Vitalik Buterin: Ethereum 2.0 | Lex Fridman Podcast #188

The following is a conversation with Vitalik Buterin, his second time on the podcast. Vitalik is the co founder of Ethereum and one of the most influential people in cryptocurrency and technology broadly defined. Quick mention of our sponsors Athletic greens, Magic Spoon, Indeed Four Sigmatic and BetterHelp. Check them out in the description to support this podcast. As a side note, let me say that Ethereum, Bitcoin and many other cryptocurrencies have been taking a wild ride of prices going up and down in the past few months. To me, the prices were never as important as the ideas, both technical and philosophical. Cryptocurrency has the potential to empower billions of people to participate in the global economy in a way that resists the manipulation by centralized power. Also, with smart contracts, layer two technologies, data pools, NFTs, and of course integration of artificial intelligence into the whole thing, we have the opportunity to build tools and worlds that transform physical and digital life as we know it, hopefully minimizing the suffering in the world and maximizing the fun. This is the Lex Friedman podcast and here is my conversation with Vitalik Buterin. Let's first talk about Shiba Inu, if we can. Also known as Shiba Token Code Shib for context, Shiba Unu was created in August 2020 modeled off of Dogecoin by the anonymous founder known as Ryoshi. On May 10 this year it had a market capitalization of over 13 billion. And maybe you can explain this, but in a crazy move, you were given half of Shib total supply. You burned aka destroyed 90% of it, that's worth $6.7 billion, and you donated 10%, that's worth 1.2 billion at the time to an India COVID 19 relief fund saying, you don't want to be the locus of this much power. This is fascinating. Why? And how were you able to walk away from this much money and this much power?

So I should probably start by giving some of the backstory around these coins and this concept of giving me coins. So first of all, Shiba Inu, as you said, is this kind of knockoff of Dogecoin, right? And Dogecoin was this initial kind of fun coin that was created back, I think around 2014 or so, and it was just created by Jackson Palmer and put it out as a joke for a couple of hours and a community formed around it. And at the beginning people didn't take it very seriously. I actually remember putting about $25,000 into Doge sometime around 2016, and I just remember thinking to myself like, okay, how am I going to explain to my mom that I just invested $25,000 into dog coins? What even are dog coins? Like, interesting thing about this coin is that there's a logo of a dog somewhere. But of course, that ended up being one of the best investments I've ever made, and it did really well. And then at the end of 2020, Elon Musk, of course, started talking about dogecoin, and the market cap just shot up to about $50 billion. Actually, it shot up multiple times, right? Like, the first time, it went up from about 0.8 cents to about 7 cents. And this just happened all in one day. And I remember this was when still in Singapore in the middle of COVID and I saw that the price just went up by a thousand percent. And I was like, oh, my God, my Doge is worth, like, a lot. And so I immediately called up some of my friends and told them to drop everything and scramble. And I sold half the Doge and I got $4.3 million, donated the proceeds to give directly. And a few hours after I did this, the price dropped back down from about 7 cents to 4 cents, right? So I managed to sell the Doge at the top. And I remember just feeling like I was such an amazing trader. But then, of course, you know, the price went up from 4 cents, then to 7, and then 50. And just like Doge becoming this big phenomenon where there's even a lot of people that have heard of Doge that have not heard of Ethereum, is just like something even I wasn't predicting, right? And so after that, of course, you know, we have Doge, and then people are thinking, well, if the leading dog token is worth $50 billion, then surely the second largest dog token deserves at least 7 or 8 billion. I feel like that's kind of what the mindset of these Shiba people is. So then of course, they did this other gimmick, right, where they gave me half the Shiba Token supply. They were actually not the first projects to do this. So around the end of 2020, there was this weird project called Teller. It's like T E L L O R. I think they're a chainlink competitor or something like this. But I remember they just dumped $50,000 worth of their token into my wallet. And then they had their Twitter army just basically run around saying, look, look at Vitalik's wallet. Vitalik holds Tellers. He's one of us. He's a supporter. And as soon as I discovered this, I just publicly sold the Teller tokens on Uniswap. And this created a bit of a Twitter splat. Now, the Shiba people were more clever. The Shiba people, instead of dumping to that wallet, they dumped to my cold wallet. So in cryptocurrency, there's this concept of cold wallets and hot wallets, basically the thing that actually owns your money is this 80 digit number called a private key. And a hot wallet is when that private key is just stored in memory on your computer, on your phone, really easy to access. Cold wallet means it's either written down on a piece of paper or it's on a computer that's just never accessed the Internet. So cold is very inconvenient. But cold is also much more secure because even if that computer has some viruses on it, it's air gapped, it's not actually going to be able to upload it. So this cold wallet and all the money is out of the cold wallet. So it's safe for me to talk about my setup now. Right. But it was a laptop that was sitting in Canada and I also had two pieces of paper where I wrote down two numbers on those two pieces of paper. One was with me, one was in Canada. And if you add those two numbers together, you get the private key. So because of COVID travel restrictions and you know, this is this cold wallets in Canada, like it's very difficult for me to actually access it. Right. And I'm not sure if they knew this, maybe they just got lucky. But basically they sent a lot of these dog tokens into this wallet where it was very difficult for me to access it. But then I saw these dog tokens, I saw more and more people talking about them. And then at some point I realized, realized that, hey, these things are worth billions of dollars. And there's lots of really good things that you could do with that amount of money and it would actually be a waste to just see it go. So I made the decision that I would actually power through and figure out how to safely basically get my private key. I actually had to call up my family, tell them to read out their number off of their piece of paper. I entered that into a fresh laptop that I bought from Target. Then I put in my other number on my piece of paper, added the two numbers together on the computer, there's the key. And at the same time just scrambled for two days, setting up a new wallet where I could move my eth2 safely, getting people to be multisig partners, just doing all sorts of stuff that 10 years ago, you would expect to just be part of a cyberpunk science fiction novel, but now it's all real.

So you're doing this all by yourself?

Essentially most of it by myself. So I needed.

Because you have to keep it secret.

Right. I needed my family to actually go and read the number on their piece of paper, and then I am in my new multisig wallet. There's other people that are signatories, but I'm obviously not going to reveal any details beyond that. So I did this and I actually managed to get the private key, make the first transaction. That would just move all my ether to the multisig wallet so it's safe. And. And then second transact, put the private key on my main computer, then started going in and just selling some of the dog tokens and then just giving them to these different charities. Now, at the time, I actually did not even have any idea of how much you would be able to get. Right. Because on paper, the dog tokens are $7 billion, but in reality, it's a very illiquid market. Are you going to crash it by. After you sell 1 million worth, you're going to crash it after 10 million, might you actually be able to get like an entire 200 million? I had no idea. So I definitely was just of the mindset, like, okay, I'll sell a bit, maybe get some eth and then donated some eth to give. Well, donated some to other groups and then, okay, have some dog tokens. Like, I don't have an easy ability to sell more myself, but then I'll just, like, give them to these groups and hopefully they'll do good things with them. I actually donated 20% and dumped 80%. Yeah. So the COVID India group got one batch, and then there's another group that got another batch. And I don't want to say who they are because I think they want to announce themselves at some point. Sure, yeah. But you can see the fact that these transactions were made on the blockchain. But it was just very interesting and unexpected and just an insanely crazy situation.

It's been a couple of weeks. First of all, thank you for helping me hang up some curtains. This is a first for the podcast and shows that you're truly a special person to be willing to help. But now, a couple of weeks later, do you regret any aspect of that decision?

