Episode 37: Mara Pool and Mining Censorship

Aaron van Wirdum Live from Oranjestad and Utrecht, this is the Van Wirdum Sjorsnado.

Sjors Provoost Hello.

Aaron van Wirdum Hey Sjors.

Sjors Provoost Welcome to this remote session.

Aaron van Wirdum We're doing it remote. I'm on my way to Miami, my long trip to Miami. I'm in Aruba at the moment. And you're still in Utrecht.

Sjors Provoost Yep. Yeah, I'm enjoying, well, cloudy skies, rain and cold. And you're going for a swim in the tropics.

Aaron van Wirdum Have fun with that.

Sjors Provoost Exactly. Have fun staying cold.

Aaron van Wirdum Have fun staying cold and wet, Sjors. So, yeah, the audio quality is not as great as usual, but hopefully it's good enough. Today we are going to discuss miner censorship and OFAC compliance more specifically.

Sjors Provoost Yes.

Aaron van Wirdum Sjors, oh, I don't have my notes at me. Can you explain the context here? There was a new mining pool.

Sjors Provoost I think they've been around for a while, right? I forgot their name. But they basically have been bragging on press releases and social media that they are a compliant pool. They comply with all U.S. regulations, which is quite amazing because you have no idea how many regulations the U.S. has. But in particular, they want to be compliant with sanction lists.

Aaron van Wirdum It's Marathon Digital Holdings. And I think, is that also the name of the mining pool? Maybe, maybe not. Mara Pool is the name of the mining pool.

Sjors Provoost Mara Pool, that sounds familiar.

Aaron van Wirdum Yeah, that's the one. I think there are two mining pools that are doing this now, actually. I think there might be another one. But anyways, yeah, Mara Pool is one of them. And they mined their first. We're recording it on Thursday, May 6th. And they recorded their first clean block today.

Sjors Provoost Yeah, well, I mean, the only thing that technically happened is there was an OP_RETURN message in the coinbase saying OFAC compliant.

Aaron van Wirdum Right. Okay, so let's talk about this. What does this mean? What can miners do to be compliant? Should we consider it censorship? How should we look at this? So let's first start with the facts. Yeah, miners, of course, include transactions in blocks. How can they actually, let's just call it censorship. Sjors, what do you think?

Sjors Provoost I mean, sure. In this case, let's consider what they could censor, right? Not what they actually did censor. So a block consists of transactions, and transactions spend an existing UTXO, and they usually create a new UTXO. And so you can censor either, you can say, well, we don't want these UTXOs to be spent. So you censor based on the input of a transaction, or you can censor based on where the funds are going, the output of a transaction. So you can say, well, if it goes to this address, then we don't mine it. Those are the two possibilities.

Aaron van Wirdum Yeah, so a transaction is sent over a network, and then usually a miner... will pick, you know, it's mining pools generally that are actually creating blocks. And usually they will just create the set of transactions. They'll pick the transactions that pay the most fees.

Sjors Provoost Right, that would be the default way to do it.

Aaron van Wirdum Assuming there are more transactions in their mempool than fit in a single block, they can't fit them all in, so they'll select the transactions that pay them the most fees. It's a little bit more complicated than that. We've discussed that in a previous episode where the transaction selection is actually based on how they can make the most aggregate fees. But that's not the point here. We're not going to get into that. Usually the default way is a miner just picks the transactions that pays them the most fees. However, in this case they can exclude certain transactions for particular reasons, in this case because they want to be compliant with what is called the “oh fuck” list. It’s yours.

Sjors Provoost Yeah, and I think you're pronouncing it in a very specific way that suggests a slightly different spelling. But it stands for, you know, the Office of Foreign Asset Control or something like that. So this is interesting because in the promotional material of that particular pool, they were bragging about complying with all U.S. regulations, including all anti-money laundering regulations and OFAC. And for the legal nerds, which I'm not, but I find this stuff somewhat interesting, the OFAC is a very, very narrow thing. It's about basically blacklisting foreign dictators and whoever the U.S. doesn't like. And most of the people in the U.S. doesn't like, nobody likes. But in this case, a couple of years ago, the OFAC published, I think, a list of Bitcoin addresses saying, well, these addresses are associated with bad people. And so my guess is, but we don't know because all we've got is this OP_RETURN message. My guess is that they made sure that this particular block did not contain a transaction to or from that address.

