that's not really an issue. And it's also, the scale is of the whole graph of CoinJoins, right? It's not just a particular CoinJoin that needs to scale. It's rather, how does everything connect? So in this regard, the privacy assurance, the whole is greater than the sum of its parts when CoinJoins are used pervasively. So I think that argument is, you know, like saying that CoinJoin protocols don't scale, that's something that we've already refuted in practice sufficiently well.
1:15:59Whether or not it scales as a technology for privacy in general, I mean, just think of the extreme example. So Bitcoin is made out of blocks and blocks contain transactions. That's our fundamental scaling constraint, right? This is why layer two protocols are so important for scalability, because the nuanced perspective that a UTXO is more than just a coin. It's a representation of the interest of a potentially large set of users. It's more like an entitlement to use the blockchain as a court system with various costs associated with it.
1:16:48In the extreme example, or the extreme case, you can just imagine the whole blockchain, like every single block is just one giant CoinJoin transaction. That is no more unscalable than what we have today and yet substantially more private. And regardless of what layer two scaling technology you're going to do, you still need to anchor those layer two solutions in actual UTXOs and actual Bitcoin transactions that appear on chain, right?
1:17:24Right? If you cannot participate in the encumbrance of a UTXO that you're entitled to a share of, then you can't enforce using the blockchain and the consensus rules, you can't prevent other parties to that UTXO from taking your money from you. So we're always going to need layer one to work. And in my opinion, it's clearly better to do that with CoinJoins. On the very long term, we have discussions about cross-input signature aggregation and other economic incentives to make CoinJoins more scalable than just transacting on your own.
1:18:12But even now, it's already possible to kind of align incentives in a similar way, which is... So imagine this is not what's going to actually ship, but it's technically possible, and there's some nuances with the security proofs for WabiSabi credentials, but that's something that is solvable using either proving the security in a more satisfying way or... Just to summarize, there's a proof of security for a non-adaptive adversary under the Decisional Diffie-Hellman assumption, and the non-adaptive adversary is kind of a very strong assumption, like you don't want to make that assumption for unforgeability.
1:19:05And there is a more standard proof of security in terms of the adversary assumptions, but under the generic group model. So in a way, it's like you have two independent proofs of the security of unforgeability, neither of which is fully satisfying. Right, but other credential systems could allow stronger assurances to be made, and there's other reasons to maybe prefer them, like public verifiability or threshold issuance. But regardless, just assume that your CoinJoin coordinator, instead of needing to... Like, let's say you register some Bitcoin, and instead of registering with the decomposition of you through standard values, like your entire balance, you could donate some of it as extra mining fees to the transaction, right?
1:20:03Like maybe it's so close to the dust threshold, there's no point. Or maybe... Like your privacy is worth more to you. But in principle, what you could do is you could ask a coordinator to issue you a credential, which is not just for the denial of service guarantee for that particular attempt at constructing a CoinJoin. Rather, if this CoinJoin transaction goes through successfully, after it confirms, the coordinator can issue you some fee credits, basically, like an anonymous credential, which you can redeem at a later time, right? You've pre-purchased essentially a token that, if you can withdraw it as actual Bitcoin, that kind of makes the coordinator into a money transmitter.
1:20:59So, but if you restrict these tokens so that you can only use them to pay for the coordination and mining fees, well, now there's a cost reduction incentive, right? Like, if you're willing to trust a coordinator with, you know, a few sats worth of tokens, and I mean, it could do an exit scam, right? But maybe you just set it to a risk exposure that you're happy with. Now you could be making payments from CoinJoins, or you could be remixing standard values, and having exactly the same input value and output value on both sides. You could have like multiple users redeeming these credentials in one go.
1:21:51So like, one input from the coordinator could be paying for everybody's like mixing fees. Or in the payment case, like maybe this saves you from bringing in another input, right? The more you're willing to trust the coordinator, the more combinations, like the more payment amounts become possible for you to make without bringing in an extra input. You only need to cover the payment amount, and all of the fees were already prepaid in some prior CoinJoin. So in a sense, there's already reason to believe that it's possible today, with the consensus rules of today, to provide a cost-saving incentive for people who might consider using CoinJoins.
1:22:46Once you want to ensure that the privacy is robust, things get a little bit more complicated. Right, it's not such a clear trade-off anymore. But nonetheless, it's still a lever for scalability. Okay, so, sorry, I keep giving you very detailed answers for seemingly straightforward questions. So the other part of this is the censorship resistance property, right? Like what is the biggest risk of CoinJoins? I mean, when you do a CoinJoin, everybody can see that you're doing a CoinJoin. And the only way that you can have plausible deniability and claim, "Oh, that's not a CoinJoin," is if it's not actually giving you very much privacy at all, right?