I'm sure there are some things that I probably could have done better. I was actually talking to some of these charities and I was impressed by just how much money they managed to get out of selling some of these coins. So I probably could have done better by just talking more with the traders and actually ensuring that they can do a better job of maximizing the value of, of all of them. But like, you know, it was a very stressful time and I did have to act quickly. Like, I did manage to make a lot of the donations before a few days before the great crypto crash happened. So it was. It's difficult to like, obviously there's parallel universes in which I did better, but at the same time there's also lots of parallel universes where because I hesitated more and tried to spend more time thinking I missed the opportunity. So on net, it's like a luck of the draw and I'm just happy that everything was able to turn out as well as it did.

But psychologically, you mentioned stress.

How hard was was stressful, right? I think, well, one of the really stressful parts was just the fact that I had to basically move all of my funds, including the 325,000 ether, from one cold wallet into another hot wallet or sorry, into another multisig wallet. And maybe the multisig wallet had a bug in it. Maybe there's some mistake I'll make in the middle that causes the funds to get lost. That part was stressful. And I was definitely stressing out for two days triple checking the new wallet. I even did a bit of an audit of the code myself. I wrote my own JavaScript dapp to make confirmations because gnosis safe didn't work with the status wallet. Well, so there was definitely. That whole thing was definitely a bit of a marathon. I was also kind of definitely a bit worried about or uncertain, I guess, how the public and you know, including the coin communities would perceive the whole thing. But I was actually impressed. Like I. For every poster that was saying like, no, you know, why did Vitalik Rug pull on as he was. His wallet was supposed to be a burn address. There's like 10 people that were like, oh, I thought I was just in this because it's a fun pyramid gambling thing. But instead I ended up being part of this great public good thing for humanity and that's even more amazing. So the amount of that I got was very impressive. So all in all, I think the dog people did great.

The dog people. Is there something you can extend to the bigger picture of it in the principles you apply to making this decision? Is there some principles, philosophies that you apply also to the decisions you make around Ethereum?

I think a Big one for me is just this idea that crypto isn't just an opportunity to give people slightly better ways to save value in all of these things. It's also an opportunity to basically create these new digital institutions that could serve the public good in new ways. And that's something that I've been interested in for a long time. I actually even have this article in Bitcoin magazine back in 2014 where I basically suggested this idea that you would have coins that represent causes, and people would just buy and accept those coins because they support those causes. So I think it's called markets, institutions and currencies, a new form of social incentivization or something like that. And I'm sure you can find it and throw it in the links. So that was interesting to kind of see becoming real. In general, I think public goods are very important. And on the Internet, public goods are even more important. Right. Like every single Lex Friedman podcast is just on YouTube and anyone can go and see it. There's no way for you to sell it so that some people can see it, but other people can't see it. You could do that, but then you'd obviously be reducing your impact. So thank you for making the amazing Lex Friedman podcast still freely available.

Well, that's actually a tense thing, is how do you do it in a way that's not controlled in a centralized fashion? Because YouTube feels free and open, but it nevertheless is one company making centralized decisions. And the first time I realized YouTube is not forever is when a lot of the Joe Rogan Experience library was pulled from YouTube as part of the Spotify deal. And he made me realize we need to. It's like the realization the fiat money is centralized is realizing that this is not forever. And you might want to come up with schemes to distribute it, to decentralize the control of it in the way that audio for podcast is with just an RSS feed.

Exactly. One of the kind of philosophical things that I hope to achieve is kind of decouple the concept of public goods, which are incredibly important and are the lifeblood of modern civilization, from the idea that there is or can be one central organization that represents the public and perfectly understands and can impose their idea of what is the good. When people talk about public goods, it just often comes with this baggage of either centralization or conformism. And I think it doesn't have to. Right. Often the most important public goods are the ones that are created by the crazy individualists that disagree with everyone else. So trying to make this kind of synthesis where you combine the values of decentralization and the values of open source. But you're not naive about it and you realize that for these things to be produced, there needs to be a way for it to be sustainable. There needs to be some way of supporting people who are working these projects. Projects. But at the same time, you want to avoid that turning into a vector of centralization, like trying to sort of get all of the good things without the bad things. To me, that's a big part of sort of what my grand experiment in crypto is about. And we are doing things in different kinds of things for this, right? Like there's the Gitcoin grants, quadratic funding in the Ethereum ecosystem. There's obviously these dog coins that just happens, I guess, accidentally. There's other projects that like, for example, Uniswap has their Uniswap DAO that just has a huge amount of funding and we haven't seen yet how that's going to be deployed, but it could be potentially deployed to do lots of really good and amazing things.

Do you see Ethereum as essentially a mechanism to fight for social causes?

I definitely see Ethereum as being a mechanism to fight for definitely some specific things that are social causes. Like just the fact of creating an open financial system that anyone can participate in, no matter where they are in the world. That's a social cause. Just giving people the ability to organize and create projects. Even if it's five people in five different countries, that kind of inclusiveness, I think that's a social cause and it's a core crypto value. But then at the same time, the other important part of the magic of Ethereum that you have to balance that against, is that it is also this open platform where ultimately the things that are on Ethereum is just the things that the community makes of it.

Well, you kind of briefly opened the door, so let's go there. When it comes to government regulation of crypto, what's the best case scenario? What's the worst case scenario in terms of, as you've kind of mentioned, Ethereum challenges the power centers of the world. And how do you see the interplay between governments and this new technology that resists centralized power? Best case and worst case.

The best case is that blockchains continue to prosper and we figure out scalability so that people can actually start doing things on block. All of the amazing use cases that people have been talking about instead of today, where a lot of the great stuff gets priced out because transaction fees are at $5 to $10. And then we see a lot of different amazing applications happening on blockchains. It could be daos creating new ways for people to interact and organize with each other, new ways for artists to get funded and all sorts of these amazing things. And there's just enough public support and just enough people that see that, look, crypto is clearly doing a lot of good things. And there are definitely areas where there's tensions, but in those areas where there's tensions, there could be some kind of creative and interesting approaches that get figured out. The concept of corporate taxes, for example, that would disappear as a revenue stream. If theoretically corporations just all get replaced by daos. But maybe there's some other creative way by which dao like daos themselves can kind of be co, you know, have some kind of encoded governance that ensures that they have at least some of us, some kind of bias towards serving the global public good. And you know, maybe it does enough of daos can do enough of that that people are happy with it. And you know, there are going to be things that people are unhappy about. There's always going to be the people that wants to surveil everyone. But if the kind of effect of crypto from just empowering people is greater than that and greater than that in a way that people can just easily see, then that would be a good scenario, right? And we'll just become kind of incorporated and accepted the same way as happened with the Internet. But the worst case scenario would of course be just like people suddenly flipping and going into moral panic mode and just, oh my God, this technology is used by insert bad group of the day. And then I don't think governments have the ability to ban crypto to the extent of just preventing blockchains from existing, but they definitely have the ability to really marginalize it. Right? Like if you just ban all exchanges and ban all links from the fiat ecosystem to crypto, and you know, you ban all kind of mainstream employers from accepting or paying in cryptocurrency, then like you could, you can successfully turn it into a fairly kind of niche countercultural thing that has much less impact than it otherwise would. So that's somewhere between the good scenario and the bad scenario. I'm obviously hoping for the good.

Well, that's interesting. Also the tension between governments and companies. Like if you have a bunch of billionaires or a bunch of companies like Tesla investing in bitcoin, and then governments resisting that, it's interesting who wins out in that worst case scenario. And almost when companies and rich, quote, unquote, Respectable people embrace cryptocurrencies. Bitcoin, Ethereum, so on. Even the dog coins. It almost sends a signal to everybody else that this is, this is a revolution that's here to stay. On this one little tangent that you brought up, this is almost an outdated idea, but it's still with us, which is cryptocurrencies are used for illegal activity, for drugs, for crime and so on. Is there some sense that worries you that if cryptocurrency, if ethereum runs the world, then crime, making money from crime will be easier?