Aaron van Wirdum Yeah, I think there were like two addresses, right? You say a list, but it was really this two, if I'm recalling correctly.

Sjors Provoost Yeah, last time we checked it was two. Anyway, the point is the odds that those addresses were used at all are practically zero. So they could have just published this OP_RETURN message randomly in some block without even checking anything. And as far as we know, for example, on forkmonitor.info, you can click on a block. If you know the hash, you can still look at an old block too. And it tries to track what transactions should have been in a block based on the fees, the sequence of fees, and then which transactions were actually in a block. And the difference between that might be censored or might just be random, like some timing problem or whatever. and as far as I know there was nothing censored and there was definitely nothing censored that would be on the OFAC list.

Aaron van Wirdum Right. Yeah, so right.

Sjors Provoost They didn't actually censor anything as far as we know.

Aaron van Wirdum Right. So right now, as far as we know there or last time I checked anyways there were like two addresses on this OFAC list so these two addresses you can send any coins to or from but of course this could be expanded over time and over time there are more and more addresses on this list and that would mean that at some points they might actually censor transactions that include relevant addresses that are on this list.

Sjors Provoost Yeah, or it could be much worse, right? For example, I think there was a settlement between BitGo and some U.S. department, and I think that was about OFAC very specifically. They were accused of allowing people in Cuba to make transactions, and that entire country is on the OFAC list. And basically, the way they had to comply with OFAC was a lot more than just blocking a certain address because OFAC does not say these Bitcoin addresses belong to Cuba. It's up to you to decide which Bitcoin addresses belong to Cuba if you're a service provider like BitGo. So you can lazily say we're OFAC compliant, but OFAC might not agree with that at all, especially if you're making a claim like that. So I don't know why they're kicking this hornet nest.

Aaron van Wirdum Right.

Sjors Provoost Because now OFAC might come back and say, hey, we actually did an investigation of all these addresses, and we find that actually there are two that went to Cuba. And now you guys go to jail for lying. I don't think you can go to jail for lying about that, but, you know.

Aaron van Wirdum Let's… let's… get into what does… what does… this actually achieve? What does it actually mean if a miner like this, a Mara Pool, is keeping transactions out of their blocks? How would this affect? Let's say there is a transaction that's sent to one of these addresses, for example.

Sjors Provoost Yeah, then they would miss out on, I don't know, 100 Satoshi's worth of fees, and the transaction would get into the next block.

Aaron van Wirdum Yeah, so basically in that case, there would be a little bit of delay for the transaction. And maybe it could have some weird effects on the transaction fee. I was, you know, thinking. maybe it would have gotten in that block and just right after that block, there's sort of a flood of new transactions on the network for whatever reason, and then it's excluded in the next block because the fee is not high enough. So maybe if you want to be safely, sorry, and you're sending a transaction to an address like this, it would be better to pay a little bit more fees. But in general, if we're painting with broad strokes, we can say the only thing it really achieves is a little bit of delay, right?

Sjors Provoost Probably not even that because the odds of that block being mined by that specific pool are also very small because there are so many other pools.

Aaron van Wirdum Well, there's a little bit of a difference. So there's a little bit of delay and therefore I would argue it's a little bit of an infringement on fungibility. Fungibility is this idea that one Bitcoin is one Bitcoin. It doesn't matter which Bitcoin you'll have in the same way that it doesn't matter, you know, $1 bill is $1 bill. You can just change $1 bill for another one. It'll pay you the same thing at every store. And ideally we'd have that for Bitcoin. But if there are some transactions, some, you know, Bitcoins that are going to be delayed more than others, even though that is a very mild inconvenience, it's already harming fungibility at least a little bit.

Sjors Provoost I don't know about that because I can't measure it. I just can see how this goes down a slippery slope to something that you will start noticing very, very badly.