1:23:35PayJoins do that. PayJoins give a very different trade-off, and I think that's why they're such an important and complementary tool. But ultimately, when you use a PayJoin to obtain privacy, you're not obtaining any meaningful privacy from your counterparty.
1:23:57You're not... It doesn't really scale in the same way. So it's a very different tool for a very different threat model. And I think the key observation is these things are not mutually exclusive. Like, I would like to see a world where coin swaps are done, you know, are redeemed into CoinJoins or funded from CoinJoins, and Lightning channels are updated with dual funding and splicing, and every CoinJoin output could have been a statechain output that was swapped, you know, using Mercury statechains, like, swapping features, and where PayJoins are happening before and after the, you know, the obfuscating, like BIP 78 or 79, I don't remember.
1:24:51Like that kind of... Right, so all of these things, they add up together, and their combination is fungibility in practice, right? Like fungibility is when it's no longer reasonable to impose the burden of discriminating on the basis of coin histories on the economy at large. And even if it's not perfect on a technical level, in practice, if you just... It becomes sort of the status quo that everybody knows that you just don't check, you know, Bitcoin histories anymore because there's just too much stuff going on.
1:25:39Well, that's a world in which CoinJoins are completely legitimized and not stigmatized, right? This is not the world that we're in right now. And it's a very kind of delicate situation where it's a use it or lose it proposition. The costs of coin joining are quite high today. The systems are limited and you're making a target of yourself, right? Like by deciding to opt into privacy overtly, you're making a statement. You're saying on the blockchain, "I deserve privacy and I'm claiming it." And I'm claiming it by joining this crowd of other people who think like me.
1:26:24There's a systemic risk there, which is that we could be fragmenting the whole ecosystem, right? Like we could end up in a situation where some of the UTXO set is CoinJoined coins and some of the UTXO set is KYCed coins, and the two never interact with each other. I think that would be a terrible, terrible outcome. And there is also an individual risk, which is like, let's say I withdraw from Coinbase and I CoinJoin for my privacy, they might close my account because they just don't want... They're trying to reduce their risks and compliance is expensive, and they're a bullshit company run by an idiot, sorry.
1:27:10But their priorities are very much compatible with the broken authoritarian rent-seeking fiat mentality that Bitcoin is supposed to even address. So like because of this destructive, short-sighted profit-seeking motive, where compliance is a priority. And, I mean, you can be fully compliant and claim your privacy with CoinJoins, but it's more work for them. And because not enough people are asserting their rights, CoinJoins are stigmatized, and you have situations like BlockFi. I think they've actually confiscated people's money, or at least closed their accounts and sent their money back. I don't know which. Just on the basis of fear from regulators, right?
1:28:06So it's this very draconian... It's a surveillance atmosphere that is categorically way worse than a bank, right? Where because of this stigma and fear-based approach, the burden of proof to demonstrate that you're not a terrorist and you're not a pedophile and you're not, you know, all these horrible things that they... Statistically insignificant in the economy as a whole, but are very easy to weaponize into a tool of political control, because we're all like stupid monkeys that are very easily controlled through fear.
1:28:58Like because of these perceptions, because of our inherent biases, because our monkey brains are kind of like easily manipulatable, we're living in a sort of very delicate situation right now where the authoritarians haven't really caught up with this thing yet.
1:29:22They don't quite understand what's going on. And if we don't exercise our right to privacy on the blockchain today or in the coming years, I think we may end up in a world where Bitcoin is kind of stuck, right? It's at the interface to the rest of the economy, there will be insufficient privacy. Everybody will be expected to use like custodial things. And if you ever opt in to a covert privacy tool, the burden of proof is still... When you try and, I don't know, make a purchase or something with a compliant merchant or trade on an exchange, you will at that point be unable to prove that you legitimately obtained these coins.
1:30:15Right, that's something that technically you can do today with CoinJoins, with anything. If an exchange says, "Oh, we don't allow deposits from CoinJoins," you can tell them, "Bullshit, I can prove the coins. I just don't want to prove it to all the rest of the world. I can show you the full trail from when I withdrew all the way up to here with proofs that I own the keys." And the question is, are you broadcasting this to the world or is this a bilateral thing between you and that compliant exchange? And this is a business priority for them right now. They don't want to deal with this stuff because they're much happier...
1:30:57Just, you know, taking the money, taking the trading commissions or whatever from the people who don't really see this as a politically adversarial situation. So I think these systemic risks are much deeper than scaling, and I think it's a much more complicated interaction between the overt and the covert privacy technologies. And I'm very worried about a future where, because the overt tools of privacy are stigmatized, either they are their own, like the currency splits in half, or they're not even used at all. And by extension, the covert privacy tools are not really used either.