There's always that possibility. At the same time, I think if you look at the world as a whole and the way all the other technological trends are going in person, surveillance is just going up every year, right? If you commit a crime in meatspace, it's getting harder and harder to get away with it. If he wants to do something, and this is something that's just happening as a result of just better technology, information and information transparency, a lot of it's hard to prevent even if you really tried. So the world where things go dark to such an ext, as the police hawks sometimes like to say, to such an extent that, oh my God, the criminals are committing crimes with impunity and we can't see anything, that just seems unlikely. But on the other hand, the world where there just is no privacy, for example, or the world where there just is no ability to kind of act outside of the confines of mainstream institutions, that's some, that's something that's more realistic and that seems like something that could lead to a lot of scary things, right? And even from a government's point of view, right. I think governments, over the last few years, a lot of them, they're very worried about sovereignty. They're worried about if their country is economy, is social environments. They're just completely dependent on basically foreign tech companies controlled by foreign governments. You know, governments are not on team government, right? It's like, you know, the Indian government is on, you know, Team India, you know, the Russian government is on Team Russia and so forth, right? So, you know, they don't want the US to be able to like have this big backdoor into everything. So, I mean, I do think that a balance is needed, but at the same time, I do think, I guess I definitely like worry more about the possibility that just without things like crypto, kind of acting outside of institutions becomes too impossible. And I don't even necessarily mean outside of governments, even just outside of corporations becomes too impossible. And there's just terrible things that come as a result if things going in the other direction, it obviously is a risk. But at the same time, I think in the long term, a crypto can potentially even offer defenses as much as attacks against that sort of thing.

Yeah. Throughout history, many of the most destructive things came from centralized institutions versus sort of from the people operating in the shadows. And I've been talking to a bunch of psychedelics folks that people doing researchers like Rick Doblin in Johns Hopkins. There's a lot of exciting research on psychedelics. And one thing you could say about operating at the edge of legality, it could actually accelerate the adoption of particular things, like whether it's marijuana or psychedelics, they can help people out. It almost accelerates the policy. It forces the policy to catch up to where the people stand. So there's a positive way of doing things that are in the gray area of legality and creating a market that allows people to, in a safe way, be able to participate in this gray area of legality.

The other thing to keep in mind, of course, is that the set of the kinds of things that just payment processors, as companies try to restrict you from, is much larger than the set of things that's illegal. Right. Part of that is because they want to be super conservative. And the more layers you have, the more they're conservative because they're scared of what the layer below them will do to them. Sometimes they have their own moral opinions of various kinds. They go after lots of people. They make life really hard for sex workers, for example. Psychedelics, as you mentioned, there's a lot of activity, even including stuff that is totally legal. There's this, you know, shadow like PayPal, credit card governments, or whatever you want to call it, and that makes it just hard to participate in this stuff. So I think, like, reducing the number of intermediaries is definitely normally a good thing.

All right, let's talk about one of the most exciting technologies, like technically, philosophically, like socially, financially, in every way, which is Ethereum 2.0. There's a million things to talk about, but at the step, one is probably a good thing to do, which is can you briefly summarize your vision, how Ethereum 2.0 will make Ethereum more scalable, secure, and sustainable?

Sure. So I think recently we've actually been kind of deemphasizing the ETH 2.0 branding, I guess. So the reason behind that was that originally we envisioned something more like a big grand event where all the good things would happen at the same time, and it would be a new Blockchain, and it would be a new protocol and people would have to take a lot of effort to migrate over. But later we've slowly changed the roadmap over to something that's much more incremental. So proof of stake happens kind of over time, and then sharding gets added over time, and all these features get added over time. And so the experience for just a regular Ethereum user still feels very seamless. Right. It's like maybe a little bit more complex than the hard forks that we've already done from a user's point of view, but not by that much. So the big two things that are happening, these are what used to be considered the two flagship features of E 3.0, and now they're just the flagship features of the next evolution of Ethereum as proof of stake and sharding. So proof of stake is a consensus algorithm, or consensus mechanism, I should say. The difference is that an algorithm is something that you run by yourself. A mechanism is interactions between people, and it could even include incentives and all of that. So the consensus mechanism by which nodes in the network agree on which blocks came in, which transactions came in, in what order, make sure that once a block gets accepted, it can't get reverted, and all of these things that we expect from a blockchain. So existing blockchains, including Bitcoin, including the Ethereum of today, and including a lot of them, they use proof of work. The reason why we need proof of anything is because they serve this function that I call economic civil resistance. So that's obviously a big word for, especially if you've never heard of Sybils before. But the basic idea is that you have a network and you have lots of computers that agree on which block to accept, and sometimes you get two blocks that get published at the same time, and you just have to agree in an order. So there has to be some kind of voting game. But then the question is, well, in this voting game, who gets a vote? Who gets to participate? Now, you can't say one person won vote. The reason why you cannot say one person won't vote is because you need some kind of authority or some kind of mechanism to say who the humans are. And if you don't have that, then a bad guy could just come in with a virtual machine or with a computer that has on it 10 billion virtual machines that have 10 billion virtual nodes, and then just say, look, I'm 99% of the network, I should control everything. So to prevent this, what proof of work and proof of Stake both do is they basically say, well, the weight of your vote, like how much influence your votes have in the consensus is proportional to what quantity of economic resources you bring in. So in the case of proof of work, you prove what economic resources you have, because your economic resources are computers. And you prove that you have them by just running them 24,7 using these hash algorithms. So this does solve the problem because in order to attack the network, you have to come in with more computers and more money invested into computers and electricity than the rest of the network puts together. And that's extremely expensive. In proof of stake, instead of relying on people with computers that are just constantly Cranking out hashes 24,7 as your unit of economic resources, you just use holdings of coins inside the system. So all of these blockchains, they have some kind of coin in them. Bitcoin has Bitcoin, Ethereum has ether, they all have a coin. So why not just use that as the economic resource that you're using to measure participation? So that's the core distinction between proof of work and proof of stake. I like proof of stake and I've liked proof of stake for many years, basically because it just requires much less ongoing resource consumption. With proof of work, you have to actually go and buy these physical computers. And these days they have specialized hardware, asics, application specific integrated circuits. You have to go produce them, you have to go buy them. And unless you have millions of dollars, you have to buy them from one of these other people who creates them. And those other people often end up taking a huge cut of the profits themselves. And then you have to plug them in, you have to just burn all of this electricity that's just them running 24, 7. So it consumes a huge amount of energy. And not just energy, it also just to create the hardware, people focus a lot on energy, but actually about half the cost of proof of work mining is the cost of the hardware. So hardware is a very big deal too. And you need this really big and powerful, very specialized hardware, the kind that fills up these big warehouses. So proof of stake, you don't really need that much electricity, you just need just a little bit to run a regular computer. You can run proof of stake validators on computers that you already have. So it's just much less resource intensive. And this is good for a few reasons, right? One is the kind of environmental rationale that you're not breaking the environment. The second is that you're not taking away electricity and other resources from other people. I think just today I Saw a story about Iran wanting to shut down some bitcoin mining because it was just grabbing up so much electricity that it was outbidding the nearby towns and they just didn't have enough. And then there was Chia, the one that's doing proof of hard disk mining basically is just grabbing up so many hard disks there's a shortage. So that's the second reason. And then the third, more selfish reason is that because participating in consensus does not require so much energy expenditure, you don't need to pay people as much to participate. Right? So like Bitcoin and Ethereum, they both issue somewhere around 4% of the total supply every year right now to miners. So ethereum is about 4.7 million ether and the current supply is about 115 million. But with proof of stake, we expect it'll be somewhere between 500,000 and 1 million per year. So, so that means, you know, the supply doesn't have to increase so quickly.