Aaron van Wirdum Yeah, the good thing to finish off that thought is that the counter argument there is that pools that are doing this, like this Mara Pool, they are also making less in fees. Miners that will just include every fee into their block, they're making the optimum amount of fees. If you're not doing that, you're necessarily going to... earn less fees so maybe on the long run on the in the long term, you'd be out competed by miners that will just include all transactions

Sjors Provoost yeah but not but not if you only limited to these two addresses on the OFAC list, then it won't have any difference for the next million years.

Aaron van Wirdum Then the difference will be very small.

Sjors Provoost But the problem is, let's say you really want to be OFAC compliant. Now, the best way to be OFAC compliant as a mining pool is to mine empty blocks. And even then, you might not be OFAC compliant. Because let's say if you mine empty blocks, then you could say, well, if a transaction comes from a known exchange and we know they've KYC'd everybody, both the from and to address are KYC'd, then you allow those transactions in, something crazy like that?

Aaron van Wirdum An empty block doesn't have any transactions in [unclear].

Sjors Provoost No, I was saying like empty block is an extreme example, but you'd start with an empty block and then you add everything that's whitelisted.

Aaron van Wirdum Right.

Sjors Provoost But even an empty block might not be OFAC compliant.

Aaron van Wirdum Why?

Sjors Provoost Because you're a mining pool, so you don't know who the miners are in your pool. So if the miner in your pool is some guy in Iran and then he's using electricity from a local Iranian power plant, you're still violating OFAC, at least, you know. You probably should not claim that you're not violating OFAC. Let's put it that way.

Aaron van Wirdum Right. You should first need to KYC all of your hashers as well.

Sjors Provoost Yeah, and it's possible maybe they're already doing that. So that's a far, far greater censorship problem than those two addresses that they're not mining, the fact that they're KYC-ing their miners, because even that's not necessarily normal with pools.

Aaron van Wirdum Yeah.

Sjors Provoost Right?

Aaron van Wirdum Now let's get into a much bigger... Yeah, it sucks to record a podcast at distance, Sjors, because now I'm not sure if you were done talking or not.

Sjors Provoost Well, I was trying to go to what this could look like if it gets out of hand a little bit more.

Aaron van Wirdum Right. Well, that's where I wanted to go as well. And let me first mention this. So right now we were talking about just a single pool. And that's something Bitcoin can handle, I think. And I think we both agree for the reasons we just explained. Now, it becomes a real problem if... miners and mining pools that do this have a majority and they actually start to collude to reject not just transactions… that include not just… transactions but also blocks that include certain transactions right?

Sjors Provoost Exactly. So the building on top of a block that is not compliant.

Aaron van Wirdum Yeah if a majority of miners start doing that then your transaction isn't just going to be delayed. It's just going to be un-minable. It's just going to be delayed forever because, first of all, no miner is going to… well, a majority of miners isn't going to put it in a block. And then even if one of the minority of miners puts it in the block, that minority knows that block is going to be orphaned by the majority. So more likely than not, the minority will not even put it in the block in the first place just to make sure they're not going to be orphaned.

Sjors Provoost Precisely, so you get self-censorship in addition to being compliant.

Aaron van Wirdum Moreover, I think this would start to become a very real effect way before there's a majority actually doing this because you have this selfish mining dynamic where it probably starts being a real problem at like, I don't know, what is it, 30%, 33% where minority of miners are probably just going to comply with the censorship because otherwise they'll have their blocks orphaned.

Sjors Provoost So I'm going to make a little prediction here about a new company that's going to come into existence sooner rather than later. And that is an insurance company. And what this insurance company does is if you are a miner, you can buy insurance with them that they will pay for any block you need to orphan. So if you, you know, because it's very expensive for you to orphan a non-compliant block. But if the insurance company just pays for, you know, every time you don't end up in the longest chain, well, then that removes the risks for you. And that might be economical if on the other side of this equation is the U.S. government giving you a billion dollar fine or putting you in jail. You know, you might be willing to sacrifice 0.1% of your profits. And the insurance would ultimately become extremely cheap. Because as soon as you get this self-censorship dynamic, there really just is hardly any blocks that need to be orphaned. So the insurance never has to pay out. So this whole insurance scheme could be paid by some venture capitalist who wants this future. Basically insures miners at a loss until the dynamics are there. I can make this even more gloomy if you like.

Aaron van Wirdum Go for it, Sjors.