1:31:45And Bitcoin can't actually function to improve things in that kind of scenario, I think, because even though it might be inflation resistant, it might be, strictly speaking, censorship resistant on like a protocol layer, if everybody who's trying to save in it or use it needs to live in fear of financial censorship, not on the blockchain, but at the boundary, or needs to protect their personal safety in like much more elaborate ways at a much higher cost, like, what are we even doing here, right? Like, what's the point if we just live with this like...
1:32:31Panopticon, this like new tool of oppression? The whole point was to use cryptography to kind of equalize the balance of forces and restore the very ancient, like, decentralized nature of money as a protocol for mediating physical energy transactions without central coordination and in a bilateral way with no intermediaries. And right, so to restore that thing, we need fungibility and we need privacy. And I don't see how that emerges.
1:33:21Like even if CoinJoin is not the way to achieve that, I think it's an integral component. It's definitely not sufficient on its own. We need the other covert tools. But it's all of this stuff that works together, and I think it's much more to do with the politics of it. It's not really a technical thing, right? Yeah, I hope that's, again, a rambling but hopefully a satisfying answer.
1:34:27Ad voiceoverVaultoro is the exchange where you deal with honest and hard money. There is no fiat on-ramp or off-ramp, and you get to diversify your Bitcoin portfolio into gold or silver when you sense that a bearish moment is coming. Also, you can instantly trade your gold for Bitcoin to buy the dip. And if you're into gold custody, Vaultoro can also send you the gold that you own directly from their insured... Swiss vaulting facilities. Vaultoro was launched in the aftermath of the Mt. Gox hack. So since 2015, they have published monthly glass books to prove that they own all the gold reserves and all of their customers' money. Sign up today by going to vaultoro.com/bitcointakeover.
1:35:45Keep in mind that this is not financial advice and you are responsible for your own decisions. Wasabi Wallet is the perfect Bitcoin privacy wallet. It's free, it's open source, it's available on Windows, macOS and Linux, and it offers groundbreaking Chaumian CoinJoins, which mix your Bitcoin. Even if you do not use the CoinJoin feature, you still benefit from a trustless experience with block filters, a hidden IP address via Tor, and easy management of your wallet outputs.
1:36:16After you deal with KYC exchanges like Coinbase, like Kraken, Binance, Gemini, or Bitfinex, you can remove the association between your identity and your Bitcoin address by performing a few rounds of CoinJoins. To find out more about the privacy benefits and limitations of CoinJoins, listen to Bitcoin Takeover podcast season 6 episode 6 with Max Hillebrand. And if you want to give Wasabi a try, go to wasabiwallet.io and download the wallet for free. Wasabi Wallet, a Bitcoin privacy wallet for the citadels.
1:37:04Bitcoin Reserve is Europe's destination for buying Bitcoin simply online. It's the first Bitcoin-only venue in Europe that allows serious beginners as well as professionals to make daily purchases larger than 1,000 euro per day from bank account or via credit card. Bitcoin Reserve, where Europe buys Bitcoin.
1:37:42Vlad CosteaYes, and something that I asked you last week, and you said that the short answer would be no. So as far as I know, we're going to have to hard fork at some point just to fix an existing bug in the Bitcoin code. And maybe that we are delaying that just because we are hoping that we could fix the issue. But let's say that we had to do an emergency hard fork for whatever reason, and we had to move to a different chain and move the UTXO set just like, I guess, has been done a couple of times in the history of Bitcoin, but not in the last eight years or so.
1:38:20Would you be in favor of adding some sort of privacy mechanism? And it doesn't necessarily have to be confidential transactions or zero-knowledge proofs as in zk-SNARKs, or what else do we have? Something like Grin, like the Mimblewimble protocol, or something like ring signatures that we find in Monero. Do you think that any of this is appealing and desirable to get integrated into Bitcoin to help it become fungible?
1:38:50Yuval KogmanNo. There's several reasons for this. So the first one only applies to Monero and Zcash, and that's the problem that you can't really prune those blockchains, right? There's the cost of validation is strictly increasing in time, and it's not just like the size of the history is increasing with time. It's the UTXO set is increasing and monotonically increasing in time. It never can really shrink unless somebody is actually caught double spending, in which case, I mean, their transaction is not valid anyway. So I think that that is a fundamental issue that is better resolved on layer two, right?