So one of the pros that people sort of argue for the proof of work is that it is secure because it's much more difficult to sort of, as you've highlighted, is difficult to participate. What are your thoughts about the security of the proof of stake mechanism? Is there ways to make it secure?

So I think proof of stake is very secure because in order to be able to attack the system, you need to have basically as much stake as the rest of the network. So that means right now, for example, we have 5 million eth staking. So you have to come up with 5 million eth and then join the network and then the other. So 5 million eth is a lot, right? It's like, how much is it now? Like $15 billion. So that's actually more than, I believe, the cost of attacking the bitcoin network. And then the second thing is that recovering from attacks is much easier in proof of stake than in proof of work. Right? Because in proof of stake you have like. First of all, we have, for many kinds of attacks that you do against this network, we have this concept of like automatic slashing, right? Which basically means that in order to revert a finalized block, so if there is one block that's accepted by the network, and you try to convince the network to kind of revert that block and accept a different block, in order to make that kind of attack, you basically have to have your validators, a big portion of your validators, sign to conflicting messages. And this is something that once these messages are on the network, you can go and prove, look, these people did It. And so we have this feature in the protocol called slashing where you basically take all these people who provably misbehaved and you burn their coin. Coins and you don't burn anyone else's coins. Now there are other cases like for example, if instead of reverting blocks, the attack just tries to censor everyone, then everyone who got censored would just basically create the minority chain. And then the community would basically have to do a soft fork. They would just have to say, look, this chain is clearly attacking us, this chain is the one not attacking us. And so we're going to join this chain. And then what happens is that on that new chain the attackers also lose a lot of coins. Right. So the difference between proof of stake and proof of work is that in a proof of stake system you can identify specific participants and you can say this isn't like a human going in and saying, I don't like you, I don't like you, I don't like you. This is like automated, right?

So the slashing process is automated?

Yes.

Is there ways it can go wrong? So that's a painful process where the coins are burned?

It is painful, yes. I think the one big unknown of course is if an attack actually happens, and if an attack happens that requires the community to actually choose one of these minority forks, then what would the community actually successfully coordinating on this look like? We can talk about it and we can write science fiction novels about it, but until it's happened, you don't really know the details of what it looks like and how difficult it is.

What are the channels of communication for the community? If you can enlighten me a little bit. Like what? You know, in many ways in the political realm, Twitter is often used as a way to kind of have these emergent phenomena of large groups of people coming to a consensus about a particular idea. And then there's battle for consensus. In the Ethereum community, how do people, what are the sources of natural language based communication that have an emergent belief structure that you would say, or is it all through money, is it all through trading that the communication happens?

There's definitely talking as well. We have to agree on protocol changes somehow. There's Twitter, there's Reddit, there's GitHub, there's all of the various Ethereum forums, Ethereum magicians, Ethereum research. There's just in person communication. Then there's just kind of the hidden web of everyone talking to everyone on telegram or signal. So it's like some of everything. Right. But I think the thing to emphasize around can you actually come to consensus on whether or not to fork the chain? Because the attacker is censoring everyone, just for example, is everyone who's running a node is going to see almost the same thing. They're going to be off by a few seconds and maybe they'll be off by a few. They'll disagree by a few minutes. But if it's a serious attack, you know, people are going to know, right? It's not like one of those things where, you know, oh, we're trying to agree on like, I don't know, did Epstein kill himself? Or like some random political facts where, like, in reality no one knows a single thing about what's actually going on and they're all speculating. It is much more visible. Right, so we do have that. But at the same time, I'm happy to admit that these are fairly untested mechanisms, but at the same time they're also untested mechanisms in proof of work. Right? And in proof of work, it's even harder because in proof of work you don't have the ability to identify and say, I'm going to. These miners attacked. And so we're not going to let these miners in. These miners did not attack, so we're going to keep them in. You have to pretty much either take out none of the miners or you do a fork that changes their proof of work algorithm, which takes out all of the miners. So the economics of recovering from attacks and proof of work, at least to me, actually do seem more unfavorable. But I'm sure the proof of work people you talk to will give very different and contradictory opinions, and that's totally fine and amazing.

Some people describe MEV minor extractable value as an existential risk to Ethereum. What is mev? How important is it to solve mev? If it's important, what ideas do you have?

Sure. How about after this one we'll also talk about charting because that's amazing and it's.

We'll return back to sharding, which is. We'll return to the big picture of the scaling problem, as you mentioned.

I love this conversation, you know, depth first search instead of breadth first. So basically, okay, EBV minor extractable value. It is not different in proof of work and proof of stake, right? So if you want to call it block proposer extractable value, like it sounds less sexy, but we could call it BPEV instead of mav, who cares?

But, so this is a problem in both proof of work and proof of stake.

Yes. So the basic idea is that if you have the ability to choose which transactions go into a block and in what order, then you have the ability to take advantage of that position for economic gain in a lot more ways than just collecting transaction fees. Right. Like, for example, there's decentralized exchanges on chain like Uniswap. And let's say the price of ETH versus USDC was 2,700 the previous block. But then there was a bit of a market drop and now it's 2680, where you can go on Uniswap and you can just gobble up the entire part of the automated order book that's between 2,726 80. And then at the same time, you run a bot and you buy some eth back at 2680 and you've just made about $10 of profit. Right. So. Or well, $10 times, you know, whatever the depth is. Right. So and so there's lots of little things like that. There's also things that involve like front running other people's transactions. So one example of this would be that if someone sends a transaction that says buy me 5 ETH for whatever price that you can get then, but with a maximum of, let's say, $15,000, then you can go and you can send each, put a transaction right in front of that transaction, and you can buy up that ETH first and then you resell it to him at $15,000 minus one. So there's.

You get to make a little bit of money.

Exactly. So there's a lot of these different, like arbitrage, front running, back running, these different tricks that allow block proposers to.

To get some percentage on top. Like overhead.

Exactly. And the reason why this is a challenge is because, first of all, it sometimes degrades user experience because users get less favorable trades. But there are sometimes ways to mitigate that for applications. Sometimes it's not that bad. But the bigger risk that I think some people consider more existential is that there's just much more economies of scale in figuring out how to extract all this revenue. Because if you're just collecting transaction fees, there aren't really economies of scale. There aren't really benefits to centralizing. Right. Because it's a very simple formula. You just grab up the transactions that pay you the most. But with mev, there's all these sophisticated algorithms, and if you have lots of money, then you can hire really smart people to make amazing algorithms and then you can use the other half of your money to Get a lot of mining power or a lot of stake, and you get a lot of opportunities to use your even better algorithms. So there is this risk that as a result of this mining is basically or even validating proof of stake is going to centralize. So I think the ecosystem's best reply to this sort of risk, and it's the direction where projects like Flashbots are going already, is if you can't eliminate the centralization, then you try to firewall it, right? And the way that you firewall it is you basically say, we're going to try to deliberately create a marketplace where people can just do the complicated work of creating what are called bundles, like bundles of transactions that are very profitable. And then at the other side of the market, you just have block proposers or miners that are just dumb nodes, and they go and ask what are called searchers, the bundle creators. And they just ask like, hey, how much can you give me if I put in your bundle? And then they just take the highest offer. So you sort of separate out the task and you have the easy part and then you have the hard part. And you have this special class of actor called a searcher that does the hard part and then the easy part. The people doing the easy part, which is just miners and validators, they kind of just talk to all the different people doing the searching and they just accept the highest bidder. This is also just an interesting example of economic design philosophy. Sometimes you can't just make centralization go away, sometimes it's inevitable. But no, at least you can try to kind of contain it. You can direct it, or you can even sort of firewall it away from core consensus, the parts that really do need to be decentralized.