Sjors Provoost Yeah, so I briefly mentioned the idea of whitelisting transactions. So a compliant pool or even a pool that doesn't want to be compliant but just doesn't want to risk the orphaning. That's in the later stage. But in the first stage, if you just want to be the compliant pool, you probably want to have some API for all the exchanges. And every exchange publishes a feed of every transaction they're broadcasting with some sort of score next to it. You know, the maximum score would be like, yeah, we thoroughly KYC'd this person. Like, we anally probed them, checked them for COVID. Like, we really know this transaction is legit. And you only mine what's in that RSS feed depending on your risk tolerance. That's a problem. And then the other side is how many pools would be compliant. And my worst case scenario there would be... For example, the U.S. might tell every miner, physical miner in the country, like, hey, you're welcome to mine, but there's a couple of conditions. One, you have to use renewable energy. Okay, that's fine. And the other is you have to use one of the OFAC-compliant pools. And you can self-regulate this thing, and then you get this weird self-regulation stuff where people make up rules that are really bad. So that's… And then you think, okay, that's just the U.S. pools. That's still a minority, although the U.S. pools are getting bigger. And then the European pools join because Europe likes this idea. And then, you know, as part of some goodwill gesture from China, they also join this thing. And now we have, you know, basically an international cartel of compliant miners that are using compliant pools. And then the question is, like, is the Bitcoin communities going to roll over and accept this? Because the number might still go up. And a lot of big Bitcoiners, for them, the number will go up. And is this really their fight? I don't know. Even if it's not their fight, I don't know what the defense would be.

Aaron van Wirdum Yeah, that's an interesting point you bring up. I think there are several potential responses and there are a number of potential ideas that you could use in this case, I think Adam Back has been speaking about this concept called committed transactions. I think that's what the term is.

Sjors Provoost Yep.

Aaron van Wirdum I think there may be other ideas out there.

Sjors Provoost The idea basically being, in general, you make it physically impossible to do this type of censorship because you don't know what's going to be in a block. Right?

Aaron van Wirdum Sorry, what's that?

Sjors Provoost The idea is to make it impossible because... you can't see what is in a block. Like the transactions are blinded in a way.

Aaron van Wirdum Yeah, it's only revealed later. But it's an interesting idea that raises new questions like how would Lightning handle with something like this? If a penalty transaction is broadcast, how would you manage that?

Sjors Provoost Yeah, and you can get into this. The biggest problem with this type of regulation in general, like we just talked about OFAC, but any kind of AML stuff, is the reverse burden of evidence. So if your transactions are cloaked, well, then you need to prove that they don't contain any sanctions transactions and otherwise you're not allowed to mine them. So it's hard to overcome that. What I'm a little bit more optimistic about is simple things like input aggregation and also just Lightning because there you can say, well, we had to aggregate all these transactions and so we can't really see what it's what because otherwise the fees are too high. And the same with Lightning, right? You cannot publish the details of every Lightning transaction. Because that simply doesn't fit on the blockchain. So we could get lucky that sort of the forces of nature in the sense of transaction capacity pushes the information, like makes it impossible to have the information you need for complete censorship. And maybe we get lucky that way.

Aaron van Wirdum Sure, yeah, better privacy itself is a potential solution.

Sjors Provoost But I mean, the combination of privacy and the incentive for privacy being financial, rather than direct. Because if you say, I want my privacy because I'm a citizen and I have rights, well, then you're suspicious, obviously. I'm being sarcastic here. But if you're saying, well, I don't want privacy, but Lightning is just cheaper, well, then it's fine because you should be frugal.

Aaron van Wirdum Yeah, there's also the Eric Voskuil argument who says if your transactions are being censored, that just means you need to start paying more fees. There are ultimately very radical proposals, which is sort of, if nothing else works, we can always change the proof of work with a hard fork, something like that.

Sjors Provoost And then what?

Aaron van Wirdum Well, then at least...

Sjors Provoost We get all the problems that altcoins have with random proof of work. Yeah.

Aaron van Wirdum At least it will probably work for a little while again and figure out a solution in the meantime.