1:39:47It's better to have layer two privacy and fungibility but keep the UTXO set actually prunable. Grin or Mimblewimble has like an interesting property where it can kind of fix that, but there are other trade-offs. And the other main trade-off that applies to it as well as Monero and Zcash is the question of inflation resistance. I think this is a crucial property that to me is more important than privacy, because privacy can be fixed under this constraint.
1:40:30You can have basically cryptographically perfect privacy, but the risk that there's such a critical protocol bug, as you described, is substantially higher in that world. And, you know, case in point, both Monero and Zcash had subtle cryptographic bugs relating to inflation resistance.
1:41:03I think the systemic risk is much higher, and as an ecosystem, we should resist that. This is, by the way, like the reason WabiSabi is called WabiSabi, apart from being kind of a pun on Wasabi. It's like this Japanese aesthetic philosophy of finding beauty in imperfection. So, like, if we recognize that Bitcoin is imperfect, but so are Zcash and Monero, right? They make different trade-offs. We can still find privacy or fungibility in there, right? Like I said before, the satoshis are fungible. Individual transactions are fungible. Perfectly so.
1:41:52It's just a matter of like, how do we use this consensus protocol to actually create fungibility as a social phenomenon? I think that's not a hard problem, and I think the risk, the cost-benefit analysis is, when you account for the risk, is very much in favor of being very conservative on that layer. And in particular, like as a response, like if we do an emergency hard fork, like nobody has the time during an emergency hard fork to start reviewing a completely new, like alternative consensus layer with novel cryptography and new cryptographic assumptions and a new crypto implementation.
1:42:38And like, if we fucked up the very simple version, where, you know, it's just numbers, just integers, like now we need to make all of this work in a much more delicate way, like that doesn't add up to me. And I think the way to deal with this is exactly all of those technologies that can augment and go on top of it. So we have Lightning. Lightning, or payment channels more generally, can suppress information about individual economic transactions and reduce the flow of sats on the blockchain to a settlement network.
1:43:27In the future, hopefully, channel factories and other similar constructions will allow this to represent not just the aggregate flows between two parties like a payment channel does today, but a whole pool of money. And there, I mean, you can have very good privacy depending on how you negotiate the cooperative protocol between, right? You don't need a global mechanism, a global consensus protocol that ensures privacy in order to obtain privacy within the payment channel technology. I think that that's a much more composable and future-proof way of doing it.
1:44:17It's acknowledging that, you know, Satoshi made some difficult trade-offs, but it appears that they are very appropriate. They are like the time was right for a cryptocurrency, but it was too early, right? Zero-knowledge proof technology is still advancing at a staggering rate, partly because of this renaissance in cryptography. And, you know, Zcash and Monero have already evolved substantially at the protocol layer and they're still iterating. They're still like making things more efficient. They're still like improving the guarantees. And so I somehow doubt that this will ever settle on a fully satisfactory, right?
1:45:11Like, at some point, maybe the marginal cost of adding additional cryptographic assumptions to Bitcoin will be sufficiently small, and maybe the formal verification will improve substantially and we can just write a proof of the consensus protocol being valid and verify that its implementation is correct and so on and so forth. This seems to be like decades out at best to me. Maybe in that world, the trade-off between privacy and inflation resistance is going to look different to me. But until such time, I think it's just an unreasonable risk.
1:45:58And I think it's not broken enough, right? Like we can fix it. We have solutions. It's just a matter of changing the incentive landscape so that the disincentive of using those things goes away, right? Like, if we can remove the stigma from claiming privacy, if we can reduce the costs, if we can improve the user experience and make things more simple, then we will have privacy and we will have fungibility and we don't need any consensus changes. Yeah, I think that's... Did I forget anything in your question?
1:46:44Vlad CosteaNo, I don't think so. I like your pragmatism. And I like that you obviously looked into all of these privacy mechanisms and you're very much aware about their limitations and why we should be cautious, as I think that there are people right now who are very much in favor of confidential transactions because they have a very simple design. They have been around. They are well tested. You have them on projects like Blockstream's Elements and Liquid. But at the same time, they don't really scale.
1:47:19So there seems to be this compromise between privacy, which tends to add layers of obfuscation, which take a lot of data, and the data takes a lot of block space, and it will not scale under Bitcoin's decentralized framework that we have established that we should pursue. So, yes, we seem to run into this issue all the time when it comes to attaining privacy. Also, we have mentioned before with CoinJoins that obviously not everyone can do a CoinJoin right now, because obviously the demand for block space would be huge.
1:47:56And it goes back to the same conversation that we have about Lightning channels, when we say, "Okay, Lightning is great, but if everyone around the world was to open a channel, it would take several years of nothing else but channel openings on the Bitcoin blockchain to accommodate everyone to use Lightning Network as a first-class citizen and not a custodian." So there is a lot to figure out.