But you don't see it as an existential risk. It's just a little. It's a bit of a problem that has to be constantly dealt with.

It's a risk. There's obviously a risk that it's a very severe problem and that even this flashbots approach has some fatal flaw or whatever. But we're definitely approaching it with the mindset of this is a problem. And yes, we do have to do some work to solve it, but we're doing it, and so far it's being solved.

Okay, let's talk about the other really fascinating part of the future of Ethereum. Let's not call it Ethereum 2.0, but the future of Ethereum that also may require a hard fork.

I don't know.

You can correct me on this is. Well, broadly, ideas for Scaling.

Yes.

And more specifically, sort of layer two or layer one and two intersection ideas of how to achieve scaling. And at the core of that is the idea of sharding. So first, what is sharding?

Okay, so there's two major paradigms for scaling blockchains, right? As you said, layer one and layer two. And layer one basically means make the blockchain itself capable of processing more transactions by having some mechanism by which it can do that, despite the fact that there's a limit to the capacity of each participant in the blockchain. And then layer two says, well, we're going to keep the blockchain as is, but we're going to create clever protocols that sit on top of the blockchain, that still use the blockchain and then still kind of inherit things like the security guarantees of a blockchain. But at the same time a lot of things are done off chain and so you get more scalability that way. So in Ethereum, the most popular paradigm for layer two is rollups, and the most popular paradigm for layer one is sharding.

So one way to achieve layer one scaling is to increase the block size. Yes, hence the block size wars quote, unquote. And you actually tweeted something about.

People.

Are saying that Vitalik changed his mind about the. He went from being a small, I.

Went from being big to small.

But you said I've been a medium blocker all along. So maybe you can also comment on where. On the very basic aspect, before we even get to sharding of where you stand on this block size debate.

Sure. So the way that I think about the trade off is I think about it as a trade off between making it easy to write to the blockchain and making it easy to read the blockchain. When I say read, I just mean have a node and actually verify it and make sure that it's correct and all of those things. And then by write I mean send transactions. So I think for decentralization it's important for both of these tasks to be accessible and I think that they're about equally important. If you have a chain that's too expensive to read, then everyone will just trust a few people to read for them. And then those people can change the rules without anyone else's permission. But if on the other hand, it becomes really expensive to write, then everyone will move on to basically second layer systems that are incredibly centralized and that takes away from decentralization and self sovereignty as well. So this has been my viewpoint pretty much the whole time, right? It's that you need this balance and going in one direction or the other direction is very unhealthy. In the bitcoin case, basically what happens was that bitcoin, originally at the very beginning, it didn't really have a block size, it just had an accidental block size of 32, or block size limit of 32 megabytes, because that just happens to be the limit of the peer to peer messages.

But then I didn't even know that part.

Yeah. But then Satoshi back in 2010 was worried that even 32 megabyte blocks would be too hard to process. So he put the limit down to 1 megabyte. And I think the by put you.

Mean sneaked in there.

Yeah, just like made an update to the bitcoin software that made blocks bigger than one. I think it's a million bytes invalid. And I think the impression that most people had at the time is that this is just a temporary safety measure. And over time, as we become more confident in the software, that limit would be raised somewhat. But then when the actual usage of the blockchain started going up, and then started going up, first to 100 kilobytes per block, then to 250 kilobytes per block, then To 500 kilobytes per block, they started kind of coming out of the woodworks this opinion that no, that limit should just not be increased. Then there are all of these attempts at compromising. First there's a proposal for 20 megabyte blocks. Then there was the 2, 4, 8 proposal, which is a bit ironic because the 248 proposal started off being a small block negotiating position. But then when the big block people came back and said, hey, aren't we going to do this? They're like, oh no, we don't want the block size increases anymore. So There were these two different positions. The spall blockers, I think they valued 1 megabyte blocks for two reasons. One is that they just really, really believe in the importance of being able to read the chain. But two is that a lot of them really believe in maintaining this norm of never hard forking. So the difference between a hard fork and a soft fork is basically that in a soft fork, any block that's valid under the new rules was still valid under the old rules. So if you have a client that verifies according to the old rules, then you'll still be able to accept the chain that follows the new rules. Whereas with a hard fork, you have to update your code in order to stay on the chain. And they have this belief that soft forks are kind of either less coercive than hard forks, which, by the way, I completely disagree with. I actually think soft forks are more coercive because basically they force everyone who disagrees to go along by default, but. Or they have this opinion that there's like, it's more difficult to abuse soft forks to do really mean things or that completely violate people's expectations, like increasing the supply, which is like, you know, I think there is some truth to that. So because of, you know, these reasons, they just say we're only going to do soft forks and we want to just not do any hard forks. And they eventually discovered this idea called segregated witness that allows for a very tiny block size increase to the equivalent of about 2 megabytes with a soft fork. It's just a really weird and devious trick. Basically what they do is they take the signatures of transactions and then they put them outside of the block, and then they add an extra rule that says that for a block to be valid, the block has to come with a separate, basically, extension block that contains all of the transaction signatures. So when you measure it according to the old rules, like, hey, it adds up to less than a million, but actually there's this extension block that the old protocol doesn't even know about.

It's a hack that seemed to work in a small way. Extent.

Exactly. Small block side was happy with these very low levels of block size. And then the big block side wanted to expand to at the very least go to 4 megabytes and maybe go maybe 8, 20. There's disagreements within there as well. I definitely was favoring the big side the whole way through, as you can probably tell.

Even though. So the argument against the big is that it makes things more centralized.

Yes, because fewer people can run a node that verifies the chain. And also because any of these things would require a hard fork. And hard forks are inherently risky.

Do you think there's truth to that?

I'm pro hard fork. I think hard forks are actually, in a political, economic sense, they're better than soft forks.

Well, I think that's a beautiful principle, as stated, that soft forks may be more coercive than hard forks. This is not just about cryptocurrency. This is about politics and life. That's fascinating. So you're okay with hard forks. In fact, you think hard forks is the right way to make changes, because then everybody's forced to make a decision, do you accept this change or not? As opposed to ideas being sneaked in behind the door and the decision is forced on you.

Exactly. Yeah.

Okay. But hard forks, some people say this is when they talk about sort of ethereum, is there's some aspect to a hard fork where you're trying to upgrade, what is it? Airplane while it's flying.

And I think soft forks are also upgrading an airplane while it's flying, but.

It'S a smaller upgrade.

There's some truth to that. There's definitely a bit more risk of a split as a result of a hard fork than as a result of a soft fork.

And this split is highly undesirable, right?

Well, it depends. Like if it's a split because of a bug, then that's horrible. If it's a split as a result of political differences, then I think like a split is better than, you know, one side being forced to basically just like, suck it up and accept the majority position, even if it really hates it.

Well, there's also political connections throughout the history of the United States. It's like sometimes groups of people that strongly disagree with each other should be forced to. To work it out. Even when a split seems like an easy thing in the short term, it depends.

Well, for blockchains in particular, the costs of people being able to peacefully go off and do their own thing are much lower. Right. Okay. If you have a country and you have two groups, then often enough fighting out the new rules requires a civil war, requires everyone to move and so forth. But on a blockchain, the costs are lower.

And so if you were to look at the way things worked out with the block size wars and there was a split, what is it? Bitcoin. Cash and bitcoin. Yeah. Like you looking, putting on your historian hat you mentioned offline, you like Dan Carlin. So if Dan Carlin were to do an episode on the buck size wars, do you think it could have turned out better or do you. Are you okay with the way it turned out?