Sjors Provoost Who knows? Maybe it's not solvable. I hope that things will end up well, but the universe does not guarantee that this thing will work. So one scenario might be that we just end up with a completely non-private money, but at least it's hard money. So we won one battle, and then maybe a hundred or a thousand years from now somebody comes up with a new system that's also private. Or who knows, maybe governments will at some point respect privacy. I mean, radical things have happened in the past.

Aaron van Wirdum Yeah, I think that's an important part of it. People are just in general, not just governments. I mean, governments work for their people. So I think in general, people are just going to have to become more comfortable with the fact that people should have privacy, including privacy and transactions, which is an uphill battle. But that's probably part of the battle here. But yeah, I think we've now spent 20 minutes speaking about this at distance. And I think, I would say this is actually the biggest problem that Bitcoin hasn't quite solved yet. And would you agree with that?

Sjors Provoost Yeah, I think this is probably the biggest threat to Bitcoin.

Aaron van Wirdum Yeah.

Sjors Provoost Period.

Aaron van Wirdum And we're seeing concrete, you know, it's a very, very subtle, small start of it. But, you know, it's now officially a thing, I guess I would say.

Sjors Provoost I guess it was a thing when OFAC published those two addresses. And then it was another thing again when two pools, you know, from a marketing point of view said, hey, we're compliant. We like this. And now it's another thing because somebody put in an OP_RETURN and otherwise probably did not do anything then to publish this OP_RETURN. And then the next time, I mean, the problem with OFAC, of course, is the list is too small. So it doesn't, you can't really do anything, but he could probably be… you know, censor a lot more transactions. If you just use a basic filter with some of these chain analysis companies, then a darknet might really, you might really be able to see it. And then if the pool brags about that, saying, hey, look at that. You know, we stopped thousands and thousands of darknet transactions. Look at how amazing a force for good we are. They wouldn't have stopped it, right? They would have just delayed them by one block. But then they can probably sell those coins at a profit, the coinbase rewards. That's kind of the dynamic, I think. It's just a gradually increasing thing. I also made the intransigent minority argument earlier to you, but I don't know if we want to rehash that here.

Aaron van Wirdum Yeah, don't you mean the intolerant minority?

Sjors Provoost Yeah, I think there's two ways to do it. The intolerant minority, yeah. But my kind of working theory, and I could be wrong, and I know you don't agree with it, so that's good news, is that when you have an intolerant minority that is very strict about one specific thing where the rest of the people don't care, and I think that minority here is the government, and indirectly the government. So if you are a compliance person at a bank, and you want to open a, you know, and somebody comes to you and says, hey, I want to start a mining pool. Can I borrow some money from you? Then you might say, well, we really want you to be compliant because I'm the compliance officer and that's really all I do. And essentially, you know, you're a government person when you're a compliance person. Then you can say, well, we're going to give you the loan, but only if you join one of these censorship pools. Otherwise, we're not going to give you the loan. So you could still get a loan, but it's going to be at a higher interest rate. So it might not be worth it for you. It might be better for you to just censor those transactions and get a lower interest rate. And then, you know, when you are violating any of these rules, like the OFAC law, and you are probably doing something that's not allowed, the government can throw you in jail. So they're very intransigent about that. They're willing to use any level of violence, basically, to stop you from doing this. But on the flip side are hundreds of millions of potential investors who just want to buy coins, and they're happily uploading passports. So you have a very forceful group of people, like the government, like compliance officers, who really want you to not have privacy. And then you have a lot of people who would just go along with that. And it's very diffuse, because it's every bank out there. It's not just one big bank, it's every local bank branch.

Aaron van Wirdum Yeah.

Sjors Provoost You go to some small town and you want to open a business account. It's like, well, are you compliant?

Aaron van Wirdum Yeah, so we had this discussion just before recording the podcast, indeed. And I'll give you that this is sort of an open question, maybe. But I think the intolerant minority thing really works, really works well if the rest, everyone else, just doesn't really care. That's when the intolerant minority thing works. While I think in this case, there is at least going to be a minority that's going to be intolerant in the other direction. That's going to say... over my dead body that any transactions are going to be censored.

Sjors Provoost Yeah, I agree. There's probably at least a few thousand people.