1:48:24Yuval KogmanI would like to push back a little bit on this because... So ring confidential transactions like in Monero, or the approach to like shielded transactions in Zcash, or Zerocoin also has this. In a way, confidential transactions is much easier to scale, because that's just a constant overhead, right? Like, now every output takes, you know, 10 times as much data, right? There's like a range proof in every transaction or, right, it's not literally 10 times, but it's within a constant factor.
1:49:00That's a problem that we can address much more easily using technologies like Lightning, right? Like we can amortize that cost and we can make it more cost effective. The reason I think Zcash and Monero are fundamentally less scalable is that in time the complexity of running a full node increases and keeps increasing.
1:49:34And it increases scaled to the history, not scaled to the current usage, right? So with confidential transactions, you can still run a pruning node, right? You don't need to save all of the history. You just need to observe it happening to be satisfied that there is no inflation going on. So that's the first bit of pushback. And the second bit of pushback is like, I mean, confidential transactions, yes, it's a very conservative technology. And even if we account for a post-quantum adversary, like we can find ways of like... This goes back to what I was saying earlier about commitment protocols.
1:50:20There's two basic kinds of cryptographic commitments. You can have commitments that are computationally binding and perfectly hiding. So what that means is there's not enough information there to fully constrain what the commitment data is. And if you have a post-quantum adversary, that adversary can open the commitment to basically any value, right? Like, if it's basically like you have the sum of two public keys, and because it's two public keys, there's enough degrees of freedom there that like, instead of trying to break the public key, a Pedersen commitment says, like, a post-quantum adversary can make it look like it's a sum of any two arbitrary keys.
1:51:13Whereas a perfectly binding and computationally hiding commitment scheme, that's something that's more analogous to just a single private key. So ElGamal commitments can do that. And you can make a confidential transaction implementation that uses ElGamal commitments instead of Pedersen commitments. What that does is it gives the post-quantum adversary the ability to undo the privacy after the fact. So there's good reasons to be skeptical about whether or not we'll ever have post-quantum adversaries in the real world, but, you know, it's a contingency that we should be prepared for.
1:51:53And to me, the systemic risk of confidential transactions is that, supposing this happens, like if there's a bug, we might only find out about it after the fact. And even if we can audit the supply in retrospect, malfeasance may have already happened, and therefore the value of this as a currency might fall apart. And secondly, the trade-off is basically like either the quantum adversary is able to print money out of nothing, or it's able to break everybody's privacy after the fact. Neither of these seems very robust to me. And that's why I'm much more optimistic about like... I think the whole scaling versus privacy thing is a complete false dichotomy.
1:52:45The best privacy is having the least amount of information on the blockchain. So that is entirely aligned with scaling. And I think the way we solve both together is we need to push as much of the economic activity down to second and third layers, and we need to be liberal about, you know, our trust assumptions, right? Like, "Bitcoin is the currency of enemies," goes the trope. So recently, I came up with a kind of, I guess, smart-ass tweet, which is like, layer one Bitcoin is for enemies, right? The layer two, the entire value proposition comes from the fact that you can coordinate, right?
1:53:37You can, by cooperating, reduce everybody's costs, improve everybody's privacy. And you only need to go to layer one as a path of last resort. It's more of a deterrence. And so using Bitcoin doesn't mean using layer one Bitcoin with these very pessimal assumptions and prohibitively high costs. And, I mean, yeah, if we need to wait for everybody to open a channel on their own, then we need to like take the block size, make it hundreds of times bigger, and even then it's going to be a very expensive and constrained system. I think that's the wrong question.
1:54:26The question is like, okay, how can we represent the financial interests under reasonable risk and trust and custody and privacy assumptions, so that a single UTXO can anchor the property rights of a very large number of economic actors?
1:54:46And that's perfectly achievable, right? Lightning demonstrates that this is possible. Coinswap demonstrates that this is possible. Coinswap is basically just, you know, a kind of privacy-optimized and not efficiency-optimized payment channel technology. The two can, you know, grow together. You can put Chaumian e-banks on top of that. You can make all of these things use CoinJoin at the base layer, right? There's no additional cost to doing everything collaboratively with other users. So we can address these issues in a way that's not framing it as like a zero-sum trade-off, scaling versus privacy. The reality is much more sophisticated and much more convex, much more positive-sum.
1:55:34We can get more value out of the system if we just use it in a more like mentally flexible way. And by the way, this is why I'm so excited about Taproot, is it eliminates so much of these, like you don't pay for the costs of these trade-offs upfront like you currently needed, or up to recently needed to do. Now with complicated Tapscript spend paths, which by design you're never really meant to use those, right? Those are all paths of last resort.