I'm definitely disappointed with what happened with the block with the big block side. I think the source of my disappointment is that, like, one of the things that you notice when just looking at like this, political disagreements generally, especially when you have environments where, you know, they're authoritarian or like single party dominated and then there's some opposition party, and the opposition often has very legitimate grievances. But at the same time, the thing you notice is that often enough the opposition just sucks. Right. It just doesn't have political capacity. It doesn't have the ability to come up with policy because its entire culture is Designed around resisting much more than it's designed around actually debating serious policy trade offs. And I worry, or I guess not so much worry, because it's already happened. I unfortunately think that bitcoin cash ended up being a victim of this. First there was a split with bitcoin cash, and then, of course, Craig Wright came in. And Craig Wright was this basically scammer who just keeps on pretending that he is Satoshi Nakamoto, the inventor of bitcoin. Hey, Craig Wright's legal team, do you hear me? Yes. I still think your client is a scammer. So sue me.

This is definitely going to be depth first search, because I got to ask you about Greg, because these people have been contacting me, and I'm trying to figure out what is up with this human being. So for people who don't know, there's somebody who is.

Let's start.

There's Satoshi Nakamoto, who is the creator of bitcoin, who's anonymous. And actually, most really big people in the cryptocurrency space do not like. Like yourself and others do not dare claim that they are even for funds. Satoshi Nakamoto. In fact, if Satoshi Nakamoto is still alive and is like, if, say you were Satoshi Nakamoto, it seems like the thing he would do is probably or she is trying to remain anonymous. On the flip side of that, there's a guy named Craig Wright who continually keeps claiming that he is in fact Satoshi Nakamoto and keeps suing a lot of people. So on him, if we could just linger on him. What do you make of this character? What are we supposed to make of this character? Should he be ignored? Is there any possible truth to his claims? What do you make of him?

The analogy that's at the top of my head will get a bit political, but that's fine. You've had Michael Ballas. So I guess I view Craig Wright as being kind of like a Donald Trump figure in that he's not very intellectual. But I think he gets a big audience because he says things that play to the resentments that people have, and he says things that people want to hear. In the wake of this block size war, the big blockers did feel very disenchanted. They felt that bitcoin always had this vision that we were supposed to just keep increasing the block size. And bitcoin is peer to peer cash. It says so in the white paper. And then this elitist clique of core devs just came in and said, no, no, no, we're going to impose this totally different vision. And if you ever want your scalability, you'll have to wait for us to create this totally unproven, fancy technology called the Lightning Network that works under completely different principles. And they were very angry at this. And I think a lot of that anger is justified. But at the same time, when people are in that mental state, like, it's very easy for you to just kind of like latch on. And if you find someone who expresses anger at the same things that you're angry at and also seems like someone who's strong and seems like someone who might be good to rally around, it's very easy to just get behind that.

But that extra part about where he's Satoshi Nigamura, I don't understand why that's necessary.

I think that he could have done it without that. But that, I mean, that just. It's a marketing strategy. Like, it sort of gives him more salience. Like there's other big block personalities, right? But what's the difference between. With Craig, right? He's not just a big block personality, he's potentially satoshi. And he did say all the big block things, right? Like he talked about how, oh, the concept of a fee market is fundamentally economically wrong and it should be a free market and you should be able to have blocks as big as you want. Like, he repeated all the talking points and so a lot of people were kind of sucked into that. So he unfortunately was able to basically dominate a big part of the bitcoin cash community for a long time. And then eventually, of course, more and more people started to catch on. He would just say technical things that are completely wrong. One example of this that I remember is that he mixed up the concept of 256 bits and 2 to the power of 256 bits. So the difference is, it's like the difference between 80 and the concept of 80 digit numbers. And because of this, he made this argument that said that bitcoin's elliptic curve is friendly to cryptographic pairings. You don't have to understand what that is, but if you want to know, I have articles on both at Vitalik ca. But, but basically he made this technical argument that really hedged on this point. And then when people pressed him on it is like, yes, what? No, no, like what? Look, exactly. The height is like what, two to the 256 bits. That's a very tiny amount of information. No, no, no, no. 2 to the 256 bits is more than the amount of information in the universe. And he equivocated and kind of preyed on people's inability to understand that mathematical nuance. And I called him out. And eventually I even called him out in person at this conference in Seoul. I just stood up and asked, hey, conference organizer, why are you letting this fraud speak at this conference? I remember even some big blockers at the time getting angry at me. But eventually they did get rid of him. And basically Craig Wright was forced to split off because the rest of the community refused to accept some network change that he wanted to. And so then there was the BCH and bsp. And then in the bitcoin cash community, there was this drama of are they going to add a developer fund where they redirect 12 and a half percent of the revenue from the miners to the devs? And according to the libertarian non aggression principle, is just technically theft.

Like his understanding of the technical depths of cryptocurrency was lacking in a way that you, Satoshi Nakamoto, certainly would not.

Yes, exactly. But the point is that even after Craig Wright got a kick expunged, the bitcoin cash community kept having these disagreements. Right? And now after this development funds dispute, there was a further split between bitcoin cash and abc. The branching tree continues to extend. Right.

So in that way, it's disappointing to see those kinds of splitting there was never resolved.

It is. I would have definitely wanted to see more of a kind of principled coin that tries to be bitcoin but follows consistent big block values. But I don't know, maybe I should just stop expecting projects that I have no involvement in to care at all about what my values are. And maybe Ethereum just is.

I think you have a powerful voice and you can inspire other projects to live up to their best possible selves. Okay, so that's the layer one approach. The other layer one is within Ethereum is the idea of sharding.

Yes.

What the heck is sharding? Okay, what does the future of sharding look like?

Right, so to summarize, that big long tangent that we just went, it's a beautiful tangent. It's an amazing tangent. And I think, like, crypto is just one of the most underrated aspects of crypto is I think how you can like analyze the, you know, the sociology and the politics and the anthropology and. Yeah, and I'm sure Dan Carlton would have fun again exploring the space at some point. But the core trade off is that if you scale blockchains the dumb way, just by increasing the parameters, then eventually you just make it harder and harder to participate as a node, and you end up with a system where there's like 20 computers running the whole thing, and it's just very centralized. So Sharding basically says, well, instead of just increasing the parameters, what we're going to do is we're going to change the blockchain architecture in such a way that each individual node in the blockchain only needs to store a small portion of the data and only needs to process a small portion of the transactions. So you can think about it as being inspired by BitTorrent. On BitTorrent, there's no such thing as a BitTorrent full node that has every movie. The work is split up among a huge number of computers. That makes sense. That's the only sane way to scale a system like that. And if they actually tried making a version of bitsour that required full nodes that store every movie, then it would have zero censorship resistance, and it would just be dead in an instant. So the challenge with taking that model and applying it to blockchains is that blockchains aren't just about spreading data around. They're about agreeing on exactly what data was spread around and ensuring that everything that you agree on actually is correct. And so you have this paradox where, let's say you want to have a system that supports 10,000 transactions a second, but each computer in the network can only personally verify 100 transactions a second. So how can each computer get a guarantee about the other 9,900 without actually going and verifying them themselves? And it turns out that there are a bundle of different tricks that can do that, right? So one of them is just random sampling. So the idea behind random sampling is, let's say, for simplicity, this is a proof of stake chain, and you have 10,000 validators, validators are stakers. And for simplicity, we'll assume they all have the same number of coins. If someone has more coins, we'll just kind of split them up and pretend they're 10 stakers. Then you do some random shuffling, and you basically say, these random hundred validators are assigned to validate this block. These random 100 validators are assigned to validate this block. These random hundred validators are assigned to validate this block. And so each individual computer only gets assigned to validate a small piece. But then the way that the information about what's valid gets passed around is that when these hundred participants validate a block, they all sign a message basically saying, yes, we agree that this block is valid. And then they combine that signature into one, and then they broadcast that signature, and then everyone else, instead of verifying the blocks directly, just verifies that signature. And so if I see the signature, I'm not directly convinced that that block is valid, but what I am convinced of is that out of this committee, of this randomly selected group of 100 validators, let's say at least 70 of them agree that this block is valid. And so if I trust that the majority of these participants are all honest, then because it's all randomly selected, the attacker can't just force themselves into one committee. And so the attacker is going to be evenly spread out too. And so if the entire set of validators is mostly honest, every committee is going to be mostly honest. And so bad blocks are not going to go through. So that's one simple form of sharding. There's also other more clever things that you can do. So, for example, there's this concept of ZK snarks called zero knowledge proofs. So this is the idea that you can make a cryptographic proof that says, I verified or I ran some complex computation on this piece of data and I got this answer. And so if you make these kinds of proofs, then if you see a ZK snark that says some block is valid, then you're convinced that that block is valid. And even if everyone in that committee is evil, they have no way of making a valid proof for a bad block. Right. Because the proof itself, it is a proof that you did the computation where that proof is much easier to verify than just running the computation yourself. There's once again super awesome mathematical cryptographic magic behind making ZK snarks work.