Aaron van Wirdum We will consider that a 51% attack, and we will take measures against that. It's actually interesting that you bring this up also because, you know, we've just done a couple of episodes on Taproot activation and the UASF debate is relevant again. And the UASF is in a way an intolerant minority as well. And the reason that I think it works in the UASF context is because everyone wants Taproot. Everyone wants Taproot. There's no one against it. There's only debates basically on how to activate it. And I think when it comes to activation… I think, but I could be wrong, and I'm willing to hear why I would be wrong, but I think when it comes to the activation part itself, that's where the intolerant minority thing works. Because people have different opinions, but will they really sort of reject the whole idea if they get Taproot, or will they just be like, well, it wasn't our preference, but at least we get Taproot, so it's fine. That's sort of what I… that's sort of why I think the intolerant minority would work in that case.

Sjors Provoost Right. But the intolerant minority in general as a mechanism does not… may not require that. You just… all you need for that in general is a very fanatic group of people that is spatially dispersed. I mean, that's at least what Taleb says. Every neighborhood, every city, every neighborhood is like one Jewish family, and so the entire world is drinking kosher food… drinking kosher drinks. not necessarily eating kosher food. That's all you need. And it's not that the rest of the world has to be pro-kosher food. That's not at all the case. In Taproot, it is the case, right? Everybody says Taproot is great. But that part isn't necessary in general for this mechanism. And because of this geographic dispersion where in every place on the planet there is a compliance officer working on behalf of some government really insisting on you not having privacy, that's why I'm seeing the same dynamic. It's not, but it's very, it may be different from the UASF story.

Aaron van Wirdum Well, there's another interesting thing that's relevant. And we should dive in this deeper another time bit maybe, but I'm going to mention it briefly. It's this idea that 51% attacks and soft forks in the way they work are very similar. Right? They’re, they’re sort of…

Sjors Provoost Yeah, so a soft fork is like a consensual 51% attack.

Aaron van Wirdum What's that? Yes, exactly.

Sjors Provoost [unclear] Yeah.

Aaron van Wirdum So and I think that's actually one of the arguments that some people have been making in the Taproot debate is that what makes it consensual is that anyone who would want to opt out can opt out. And this requires that there is a signal on the blockchain which opponents can use to opt out. While 51% attack doesn't have this signal on the blockchain, so it doesn't really allow opponents to opt out of the upgrade, the upgrade in quote marks. So when it comes to a censorship pool like this, if they would get a majority, but they're never going to broadcast the signal on the blockchain, then they're not really leaving opponents of this change a way to opt out cleanly. And that's what makes it a 51% attack. Well, if they somehow would make a clear signal on the blockchain that there's going to be new rules, then that would allow you and me to say, no, no, we don't want these rules at all. We're going to use the signal to actually fork off to a different chain where there is no censorship. And that would sort of make it a soft fork because now we can opt out. This is something we need to revisit sometimes.

Sjors Provoost Yeah, and your prediction market solution is not going to help you either because the prediction market itself would have compliance problems.

Aaron van Wirdum No, it would actually work perfectly, especially in this case. It would work in the software case, but it wouldn't work in the 51% [unclear] case. And that's exactly the difference that I'm going to explain to you in another episode in more detail.

Sjors Provoost Yeah, sounds good. All right. So that was just one OP_RETURN message in a pool, which probably didn't censor anything. And, you know, here we are.

Aaron van Wirdum Here we are.

Sjors Provoost [unclear] later.

Aaron van Wirdum Sjors, I'm going back to the pool, I think.

Sjors Provoost Yeah, have fun in your pool. I will keep enjoying this cold weather here.

Aaron van Wirdum A very different type of pool. I'm going to the swimming pool.

Sjors Provoost Oh, that's terrible. I hate swimming pools. No, they're great. Is it an infinity pool? [unclear]

Aaron van Wirdum Oh, they've got a whole ton of pools here.

Sjors Provoost Okay, my favorite pool is like an infinity pool right next to the ocean. So you don't have to get your feet wet with the sand, but you do get to enjoy the view of the ocean. Those are perfect.

Aaron van Wirdum No, they don't have that. They do have infinity pools, it looks like.

Sjors Provoost All right. Well, everybody, thank you for listening to this somewhat strange episode of the Van Wirdum Sjorsnado.

Aaron van Wirdum There you go.