1:56:14The happy path is if everybody just agrees and uses the key spend path. But the option of using that as a deterrent is what secures all of these much more optimistic ways of concurrently, you know, actually performing, like, mediating economic activity where the value is actually created, right? Like actual material resources, actual energy in the real world, actual creativity, right?
1:57:00There's nothing inherently valuable about Bitcoin itself. It's just, you know, a database with some numbers in it that we all believe has meaning to the humans. And the humans and the thermodynamics is where the notions of value come from. Any currency is just a tool to let us all make the most out of that, to enable positive-sum interactions where they were not previously possible. And yeah, that's, I think, a healthier perspective on this whole scaling versus privacy false dilemma.
1:57:45Vlad CosteaYes, I'm very happy that I asked you this question specifically. And now let's get to your personal work because you seem to be very much into solving privacy and trying to find a way in which it also becomes scalable. So what are you going to be up to in the coming months or possible, you know, coming years? And also, how can people check out what you're up to? Like, do you have a public GitHub repository? Do you post about your activity on Twitter?
1:58:14Yuval KogmanI mostly shitpost on Twitter. I do try and engage in conversations. I find Twitter to be super... It's a truly despicable, toxic hellhole of a software. A piece of software that seems to be designed to get the worst out of people, right? It's almost constructed for just toxic interactions, and yet somehow people are on average just cool and nice, and a lot of people use Twitter to like share ideas. So in spite of these challenges with the system, I do find that there are many deep and interesting conversations going on Twitter, and I try to participate in some of them. So yeah, that's a good place to kind of, like if you want to intermingle memes with my meat brain, like that's a good forum to do that.
1:59:20github.com/nothingmuch is where, you know, I have a bunch of repositories, but most of that is like old stuff. I've been far less active as a coder in the last decade. I don't know, I mean, I'm not really sure, I'm not good at presenting myself to the world at large, but I do have a bunch of projects that I think are important for me and humanity that I would like to work on.
2:00:04Bitcoin privacy is one of them. More generally, like, I think there is so much historical baggage in the way we approach software in general. It's like deeply embedded in our devices, in our operating systems, in, you know, this box metaphor of software, where typically an application is like this thing that you open up, right? What's opening there? That's a fiction, right? And then after you've opened this like make-believe box, the program invites you to put your data inside of the box. And that's another fiction. And then if the makers of the software are kind enough to you, then they give you the insight, right?
2:00:50Like you're allowed to look at the data differently or maybe exchange data with other users. And then at the very end, if it's not like a hostile system, then you're even allowed to take your data back out. I think this is a very violent and oppressive way of interacting with these like miracle devices, right? The computers are like extensions of the human mind, and their potential is almost limitless. Like it's so far beyond the human imagination, and yet we're using it for bullshit like, you know, showing ads to people or surveying their activity and making them solve CAPTCHAs just because, like, we have this oppressive financial system where, like, everybody's assumed to be a money launderer by default instead of a sentient being with creative urges and curiosity and interest.
2:01:50So I think there's a lot of stuff that's really deeply wrong in the way we approach software. It lacks empathy and it lacks respect. And it's also just technically like anachronistic. The physical constraints of computing of the middle of the last century no longer apply, and yet we're still doing things as if, you know, the hardware is the expensive thing and human time and attention is the cheap thing. So yeah, I invite everybody to question these controversial things I just said, and I'll do my best to give a reasonable answer. But I think right now it's best if I shut up about that, because otherwise this might go on for several hours more.
2:02:40And yeah, I'm interested in Bitcoin. I spend most of my time like studying Bitcoin-related things. I'm actively like working on Wasabi and WabiSabi right now.
2:02:57There's some stuff to still implement. Some of the hard questions still need to be resolved. So hopefully before too long we'll be able to come up with a mainnet release that everybody's happy putting their names behind. And I think the future for that... So Wasabi 2.0 is designing to be a very simple wallet to make privacy accessible for everybody, but WabiSabi is a fully general protocol that you can do many, many interesting things. And I'm very hopeful that in the coming years, we can integrate that into more wallets, and hopefully wallets that also do like interesting layer two scaling and privacy technologies as well.
2:03:47So yeah, I have much more ideas and ambitions than I have time. And most of that has to do with, like, I just want to think more nice thoughts before I die. And life is finite, and this like consumer culture, like bullshit culture, is very intrusive. It keeps demanding attention for things that I see little to no value in. So everything I try to do has, you know, is in a sense a kind of defensive response against that ongoing attack on our collective consciousness.