It gives you a little bit of a leg up over the 51% honest assumption. So it's a little hack that improves upon the random sampling thing.

Exactly. And there's other hacks. Right? There is another hack called data availability sampling that allows you to make sure that the data in the blocks was actually published. But basically, if you stack a couple of these tricks on top of each other, you can create a system where I, as an individual participant can be convinced that everything that's going on in this distributed blockchain thing is correct without actually personally checking more than a percent of it. So that's charting, that's sharding.

But as I understand, maybe correct me on this, is in the space of Ethereum, the sharding happens on some fixed number. Like the split is on some fixed number. I think it's 64 is the currently sort of proposed number. So how does that help scaling? Is it just a fixed constant scaling by 64, and is that a way to achieve those crazy. The crazy amount of scaling that seems to be required to use cryptocurrency for purchasing, so competing with credit cards and Visa and so on.

Right. So first, I think the 64 can be hard forked up over time. So we have set it so that there's theoretically space in the data structure for 1,024 shards. It's just that 64 of them are turned on. There are challenges with having more shards because you have to have logic that just checks and manages all of those shards. And if there's too many of them, then that becomes too expensive. But even still, you can improve quite a bit. And then the other thing that we're doing is if we're getting maximum scalability by combining rollups and sharding. So this might be a good time to talk about rollups.

What are rollups? Now we're moving into layer two ideas.

Yes. So the idea behind a rollup is basically that. So instead of just publishing transactions directly on chain and having everyone do all of the checking of those transactions, what you do is you create a system where users send their transactions to some central party called an aggregator. Well, theoretically you can have a system where the aggregator switches around or anyone can be an aggregator. So, you know, it's still like permissionless to send things. Then what the aggregator does is they strip out all of the transaction data that is not relevant to helping people update the state. So when I say the state, this is like, this is a very important kind of technical term for blockchains. I mean, account balances code, things that are memory, internal memory of smart contracts. Like basically everything the blockchain actually has to keep track of and remember, right? So you just put in, you take all these transactions, strip out all the data that's not relevant to telling people how to update the state, and then you take the data that's needed to update the state and then you really compress it. Right. So like for example, if we say, you know, I, Vitalik, have an account that's 0x AB5, 8, blah blah, blah, blah, blah, and it's 20 bytes. Well, instead we can say, well, I have an account that is number 1-877-4224 in the tree, and that goes down from 20 bytes to just an index and a position which is three bytes. So you use all sorts of these fancy compression tricks and you basically just instead of publishing all these transactions, you publish this tiny compressed blob. So the amount of data that goes on chain goes down by maybe about a factor of 10. And then the second thing is that you don't do the computation on chain, instead you do the computation off chain. And there's one of two ways to do this. One is called a ZK rollup, which is you just provide a ZK snark that basically says, hey, look, I did this computation and I have this proof that here's some hash of the result and it's correct. And then you stick it on chain. And everyone verifies this one proof instead of verifying all these transactions. And then the other approach is called an optimistic rollup, which is basically made of this scheme where first someone says, hey, this is what I think the result of applying these transactions is. And then someone else can say, I disagree, the result is different. And only if two people disagree do you actually just publish all of the data and run that whole block on chain. So if there's disagreements, then you just run everything on chain and whoever was wrong loses a lot of money. So disagreements are very rare and they're very expensive. And in SDK rollup, you don't even rely on this challenging game at all. You just rely on a proof. So the core principle is basically that instead of lots of transactions and everyone verifies every transaction, it is. You take the transactions, you strip them down and compress them as much as possible, then stick that on the blockchain. You do need to stick something on the blockchain just so that everyone else can keep up to date with the state, so they know what all the contracts are, what all the balances are and all of this. But it's a very small amount of data. And then you use one of these other off chain games, could be this optimistic game, could be a ZK snark to just prove that somebody out there did the computation and the result is correct. So you're pushing 90% of the work off chain and then, well, 90% of the data and 99% of the computation off chain. And then you still have 10% of the data and 1% of the computation computation on chain. And so your scalability goes up by a factor of about 100. So these systems are already alive for some applications. So there's something called loop ring, which is just a ZK roll up for payments, right? So you can have assets inside of the loop ring system and you can go around and transfer them. But what you. And you get like much lower transaction fees, right? Like instead of $5, you have to pay less than 5 cents. But the only problem is that this only supports a couple of applications right now. Like making one that supports anything that you can do on Ethereum just takes a bit more work, but that's being done as well, right? So like within a few months I'm expecting fully Ethereum capable rollups to be available as well. So roll ups just summarizing, do most of the work off chain, put only a little bit on chain, factor of 100, scaling, sharding another factor of 100, scaling hundred times 100, factor of 10,000. Hundreds of thousands of transactions a second. And there's your scalability.

Okay, so you achieve scalability, you can do a large number of transactions very quickly and the cost of doing those transactions are much lower. You wrote that in the long term ZK roll ups are going to win in terms of layer two technology. Specifically, you wrote in general, my own view is that in the short term, optimistic roll ups, as you were saying, are likely to win out of general purpose EVM computation and ZK rollups are likely to win out for simple payments exchange and other application specific use cases. Just as you were saying. But in the medium to long term, ZK rollups will win out in all use cases as ZK snark technology improves. Why do you think ZK roll ups are going to win the big picture battle over layer two technologies?

So I think ZK rollups, once you accept that the technology works, are just conceptually simpler and they have nicer properties. The reason is that they do not have this concept of a challenge game. As I mentioned, in an optimistic rollup, the way that you ensure that the results are correct is that you let one person submit and they just submit with no proof. They just say here's what I think the result is. And then if someone else disagrees, they make their own submission. And then if you have two disagreeing submissions, then you actually publish it on chain and you see who's right. But for this to work, you need to actually wait for someone to disagree. So for example, if I have an asset inside of an optimistic rollup and I want to withdraw it, then I actually have to wait a week to withdraw it. Because if a block that contains my withdrawal turned out to be invalid, then there needs to be space for someone to disagree with it. Whereas with a CK rollup you don't need time for disagreement because you just have a proof, right? As soon as a block is submitted, there's a proof and you know it's Correct.

So if disagreements, especially in the long term, are sparse, then you don't want to do the optimistic, the game theoretic thing. You want to do the ZK snark.