2:04:38Vlad CosteaAlright, I just followed you on GitHub and I noticed that since 2008 when you started your account, yes, 2008, you have committed to 231 repositories, which is huge. Used to be very active, like 2008, 2009. Even this year, you have like a nearly full board, just working every day. It seems like you have been very proactive. So, yeah. One last question that I have for you. Are there any projects or any people that you follow and to whom or to which you look up?
2:05:20Yuval KogmanOoh, there's so many, so many lists. I think the Rust Bitcoin ecosystem is very exciting. There's beautiful work being done there. I think Bitcoin Core is a project I've only contributed a very silly comment patch to and helped debug this like weird edge case. So I think my most significant contribution there is basically just a comment on a GitHub issue. But I try to follow it, and there's really amazing stuff. I'm super excited for, like, Russell Yanofsky's multiprocess Bitcoin stuff. Oh, it's just, you know, there's privacy projects. Like, I think Nym is really cool.
2:06:13And I think everybody is criticizing it in, like, sort of weird ways, like, why don't you just use Lightning? I mean, they have, I think, very, like, it's one of the first projects that I actually believe has a token that makes sense on some level, and I am really hopeful that they stay true to the... like, they don't squander this opportunity with any sort of weird economic decisions or something, but it appears to be moving in a very nice direction.
2:06:58I am obsessed, utterly obsessed, with content-addressed systems, so anything having to do with Merkleized data is interesting. There's just so much going on in open source, so much to keep up on. So if you go into my GitHub profile, like most of what you see there is dead code from a prior life where I was very active in the Perl community. And I think none of that code is really meaningful or valuable anymore. The bits that are valuable are very thankfully maintained by other people who are doing a much better job than me.
2:07:41And yeah, I've kind of abandoned a lot of that stuff. And I mean, you said I have a lot of contributions. That's a very variable thing, right? Like I have years of silence in there as well. But if you're interested in what I'm interested in, then maybe the most valuable thing on my GitHub profile is actually my starred repositories. There's literally thousands of intriguing or interesting or like just projects that I think are cool. And I think I spend most of my time reading research papers or reading source code these days.
2:08:19So there's just a lot of good stuff going on out there. A lot of good people with good ideas about how to make the world a better place and to actually put the computers in the service of human creativity. And yeah, there's no one answer for this. It's in Bitcoin and it's much larger than Bitcoin. And yeah, this is kind of a cop-out, but I'm just... my mind is kind of like overflowing with all this like just stuff that I think is cool and I appreciate. And I don't even know where to begin.
2:09:03Vlad CosteaWhich is fine because you do a lot of useful and interesting work in the field. And I very much look forward to see how you solve some of the issues in the WabiSabi protocol and what your take on that one is going to be. So thank you very much, Yuval. I very much appreciate your time and your inputs. I think that most people will not get to this point to listen for two hours. So maybe that we should give them some sort of reward or shout out.
2:09:33So let me think of something. So how about you tweet at me and say, "Yuval is awesome." And if you do that, I'm going to find a way to reward you, maybe send you something. I'll see. I'll figure it out along the way. So thank you very much, Yuval, and I'll talk to you later.
2:09:57Yuval KogmanThanks so much. And yeah, thanks for indulging, both you and the audience. I hope I didn't test your patience too much and that I was able to share, you know, from like my perspective on what is beautiful or important about the world. Yeah. Thank you.
2:10:56Ad voiceoverVaultoro is the exchange where you deal with honest and hard money. There is no fiat on-ramp or off-ramp and you get to diversify your Bitcoin portfolio into gold or silver when you sense that a bearish moment is coming. Also, you can instantly trade your gold for Bitcoin to buy the dip, and if you're into gold custody, Vaultoro can also send you the gold that you own directly from their insured Swiss vaulting facilities. Vaultoro was launched in the aftermath of the Mt. Gox hack, so since 2015, they have published monthly glass books to prove that they own all the gold reserves and all of their customers' money. Sign up today by going to vaultoro.com/bitcointakeover. Keep in mind that this is not financial advice and you are responsible for your own decisions.
2:12:17Wasabi Wallet is the perfect Bitcoin privacy wallet. It's free, it's open source, it's available on Windows, macOS and Linux, and it offers groundbreaking Chaumian CoinJoins, which mix your Bitcoin. Even if you do not use the CoinJoin feature, you still benefit from a trustless experience with block filters, a hidden IP address via Tor, and easy management of your wallet outputs. After you deal with KYC exchanges like Coinbase, like Kraken, Binance, Gemini, or Bitfinex, you can remove the association between your identity and your Bitcoin address by performing a few rounds of CoinJoins. To find out more about the privacy benefits and limitations of CoinJoins, listen to Bitcoin Takeover Podcast Season 6 Episode 6 with Max Hillebrand. And if you want to give Wasabi a try, go to wasabiwallet.io and download the wallet for free.
2:13:23Wasabi Wallet, a Bitcoin privacy wallet for the citadels. Bitcoin Reserve is Europe's destination for buying Bitcoin simply online. It's the first Bitcoin-only venue in Europe that allows serious beginners as well as professionals to make daily purchases larger than 1,000 euro per day from bank account or via credit card. Learn more about Bitcoin Reserve by heading to bitcoinreserve.com. Bitcoin Reserve, where Europe buys Bitcoin.
2:13:23================================================================================ Production notes: Remote recorded conversation in English between Vlad Costea and Yuval Kogman, with a short music intro and repeated sponsor voiceover blocks (Vaultoro, Wasabi Wallet, Bitcoin Reserve) at roughly 00:28, 01:34 and 02:10. Correction: anthropic/claude-opus-5.5 via NanoGPT in 7 windows with a whole-episode speaker/glossary plan, the show notes, the Qwen3-ASR full pass (179 chunks) and 46 Scribe v3 re-checks of doubtful spans; second opinion: /workspace/bitcoin-takeover-best-quotes/s10-transcripts/s10e7.txt (read-only); [unclear] resolve pass: 2 of 2 flagged paragraphs re-checked. Automatic checks: word count vs raw, timestamps, seam duplication. Fixes: Removed hallucinations (♪, "Thank you." at 01:49/02:09/06:20, "This is a very important topic." x2, "This is a video that was made by a friend of mine..."); restored dropped speech from Qwen/Scribe (Yuval's "Hello, hello", "He is a researcher", "you need to protect your business interests", "personal safety part, which is something", Amiti passage, "lines of evidence... network layer", "naive maybe... cypherpunk"); fixed Yuval/Chaumian e-cash/nothingmuch/xpub/heuristics/light clients/Bloom/"Reddit"→"Right?"; reattributed "sorry" and "Yeah. Hmm." to Yuval; split 07:28 paragraph; Removed hallucinations ("This is a very important topic...", "Thank you for watching", "Episode:", "Video by Yannick Mahé", "Thank you", "Golden Bitcoin of Altoro" → jingle "Gold and Bitcoin on Vaultoro"); restored dropped speech from second opinion/Qwen/Scribe (slap on the wrist, cash used by criminals, Hal Finney, federated ones, Selgin banking, Vlad's paper bills question); relabelled ads as Ad voiceover and 00:32:12 as Vlad; fixed Vaultoro, Chaumian, CoinJoins, Opendime, Tangem, Fedimint, Cashu (unclear), WabiSabi, eCash, David Chaum, nopara73, Max Hillebrand, Monero/Zcash; Removed hallucinations ("This is a podcast...", "This is a very important topic", "This is a video that I made...", "Video by Yannick Mahé", "Episode:", "I'm going to show you how to do it", stray "Thank you"); restored dropped speech from second ASR (withdrawal/anonymity set passage, "place in the binary representation", "hold that thought", "$5 on the table", "bundling together of bits", Nym network, "unspent coins and so on", "inflation protection", "use case... eCash system"); split "I think Adam Back was" to Vlad and "Yes, and" to Vlad; fixed Patterson→Pedersen, Wabisabi→WabiSabi, TOR→Tor, coin join→CoinJoin, NIM→Nym; Removed hallucinations ("Episode: Yuval Kogmanpresented...", "This is a video that needs to scale", stray "Thank you.", "Okay."); restored dropped words from second opinion/Qwen ("is kind of going for right now", "a balance that is convex", "Whether or not it scales as a technology for privacy", "the output... take all of the input amounts, compute the greatest", "what the structure"); fixed "perfectly fine"→fungible, bloat→load, 1-2-5 series, Adam Nopara→nopara73, SharedCoin, coin joint/coin-chain→CoinJoin, WabiSabi, statechain; split long turn at 00:04:20; split Vlad's "Well, obviously." at 01:12:03 as Vlad; Speaker fixed to Yuval Kogman throughout (Unknown speaker at 01:34:12); ad blocks relabelled Ad voiceover; hallucinations removed ("This is a video that I made...", "Thank you.", "Okay.", "Golden Bitcoin of Altoro"); dropped speech restored from second ASR (encumbrance passage, exit scam, consensus rules line, rent-seeking fiat, "I just don't want to prove it to all the rest", "even if CoinJoin is not", Vaultoro gold-for-Bitcoin line, Wasabi trustless/Tor line, Max Hillebrand); jargon fixed: WabiSabi, CoinJoins, PayJoins, statechains, Mercury, unforgeability, Chaumian, V