Right? The ZK stuff is just in a ZK rollup. You can withdraw immediately. You don't have to worry about the economics of proving as much. There's just fewer issues. The reason why ZK rollups are not winning everywhere today is because ZKsdarks is still a crazy new technology. This is something that 10 years ago it existed only in theory and there was none in practice. Then eight years ago, people were just getting excited about it in Bitcoin conferences for the first time starting four years ago or three and a half years ago. Even that was the first time you were able to make any ZK snark based anything on Ethereum. And then people started making them and ZK technology has only really become efficient enough to do a lot of things within the past maybe one and a half years. So it's. It's new technology, it's crazy technology. It's admittedly scary technology. If you want to learn more, I also have an article about this on Vitalik ca.

It's actually really, really good. Most of your writing, it's. It goes, it's technical, but it's accessible. I highly, highly recommend to check out Vitalik's articles and blogs, whatever, whatever you call them. It's brilliant summary of the work. Actually Ethereum documentation period is really good. I think that's somewhat craz crowdsourced. That documentation is really accessible and brilliant. But let me ask about sort of other approaches to Layer two, like side chains. So the one popular one is Polygon. What are your thoughts about Polygon, which is layer two network? Is it positive? Is it negative for Ethereum? Is it both? Does it have a future which is its own chain, but it's using Ethereum? It's based on Ethereum, essentially, or maybe you can describe what it is.

So I think there's a really big and important difference in security models between rollups and sidechains, which is basically that roll ups inherit from the security of Ethereum. Right? So if I have coins inside of Loopring or Optimism or Arbitrum or Zksync, then even if everyone else in the world who is participating in these ecosystems hates me and wants to steal my money, I can still personally make sure that no matter what happens, I get my money out. It might be a bit expensive for me to get my money out and I have to do transactions on the main chain, but I'll be able to do it. Whereas in Polygon, which is a sidechain. And so instead of being secured by Ethereum, it's also in part secured by its own proof of stake consensus with its own token. So if 70% of the whole or even 51% of the holders of Polygon tokens wanted to take my money in Polygon, they can, and to be fair, the supply I don't think is even that widely distributed. Right. So potentially this idea of 51% of the token holders coming together and stealing everything, it's not impossible.

Where does the scaling of Polygon come from? Why is it it able to process much more transactions than the Ethereum main chain? What's the idea there?

I think in part, I'm not sure exactly what its capacity level is, but I imagine it has a higher capacity because it's a bit more willing to take centralization trade offs. And then another thing is that if the Ethereum ecosystem, even if it did not do that, if you think about an Ethereum ecosystem hypothetically scaling with sidechains, then it you would have 100 copies of Polygon and they would each have their own tokens, they would each have their own chains. And so even if each one of those chains was only as scalable as Ethereum, the total sum of them would still be 100 times more than Ethereum. Now the thing that I want to say in Polygon's favor, just to be very fair to them, I definitely really respect the work that they're doing. So start a bit with that word of not criticism, caution. Right. It's that they made this kind of deliberate trade off for very pragmatic reasons, which is that the Ethereum ecosystem needs to scale now and there are applications that want to do something now. And if there aren't Ethereum friendly options for them, then they're not going to just wait peacefully and do nothing for 12 months. They're going to go to either Binance Smart chain or some other system or potentially something that just has totally no alignment with Ethereum values whatsoever. But whereas with Polygon, the best thing that you can say in Polygon's favor and against optimism is that optimism is not live and Polygon is live. It just takes more work to create a system that has these extra roll ups, security features. And so Polygon just said we're going to be the system that makes a pragmatic trade off. We're going to go functionality first and then we can talk about adding back the security later. So I've talked to them and in principle I think they're very open to the idea of adding more security and becoming a roll up, or at least adding a polygon chain, let's say roll up at some point in the future, which is definitely something I think they absolutely should follow through on. But the fact that they exist now and so applications can kind of bootstrap now on a chain that even though its security isn't perfect, at least it exists and people can go use it and then over time the chain matures as the applications mature. I think a very reasonable strategy and I'm definitely really happy that they're part of the ecosystem.

Yeah, it's kind of interesting. The history of cryptocurrency has this tension of really good ideas that are hard to implement, so they take longer to implement and ideas that are not as good but are faster to implement. This is like the story of you have JavaScript that basically took over the world because it was quick to implement within 10 days and then later kept fixing itself. I don't know what to make of that. Sort of. From an engineering perspective, I'm more and more becoming comfortable and accepting the fact that our whole world will run a technology that's not as good as it could have been just because the crappy solution is faster to implement and it sticks. What do you make of that tension?

I think the compromise that we've been taking within Ethereum is when we have to take the crappy solution, we look for crappy solutions that are forward compatible with becoming good over time. Time, right. Like when you build the quick and dirty thing, you would still already have ideas in your head about what the more complete thing with all the security features added on would look like. Right.

Even if it requires a hard fork.

Yes. For example, with Sharding, I think it's likely that the first version of Sharding that comes out is not going to have ZK snarks and data availability sampling, for example. But we know what these technologies are. We feel like we have wrapped our heads around them and so we know how to build a system where we can put all the pieces in place so that it becomes very easy to bolt those components on in the future. So if you do things that way, then at the beginning you could have your system that has the functionality but say has less security or less sustainability or less of something else. But then over time it's designed in such a way that it has this easy on ramp to adding those things. And if you don't think explicitly about being future compatible, then you do often end up with a quick and dirty solution that backs you into a corner. There are definitely cases where I think the Ethereum ecosystem has suffered from that and we have had to expend pretty significant effort on, for example, removing features that we didn't realize that we actually can't sustain. Like one big example is just increasing the gas costs. So like making some operations more expensive when they. Because they should be expensive because they actually take a lot of time in the process. So that's making some things more expensive, kind of like taking some functionality away. So if you can be cognizant of where you're likely going into the future, and if you don't know, even be cognizant of both the most likely paths that you'll take in the future and thinking about your roadmap and coming up with a roadmap where you know that if you want to do either of those things, then you have a clean path toward it. That's probably the best kind of practical way to get the best of both worlds that we have.

Okay, let's talk about this wonderful process of merging. Okay, so there's the main net, which is the Ethereum 1.0 chain, or the chain, what should we say? The chain that uses proof of work as a consensus mechanism. And then there is what is it called? The beacon chain that uses the proof of stake mechanism. And I believe the beacon has been deployed successfully, is working. So that was in December 2020. There's a bunch of questions around that that's fascinating as well. But I think the most fascinating question is about merging those two. When do the two chains, one that's proof of work, one that's proof of stake merge. And what are the most difficult parts of this process?

Right, so as you've said, the way that we have set up this proof of stake transition is that at first the proof of stake chain just launches on its own. Right? And this was the thing that happened in December. And the proof of stake chain has been running for close to six months now. I mean, by the time people watch this, it might actually be six months, but it isn't actually coming to consensus on anything except for itself. Right? So the idea behind that is to just give the proof of stake chain time to mature, time for people to build the ecosystem around it, time to make sure that there aren't any bugs and just like prove to the community that no proof of stake actually is real and a full transition is realistic. Because the thing that you're transitioning to already exists and already works. And then at some point in the future, you have this event called the merge, where you basically take the activity that's being done inside of the proof of work chain and you actually move it over into the proof of stake chain. So you get rid of the proof of work side completely. So the way that the merge will work is it's definitely gone through a few different iterations. The earlier versions of this actually required more work for users and more work for clients. It was much more like, oh, there's this new chain, there's this old chain, and then everyone has to migrate from the old chain to the new chain, and then at some point we'll forget about the old chain.